- Email: [email protected]
The power of language: Functional brain network topology of deaf and hearing in relation to sign language experience
Accepted Manuscript The power of language: functional brain network topology of deaf and hearing in relation to sign language experience Michel R.T. Sinke, Jan W. Buitenhuis, Frank van der Maas, Job Nwiboko, Rick M. Dijkhuizen, Eric van Diessen, Willem M. Otte PII:
S0378-5955(18)30235-1
DOI:
https://doi.org/10.1016/j.heares.2018.12.006
Reference:
HEARES 7660
To appear in:
Hearing Research
Received Date: 31 May 2018 Revised Date:
8 December 2018
Accepted Date: 12 December 2018
Please cite this article as: Sinke, M.R.T., Buitenhuis, J.W., van der Maas, F., Nwiboko, J., Dijkhuizen, R.M., van Diessen, E., Otte, W.M., The power of language: functional brain network topology of deaf and hearing in relation to sign language experience, Hearing Research, https://doi.org/10.1016/ j.heares.2018.12.006. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
ACCEPTED MANUSCRIPT 1
The power of language: functional brain network topology of deaf and
2
hearing in relation to sign language experience
3
Michel R.T. Sinke 1 *, Jan W. Buitenhuis 1, Frank van der Maas 3,4, Job Nwiboko 4, Rick M.
5
Dijkhuizen 1, Eric van Diessen 2 § and Willem M. Otte 1,2 §
6
§ Shared last author
7
Biomedical MR Imaging and Spectroscopy Group, Center for Image Sciences, University Medical Center Utrecht and Utrecht University, Utrecht, The Netherlands
9 10
SC
1
2
M AN U
8
Department of Pediatric Neurology, Brain Center Rudolf Magnus, University Medical Center Utrecht, Utrecht, The Netherlands
11 12 13
3
Reabilitação Baseadana Comunidade (RBC) Effata, Bissorã, Oio, Guinea-Bissau 4
CBR Effata, Omorodu Iseke Ebonyi LGA, Ebonyi State, Nigeria
TE D
14 15
AC C
18
EP
16 17
RI PT
4
19
* Corresponding author
20
Michel R.T. Sinke, MSc
21
Biomedical MR Imaging and Spectroscopy Group, Center for Image Sciences, University
22
Medical Center Utrecht, Yalelaan 2, 3584 CM Utrecht, The Netherlands.
23
Email: m.r.t.sinke @umcutrecht.nl. Phone: +31 30 253 5569; Fax: +31 30 253 5561
24
ORCID ID: https://orcid.org/0000-0002-8185-9209
25 26 1
ACCEPTED MANUSCRIPT Abstract
28
Prolonged auditory sensory deprivation leads to brain reorganization. This is indicated by
29
functional enhancement in remaining sensory systems and known as cross-modal plasticity. In
30
this study we investigated differences in functional brain network topology between deaf and
31
hearing individuals. We also studied altered functional network responses between deaf and
32
hearing individuals with a recording paradigm containing an eyes-closed and eyes-open
33
condition.
34
Electroencephalography activity was recorded in a group of sign language-trained deaf (N =
35
71) and hearing people (N = 122) living in rural Africa. Functional brain networks were
36
constructed from the functional connectivity between fourteen electrodes distributed over the
37
scalp. Functional connectivity was quantified with the phase lag index based on bandpass
38
filtered epochs of brain signal. We studied the functional connectivity between the auditory,
39
somatosensory and visual cortex and performed whole-brain minimum spanning tree analysis
40
to capture network backbone characteristics.
41
Functional connectivity between different regions involved in sensory information processing
42
tended to be stronger in deaf people during the eyes-closed condition in both the alpha and
43
beta frequency band. Furthermore, we found differences in functional backbone topology
44
between deaf and hearing individuals. The backbone topology altered during transition from
45
the eyes-closed to eyes-open condition irrespective of deafness, but was more pronounced in
46
deaf individuals. The transition of backbone strength was different between individuals with
47
congenital, pre-lingual or post-lingual deafness. Functional backbone characteristics
48
correlated with the experience of sign language. Overall, our study revealed more insights in
49
functional network reorganization caused by auditory deprivation and cross-modal plasticity.
50
It further supports the idea of a brain plasticity potential in deaf and hearing people. The
AC C
EP
TE D
M AN U
SC
RI PT
27
2
ACCEPTED MANUSCRIPT 51
association between network organization and acquired sign language experience reflects the
52
ability of ongoing brain adaptation in people with hearing disabilities.
53
Keywords
55
Cross-modal plasticity; Deafness; ASL; EEG; Functional networks; Minimum spanning tree
RI PT
54
56 57
Abbreviations
59 60
ASL = American Sign Language; CBR = Community-Based Rehabilitation; EEG = Electroencephalography; MRI = Magnetic Resonance Imaging; MST = Minimum Spanning Tree
AC C
EP
TE D
M AN U
SC
58
3
ACCEPTED MANUSCRIPT 1. Introduction
62
Impairment or loss of hearing interferes with many activities in daily life, specifically limiting
63
communication with others. This could easily lead to social isolation. The prevalence of this
64
serious disability is greatest in middle- and low-income countries (Durkin, 2002; Stevens et
65
al., 2013; WHO, 2014). While in the United States about two out of every 1,000 children are
66
born with disabling hearing loss (Vohr, 2003), this number is considerably higher in Sub-
67
Saharan Africa where about two percent of the children is born with disabling hearing loss
68
(WHO, 2012). Many of these children have profound hearing loss resulting in absolute
69
deafness. Infectious diseases are a major cause of deafness in these regions (Mulwafu et al.,
70
2016).
M AN U
SC
RI PT
61
Prolonged periods of sensory deprivation often leads to extensive reorganization in the
72
brain. This reorganization is caused by compensatory and cross-modal plasticity (Bavelier and
73
Neville, 2002; Merabet and Pascual-Leone, 2010; Ptito et al., 2001). Brain reorganization
74
after auditory deprivation has been mapped by different functional neuroimaging modalities,
75
such as positron emission tomography, functional near-infrared spectroscopy, functional
76
magnetic resonance imaging (fMRI) and electroencephalography (EEG) (Buckley and Tobey,
77
2011; Dewey and Hartley, 2015; Doucet et al., 2006; Finney et al., 2001; Yoshida et al.,
78
2011). Consequently, deaf are better than hearing people at detecting visual stimuli (e.g.
79
Almeida et al., 2015; Bavelier et al., 2006, 2000; Bosworth and Dobkins, 2002; Brozinsky
80
and Bavelier, 2004; Dye et al., 2007; Finney and Dobkins, 2001; Hauser et al., 2007; Neville
81
and Lawson, 1987a, 1987b) and show increased tactile sensitivity as well (Auer et al., 2010;
82
Levänen and Hamdorf, 2001; Meredith and Lomber, 2011). Accordingly, the auditory cortex
83
of deaf is found to be responsive to non-auditory stimuli (e.g. Almeida et al., 2015; Auer et
84
al., 2010; Buckley and Tobey, 2011; Doucet et al., 2006; Finney et al., 2001; Karns et al.,
85
2012; Meredith and Lomber, 2011; Neville and Lawson, 1987b; Scott et al., 2014). For
86
example fMRI and positron emission tomography studies showed that the cortical auditory
AC C
EP
TE D
71
4
ACCEPTED MANUSCRIPT and association areas of deaf people are responsive to visual motion stimuli. These regions
88
include the planum temporale (Petitto, 2000; Sadato et al., 2005; Shiell et al., 2016) and
89
primary auditory cortices, like posterior superior temporal gyrus (Almeida et al., 2015; Ding
90
et al., 2015; Karns et al., 2012; Li et al., 2015) and Heschl’s gyrus (Karns et al., 2012; Meyer
91
et al., 2007; Scott et al., 2014; Smith et al., 2011). Cortical reorganization (Campbell and
92
Sharma, 2014, 2013) and responsive auditory cortex to visual stimuli in deaf have also been
93
described with functional brain data recorded with EEG (e.g. Buckley and Tobey, 2011;
94
Doucet et al., 2006; Neville and Lawson, 1987b, 1987a). While this reorganization occurs
95
inevitably as a result of profound deafness, cross-modal plasticity also strongly relates to the
96
acquisition and use of sign language (Meyer et al., 2007; Pénicaud et al., 2013). Furthermore,
97
the extent of cross-modal plasticity is dependent on the age of onset and the duration of
98
deafness (Brotherton et al., 2016; Li et al., 2013). Although stimuli- and task-based
99
approaches have provided valuable insights in cross-modal plasticity, they do not capture the
100
mutual dependency of different functional brain regions as well as the integrative nature of
101
the human brain (Hackett, 2012; Stam and van Straaten, 2012).
TE D
M AN U
SC
RI PT
87
The human brain forms a complex integrative network, which consists of spatially
103
distributed, but functionally connected regions that continuously interact with each other
104
(Bassett et al., 2018; Bassett and Sporns, 2017; Bullmore et al., 2009; Bullmore and Sporns,
105
2009; van den Heuvel and Hulshoff Pol, 2010). As such, functional brain connectivity and
106
reorganization can be better understood when these processes are not studied in isolation.
107
Brain network analysis explicitly takes the interdependencies between functionally connected
108
regions into account as it shifts emphasis from specific locational changes to global
109
topological alternations. The functional network topology, and potential reorganization
110
therein, can be effectively mapped with several network metrics (Bassett et al., 2018; Bassett
111
and Sporns, 2017; Bullmore and Sporns, 2012). Classical graph analysis describes the human
AC C
EP
102
5
ACCEPTED MANUSCRIPT brain as a collection of nodes (i.e. functional brain regions such as the auditory or visual
113
cortex) and edges (i.e. the functional connections between regions), and provides quantitative
114
information on the topological properties of these networks (Bullmore and Sporns, 2009;
115
Heuvel et al., 2012; Rubinov and Sporns, 2010). The healthy human brain has been
116
characterized as a complex network that effectively combines global and efficient integration
117
with segregation of functionally specialized brain regions (Bullmore and Sporns, 2012). This
118
unique topology with high integration and segregation is defined as a small-world network
119
organization (Bullmore and Sporns, 2009; Watts and Strogatz, 1998). Deviation from a small-
120
world organization has been related to many neurological and psychiatric disorders (Bassett
121
and Bullmore, 2009; Stam, 2014). Surprisingly few studies have used network analysis to
122
examine reorganization of brain networks in deaf individuals. Pre-lingual deaf adults showed
123
increased network clustering and nodal efficiency compared to controls, whereas brain
124
networks from post-lingual deaf adults did not differ from controls (Kim et al., 2014). This
125
indicates that auditory experience might affect the morphology of brain networks in deaf
126
adults. In another study increased functional connectivity was found between regions within
127
the limbic system, a system involved in sensory information processing (Li et al., 2016).
128
Functional network hubs shifted in the deaf subjects. The small-worldness did not differ
129
between pre-lingual deaf as compared to hearing controls (Li et al., 2016). Yet another study
130
found increased functional connectivity in brain networks in deaf during rest, which was also
131
shown to be related to sign language experience (Malaia et al., 2014).
SC
M AN U
TE D
EP
AC C
132
RI PT
112
Resting-state functional connectivity measurements can be performed in an eyes-open
133
or eyes-closed condition. Opening and closing the eyes are very basic attention-directing
134
behaviors. Eyes-open is related to ‘exteroceptive’ awareness, characterized by more
135
specialized overt attention and oculomotor activity, whereas eyes-closed is related to
136
‘interoceptive’ awareness, characterized by more integrative multisensory activity and
6
ACCEPTED MANUSCRIPT imagination (Xu et al., 2014). Eyes-open and eyes-closed conditions relate to different brain
138
states (Marx et al., 2004, 2003; Zhang et al., 2015) and topological organizations of functional
139
networks (Gómez-Ramírez et al., 2017; Tan et al., 2013; Xu et al., 2014). This is even so in
140
darkness (Hüfner et al., 2009, 2008). In EEG-studies functional networks showed increased
141
global efficiency and decreased clustering during the eyes-open state, specifically in the alpha
142
band, which might be due to alpha desynchronization, i.e. a reduction in the number of
143
functional connections in the eyes-open state (Gómez-Ramírez et al., 2017; Miraglia et al.,
144
2016; Tan et al., 2013).
SC
RI PT
137
Despite the usefulness of classical network analysis in capturing brain network
146
reorganization, it has some intrinsic limitations. The classical network analysis is particularly
147
limited in comparing inter-subject networks with different network densities (van Wijk et al.,
148
2010). The network density is defined as the number of connections relative to the potential
149
number of connections. Commonly used network metrics, such as the clustering coefficient –
150
used to measure network segregation – and average path length – used to measure network
151
integration – are highly affected by the number of connections within a network (Stam et al.,
152
2014; van Wijk et al., 2010). Therefore, comparing healthy and affected (or reorganized)
153
brain networks, in a situation of cross-modal plasticity or alpha desynchronization, might
154
yield biased results (Tewarie et al., 2015; van Wijk et al., 2010; Zalesky et al., 2010).
155
Solutions have been provided. A promising alternative network characterization approach not
156
limited by the network density is the assessment of the network backbone. Network
157
backbones are robustly and efficiently operationalized by the minimum spanning tree (MST)
158
(Stam et al., 2014; Tewarie et al., 2015). An increasing number of studies have shown the
159
usefulness of MSTs in capturing subtle network changes in brain development and ageing
160
(Boersma et al., 2013; Otte et al., 2015; Smit et al., 2016; Vourkas et al., 2014). MSTs have
AC C
EP
TE D
M AN U
145
7
ACCEPTED MANUSCRIPT 161
also been useful in characterizing multiple sclerosis, Alzheimer’s disease and epilepsy
162
(Engels et al., 2015; Tewarie et al., 2014; van Diessen et al., 2016, 2014). The present study therefore investigated the effects of prolonged periods of deafness
164
on the functional brain network backbone topology. To that aim we acquired resting-state
165
EEGs in deaf and hearing individuals. Data were recorded within a unique homogeneous
166
population living in a representative rural region in sub-Saharan Africa. In this region
167
deafness is a common disability and cochlear implants are not available. We hypothesized
168
stronger functional connectivity (i.e. more integration) between auditory cortex and other
169
sensory (i.e. visual and somatosensory) cortices in deaf people, due to cross-modal plasticity.
170
Hence, we also expected differences in functional network backbone topology between
171
controls and deaf. Given the expected cross-modal plasticity as well as auditory sensitivity to
172
visual stimuli, we further expected larger shifts in functional network topology between eyes-
173
open (i.e. ‘exteroceptive’ awareness) and eyes-closed (i.e. ‘interoceptive’ awareness) in deaf.
174
We also explored whether (shifts in) functional backbone topologies were different between
175
congenital, pre-lingual and post-lingual deaf. Lastly, we anticipated a relationship between
176
functional backbone characteristics and years of American Sign Language (ASL) experience.
SC
M AN U
TE D
EP
178
AC C
177
RI PT
163
8
ACCEPTED MANUSCRIPT 2. Methods
179
2.1 Study setting and ethics
181
The study pipeline as described below is schematically visualized in Figure 1. Our study was
182
conducted at two inclusive primary schools and one inclusive secondary school, located in
183
two separate rural villages in Ebonyi State, southeast Nigeria. The schools are part of a
184
Community-Based Rehabilitation (CBR) program, which implies that all students has to live
185
in surrounding villages and communities, aiming for a full integration within the community.
186
Inclusive education means that both students with and without disabilities are allowed to
187
participate in regular classes, and are supported to learn, contribute and participate in all
188
aspects of the educational program. This inclusive educational approach is a potential strategy
189
to reduce the individual as well as shared burden of disability (Eleweke and Rodda, 2002;
190
Pförtner, 2014). The number of students with and without disabilities enrolled in the schools
191
is approximately equal. Since the integration of deaf people forms one of the main focuses of
192
this CBR program, the majority of students with disability are deaf. Standard ASL forms an
193
integral part of the educational program for more than twenty years. This sign language has to
194
be learned by all teachers and students, both the hearing and the deaf. Besides, deaf students
195
also receive speech therapy. All lessons are taught in English and if a teacher does not yet
196
sufficiently master sign language there will be an interpreter who translates spoken language
197
into sign language for deaf students and vice versa. The CBR program leads to an increasing
198
amount of people within the community that speak sign language. This implies a more regular
199
use, development and integration of sign language by deaf students in their daily lives.
SC
M AN U
TE D
EP
AC C
200
RI PT
180
Our study was approved by the organizational boards (RBC/CBR Effata), the local
201
health ministry (Izzi, Local Government Area) and the federal government (Ebonyi State
202
House of Assembly, Abakaliki [7-11-2016]) in Nigeria. The study protocol was clearly
203
explained to all students in class before they were asked to participate in the EEG recordings.
204
Written informed consent was obtained from adult participants and caretakers of students
9
ACCEPTED MANUSCRIPT 205
below eighteen years. In addition, we also obtained assent from the students below eighteen
206
years.
207
2.2 Participants
209
Table 1 shows the demographic information of all participants. We included 193 participants
210
between ten and 43 years old (mean age of 18.5 (standard deviation 6.0); gender: 103 male,
211
90 female), both students and teachers. Sign language experience of participants varied
212
between zero and seventeen years at the time of recording. We selected both hearing (n=122)
213
and deaf (n=71) participants. The pre-lingual deaf participants were all capable of lip reading.
SC
RI PT
208
M AN U
214
2.3 Data acquisition
216
We used a sixteen sensor / fourteen channel EEG monitor configured to sample at 128 Hertz
217
with a 16-bit resolution (EMOTIV Inc, San Francisco, USA), which has been validated and
218
successfully applied in several studies (Aspinall et al., 2013; Badcock et al., 2015, 2013;
219
McMahan et al., 2015; Prause et al., 2016; Yu and Sim, 2016). This wireless headset can be
220
connected to a computer via Bluetooth and is an invaluable tool to collect EEG signals from
221
participants in rural or resource-limited areas, where access to a standard EEG system is often
222
impossible or burdensome. Two sensors were preserved for reference and grounding: the
223
‘common mode sense’ (CMS; located at P3) sensor was used as the active reference for
224
absolute referencing. The ‘driven right leg’ (DRL; located at P4) sensor was used for
225
feedback noise cancelation. The electrodes were located at anterofrontal (AF3, AF4, F3, F4,
226
F7, F8), frontocentral (FC5, FC6), occipital (O1, O2), parietal (P7, P8) and temporal sites (T7,
227
T8), according to the International 10–20 system. Signal quality scores are recorded for each
228
electrode with a range from one to five (no units), with five as best quality score.
AC C
EP
TE D
215
229
Participants were seated in a comfortable chair in a sound-attenuated room where the
230
Emotiv headset was placed. Participants were instructed to keep their eyes closed for the first
10
ACCEPTED MANUSCRIPT three minutes and open in the next two minutes, or vice versa. The order of the condition
232
sequence was assigned alternatingly, so that half of the participants started in the eyes-open
233
condition whereas the other half of the participants started in the eyes-closed condition. The
234
researcher kept a log on deviations from the protocol, or unusual events in the environment,
235
that may affect the experiment. Example recordings are shown in Figure 2. As an EEG
236
contains a multitude of overlapping signal waves with distinct amplitudes and frequencies we
237
separated the signal into the five most common frequency ranges. The EEG signals were
238
band-pass filtered into the delta (0.5–4 Hz), theta (4–8 Hz), alpha (8–16 Hz), beta (16–32 Hz)
239
and gamma (32–64 Hz) frequency bands. Examples are shown in Suppl. Figure 1 and 2.
SC
RI PT
231
M AN U
240
2.4 Data cleaning and window selection
242
Time segments were removed from the recordings if i) the research log indicated a deviation
243
from the protocol, ii) the EEG signal quality score was below four for any of the channels,
244
and iii) if the absolute deviation of the gyroscope signals relative to the gyroscope signal
245
median exceeded five times the standard deviation. This threshold was based on visual data
246
inspection (See example in Suppl. Figure 3). Subsequently, the cleaned and filtered time-
247
series were cut into ten-second epochs. Functional connectivity measurements as well as
248
multiple network backbone metrics has been shown to stabilize within recordings if the epoch
249
length is six seconds or longer (Fraschini et al., 2016). In addition, multiple epochs per subject
250
further increase the stability of network backbone metrics (van Diessen et al., 2015).
251
Therefore we used multiple epochs combined with this conservative ten-second length.
EP
AC C
252
TE D
241
253
2.5 Functional connectivity
254
For each epoch a functional network was constructed, which can either be visualized as a
255
connectivity matrix or a network graph (Figure 3). Recorded time-series within each epoch
256
were used to determine functional connectivity (i.e. forming the ‘edges’ in a network)
11
ACCEPTED MANUSCRIPT between different electrodes capturing neuronal signals from underlying brain areas (i.e.
258
forming the ‘nodes’ in a network). Functional connectivity was computed and quantified with
259
the phase lag index. This is a measure of the asymmetry of the distribution of instantaneous
260
phase differences between two time series and scales between zero and one (Pillai and
261
Sperling, 2006). It is relative resistant to the influence of common sources, including volume
262
conduction and active reference electrodes. An index of zero indicates no phase coupling (i.e.
263
no functional connectivity) between time series, or coupling with a phase difference centered
264
on zero ± p radians. A non-zero index indicates the presence of phase coupling (i.e. functional
265
connectivity), where higher values indicate stronger functional connections. A more
266
mathematical description of computing the phase lag index can be found elsewhere (Stam et
267
al., 2007).
M AN U
SC
RI PT
257
268
2.6 Functional connectivity strength between auditory, visual and somatosensory cortices
270
Since we expected remodeling of the auditory cortex in deaf people, reflected as enhanced
271
functional connectivity between the auditory and sensory cortices, we characterized in both
272
groups the average phase lag index between sets of electrodes. These sets were the temporal
273
electrodes T7/T8, covering the auditory cortex, the parietal electrodes P7/P8, covering the
274
somatosensory cortex, and the occipital electrodes O1/O2, covering the visual cortex, in eyes-
275
open and eyes-closed conditions.
EP
AC C
276
TE D
269
277
2.7 Minimum spanning tree analysis
278
For each functional network a minimum spanning tree was calculated from the connectivity
279
graph G by applying Kruskal’s algorithm (Kruskal, 1956) (Figure 3B). This tree captures the
280
network’s backbone and is defined as a subset of the network nodes (forming the original
281
weighted graph G) that connects all the nodes and does not contain cycles or loops (Jackson
12
ACCEPTED MANUSCRIPT 282
and Read, 2010). Mathematically, a minimum spanning tree T minimizes the sum of the costs
283
of its edges, l(T ) = ∑ l w over the set of all possible spanning trees on G (Hidalgo et al., w∈T
2007). Since we are interested in the strongest functional connections (i.e. the network
285
backbone), we first inverted the edge weights of the functional network before determining
286
the MSTs. Subsequently, several MST metrics were calculated at the nodal or network level.
287
Although some metrics are determined at the nodal level (e.g. degree or betweenness
288
centrality), they can still be used to summarize – or indicate – specific properties for the
289
backbone as a whole. For example with the ‘maximum degree’ or the ‘average strength’,
290
where higher values may indicate higher overall connectivity. The following MST metrics
291
were calculated at nodal or network level:
292
i)
Maximum node degree (nodal): every tree was summarized by taking the maximum node degree: Smax, the node with the maximum number of connections.
ii)
Leaf number (Nleaf) (network): the number of nodes of the tree with exactly one
TE D
293 294
M AN U
SC
RI PT
284
connection to any other node (with maximum degree = 1). A higher leaf number is
296
related to increased global efficiency and integration (Stam et al., 2014; Tewarie et al.,
297
2015). iii)
has a lower bound of two and an upper bound of m = N – 1. The largest possible
299
diameter will decrease with increasing leaf number (Boersma et al., 2013; Stam et al.,
300
2014; Tewarie et al., 2015).
301 302
Diameter (d) (network): the largest distance between any two nodes in a tree, which
AC C
298
EP
295
iv)
Eccentricity (network): the shortest path length between a tree node I and any other
303
node from the tree. Eccentricity decreases when nodes become more central in the
304
tree.
305 306
v)
Radius (nodal): the smallest node eccentricity in the tree. The lower the eccentricity, the more central a node in a tree.
13
ACCEPTED MANUSCRIPT 307
vi)
weights (Hagmann et al., 2010; Rubinov and Sporns, 2010).
308 309
Strength (nodal): the tree node strength is a summation of all nodal connection
vii)
Maximum betweenness centrality (BCmax): a network hub metric which relies on the identification of the number of shortest paths that pass through a node (Rubinov and
311
Sporns, 2010). The more the passages, the higher the betweenness-centrality (i.e.
312
hubness), which is defined by bc i =
1 ( n − 1)( n − 2)
n
RI PT
310
∑
j ≠ k ,k ≠ i , j ≠ i
g jk (i ) g jk
where gjk is the
shortest path between two nodes and gjk(i) is the number of node paths that actually
314
pass through i. We summarized the tree by taking the maximum betweenness
315
centrality.
317
viii)
M AN U
316
SC
313
Closeness centrality (nodal): the inverse of the sum of all distances to other nodes (Sabidussi, 1966).
318
2.8 Statistical analyses
320
All statistical analyses were performed within a Bayesian framework. Differences between
321
groups, eye conditions and interactions between groups and eye conditions were evaluated
322
with Bayes factors. Bayes factors were extracted from Bayesian model comparisons. This was
323
done for all frequency bands separately. We determined the model likelihood of a null model
324
without an interaction effect of group and condition as well as the likelihood of an alternative
325
model with an interaction effect of group and condition for functional connectivity strength.
326
This was done for all MST metrics and for their potential relationship with sign language
327
experience. Bayes factors give the ratio of model likelihoods, indicating which model is
328
supported by the data. For example, if – given the data – a null model (M0) without an effect
329
of condition (i.e. eyes-open versus eyes-closed) on functional connectivity strength has a very
330
low probability, whereas an alternative model (M1) with an effect of condition on functional
331
connectivity strength has a high probability, this would yield a high Bayes factor (e.g. 50).
AC C
EP
TE D
319
14
ACCEPTED MANUSCRIPT This may be interpreted as very strong evidence for M1 as M1 is fifty times more likely than
333
M0 in explaining the data. Table 2 gives an overview of Bayes factors and their interpretation
334
(Raftery, 1995). Since sex (Boersma et al., 2011) and age (Smit et al., 2012) influence
335
functional network topologies, we tested whether models with strong evidence were affected
336
by sex- and age.
RI PT
332
All network analyses, statistical modeling and visualizations were performed in R
338
(http://www.r-project.org/) using the packages igraph, BayesFactor, reshape2, dplyr and
339
ggplot2. Epoch data and scripts are freely available at the Open Science Framework in
340
anonymized form (Otte et al., 2018b) and the GitHub repository (Otte et al., 2018a).
SC
337
M AN U
341
AC C
EP
TE D
342
15
ACCEPTED MANUSCRIPT 3. Results
343 344
3.1 Functional connectivity strength between auditory, visual and somatosensory cortices
346
Figure 4 and Suppl. Figure 6 show the functional connectivity strength between the occipital
347
cortex and the parietal cortex for the different frequency bands. Occipital-parietal functional
348
connectivity strength was lower in the eyes-open condition compared to the eyes-closed
349
condition in the theta, alpha and beta frequency bands. After transition from eyes-closed to
350
eyes-open, differences in functional connectivity strength were most pronounced in the alpha
351
frequency band with a significant reduction of 54.9% (95% confidence interval (CI): -68.2%
352
to -41.6%) in controls and an even larger reduction of 88.0% (95% CI: -112% to -63.6% in
353
deaf (Figure 4B). A decrease in functional connectivity strength from eyes-closed to eyes-
354
open was also present in the beta frequency band, with a reduction of 27.7% (95% CI: -37.3
355
to -18.0) in controls and 36.1% (95% CI: -49.2 to -23.0) in deaf. These reductions in
356
functional connectivity strength in the alpha and beta band were supported by Bayes factors
357
of respectively >100 (i.e. labeled as ‘extreme evidence’) and 2.6 (‘moderate evidence’)
358
(Table 3).
EP
TE D
M AN U
SC
RI PT
345
Similar trends were seen for functional connectivity strength between the parietal
360
cortex and temporal cortex (Suppl. Figure 4 and Suppl. Figure 7) as well as between the
361
occipital cortex and temporal cortex (Suppl. Figure 5 and Suppl. Figure 8). However, for
362
those connections no significant differences were found between controls and deaf in
363
functional connectivity reduction from eyes-open to eyes-closed, as indicated by Bayes
364
factors of <1 (Suppl. Table 1 and Suppl. Table 2). The parietal-temporal delta-band
365
connectivity strength was significantly more reduced from eyes-closed to eyes-open in deaf,
366
while no reduction was found in controls (Suppl. Figure 5), as indicated by the high Bayes
AC C
359
16
ACCEPTED MANUSCRIPT 367
factor (Suppl. Table 2). All evidence was still present if age and sex were added to the
368
models.
369
3.2 Functional backbone differences between eyes-open and eyes-closed conditions
371
Transition from eyes-closed to eyes-open initiated visible changes in functional network
372
topology, as indicated by changes in backbone metrics. The most notable effects were found
373
in the alpha and beta bands (Table 4), which are shown in Figure 5. Overall the backbone
374
leaf number, the average and maximum strength as well as the kappa were lower in the eyes-
375
open condition as compared to the eyes-closed condition. In contrast, the diameter,
376
eccentricity, radius as well as the median and maximum closeness centrality were higher in
377
the eyes-open condition compared to the eyes-closed condition. Interestingly, in contrast to
378
deaf, controls did not show a transition effect on some network metrics in the alpha band (i.e.
379
leaf number, eccentricity, radius and diameter).
M AN U
SC
RI PT
370
TE D
380
3.3 Larger functional backbone modifications in deaf
382
Some functional backbone characteristics were different between deaf and hearing controls
383
(Figure 5). In both the alpha and beta band, functional backbone strength was stronger in the
384
eyes-closed condition and weaker in the eyes-open condition in deaf as compared to hearing
385
controls. In the alpha band the leaf number was lower in deaf, whereas the diameter and
386
radius were larger in the eyes-open condition.
AC C
387
EP
381
Several functional backbone characteristics showed larger shifts in deaf than controls
388
when comparing eyes-closed to eyes-open (Figure 5). For both the alpha and beta band,
389
evidence was found for a larger decrease in the average and maximum backbone connectivity
390
strength in deaf (as indicated by the Bayes factors in Table 4). Furthermore, for the alpha
391
band there was moderate to strong evidence for a larger increase in both betweenness and
392
closeness centrality in deaf. Anecdotal evidence was found for larger shifts in leaf number and 17
ACCEPTED MANUSCRIPT 393
diameter – in the alpha band – as well as in closeness centrality – in the beta band – in deaf
394
(Table 4). All evidence was still present if age and sex were added to the models. We found that the decrease in average connectivity strength (Suppl. Figure 9) and
396
increase in closeness centrality (Suppl. Figure 10) from eyes-closed to eyes-open was
397
different across the congenital, pre-lingual and post-lingual deafness types. This is supported
398
by the Bayes factors in Suppl. Table 3.
RI PT
395
399
3.4 Relation between backbone characteristics and American Sign Language
401
We investigated the relationship between sign language experience and functional backbone
402
characteristics. Initially, with no distinction made between deaf and controls. Sign language
403
experience was related to altered backbone characteristics in the delta, theta and alpha band.
404
The most pronounced effects were found in the theta band (Table 5), which are visualized in
405
Figures 6 and 7. More specifically, for the theta band an increase in ASL experience was
406
related to a higher average backbone connectivity strength and a lower closeness centrality for
407
the eyes-closed condition (Figure 6). Furthermore, strong evidence was found for a positive
408
relationship between sign language experience and higher backbone connectivity strength for
409
the delta band (Table 5). Figure 7 shows the relationship between sign language experience
410
and functional backbone characteristics for deaf subjects only. Again most pronounced effects
411
were found for the theta band (Table 6) with a positive relation between sign language
412
experience and higher average backbone connectivity strength. All associations were still
413
present if age and sex were added to the models.
AC C
EP
TE D
M AN U
SC
400
414
18
ACCEPTED MANUSCRIPT 4. Discussion
416
The present study used resting-state EEG to map functional network backbone differences
417
between deaf and hearing people. We showed that transition from eyes-closed to eyes-open
418
was associated with changes in functional connectivity strength between occipital and parietal
419
lobes as well as changes in functional backbone topology. In both deaf and hearing, these
420
changes occurred especially in the alpha and beta frequency bands. Moreover, the difference
421
in functional connectivity strength as well as in functional backbone characteristics between
422
eyes-closed to eyes-open tended to be larger in deaf as compared to hearing controls.
SC
RI PT
415
423
4.1 Functional cortical remapping
425
We examined functional connectivity strength between sensory regions involved in auditory,
426
visual and somatosensory information processing. With some evidence we can state that
427
functional connectivity appears stronger in deaf, specifically in the alpha and beta bands. This
428
might indicate increased integration between the different sensory cortical regions in deaf.
429
Our results are in agreement with previous fMRI studies that also found increased audiovisual
430
connectivity in deaf (Bola et al., 2017; Li et al., 2016, 2013; Shiell et al., 2014). With the
431
shifts from an eyes-closed to an eyes-open condition we found a reduction in functional
432
connectivity strength. This is expected given the shift from more ‘interoceptive’ awareness,
433
characterized by integrative multisensory activity, towards more ‘exteroceptive’ awareness
434
focused on attention and oculomotor activity (Marx et al., 2004, 2003; Xu et al., 2014).
435
Specifically for the alpha band, this change may be related to ‘alpha desynchronization’
436
during the eyes-open state (Barry et al., 2009; Gómez-Ramírez et al., 2017; Xu et al., 2014).
437
Moreover, we found this reduction in functional connectivity strength from eyes-closed to
438
eyes-open to be larger in deaf than hearing controls. This may be explained by long-term
439
auditory deprivation and related cross-modal plasticity mechanisms (Bavelier and Neville,
440
2002; Merabet and Pascual-Leone, 2010). It may be explained more specifically by the
AC C
EP
TE D
M AN U
424
19
ACCEPTED MANUSCRIPT enhanced sensitivity of the auditory cortex to non-auditory stimuli (e.g. Almeida et al., 2015;
442
Auer et al., 2010; Brozinsky and Bavelier, 2004; Dye et al., 2007; Finney and Dobkins, 2001;
443
Levänen and Hamdorf, 2001; Neville and Lawson, 1987a, 1987b) and increased audiovisual
444
connectivity in deaf (Bola et al., 2017; Li et al., 2016, 2013; Shiell et al., 2014) together with
445
the reduced integrative multi-sensory activity in the eyes-open state (Gómez-Ramírez et al.,
446
2017; Xu et al., 2014).
RI PT
441
447
4.2 Functional backbone differences between eyes-open and eyes-closed conditions
449
We found functional backbone differences between the eyes-open and eyes-closed conditions
450
in both deaf and hearing subjects, mainly in the alpha and beta bands. During the eyes-open
451
condition the functional network backbone showed an increased diameter and eccentricity
452
combined with a decreased leaf number and connectivity strength, although hearing controls
453
did not show differences in diameter and leaf number in the alpha band. These network
454
metrics indicate that the functional backbone topology in the eyes-open condition was more
455
chainlike (i.e. less integrated and with reduced global efficiency), whereas during the eyes-
456
closed condition the topology was more star-like (i.e. more functional integration and
457
increased global efficiency) (Stam et al., 2014). These results also nicely fit with the reduced
458
connectivity and integration in the ‘exteroceptive’ state, as well as the alpha
459
desynchronization, during the eyes-open condition (Barry et al., 2007; Chen et al., 2008;
460
Gómez-Ramírez et al., 2017; Marx et al., 2004; Xu et al., 2014). Other studies reported
461
increased global efficiency and decreased clustering of functional networks in the alpha band
462
(Miraglia et al., 2016; Tan et al., 2013) and beta-band (Gaál et al., 2010; Knyazev et al., 2015)
463
in the eyes-open condition as compared to the eyes-closed condition, which seem to partly
464
contradict our findings. This could potentially be explained by the distinct age groups. One
465
study showed similar effects as our study in participants between 18-35 years, but the
AC C
EP
TE D
M AN U
SC
448
20
ACCEPTED MANUSCRIPT opposite effects in participants between 51-80 years (Knyazev et al., 2015). This pattern is in
467
line with (Miraglia et al., 2016). However, two other studies also showed contrasting results
468
in younger participants (Gaál et al., 2010; Tan et al., 2013). These discrepancies might be due
469
to methodological differences as two studies (i.e. Miraglia et al., 2016; Tan et al., 2013) used
470
a large number of EEG electrodes or regions of interest (i.e. respectively 128 and 84)
471
compared to our study. Increased network sizes, with potential higher densities, may arguably
472
affect alteration of topological network characteristics if analyzed with classical network
473
metrics (van Wijk et al., 2010). Network size also affects backbone metrics (Tewarie et al.,
474
2014), which may explain the discrepancy between our study and Gaál et al. (2010).
SC
RI PT
466
M AN U
475
4.3 Larger functional backbone modifications in deaf
477
We found some backbone differences between deaf and hearing subjects. Functional
478
backbones in deaf showed higher strength in the eyes-closed, while lower strength during the
479
eyes-open condition, in both the alpha and beta band. Also in the eyes-open condition, the leaf
480
number was lower, while the radius and diameter were larger, which means that the functional
481
network topology in this state is different between deaf and hearing. More specifically, the
482
functional backbone of deaf shows less integration (i.e. reduced global efficiency) and
483
increased clustering. Previous network studies in deaf also showed increased clustering and
484
local efficiency, although the small-worldness was preserved (Kim et al., 2014; Li et al.,
485
2016).
EP
AC C
486
TE D
476
Furthermore, we found larger functional backbone shifts in these functional backbone
487
characteristics, from the eyes-open to the eyes-closed condition in deaf people. The patterns in
488
functional backbone strength correspond with changes we found in functional connectivity
489
strength between the different sensory cortical regions. They additionally indicate large-scale
490
functional connectivity changes, i.e. beyond the sensory cortices, since the functional
21
ACCEPTED MANUSCRIPT backbone topology is more altered in deaf when going from eyes-closed to eyes-open.
492
Accordingly, these findings may also be explained by transition from ‘exteroceptive’
493
awareness to ‘interoceptive’ awareness (Marx et al., 2004, 2003; Xu et al., 2014) together
494
with alterations in inter-regional connectivity – partly due to alpha desynchronization (Barry
495
et al., 2009; Gómez-Ramírez et al., 2017; Xu et al., 2014) – combined with auditory
496
deprivation and cross-modal plasticity (Bavelier and Neville, 2002; Merabet and Pascual-
497
Leone, 2010).
RI PT
491
We also found that the functional backbone connectivity strength differed between the
499
different forms of deafness. Functional backbone connectivity might depend on whether
500
people are born deaf or acquired deafness later in life. Both congenital and pre-lingual
501
deafness showed similar patterns as hearing controls, whereas post-lingual deaf showed the
502
largest deviation, which is in agreement with previous findings. It has been shown that cross-
503
modal plasticity is dependent on the age of onset and duration of deafness (Brotherton et al.,
504
2016; Li et al., 2016, 2013; Sadato et al., 2004). However, given the limited number of post-
505
lingual deaf in our study as well as the complex relationship of cross-modal plasticity with
506
other factors, such as age of onset, duration of deafness and sign-language experience, it is
507
impossible to be conclusive.
M AN U
TE D
EP
508
SC
498
4.4 Relation between backbone characteristics and sign language experience
510
Sign language comprehension have shown to be related to activation of brain regions which
511
are normally considered to be involved in unimodal (e.g. speech or sound) auditory
512
processing, such as frontal and temporal regions (Li et al., 2016; Liu et al., 2017; Malaia et
513
al., 2014; Meyer et al., 2007; Neville et al., 1998; Nishimura et al., 1999; Petitto, 2000; Sadato
514
et al., 2005). Language comprehension requires higher cognitive functions. Increased use and
515
experience in sign language may therefore arguably enhance functional integration between
AC C
509
22
ACCEPTED MANUSCRIPT different brain regions as well as cross-modal plasticity (Nishimura et al., 1999; Petitto, 2000;
517
Sadato et al., 2005). Accordingly, we found a relation between years of sign language
518
experience and functional backbone characteristics. This was found most prominently in the
519
theta band, but also in the delta and alpha bands. In both deaf and hearing, increasing
520
experience in sign language was related to a higher average backbone strength, suggesting
521
increased global efficiency and integration. The decreased closeness centrality suggests more
522
segregation and local clustering. Our reported relations with sign language experience are in
523
line with previous examinations of functional network differences between hearing signers
524
and non-signers. That study found sign-language comprehension to be related to increased
525
local efficiency, small-worldness and modularity (i.e. segregation) (Liu et al., 2017). To our
526
knowledge no network studies were performed on sign-language comprehension in deaf.
527
However, other connectivity studies in deaf showed enhanced functional connectivity
528
between brain regions which are specifically recruited for higher cognitive functions, such as
529
comprehension of sign language (Li et al., 2016; Malaia et al., 2014). Further research is
530
needed to elucidate the mechanisms of reorganization in functional brain networks in relation
531
to sign language acquisition and experience.
SC
M AN U
TE D
EP
532
RI PT
516
4.5. Differences between frequency bands
534
Our study shows that the effects of deafness, eyes-condition or sign-language experience are
535
limited to specific frequency bands, which suggests that auditory deprivation does not alter
536
functional networks as much in all frequency bands. Different brain networks and behavioral
537
functionalities are related to distinct frequency bands (Başar et al., 2000; Buschman and
538
Miller, 2007; Klimesch, 1999; Wróbel, 2000; Wrobel et al., 2007). Cognitive tasks that
539
involve working memory, such as sign language, are mostly related to alpha and theta activity
540
(Klimesch, 1999; Stam and van Straaten, 2012). This may explain the relationship we found
AC C
533
23
ACCEPTED MANUSCRIPT between sign language experience and functional backbone topology in those frequency
542
bands. Furthermore, both the alpha and beta band are related to attention, including visual
543
attention (Stam and van Straaten, 2012; Wróbel, 2000; Wrobel et al., 2007). These bands alter
544
by transition from eyes-closed (i.e. ‘interoceptive’ awareness) to eyes-open (i.e.
545
‘exteroceptive’ awareness) (Barry et al., 2009, 2007; Gómez-Ramírez et al., 2017). Overall, it
546
seems therefore that reorganization in topology of functional brain networks in deaf is
547
strongly associated with cognitive functioning as well as with attentional state. Further
548
research is nonetheless needed to investigate this into more detail.
SC
549
RI PT
541
4.6 Advantages of study design and tools
551
Our study used the unbiased MST approach (Tewarie et al., 2015) for network analyses. The
552
methodology illustrates that acquired backbone metrics are highly suitable in exploring the
553
topology and connectivity of brain networks and cross modal neuroplasticity (Engels et al.,
554
2015; Tewarie et al., 2014; van Diessen et al., 2016). Our study yielded a unique dataset of
555
subjects with intermediate to long periods of auditory deprivation combined with different
556
levels of sign language experience. Such a dataset would be difficult to acquire in Western
557
countries with well-established health-care systems and many people with hearing disabilities
558
equipped with a cochlear implant. Cochlear implants have shown to be related to functional
559
cortical reorganization (Strelnikov et al., 2010) and almost normal developing auditory
560
language processing (Hammes et al., 2002). In our study population, none of the participants
561
had a cochlear implant. Lastly, our study shows the usefulness of a portable EEG device.
562
These mobile devices are invaluable tools for use in rural or resource-limited settings. It
563
enabled us to acquire recordings from participants in a country where auditory deprivation –
564
and neurological disorders in general – are more prevalent than in Western countries, but
AC C
EP
TE D
M AN U
550
24
ACCEPTED MANUSCRIPT 565
where neuroimaging research is often impossible or burdensome due to geographic
566
conditions, poor health-care infrastructure and high levels of poverty.
567
4.7 Study limitations and future directions
569
Our study has limitations. The neural brain activity was measured at a limited amount of scalp
570
locations. EEG also lacks information from deeper brain structures. The EEG signals are
571
linear combinations of the neural generators they project to the scalp location of the
572
electrodes. The precision of functional network mapping is consequently reduced. Other
573
neuroimaging techniques, such as fMRI, might therefore be more capable to capture different
574
activation patterns of the whole brain at both cortical and sub-cortical level. The use of
575
wireless headsets for EEG recording may have increased the noise in the EEG signal, and
576
affected the backbone computations. However, recent studies have shown similar
577
performance of wireless headsets compared to standard EEG hardware (Badcock et al., 2013;
578
David Hairston et al., 2014; Schiatti et al., 2016).
TE D
M AN U
SC
RI PT
568
History taking in our study might be affected by a recall bias. Patient records (i.e.
580
dates of births) in Nigeria are not stored like they are in modern Western countries. Many
581
people living in rural Nigeria do not exactly know their birthday. This bias may have
582
increased the noise-level in the regression analysis. Apart from backbone analysis future studies may try alternative promising network
AC C
583
EP
579
584
analysis techniques such as dynamic functional connectivity modeling (Avena-Koenigsberger
585
et al., 2017; Breakspear, 2017), Bayesian exponential random graph models (Caimo and Friel,
586
2011; Sinke et al., 2016), mixed-effect models (Simpson and Laurienti, 2015) and Gibbs
587
distribution models (La Rosa et al., 2016). These techniques also enable unbiased comparison
588
of networks differing in size and density but may capture more subtle differences between
589
groups. Combining techniques might further elucidate the role of specific brain areas in
25
ACCEPTED MANUSCRIPT 590
functional network alterations in normal and sensory lacking conditions. This may ultimately
591
improve our understanding of neuroplasticity occurring after auditory and other types of
592
sensory deprivation.
593
5. Conclusion
595
In conclusion, we were able to detect functional backbone differences between eyes-closed
596
and eyes-open conditions as well as larger shifts in functional backbone characteristics in deaf
597
as compared to controls. Effects that are presumably a consequence of auditory deprivation
598
and cross-modal plasticity. Subtle differences were seen between different forms of deafness.
599
Our study demonstrated functional network backbone characteristics to be related with
600
increasing experience of sign language. Our study provide original insights into the
601
organization and reorganization of functional brain networks derived from EEG data, both in
602
deaf and healthy people. Our results further underpin the notion of brain-wide neuroplasticity
603
mechanisms and global network reorganization in the cortex of deaf, which emphasize the
604
importance to study the brain – as well as cross-modal plasticity – from a network
605
perspective. The link between the functional network backbone characteristics and acquired
606
sign language experience reflects ongoing brain adaptation in both hearing and deaf people.
607
EP
TE D
M AN U
SC
RI PT
594
Contributors
609
MRTS, JWB, EvD and WMO were involved in the study design. MRTS, JWB, FvdM, JN and
610
WMO were involved in the data acquisition. MRTS, JWB, and WMO performed the data
611
analysis. MRTS, JWB, FvdM, JN, RMD, EvD and WMO interpreted the data and wrote the
612
manuscript.
AC C
608
613 614
Funding
26
ACCEPTED MANUSCRIPT 615
This work was supported by the Netherlands Organization for Scientific Research (NWO-
616
VENI 016.168.038), and the Dutch Brain Foundation [F2014(1)-06].
617
Disclosure/Conflict of Interest
619
None of the authors has any conflict of interest to disclose in relation to this work.
RI PT
618
620
AC C
EP
TE D
M AN U
SC
621
27
ACCEPTED MANUSCRIPT Tables
622 623 Table 1. Participant demographics. Deaf pre-lingual
post-lingual
sign language
no sign
RI PT
congenital
Hearing controls
experience
language
experience
7
65
57
13 males / 29
16 males / 6
7 males
31 males / 34
36 males / 21
females
females
females
females
18 ± 5
19.6 ± 4.4
16.6 ± 4.1
17.3 ± 5.6
20 ± 7.6
(range 10-30)
(range 10-26)
(range 10-22)
(range 12-39)
(range 12-43)
First eyes
25 open /
11 open /
2 open / 5
30 open / 35
28 open / 29
condition
17 closed
11 closed
closed
closed
closed
Sign
8.9 ± 3.5
8.9 ± 3.4
8.1 ± 3.7
2.3 ± 2.6
-
(range 3-15)
(range 1-17)
Sex
Age (years)
language
624
625
AC C
(years)
(range 1-16)
EP
experience
(range 2-16)
SC
22
M AN U
42
TE D
N total
28
ACCEPTED MANUSCRIPT Table 2. Bayes factors and their interpretations, based on (Raftery, 1995). M0: the baseline model without interaction term, M1: the alternative model with interaction term. In the subsequent tables the Bayes factors for M1 are reported only. Interpretation 100
Extreme evidence for M1
30
–
100
Very strong evidence for M1
10
–
30
Strong evidence for M1
3
–
10
Moderate evidence for M1
1
–
3
Anecdotal evidence for M1
M AN U
SC
>
1
628
629
630
–
1
Anecdotal evidence for M0
1/10
–
1/3
Moderate evidence for M0
1/30
–
1/10
Strong evidence for M0
1/100
–
1/30
Very strong evidence for M0
<
1/100
TE D
1/3
Extreme evidence for M0
EP
627
No evidence
AC C
626
RI PT
Bayes factor
631
632
633
29
ACCEPTED MANUSCRIPT Table 3. Bayes factors of the interaction effect between condition (i.e. eyes-closed and eyes-open)
635
and group (i.e. deaf and controls) in relation to occipital-parietal functional connectivity
636
strength (per frequency band).
Bayes factor
Delta (0.5 – 4 Hz)
0.04
Theta (4 – 8 Hz)
0.07
Alpha (8 – 16 Hz)
>100 V
Beta (16 – 32 Hz)
2.39 I
637
I
0.04
M AN U
Gamma (32 – 64 Hz) Anecdotal evidence, V Extreme evidence for interaction
638
643 644 645 646 647 648 649
EP
642
AC C
641
TE D
639 640
SC
Frequency
RI PT
634
650 Table 4. Bayes factors of the interaction effect between condition (i.e. eyes-closed and eyesopen) and group (i.e. deaf and controls) in relation to the network backbone metrics (shown per frequency band).
30
ACCEPTED MANUSCRIPT Gamma Alpha band
Beta band
band (32-
(0.5-4 Hz)
(4-8 Hz)
(8-16 Hz)
(16-32 Hz)
64 Hz)
Strength (max)
0.09
0.05
>100 V
1.37 I
0.06
Strength (mean)
0.03
0.19
>100 V
57.82 IV
0.07
Degree (max)
0.04
0.06
0.27
BC (max)
0.02
0.06
0.08
BC (median)
0.06
0.06
4.63 II
CC (max)
0.04
0.27
26.64 III
CC (median)
0.04
0.26
93.37 IV
Leaf
0.07
0.06
Diameter
0.08
0.06
Eccentricity
0.07
0.06
Radius
0.08
0.05
Tree-hierarchy
0.06
0.05
Kappa
0.10
0.10
0.06
0.09
0.06
0.06
0.06
0.14
1.59 I
0.09
2.61 I
0.10
2.31 I
0.08
0.09
2.42 I
0.08
0.10
0.39
0.07
0.08
0.16
0.06
0.06
0.73
0.14
0.11
0.62
0.06
0.12
SC
MST-metric
RI PT
band
TE D
Delta band
M AN U
Theta
BC = betweenness centrality, CC = closeness centrality, I Anecdotal evidence, II Moderate
EP
evidence, III Strong evidence, IV Very strong evidence V Extreme evidence for interaction effect
651 652
AC C
between group and condition.
31
ACCEPTED MANUSCRIPT Table 5. Bayes factors of the relationship between the network backbone metrics and sign language experience in all participants. Gamma Theta band
Alpha band
Beta band
band (32-64
(0.5-4 Hz)
(4-8 Hz)
(8-16 Hz)
(16-32 Hz)
Hz)
Strength (max)
20.92 III
0.73
0.07
Strength (mean)
>100 V
>100 V
66.84 IV
0.22
0.06
0.37
3.94 II
0.05
0.06
BC (median)
0.07
0.05
0.09
CC (max)
0.91
2.37 I
11.80 III
Leaf
Degree (max)
0.04
0.05
0.09
0.09
0.08
0.04
SC
MST-metric
RI PT
Delta band
0.05
0.04
0.05
62.60 IV
0.05
0.18
52.06 IV
46.38 IV
0.04
0.25
0.04
0.04
0.21,
0.05
0.05
Diameter
0.04
0.04
0.21
0.04
0.05
Eccentricity
0.84
0.06
0.13
0.05
0.05
Radius
2.40 I
0.05
0.09
0.05
0.05
Tree-hierarchy
0.04
0.04
0.11
0.04
0.04
Kappa
0.17
0.05
0.86
0.08
0.04
TE D
EP
CC (median)
M AN U
0.06
BC (max)
BC = betweenness centrality, CC = closeness centrality, I Anecdotal evidence, II Moderate
AC C
evidence, III Strong evidence, IV Very strong evidence, V Extreme evidence for relation of MST metric with sign language experience.
653 654 655 656 657
32
ACCEPTED MANUSCRIPT Table 6. Bayes factors of the relationship between the network backbone metric and ASLexperience in deaf people only. MST-metric
Delta band
Theta band
Alpha band
Beta band
Gamma
(0.5-4 Hz)
(4-8 Hz)
(8-16 Hz)
(16-32 Hz)
band (32-64
1.56 I
3.37 II
0.05
Strength (mean)
0.18
43.35 IV
0.71
Degree (max)
8.14 II
0.22
0.51
BC (max)
1.97 I
0.06
0.07
BC (median)
0.61
0.45
0.09
CC (max)
0.04
0.22
CC (median)
0.07
Leaf
0.12
0.10
0.11
0.09
0.10
0.16
SC
Strength (max)
RI PT
Hz)
0.07
0.06
0.07
2.08 I
0.06
0.07
0.99
1.75 I
0.07
0.07
0.36
0.46
0.15
0.10
0.08
Diameter
0.38
0.46
0.16
0.10
0.08
Eccentricity
0.69
0.10
0.10
0.06
0.05
Radius
2.44 I
0.10
0.08
0.05
0.07
Tree-hierarchy
0.04
0.38
0.11
0.09
0.06
5.00 II
0.44
0.50
0.21
0.09
TE D
EP
Kappa
M AN U
0.06
BC = betweenness centrality, CC = closeness centrality, I Anecdotal evidence, II Moderate
658 659 660
AC C
evidence, IV Very strong evidence for relation of MST metric with sign language experience.
661 662
33
ACCEPTED MANUSCRIPT References
663 664
Almeida, J., He, D., Chen, Q., Mahon, B.Z., Zhang, F., Gonçalves, Ó.F., Fang, F., Bi, Y., 2015.
665
Decoding visual location from neural patterns in the auditory cortex of the congenitally deaf.
666
Psychol. Sci. 26, 1771–1782. doi:10.1177/0956797615598970
668 669
Aspinall, P., Mavros, P., Coyne, R., Roe, J., 2013. The urban brain: analysing outdoor physical
activity with mobile EEG. Br. J. Sports Med. online, 1–6. doi:10.1136/bjsports-2012-091877 Auer, E., Bernstein, L., Sungkarat, W., Singh, M., 2010. Vibrotactile Activation of the Auditory Cortices in Deaf versus Hearing Adults. Changes 38, 319–335.
671
doi:10.1146/annurev.neuro.31.060407.125627.Brain
674
Avena-Koenigsberger, A., Misic, B., Sporns, O., 2017. Communication dynamics in complex brain
M AN U
673
SC
670
672
RI PT
667
networks. Nat. Rev. Neurosci. 19, 17–33. doi:10.1038/nrn.2017.149 Badcock, N.A., Mousikou, P., Mahajan, Y., de Lissa, P., Thie, J., McArthur, G., 2013. Validation of
675
the Emotiv EPOC ® EEG gaming system for measuring research quality auditory ERPs. PeerJ 1,
676
e38. doi:10.7717/peerj.38
Badcock, N.A., Preece, K.A., de Wit, B., Glenn, K., Fieder, N., Thie, J., McArthur, G., 2015.
678
Validation of the Emotiv EPOC EEG system for research quality auditory event-related
679
potentials in children. PeerJ 3, e907. doi:10.7717/peerj.907 Barry, R.J., Clarke, A.R., Johnstone, S.J., Brown, C.R., 2009. EEG differences in children between
EP
680
TE D
677
eyes-closed and eyes-open resting conditions. Clin. Neurophysiol. 120, 1806–1811.
682
doi:10.1016/j.clinph.2009.08.006
683
AC C
681
Barry, R.J., Clarke, A.R., Johnstone, S.J., Magee, C.A., Rushby, J.A., 2007. EEG differences between
684
eyes-closed and eyes-open resting conditions. Clin. Neurophysiol. 118, 2765–2773.
685
doi:10.1016/j.clinph.2007.07.028
686
Başar, E., Başar-Eroglu, C., Karakaş, S., Schürmann, M., 2000. Gamma, alpha, delta, and theta
687
oscillations govern cognitive processes. Int. J. Psychophysiol. 39, 241–248. doi:10.1016/S0167-
688
8760(00)00145-8
689
Bassett, D.S., Bullmore, E.T., 2009. Human brain networks in health and disease. Curr Opin Neurol.
34
ACCEPTED MANUSCRIPT
692 693 694 695 696 697 698 699
Bassett, D.S., Sporns, O., 2017. Network Neuroscience. Nat. Neurosci. 20, 353–364. doi:10.1038/nn.4502.Network Bassett, D.S., Zurn, P., Gold, J.I., 2018. On the nature and use of models in network neuroscience. Nat. Rev. Neurosci. 19, 566–578. doi:10.1038/s41583-018-0038-8
RI PT
691
22, 340–347. doi:10.1097/WCO.0b013e32832d93dd.Human
Bavelier, D., Dye, M.W.G., Hauser, P.C., 2006. Do deaf individuals see better? Trends Cogn. Sci. 10, 512–518. doi:10.1016/j.tics.2006.09.006
Bavelier, D., Neville, H.J., 2002. Cross-modal plasticity: where and how? Nat. Rev. Neurosci. 3, 443– 52. doi:10.1038/nrn848
SC
690
Bavelier, D., Tomann, A., Hutton, C., Mitchell, T., Corina, D., Liu, G., Neville, H., 2000. Visual attention to the periphery is enhanced in congenitally deaf individuals. J. Neurosci. 20, RC93--
701
RC93.
702
M AN U
700
Boersma, M., Smit, D.J.A., Boomsma, D.I., De Geus, E.J.C., Delemarre-van de Waal, H.A., Stam, C.J., 2013. Growing trees in child brains: graph theoretical analysis of electroencephalography-
704
derived minimum spanning tree in 5- and 7-year-old children reflects brain maturation. Brain
705
Connect. 3, 50–60. doi:10.1089/brain.2012.0106
706
TE D
703
Boersma, M., Smit, D.J.A., De Bie, H.M.A., Van Baal, G.C.M., Boomsma, D.I., De Geus, E.J.C., Delemarre-Van De Waal, H.A., Stam, C.J., 2011. Network analysis of resting state EEG in the
708
developing young brain: Structure comes with maturation. Hum. Brain Mapp. 32, 413–425.
709
doi:10.1002/hbm.21030
AC C
EP
707
710
Bola, Ł., Zimmermann, M., Mostowski, P., Jednoróg, K., Marchewka, A., Rutkowski, P., Szwed, M.,
711
2017. Task-specific reorganization of the auditory cortex in deaf humans. Proc. Natl. Acad. Sci.
712
114, E600–E609. doi:10.1073/pnas.1609000114
713
Bosworth, R.G., Dobkins, K.R., 2002. The effects of spatial attention on motion processing in deaf
714
signers, hearing signers, and hearing nonsigners. Brain Cogn. 49, 152–69.
715
doi:10.1006/brcg.2001.1497
716 717
Breakspear, M., 2017. Dynamic models of large-scale brain activity. Nat. Neurosci. 20, 340–352. doi:10.1038/nn.4497
35
ACCEPTED MANUSCRIPT 718
Brotherton, H., Plack, C.J., Schaette, R., Munro, K.J., 2016. Time course and frequency specificity of
719
sub-cortical plasticity in adults following acute unilateral deprivation. Hear. Res. 341, 210–219.
720
doi:10.1016/j.heares.2016.09.003
721
Brozinsky, C.J., Bavelier, D., 2004. Motion velocity thresholds in deaf signers: Changes in lateralization but not in overall sensitivity. Cogn. Brain Res. 21, 1–10.
723
doi:10.1016/j.cogbrainres.2004.05.002
724
RI PT
722
Buckley, K.A., Tobey, E.A., 2011. Cross-modal plasticity and speech perception in pre- and postlingually deaf cochlear implant users. Ear Hear. 32, 2–15.
726
doi:10.1097/AUD.0b013e3181e8534c
SC
725
Bullmore, E., Barnes, A., Bassett, D.S., Fornito, A., Kitzbichler, M., Meunier, D., Suckling, J., 2009.
728
Generic aspects of complexity in brain imaging data and other biological systems. Neuroimage
729
47, 1125–1134. doi:10.1016/j.neuroimage.2009.05.032
731 732 733 734
Bullmore, E., Sporns, O., 2012. The economy of brain network organization. Nat. Rev. Neurosci. 13, 336–349. doi:10.1038/nrn3214
Bullmore, E., Sporns, O., 2009. Complex brain networks: graph theoretical analysis of structural and
TE D
730
M AN U
727
functional systems. Nat Rev Neurosci 10, 186–198. doi:10.1038/nrn2575 Buschman, T.J., Miller, E.K., 2007. Top-Down Versus Bottom-Up Control of Attention in the Prefrontal and Posterior Parietal Cortices. Science (80-. ). 315, 1860–1862.
736
doi:10.1126/science.1138071
738 739 740 741 742 743
Caimo, A., Friel, N., 2011. Bayesian inference for exponential random graph models. Soc. Networks
AC C
737
EP
735
33, 41–55. doi:10.1016/j.socnet.2010.09.004 Campbell, J., Sharma, A., 2014. Cross-modal re-organization in adults with early stage hearing loss. PLoS One 9. doi:10.1371/journal.pone.0090594 Campbell, J., Sharma, A., 2013. Compensatory changes in cortical resource allocation in adults with hearing loss. Front. Syst. Neurosci. 7, 1–9. doi:10.3389/fnsys.2013.00071 Chen, A.C.N., Feng, W., Zhao, H., Yin, Y., Wang, P., 2008. EEG default mode network in the human
744
brain: Spectral regional field powers. Neuroimage 41, 561–574.
745
doi:10.1016/j.neuroimage.2007.12.064
36
ACCEPTED MANUSCRIPT 746
David Hairston, W., Whitaker, K.W., Ries, A.J., Vettel, J.M., Cortney Bradford, J., Kerick, S.E.,
747
McDowell, K., 2014. Usability of four commercially-oriented EEG systems. J. Neural Eng. 11,
748
046018. doi:10.1088/1741-2560/11/4/046018
749
Dewey, R.S., Hartley, D.E.H., 2015. Cortical cross-modal plasticity following deafness measured using functional near-infrared spectroscopy. Hear. Res. 325, 55–63.
751
doi:10.1016/j.heares.2015.03.007
752
RI PT
750
Ding, H., Qin, W., Liang, M., Ming, D., Wan, B., Li, Q., Yu, C., 2015. Cross-modal activation of auditory regions during visuo-spatial working memory in early deafness. Brain 138, 2750–2765.
754
doi:10.1093/brain/awv165
755
SC
753
Doucet, M.E., Bergeron, F., Lassonde, M., Ferron, P., Lepore, F., 2006. Cross-modal reorganization and speech perception in cochlear implant users. Brain 129, 3376–3383.
757
doi:10.1093/brain/awl264
759 760
Durkin, M., 2002. The epidemiology of developmental disabilities in low-income countries. Ment. Retard. Dev. Disabil. Res. Rev. 8, 206–211. doi:10.1002/mrdd.10039 Dye, M.W.G., Baril, D.E., Bavelier, D., 2007. Which aspects of visual attention are changed by
TE D
758
M AN U
756
761
deafness? The case of the Attentional Network Test. Neuropsychologia 45, 1801–1811.
762
doi:10.1016/j.neuropsychologia.2006.12.019
765
EP
764
Eleweke, C.J., Rodda, M., 2002. The challenge of enhancing inclusive education in developing countries. Int. J. Incl. Educ. 6, 113–126. doi:10.1080/13603110110067190 Engels, M.M.A., Stam, C.J., van der Flier, W.M., Scheltens, P., de Waal, H., van Straaten, E.C.W.,
AC C
763
766
2015. Declining functional connectivity and changing hub locations in Alzheimer’s disease: an
767
EEG study. BMC Neurol. 15, 145. doi:10.1186/s12883-015-0400-7
768
Finney, E.M., Dobkins, K.R., 2001. Visual contrast sensitivity in deaf versus hearing populations:
769
Exploring the perceptual consequences of auditory deprivation and experience with a visual
770
language. Cogn. Brain Res. 11, 171–183. doi:10.1016/S0926-6410(00)00082-3
771 772 773
Finney, E.M., Fine, I., Dobkins, K.R., 2001. Visual stimuli activate auditory cortex in the deaf. Nat. Neurosci. 4, 1171–1173. doi:10.1038/nn763 Fraschini, M., Demuru, M., Crobe, A., Marrosu, F., Stam, C.J., Hillebrand, A., 2016. The effect of
37
ACCEPTED MANUSCRIPT 774
epoch length on estimated EEG functional connectivity and brain network organisation. J. Neural
775
Eng. 13, 036015. doi:10.1088/1741-2560/13/3/036015
776
Gaál, Z.A., Boha, R., Stam, C.J., Molnár, M., 2010. Age-dependent features of EEG-reactivitySpectral, complexity, and network characteristics. Neurosci. Lett. 479, 79–84.
778
doi:10.1016/j.neulet.2010.05.037
RI PT
777
779
Gómez-Ramírez, J., Freedman, S., Mateos, D., Pérez-Velázquez, J.L., Valiante, T., 2017. Eyes closed
780
or Eyes open? Exploring the alpha desynchronization hypothesis in resting state functional
781
connectivity networks with intracranial EEG 1–26. doi:10.1101/118174
783
Hackett, T.A., 2012. Information flow in the auditory cortical network. Hear. Res. 271, 133–146. doi:10.1016/j.heares.2010.01.011.Information
SC
782
Hagmann, P., Sporns, O., Madan, N., Cammoun, L., Pienaar, R., Wedeen, V.J., Meuli, R., Thiran, J.-
785
P., Grant, P.E., 2010. White matter maturation reshapes structural connectivity in the late
786
developing human brain. Proc. Natl. Acad. Sci. U. S. A. 107, 19067–19072.
787
doi:10.1073/pnas.1009073107
Hammes, D.M., Novak, M.A., Rotz, L.A., Willis, M., Edmondson, D.M., Thomas, J.F., 2002. Early
TE D
788
M AN U
784
789
identification and cochlear implantation: Critical factors for spoken language development. Ann.
790
Otol. Rhinol. Laryngol. Suppl. 189, 74–78.
Hauser, P.C., Dye, M.W.G., Boutla, M., Green, C.S., Bavelier, D., 2007. Deafness and visual
EP
791
enumeration: Not all aspects of attention are modified by deafness. Brain Res. 1153, 178–187.
793
doi:10.1016/j.brainres.2007.03.065
794
AC C
792
Heuvel, M.P. Van Den, Kahn, R.S., Goñi, J., Sporns, O., 2012. Brain Communication. Proc. Natl.
795
Acad. Sci. U. S. A. 109, 11372–77. doi:10.1073/pnas.1203593109/-
796
/DCSupplemental.www.pnas.org/cgi/doi/10.1073/pnas.1203593109
797 798 799
Hidalgo, C.A., Klinger, B., Barabasi, A.-L., Hausmann, R., 2007. The product space conditions the development of nations. Science (80-. ). 317, 482–487. doi:10.1126/science.1144581 Hüfner, K., Stephan, T., Flanagin, V.L., Deutschländer, A., Stein, A., Kalla, R., Dera, T., Fesl, G.,
800
Jahn, K., Strupp, M., Brandt, T., 2009. Differential effects of eyes open or closed in darkness on
801
brain activation patterns in blind subjects. Neurosci. Lett. 466, 30–34.
38
ACCEPTED MANUSCRIPT 802
doi:10.1016/j.neulet.2009.09.010 Hüfner, K., Stephan, T., Glasauer, S., Kalla, R., Riedel, E., Deutschländer, A., Dera, T., Wiesmann,
804
M., Strupp, M., Brandt, T., 2008. Differences in saccade-evoked brain activation patterns with
805
eyes open or eyes closed in complete darkness. Exp. Brain Res. 186, 419–430.
806
doi:10.1007/s00221-007-1247-y
RI PT
803
807
Jackson, T.S., Read, N., 2010. Theory of minimum spanning trees. I. Mean-field theory and strongly
808
disordered spin-glass model. Phys. Rev. E - Stat. Nonlinear, Soft Matter Phys. 81, 1–18.
Karns, C.M., Dow, M.W., Neville, H.J., 2012. Altered cross-modal processing in the primary auditory
810
cortex of congenitally deaf adults: A visual-somatosensory fMRI study with a double-flash
811
illusion. J. Neurosci. 32, 9626–9638. doi:10.1523/JNEUROSCI.6488-11.2012 Kim, E., Kang, H., Lee, H., Lee, H.-J., Suh, M.-W., Song, J.-J., Oh, S.-H., Lee, D.S., 2014.
M AN U
812
SC
809
813
Morphological brain network assessed using graph theory and network filtration in deaf adults.
814
Hear. Res. 315, 88–98. doi:10.1016/j.heares.2014.06.007
Klimesch, W., 1999. EEG alpha and theta oscillations reflect cognitive and memory performance: a
816
review and analysis. Brain Res. Rev. 29, 169–195. doi:10.1016/S0165-0173(98)00056-3
817
TE D
815
Knyazev, G.G., Volf, N. V., Belousova, L. V., 2015. Age-related differences in electroencephalogram connectivity and network topology. Neurobiol. Aging 36, 1849–1859.
819
doi:10.1016/j.neurobiolaging.2015.02.007
821 822
Kruskal, J.B., 1956. On the shortest spanning subtree of a graph and the traveling salesman problem. Proc. Am. Math. Soc. 7, 48–50. doi:10.2307/2033241
AC C
820
EP
818
La Rosa, P.S., Brooks, T.L., Deych, E., Shands, B., Prior, F., Larson-Prior, L.J., Shannon, W.D., 2016.
823
Gibbs distribution for statistical analysis of graphical data with a sample application to fcMRI
824
brain images. Stat. Med. 35, 566–580. doi:10.1002/sim.6757
825 826 827 828 829
Levänen, S., Hamdorf, D., 2001. Feeling vibrations: Enhanced tactile sensitivity in congenitally deaf humans. Neurosci. Lett. 301, 75–77. doi:10.1016/S0304-3940(01)01597-X Li, W., Li, J., Wang, J., Zhou, P., Wang, Z., Xian, J., He, H., 2016. Functional reorganizations of brain network in prelingually deaf adolescents. Neural Plast. 2016, 1–10. doi:10.1155/2016/9849087 Li, W., Li, J., Wang, Z., Li, Y., Liu, Z., Yan, F., Xian, J., He, H., 2015. Grey matter connectivity
39
ACCEPTED MANUSCRIPT 830
within and between auditory, language and visual systems in prelingually deaf adolescents.
831
Restor. Neurol. Neurosci. 33, 279–290. doi:10.3233/RNN-140437
832
Li, Y., Booth, J.R., Peng, D., Zang, Y., Li, J., Yan, C., Ding, G., 2013. Altered intra- and interregional synchronization of superior temporal cortex in deaf people. Cereb. Cortex 23, 1988–
834
1996. doi:10.1093/cercor/bhs185
835
RI PT
833
Liu, L., Yan, X., Liu, J., Xia, M., Lu, C., Emmorey, K., Chu, M., Ding, G., 2017. Graph theoretical
836
analysis of functional network for comprehension of sign language. Brain Res. 1671, 55–66.
837
doi:10.1016/j.brainres.2017.06.031
840
SC
839
Malaia, E., Talavage, T.M., Wilbur, R.B., 2014. Functional connectivity in task-negative network of the Deaf: effects of sign language experience. PeerJ 2, e446. doi:10.7717/peerj.446 Marx, E., Deutschländer, A., Stephan, T., Dieterich, M., Wiesmann, M., Brandt, T., 2004. Eyes open
M AN U
838
841
and eyes closed as rest conditions: Impact on brain activation patterns. Neuroimage 21, 1818–
842
1824. doi:10.1016/j.neuroimage.2003.12.026
843
Marx, E., Stephan, T., Nolte, A., Deutschländer, A., Seelos, K.C., Dieterich, M., Brandt, T., 2003. Eye closure in darkness animates sensory systems. Neuroimage 19, 924–934. doi:10.1016/S1053-
845
8119(03)00150-2
TE D
844
McMahan, T., Parberry, I., Parsons, T.D., 2015. Modality specific assessment of video game player’s
847
experience using the Emotiv. Entertain. Comput. 7, 1–6. doi:10.1016/j.entcom.2015.03.001
848
Merabet, L.B., Pascual-Leone, A., 2010. Neural reorganization following sensory loss: the opportunity of change. Nat. Rev. Neurosci. 11, 44–52. doi:10.1038/nrn2758
AC C
849
EP
846
850
Meredith, M.A., Lomber, S.G., 2011. Somatosensory and visual crossmodal plasticity in the anterior
851
auditory field of early-deaf cats. Hear. Res. 280, 38–47. doi:10.1016/j.heares.2011.02.004
852
Meyer, M., Toepel, U., Keller, J., Nussbaumer, D., Zysset, S., Friederici, A.D., 2007. Neuroplasticity
853
of sign language: implications from structural and functional brain imaging. Restor. Neurol.
854
Neurosci. 25, 335–351.
855
Miraglia, F., Vecchio, F., Bramanti, P., Rossini, P.M., 2016. EEG characteristics in “eyes-open”
856
versus “eyes-closed” conditions: Small-world network architecture in healthy aging and age-
857
related brain degeneration. Clin. Neurophysiol. 127, 1261–1268.
40
ACCEPTED MANUSCRIPT 858 859 860 861
doi:10.1016/j.clinph.2015.07.040 Mulwafu, W., Kuper, H., Ensink, R.J.H., 2016. Prevalence and causes of hearing impairment in Africa. Trop. Med. Int. Heal. 21, 158–165. doi:10.1111/tmi.12640 Neville, H.J., Bavelier, D., Corina, D., Rauschecker, J., Karni, A., Lalwani, A., Braun, A., Clark, V., Jezzard, P., Turner, R., 1998. Cerebral organization for language in deaf and hearing subjects:
863
biological constraints and effects of experience. Proc. Natl. Acad. Sci. U. S. A. 95, 922–9.
864
doi:10.1073/pnas.95.3.922
867 868 869
detection task: II. Brain Res. 405, 268–283.
SC
866
Neville, H.J., Lawson, D., 1987a. Attention to central and peripheral visual space in a movement
Neville, H.J., Lawson, D., 1987b. Attention to central and peripheral visual space in a movement detection task: III. Brain Res. 405, 284–294.
M AN U
865
RI PT
862
Nishimura, H., Hashikawa, K., Doi, K., Iwaki, T., Watanabe, Y., Kusuoka, H., Nishimura, T., Kubo,
870
T., 1999. Sign language “heard” in the auditory cortex. Nature 397, 116. doi:10.1038/16376
871
Otte, W.M., Sinke, M.R.T., Van Diessen, E., 2018a. Functional brain network analysis with minimum
873 874
spanning trees. https://zenodo.org/record/2066281#.XAudsRNKhDU.
TE D
872
Otte, W.M., Sinke, M.R.T., van Diessen, E., Dijkhuizen, R.M., 2018b. Functional Connectivity and Brain Networks in Deaf and Hearing. Open Sci. Framew. Otte, W.M., van Diessen, E., Paul, S., Ramaswamy, R., Subramanyam Rallabandi, V.P., Stam, C.J.,
876
Roy, P.K., 2015. Aging alterations in whole-brain networks during adulthood mapped with the
877
minimum spanning tree indices: The interplay of density, connectivity cost and life-time
878
trajectory. Neuroimage 109, 171–189. doi:10.1016/j.neuroimage.2015.01.011
AC C
EP
875
879
Pénicaud, S., Klein, D., Zatorre, R.J., Chen, J.K., Witcher, P., Hyde, K., Mayberry, R.I., 2013.
880
Structural brain changes linked to delayed first language acquisition in congenitally deaf
881
individuals. Neuroimage 66, 42–49. doi:10.1016/j.neuroimage.2012.09.076
882
Petitto, L.A., 2000. Speech-like cerebral activity in profoundly deaf people processing signed
883
languages: implications for the neural basis of human language. Proc. Natl Acad. Sci. USA 97,
884
13961–13966.
885
Pförtner, K., 2014. Community-based inclusive education: Best practices from Nicaragua, El
41
ACCEPTED MANUSCRIPT 886
Salvador, Guatemala and Honduras. Disabil. CBR Incl. Dev. 25, 72–81.
887
doi:10.5463/DCID.v25i1.321
889 890
Pillai, J., Sperling, M.R., 2006. Interictal EEG and the diagnosis of epilepsy. Epilepsia 47, 14–22. doi:10.1111/j.1528-1167.2006.00654.x Prause, N., Siegle, G.J., Deblieck, C., Wu, A., Iacoboni, M., 2016. EEG to primary rewards: Predictive
RI PT
888
891
utility and malleability by brain stimulation. PLoS One 11. doi:10.1371/journal.pone.0165646
892
Ptito, M., Giguere, J.F., Boire, D., Frost, D.O., Casanova, C., 2001. When the auditory cortex turns
895 896 897
Raftery, A.E., 1995. Bayesian model selection in social research. Sociol. Methodol.
SC
894
visual. Prog. Brain Res. 134, 447–458.
doi:10.2307/271063
Rubinov, M., Sporns, O., 2010. Complex network measures of brain connectivity: Uses and
M AN U
893
interpretations. Neuroimage 52, 1059–1069. doi:10.1016/j.neuroimage.2009.10.003
898
Sabidussi, G., 1966. The centrality index of a graph. Psychometrika 31, 581–603.
899
Sadato, N., Okada, T., Honda, M., Matsuki, K.-I., Yoshida, M., Kashikura, K.-I., Takei, W., Sato, T., Kochiyama, T., Yonekura, Y., 2005. Cross-modal integration and plastic changes revealed by lip
901
movement, random-dot motion and sign languages in the hearing and deaf. Cereb. Cortex 15,
902
1113–1122. doi:10.1093/cercor/bhh210
Sadato, N., Yamada, H., Okada, T., Yoshida, M., Hasegawa, T., Matsuki, K.-I., Yonekura, Y., Itoh,
EP
903
TE D
900
H., 2004. Age-dependent plasticity in the superior temporal sulcus in deaf humans: a functional
905
MRI study. BMC Neurosci. 5, 1–6. doi:10.1186/1471-2202-5-56
AC C
904
906
Schiatti, L., Faes, L., Tessadori, J., Barresi, G., Mattos, L., 2016. Mutual information-based feature
907
selection for low-cost BCIs based on motor imagery. Proc. Annu. Int. Conf. IEEE Eng. Med.
908
Biol. Soc. EMBS 2016–Octob, 2772–2775. doi:10.1109/EMBC.2016.7591305
909
Scott, G.D., Karns, C.M., Dow, M.W., Stevens, C., Neville, H.J., 2014. Enhanced peripheral visual
910
processing in congenitally deaf humans is supported by multiple brain regions, including primary
911
auditory cortex. Front. Hum. Neurosci. 8, 1–9. doi:10.3389/fnhum.2014.00177
912 913
Shiell, M.M., Champoux, F., Zatorre, R.J., 2016. The right hemisphere planum temporale supports enhanced visual motion detection ability in deaf people: Evidence from cortical thickness. Neural
42
ACCEPTED MANUSCRIPT 914 915
Plast. 1–10. doi:10.1155/2016/7217630 Shiell, M.M., Champoux, F., Zatorre, R.J., 2014. Reorganization of auditory cortex in early-deaf
916
people: Functional connectivity and relationship to hearing aid use. J. Cogn. Neurosci. 27, 150–
917
163. doi:10.1162/jocn
919 920
Simpson, S.L., Laurienti, P.J., 2015. A two-part mixed-effects modeling framework for analyzing
RI PT
918
whole- brain network data. Neuroimage 113, 310–319.
Sinke, M.R.T., Dijkhuizen, R.M., Caimo, A., Stam, C.J., Otte, W.M., 2016. Bayesian exponential random graph modeling of whole-brain structural networks across lifespan. Neuroimage 135,
922
79–91. doi:10.1016/j.neuroimage.2016.04.066
SC
921
Smit, D.J.A., Boersma, M., Schnack, H.G., Micheloyannis, S., Boomsma, D.I., Hulshoff Pol, H.E.,
924
Stam, C.J., de Geus, E.J.., 2012. The brain matures with stronger functional connectivity and
925
decreased randomness of its network. PLoS One 7, e36896. doi:10.1371/journal.pone.0036896
M AN U
923
Smit, D.J.A., Geus, E.J.C. de, Boersma, M., Boomsma, D.I., Stam, C.J., 2016. Life-span development
927
of brain network integration assessed with Phase Lag Index connectivity and minimum spanning
928
tree graphs. Brain Connect. 8, 1–38. doi:10.1089/brain.2015.0359
TE D
926
Smith, K.M., Mecoli, M.D., Altaye, M., Komlos, M., Maitra, R., Eaton, K.P., Egelhoff, J.C., Holland,
930
S.K., 2011. Morphometric differences in the heschl’s gyrus of hearing impaired and normal
931
hearing infants. Cereb. Cortex 21, 991–998. doi:10.1093/cercor/bhq164
933
Stam, C.J., 2014. Modern network science of neurological disorders. Nat. Rev. Neurosci. 15, 683–695. doi:10.1038/nrn3801
AC C
932
EP
929
934
Stam, C.J., Nolte, G., Daffertshofer, A., 2007. Phase lag index: Assessment of functional connectivity
935
from multi channel EEG and MEG with diminished bias from common sources. Hum. Brain
936
Mapp. 28, 1178–1193. doi:10.1002/hbm.20346
937
Stam, C.J., Tewarie, P., Van Dellen, E., van Straaten, E.C.W., Hillebrand, a, Van Mieghem, P., 2014.
938
The trees and the forest: Characterization of complex brain networks with minimum spanning
939
trees. Int. J. Psychophysiol. 92, 129–138. doi:10.1016/j.ijpsycho.2014.04.001
940 941
Stam, C.J., van Straaten, E.C.W., 2012. The organization of physiological brain networks. Clin. Neurophysiol. 123, 1067–1087. doi:10.1016/j.clinph.2012.01.011
43
ACCEPTED MANUSCRIPT 942
Stevens, G., Flaxman, S., Brunskill, E., Mascarenhas, M., Mathers, C.D., Finucane, M., 2013. Global
943
and regional hearing impairment prevalence: An analysis of 42 studies in 29 countries. Eur. J.
944
Public Health 23, 146–152. doi:10.1093/eurpub/ckr176
945
Strelnikov, K., Rouger, J., Demonet, J.F., Lagleyre, S., Fraysse, B., Deguine, O., Barone, P., 2010. Does brain activity at rest reflect adaptive strategies? evidence from speech processing after
947
cochlear implantation. Cereb. Cortex 20, 1217–1222. doi:10.1093/cercor/bhp183
948
RI PT
946
Striem-Amit, E., Almeida, J., Belledonne, M., Chen, Q., Fang, Y., Han, Z., Caramazza, A., Bi, Y., 2016. Topographical functional connectivity patterns exist in the congenitally, prelingually deaf.
950
Sci. Rep. 6, 29375. doi:10.1038/srep29375
951
SC
949
Tan, B., Kong, X., Yang, P., Jin, Z., Li, L., 2013. The difference of brain functional connectivity between eyes-closed and eyes-open using graph theoretical analysis. Comput Math Methods Med
953
1–15. doi:10.1155/2013/976365
M AN U
952
Tewarie, P., Hillebrand, A., Schoonheim, M.M., van Dijk, B.W., Geurts, J.J.G., Barkhof, F., Polman,
955
C.H., Stam, C.J., 2014. Functional brain network analysis using minimum spanning trees in
956
Multiple Sclerosis: An MEG source-space study. Neuroimage 88, 308–318.
957
doi:10.1016/j.neuroimage.2013.10.022
958
TE D
954
Tewarie, P., van Dellen, E., Hillebrand, A., Stam, C.J., 2015. The minimum spanning tree: An unbiased method for brain network analysis. Neuroimage 104, 177–188.
960
doi:10.1016/j.neuroimage.2014.10.015 van den Heuvel, M.P., Hulshoff Pol, H.E., 2010. Exploring the brain network: A review on resting-
AC C
961
EP
959
962
state fMRI functional connectivity. Eur. Neuropsychopharmacol. 20, 519–534.
963
doi:10.1016/j.euroneuro.2010.03.008
964
van Diessen, E., Numan, T., van Dellen, E., van der Kooi, A.W., Boersma, M., Hofman, D., van
965
Lutterveld, R., van Dijk, B.W., van Straaten, E.C.W., Hillebrand, A., Stam, C.J., 2015.
966
Opportunities and methodological challenges in EEG and MEG resting state functional brain
967
network research. Clin. Neurophysiol. 126, 1468–1481. doi:10.1016/j.clinph.2014.11.018
968 969
van Diessen, E., Otte, W.M., Braun, K.P.J., Stam, C.J., Jansen, F.E., 2014. Does sleep deprivation alter functional EEG networks in children with focal epilepsy? Front. Syst. Neurosci. 8, 67.
44
ACCEPTED MANUSCRIPT 970
doi:10.3389/fnsys.2014.00067
971
van Diessen, E., Otte, W.M., Stam, C.J., Braun, K.P.J., Jansen, F.E., 2016. Electroencephalography
972
based functional networks in newly diagnosed childhood epilepsies. Clin. Neurophysiol. 127,
973
2325–2332. doi:10.1016/j.clinph.2016.03.015 van Wijk, B.C.M., Stam, C.J., Daffertshofer, A., 2010. Comparing brain networks of different size and
975
connectivity density using graph theory. PLoS One 5, e13701. doi:10.1371/journal.pone.0013701
976 977
RI PT
974
Vohr, B., 2003. Infants and children with hearing loss - part 1. Ment. Retard. Dev. Disabil. Res. Rev. 9, 218–9. doi:10.1002/mrdd.10082
Vourkas, M., Karakonstantaki, E., Simos, P.G., Tsirka, V., Antonakakis, M., Vamvoukas, M., Stam,
979
C.J., Dimitriadis, S., Micheloyannis, S., 2014. Simple and difficult mathematics in children: A
980
minimum spanning tree EEG network analysis. Neurosci. Lett. 576, 28–33.
981
doi:10.1016/j.neulet.2014.05.048
985 986 987 988 989 990
M AN U
WHO, 2014. WHO | Deafness and hearing loss. World Heal. Organ.
TE D
984
442.
doi:/entity/mediacentre/factsheets/fs300/en/index.html WHO, 2012. WHO global estimates on prevalence of hearing loss. World Heal. Organ. 1–15. doi:10.1002/2014GB005021
EP
983
Watts, D.J., Strogatz, S.H., 1998. Collective dynamics of `small-world’ networks. Nature 393, 440–
Wróbel, A., 2000. Beta activity: a carrier for visual attention Andrzej. Acta Neurobiol Exp 60, 247–60. doi:10.1016/j.procs.2015.07.351
AC C
982
SC
978
Wrobel, A., Ghazaryan, A., Bekisz, M., Bogdan, W., Kaminski, J., 2007. Two Streams ofAttention-
991
Dependent B Activity in the Striate Recipient Zone of Cat’s Lateral Posterior–Pulvinar Complex.
992
J. Neurosci. 27, 2230–2240. doi:10.1080/00268976.2013.805848
993
Xu, P., Huang, R., Wang, J., Van Dam, N.T., Xie, T., Dong, Z., Chen, C., Gu, R., Zang, Y.F., He, Y.,
994
Fan, J., Luo, Y. jia, 2014. Different topological organization of human brain functional networks
995
with eyes open versus eyes closed. Neuroimage 90, 246–255.
996
doi:10.1016/j.neuroimage.2013.12.060
997
Yoshida, H., Kanda, Y., Takahashi, H., Miyamoto, I., Chiba, K., 2011. Observation of cortical activity
45
ACCEPTED MANUSCRIPT 998
during speech stimulation in prelingually deafened adults with cochlear implantation by positron
999
emission tomography-computed tomography. Ann. Otol. Rhinol. Laryngol. 120, 499–504.
1000 1001
doi:10.1177/000348941112000802 Yu, J.-H., Sim, K.-B., 2016. Classification of color imagination using Emotiv EPOC and event-related potential in electroencephalogram. Opt. - Int. J. Light Electron Opt. 127, 9711–9718.
1003
doi:10.1016/j.ijleo.2016.07.074
1004
RI PT
1002
Zalesky, A., Fornito, A., Bullmore, E.T., 2010. Network-based statistic: Identifying differences in brain networks. Neuroimage 53, 1197–1207. doi:10.1016/j.neuroimage.2010.06.041
1006
Zhang, D., Liang, B., Wu, X., Wang, Z., Xu, P., Chang, S., Liu, B., Liu, M., Huang, R., 2015.
SC
1005
Directionality of large-scale resting-state brain networks during eyes open and eyes closed
1008
conditions. Front. Hum. Neurosci. 9, 1–10. doi:10.3389/fnhum.2015.00081
M AN U
1007
AC C
EP
TE D
1009
46
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
ACCEPTED MANUSCRIPT Figure 1. Schematic overview of the study pipeline. Resting-state electroencephalography (EEG) was acquired for five minutes with wireless headsets in deaf and controls. The sensor locations corresponding to the fourteen channels are shown in orange; the reference sensors are shown in blue. The first minute of data acquisition, required for acclimatization, was
RI PT
discarded, yielding two blocks of recordings – two minutes with eyes-closed and two minutes with eyes-open. The order of the eyes condition (‘open’ and ‘closed’) was alternated. Subsequently, functional networks were constructed from ten-second epochs from distinct
SC
frequency bands (i.e. delta, theta, alpha, beta and gamma), and functional network backbone metrics were determined for all bands. These network backbone metrics were related to years
M AN U
of sign language experience. Main effects between groups (i.e. deaf and control) and conditions (i.e. eyes-open and eyes-closed) as well as interaction effects between group and condition were assessed using model selection within a Bayesian framework.
TE D
Figure 2. Raw EEG time-series of one participant. The visualized time-series were acquired from a deaf male (22 years old). Vertical grey lines indicate the initial acclimatization period, the first two minutes (i.e. eyes-closed in this subject) and the second
EP
two minutes (i.e. eyes-open). The y-scaling is arbitrary. Labels of the fourteen channels are
AC C
shown on top of the time-series. Abbreviations of these channel labels are explained in the main text.
Figure 3. Schematic overview of functional network and minimum spanning tree (MST) construction. (A) EEG epochs were measured from 14 electrodes, represented as network nodes. The electrodes were placed across the left (L, yellow) and right (R, orange) hemisphere (top), two electrodes served as references (blue). The recorded time-series (bottom) were used to determine functional connectivity between brain regions. (B) The determined functional
ACCEPTED MANUSCRIPT connections can either be depicted as a weighted functional network graph (top) or an adjacency matrix (bottom) where columns and rows represent nodes and colored squares indicate functional connectivity between nodes. The color indicate functional connectivity strength (i.e. darker = stronger). (C) The weighted functional network can be used to
RI PT
determine the minimal spanning tree (MST), which only includes the strongest connections, forming a functional network backbone. This determined MST can also be depicted as a network graph (top) or an adjacency matrix (bottom). Subsequently, MST metrics, such as
SC
leaf fraction and diameter, can be determined. Abbreviations of channel labels are explained
M AN U
in the main text.
Figure 4. Functional connectivity strength between the occipital cortex and the parietal cortex. (A) The average functional connectivity, quantified with the phase lag index, between O1/O2 and P7/P8 (y-axis) is shown for eyes-open and eyes-closed conditions (x-axis) and all
TE D
frequency bands (top) in controls and deaf people. (B) The delta functional connectivity between eyes-open en eyes-closed is plotted as percentage change relative to the eyes-open functional connectivity values for each frequency band, based on the data shown in the left
AC C
EP
panel: ∆ in % = 100 × ( closed – open ) / open.
Figure 5. Network backbone comparisons between eyes-open and eyes-closed conditions for both deaf and controls. Functional network backbone characteristics in the alpha band (8-16Hz) and beta band (16-32 Hz) (top), are shown for deaf (yellow) and controls (blue) for both the eyes-open and eyes-closed condition (x-axis) and indicated by the following minimum spanning tree metrics (y-axis), (from top-left to bottom-right): diameter, maximum closeness centrality, median closeness centrality, leaf number, maximum strength, mean strength, eccentricity, radius and kappa. Error bars represent the 95% confidence intervals.
ACCEPTED MANUSCRIPT Figure 6. The relation between American Sign Language (ASL) and functional backbone characteristics. The sign language experience in years (x-axis) is related to functional backbone characteristics (y-axis) in the theta band (4-8 Hz) for both the eyes-open
RI PT
(yellow) and eyes-closed (blue) condition, as indicated by the following minimum spanning tree metrics (from top-left to bottom-right): mean strength, maximum degree, median betweenness centrality (BC), median closeness centrality (CC), leaf number, diameter, radius
M AN U
SC
and kappa. Shaded areas: 95% confidence intervals.
Figure 7. The relation between American Sign Language (ASL) and functional backbone characteristics in deaf subject only. The sign language experience in years (xaxis) related to functional backbone characteristics (y-axis) in the theta band (4–8 Hz) for both the eyes-open (yellow) and eyes-closed (blue) condition, as indicated by the following
TE D
minimum spanning tree metrics (from top-left to bottom-right): mean strength, maximum degree, median betweenness centrality (BC), median closeness centrality (CC), leaf number,
AC C
EP
diameter, radius and kappa. Shaded areas: 95% confidence intervals.
ACCEPTED MANUSCRIPT Highlights
EP
TE D
M AN U
SC
RI PT
Functional brain mapping in homogeneous population in rural Nigeria. Increased functional synchronization in deaf people in alpha en beta frequency. Synchronization effects are present in brain-wide network backbone structures. Topological network reorganization is associated with sign language acquisition.
AC C
• • • •
S0378-5955(18)30235-1
DOI:
https://doi.org/10.1016/j.heares.2018.12.006
Reference:
HEARES 7660
To appear in:
Hearing Research
Received Date: 31 May 2018 Revised Date:
8 December 2018
Accepted Date: 12 December 2018
Please cite this article as: Sinke, M.R.T., Buitenhuis, J.W., van der Maas, F., Nwiboko, J., Dijkhuizen, R.M., van Diessen, E., Otte, W.M., The power of language: functional brain network topology of deaf and hearing in relation to sign language experience, Hearing Research, https://doi.org/10.1016/ j.heares.2018.12.006. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
ACCEPTED MANUSCRIPT 1
The power of language: functional brain network topology of deaf and
2
hearing in relation to sign language experience
3
Michel R.T. Sinke 1 *, Jan W. Buitenhuis 1, Frank van der Maas 3,4, Job Nwiboko 4, Rick M.
5
Dijkhuizen 1, Eric van Diessen 2 § and Willem M. Otte 1,2 §
6
§ Shared last author
7
Biomedical MR Imaging and Spectroscopy Group, Center for Image Sciences, University Medical Center Utrecht and Utrecht University, Utrecht, The Netherlands
9 10
SC
1
2
M AN U
8
Department of Pediatric Neurology, Brain Center Rudolf Magnus, University Medical Center Utrecht, Utrecht, The Netherlands
11 12 13
3
Reabilitação Baseadana Comunidade (RBC) Effata, Bissorã, Oio, Guinea-Bissau 4
CBR Effata, Omorodu Iseke Ebonyi LGA, Ebonyi State, Nigeria
TE D
14 15
AC C
18
EP
16 17
RI PT
4
19
* Corresponding author
20
Michel R.T. Sinke, MSc
21
Biomedical MR Imaging and Spectroscopy Group, Center for Image Sciences, University
22
Medical Center Utrecht, Yalelaan 2, 3584 CM Utrecht, The Netherlands.
23
Email: m.r.t.sinke @umcutrecht.nl. Phone: +31 30 253 5569; Fax: +31 30 253 5561
24
ORCID ID: https://orcid.org/0000-0002-8185-9209
25 26 1
ACCEPTED MANUSCRIPT Abstract
28
Prolonged auditory sensory deprivation leads to brain reorganization. This is indicated by
29
functional enhancement in remaining sensory systems and known as cross-modal plasticity. In
30
this study we investigated differences in functional brain network topology between deaf and
31
hearing individuals. We also studied altered functional network responses between deaf and
32
hearing individuals with a recording paradigm containing an eyes-closed and eyes-open
33
condition.
34
Electroencephalography activity was recorded in a group of sign language-trained deaf (N =
35
71) and hearing people (N = 122) living in rural Africa. Functional brain networks were
36
constructed from the functional connectivity between fourteen electrodes distributed over the
37
scalp. Functional connectivity was quantified with the phase lag index based on bandpass
38
filtered epochs of brain signal. We studied the functional connectivity between the auditory,
39
somatosensory and visual cortex and performed whole-brain minimum spanning tree analysis
40
to capture network backbone characteristics.
41
Functional connectivity between different regions involved in sensory information processing
42
tended to be stronger in deaf people during the eyes-closed condition in both the alpha and
43
beta frequency band. Furthermore, we found differences in functional backbone topology
44
between deaf and hearing individuals. The backbone topology altered during transition from
45
the eyes-closed to eyes-open condition irrespective of deafness, but was more pronounced in
46
deaf individuals. The transition of backbone strength was different between individuals with
47
congenital, pre-lingual or post-lingual deafness. Functional backbone characteristics
48
correlated with the experience of sign language. Overall, our study revealed more insights in
49
functional network reorganization caused by auditory deprivation and cross-modal plasticity.
50
It further supports the idea of a brain plasticity potential in deaf and hearing people. The
AC C
EP
TE D
M AN U
SC
RI PT
27
2
ACCEPTED MANUSCRIPT 51
association between network organization and acquired sign language experience reflects the
52
ability of ongoing brain adaptation in people with hearing disabilities.
53
Keywords
55
Cross-modal plasticity; Deafness; ASL; EEG; Functional networks; Minimum spanning tree
RI PT
54
56 57
Abbreviations
59 60
ASL = American Sign Language; CBR = Community-Based Rehabilitation; EEG = Electroencephalography; MRI = Magnetic Resonance Imaging; MST = Minimum Spanning Tree
AC C
EP
TE D
M AN U
SC
58
3
ACCEPTED MANUSCRIPT 1. Introduction
62
Impairment or loss of hearing interferes with many activities in daily life, specifically limiting
63
communication with others. This could easily lead to social isolation. The prevalence of this
64
serious disability is greatest in middle- and low-income countries (Durkin, 2002; Stevens et
65
al., 2013; WHO, 2014). While in the United States about two out of every 1,000 children are
66
born with disabling hearing loss (Vohr, 2003), this number is considerably higher in Sub-
67
Saharan Africa where about two percent of the children is born with disabling hearing loss
68
(WHO, 2012). Many of these children have profound hearing loss resulting in absolute
69
deafness. Infectious diseases are a major cause of deafness in these regions (Mulwafu et al.,
70
2016).
M AN U
SC
RI PT
61
Prolonged periods of sensory deprivation often leads to extensive reorganization in the
72
brain. This reorganization is caused by compensatory and cross-modal plasticity (Bavelier and
73
Neville, 2002; Merabet and Pascual-Leone, 2010; Ptito et al., 2001). Brain reorganization
74
after auditory deprivation has been mapped by different functional neuroimaging modalities,
75
such as positron emission tomography, functional near-infrared spectroscopy, functional
76
magnetic resonance imaging (fMRI) and electroencephalography (EEG) (Buckley and Tobey,
77
2011; Dewey and Hartley, 2015; Doucet et al., 2006; Finney et al., 2001; Yoshida et al.,
78
2011). Consequently, deaf are better than hearing people at detecting visual stimuli (e.g.
79
Almeida et al., 2015; Bavelier et al., 2006, 2000; Bosworth and Dobkins, 2002; Brozinsky
80
and Bavelier, 2004; Dye et al., 2007; Finney and Dobkins, 2001; Hauser et al., 2007; Neville
81
and Lawson, 1987a, 1987b) and show increased tactile sensitivity as well (Auer et al., 2010;
82
Levänen and Hamdorf, 2001; Meredith and Lomber, 2011). Accordingly, the auditory cortex
83
of deaf is found to be responsive to non-auditory stimuli (e.g. Almeida et al., 2015; Auer et
84
al., 2010; Buckley and Tobey, 2011; Doucet et al., 2006; Finney et al., 2001; Karns et al.,
85
2012; Meredith and Lomber, 2011; Neville and Lawson, 1987b; Scott et al., 2014). For
86
example fMRI and positron emission tomography studies showed that the cortical auditory
AC C
EP
TE D
71
4
ACCEPTED MANUSCRIPT and association areas of deaf people are responsive to visual motion stimuli. These regions
88
include the planum temporale (Petitto, 2000; Sadato et al., 2005; Shiell et al., 2016) and
89
primary auditory cortices, like posterior superior temporal gyrus (Almeida et al., 2015; Ding
90
et al., 2015; Karns et al., 2012; Li et al., 2015) and Heschl’s gyrus (Karns et al., 2012; Meyer
91
et al., 2007; Scott et al., 2014; Smith et al., 2011). Cortical reorganization (Campbell and
92
Sharma, 2014, 2013) and responsive auditory cortex to visual stimuli in deaf have also been
93
described with functional brain data recorded with EEG (e.g. Buckley and Tobey, 2011;
94
Doucet et al., 2006; Neville and Lawson, 1987b, 1987a). While this reorganization occurs
95
inevitably as a result of profound deafness, cross-modal plasticity also strongly relates to the
96
acquisition and use of sign language (Meyer et al., 2007; Pénicaud et al., 2013). Furthermore,
97
the extent of cross-modal plasticity is dependent on the age of onset and the duration of
98
deafness (Brotherton et al., 2016; Li et al., 2013). Although stimuli- and task-based
99
approaches have provided valuable insights in cross-modal plasticity, they do not capture the
100
mutual dependency of different functional brain regions as well as the integrative nature of
101
the human brain (Hackett, 2012; Stam and van Straaten, 2012).
TE D
M AN U
SC
RI PT
87
The human brain forms a complex integrative network, which consists of spatially
103
distributed, but functionally connected regions that continuously interact with each other
104
(Bassett et al., 2018; Bassett and Sporns, 2017; Bullmore et al., 2009; Bullmore and Sporns,
105
2009; van den Heuvel and Hulshoff Pol, 2010). As such, functional brain connectivity and
106
reorganization can be better understood when these processes are not studied in isolation.
107
Brain network analysis explicitly takes the interdependencies between functionally connected
108
regions into account as it shifts emphasis from specific locational changes to global
109
topological alternations. The functional network topology, and potential reorganization
110
therein, can be effectively mapped with several network metrics (Bassett et al., 2018; Bassett
111
and Sporns, 2017; Bullmore and Sporns, 2012). Classical graph analysis describes the human
AC C
EP
102
5
ACCEPTED MANUSCRIPT brain as a collection of nodes (i.e. functional brain regions such as the auditory or visual
113
cortex) and edges (i.e. the functional connections between regions), and provides quantitative
114
information on the topological properties of these networks (Bullmore and Sporns, 2009;
115
Heuvel et al., 2012; Rubinov and Sporns, 2010). The healthy human brain has been
116
characterized as a complex network that effectively combines global and efficient integration
117
with segregation of functionally specialized brain regions (Bullmore and Sporns, 2012). This
118
unique topology with high integration and segregation is defined as a small-world network
119
organization (Bullmore and Sporns, 2009; Watts and Strogatz, 1998). Deviation from a small-
120
world organization has been related to many neurological and psychiatric disorders (Bassett
121
and Bullmore, 2009; Stam, 2014). Surprisingly few studies have used network analysis to
122
examine reorganization of brain networks in deaf individuals. Pre-lingual deaf adults showed
123
increased network clustering and nodal efficiency compared to controls, whereas brain
124
networks from post-lingual deaf adults did not differ from controls (Kim et al., 2014). This
125
indicates that auditory experience might affect the morphology of brain networks in deaf
126
adults. In another study increased functional connectivity was found between regions within
127
the limbic system, a system involved in sensory information processing (Li et al., 2016).
128
Functional network hubs shifted in the deaf subjects. The small-worldness did not differ
129
between pre-lingual deaf as compared to hearing controls (Li et al., 2016). Yet another study
130
found increased functional connectivity in brain networks in deaf during rest, which was also
131
shown to be related to sign language experience (Malaia et al., 2014).
SC
M AN U
TE D
EP
AC C
132
RI PT
112
Resting-state functional connectivity measurements can be performed in an eyes-open
133
or eyes-closed condition. Opening and closing the eyes are very basic attention-directing
134
behaviors. Eyes-open is related to ‘exteroceptive’ awareness, characterized by more
135
specialized overt attention and oculomotor activity, whereas eyes-closed is related to
136
‘interoceptive’ awareness, characterized by more integrative multisensory activity and
6
ACCEPTED MANUSCRIPT imagination (Xu et al., 2014). Eyes-open and eyes-closed conditions relate to different brain
138
states (Marx et al., 2004, 2003; Zhang et al., 2015) and topological organizations of functional
139
networks (Gómez-Ramírez et al., 2017; Tan et al., 2013; Xu et al., 2014). This is even so in
140
darkness (Hüfner et al., 2009, 2008). In EEG-studies functional networks showed increased
141
global efficiency and decreased clustering during the eyes-open state, specifically in the alpha
142
band, which might be due to alpha desynchronization, i.e. a reduction in the number of
143
functional connections in the eyes-open state (Gómez-Ramírez et al., 2017; Miraglia et al.,
144
2016; Tan et al., 2013).
SC
RI PT
137
Despite the usefulness of classical network analysis in capturing brain network
146
reorganization, it has some intrinsic limitations. The classical network analysis is particularly
147
limited in comparing inter-subject networks with different network densities (van Wijk et al.,
148
2010). The network density is defined as the number of connections relative to the potential
149
number of connections. Commonly used network metrics, such as the clustering coefficient –
150
used to measure network segregation – and average path length – used to measure network
151
integration – are highly affected by the number of connections within a network (Stam et al.,
152
2014; van Wijk et al., 2010). Therefore, comparing healthy and affected (or reorganized)
153
brain networks, in a situation of cross-modal plasticity or alpha desynchronization, might
154
yield biased results (Tewarie et al., 2015; van Wijk et al., 2010; Zalesky et al., 2010).
155
Solutions have been provided. A promising alternative network characterization approach not
156
limited by the network density is the assessment of the network backbone. Network
157
backbones are robustly and efficiently operationalized by the minimum spanning tree (MST)
158
(Stam et al., 2014; Tewarie et al., 2015). An increasing number of studies have shown the
159
usefulness of MSTs in capturing subtle network changes in brain development and ageing
160
(Boersma et al., 2013; Otte et al., 2015; Smit et al., 2016; Vourkas et al., 2014). MSTs have
AC C
EP
TE D
M AN U
145
7
ACCEPTED MANUSCRIPT 161
also been useful in characterizing multiple sclerosis, Alzheimer’s disease and epilepsy
162
(Engels et al., 2015; Tewarie et al., 2014; van Diessen et al., 2016, 2014). The present study therefore investigated the effects of prolonged periods of deafness
164
on the functional brain network backbone topology. To that aim we acquired resting-state
165
EEGs in deaf and hearing individuals. Data were recorded within a unique homogeneous
166
population living in a representative rural region in sub-Saharan Africa. In this region
167
deafness is a common disability and cochlear implants are not available. We hypothesized
168
stronger functional connectivity (i.e. more integration) between auditory cortex and other
169
sensory (i.e. visual and somatosensory) cortices in deaf people, due to cross-modal plasticity.
170
Hence, we also expected differences in functional network backbone topology between
171
controls and deaf. Given the expected cross-modal plasticity as well as auditory sensitivity to
172
visual stimuli, we further expected larger shifts in functional network topology between eyes-
173
open (i.e. ‘exteroceptive’ awareness) and eyes-closed (i.e. ‘interoceptive’ awareness) in deaf.
174
We also explored whether (shifts in) functional backbone topologies were different between
175
congenital, pre-lingual and post-lingual deaf. Lastly, we anticipated a relationship between
176
functional backbone characteristics and years of American Sign Language (ASL) experience.
SC
M AN U
TE D
EP
178
AC C
177
RI PT
163
8
ACCEPTED MANUSCRIPT 2. Methods
179
2.1 Study setting and ethics
181
The study pipeline as described below is schematically visualized in Figure 1. Our study was
182
conducted at two inclusive primary schools and one inclusive secondary school, located in
183
two separate rural villages in Ebonyi State, southeast Nigeria. The schools are part of a
184
Community-Based Rehabilitation (CBR) program, which implies that all students has to live
185
in surrounding villages and communities, aiming for a full integration within the community.
186
Inclusive education means that both students with and without disabilities are allowed to
187
participate in regular classes, and are supported to learn, contribute and participate in all
188
aspects of the educational program. This inclusive educational approach is a potential strategy
189
to reduce the individual as well as shared burden of disability (Eleweke and Rodda, 2002;
190
Pförtner, 2014). The number of students with and without disabilities enrolled in the schools
191
is approximately equal. Since the integration of deaf people forms one of the main focuses of
192
this CBR program, the majority of students with disability are deaf. Standard ASL forms an
193
integral part of the educational program for more than twenty years. This sign language has to
194
be learned by all teachers and students, both the hearing and the deaf. Besides, deaf students
195
also receive speech therapy. All lessons are taught in English and if a teacher does not yet
196
sufficiently master sign language there will be an interpreter who translates spoken language
197
into sign language for deaf students and vice versa. The CBR program leads to an increasing
198
amount of people within the community that speak sign language. This implies a more regular
199
use, development and integration of sign language by deaf students in their daily lives.
SC
M AN U
TE D
EP
AC C
200
RI PT
180
Our study was approved by the organizational boards (RBC/CBR Effata), the local
201
health ministry (Izzi, Local Government Area) and the federal government (Ebonyi State
202
House of Assembly, Abakaliki [7-11-2016]) in Nigeria. The study protocol was clearly
203
explained to all students in class before they were asked to participate in the EEG recordings.
204
Written informed consent was obtained from adult participants and caretakers of students
9
ACCEPTED MANUSCRIPT 205
below eighteen years. In addition, we also obtained assent from the students below eighteen
206
years.
207
2.2 Participants
209
Table 1 shows the demographic information of all participants. We included 193 participants
210
between ten and 43 years old (mean age of 18.5 (standard deviation 6.0); gender: 103 male,
211
90 female), both students and teachers. Sign language experience of participants varied
212
between zero and seventeen years at the time of recording. We selected both hearing (n=122)
213
and deaf (n=71) participants. The pre-lingual deaf participants were all capable of lip reading.
SC
RI PT
208
M AN U
214
2.3 Data acquisition
216
We used a sixteen sensor / fourteen channel EEG monitor configured to sample at 128 Hertz
217
with a 16-bit resolution (EMOTIV Inc, San Francisco, USA), which has been validated and
218
successfully applied in several studies (Aspinall et al., 2013; Badcock et al., 2015, 2013;
219
McMahan et al., 2015; Prause et al., 2016; Yu and Sim, 2016). This wireless headset can be
220
connected to a computer via Bluetooth and is an invaluable tool to collect EEG signals from
221
participants in rural or resource-limited areas, where access to a standard EEG system is often
222
impossible or burdensome. Two sensors were preserved for reference and grounding: the
223
‘common mode sense’ (CMS; located at P3) sensor was used as the active reference for
224
absolute referencing. The ‘driven right leg’ (DRL; located at P4) sensor was used for
225
feedback noise cancelation. The electrodes were located at anterofrontal (AF3, AF4, F3, F4,
226
F7, F8), frontocentral (FC5, FC6), occipital (O1, O2), parietal (P7, P8) and temporal sites (T7,
227
T8), according to the International 10–20 system. Signal quality scores are recorded for each
228
electrode with a range from one to five (no units), with five as best quality score.
AC C
EP
TE D
215
229
Participants were seated in a comfortable chair in a sound-attenuated room where the
230
Emotiv headset was placed. Participants were instructed to keep their eyes closed for the first
10
ACCEPTED MANUSCRIPT three minutes and open in the next two minutes, or vice versa. The order of the condition
232
sequence was assigned alternatingly, so that half of the participants started in the eyes-open
233
condition whereas the other half of the participants started in the eyes-closed condition. The
234
researcher kept a log on deviations from the protocol, or unusual events in the environment,
235
that may affect the experiment. Example recordings are shown in Figure 2. As an EEG
236
contains a multitude of overlapping signal waves with distinct amplitudes and frequencies we
237
separated the signal into the five most common frequency ranges. The EEG signals were
238
band-pass filtered into the delta (0.5–4 Hz), theta (4–8 Hz), alpha (8–16 Hz), beta (16–32 Hz)
239
and gamma (32–64 Hz) frequency bands. Examples are shown in Suppl. Figure 1 and 2.
SC
RI PT
231
M AN U
240
2.4 Data cleaning and window selection
242
Time segments were removed from the recordings if i) the research log indicated a deviation
243
from the protocol, ii) the EEG signal quality score was below four for any of the channels,
244
and iii) if the absolute deviation of the gyroscope signals relative to the gyroscope signal
245
median exceeded five times the standard deviation. This threshold was based on visual data
246
inspection (See example in Suppl. Figure 3). Subsequently, the cleaned and filtered time-
247
series were cut into ten-second epochs. Functional connectivity measurements as well as
248
multiple network backbone metrics has been shown to stabilize within recordings if the epoch
249
length is six seconds or longer (Fraschini et al., 2016). In addition, multiple epochs per subject
250
further increase the stability of network backbone metrics (van Diessen et al., 2015).
251
Therefore we used multiple epochs combined with this conservative ten-second length.
EP
AC C
252
TE D
241
253
2.5 Functional connectivity
254
For each epoch a functional network was constructed, which can either be visualized as a
255
connectivity matrix or a network graph (Figure 3). Recorded time-series within each epoch
256
were used to determine functional connectivity (i.e. forming the ‘edges’ in a network)
11
ACCEPTED MANUSCRIPT between different electrodes capturing neuronal signals from underlying brain areas (i.e.
258
forming the ‘nodes’ in a network). Functional connectivity was computed and quantified with
259
the phase lag index. This is a measure of the asymmetry of the distribution of instantaneous
260
phase differences between two time series and scales between zero and one (Pillai and
261
Sperling, 2006). It is relative resistant to the influence of common sources, including volume
262
conduction and active reference electrodes. An index of zero indicates no phase coupling (i.e.
263
no functional connectivity) between time series, or coupling with a phase difference centered
264
on zero ± p radians. A non-zero index indicates the presence of phase coupling (i.e. functional
265
connectivity), where higher values indicate stronger functional connections. A more
266
mathematical description of computing the phase lag index can be found elsewhere (Stam et
267
al., 2007).
M AN U
SC
RI PT
257
268
2.6 Functional connectivity strength between auditory, visual and somatosensory cortices
270
Since we expected remodeling of the auditory cortex in deaf people, reflected as enhanced
271
functional connectivity between the auditory and sensory cortices, we characterized in both
272
groups the average phase lag index between sets of electrodes. These sets were the temporal
273
electrodes T7/T8, covering the auditory cortex, the parietal electrodes P7/P8, covering the
274
somatosensory cortex, and the occipital electrodes O1/O2, covering the visual cortex, in eyes-
275
open and eyes-closed conditions.
EP
AC C
276
TE D
269
277
2.7 Minimum spanning tree analysis
278
For each functional network a minimum spanning tree was calculated from the connectivity
279
graph G by applying Kruskal’s algorithm (Kruskal, 1956) (Figure 3B). This tree captures the
280
network’s backbone and is defined as a subset of the network nodes (forming the original
281
weighted graph G) that connects all the nodes and does not contain cycles or loops (Jackson
12
ACCEPTED MANUSCRIPT 282
and Read, 2010). Mathematically, a minimum spanning tree T minimizes the sum of the costs
283
of its edges, l(T ) = ∑ l w over the set of all possible spanning trees on G (Hidalgo et al., w∈T
2007). Since we are interested in the strongest functional connections (i.e. the network
285
backbone), we first inverted the edge weights of the functional network before determining
286
the MSTs. Subsequently, several MST metrics were calculated at the nodal or network level.
287
Although some metrics are determined at the nodal level (e.g. degree or betweenness
288
centrality), they can still be used to summarize – or indicate – specific properties for the
289
backbone as a whole. For example with the ‘maximum degree’ or the ‘average strength’,
290
where higher values may indicate higher overall connectivity. The following MST metrics
291
were calculated at nodal or network level:
292
i)
Maximum node degree (nodal): every tree was summarized by taking the maximum node degree: Smax, the node with the maximum number of connections.
ii)
Leaf number (Nleaf) (network): the number of nodes of the tree with exactly one
TE D
293 294
M AN U
SC
RI PT
284
connection to any other node (with maximum degree = 1). A higher leaf number is
296
related to increased global efficiency and integration (Stam et al., 2014; Tewarie et al.,
297
2015). iii)
has a lower bound of two and an upper bound of m = N – 1. The largest possible
299
diameter will decrease with increasing leaf number (Boersma et al., 2013; Stam et al.,
300
2014; Tewarie et al., 2015).
301 302
Diameter (d) (network): the largest distance between any two nodes in a tree, which
AC C
298
EP
295
iv)
Eccentricity (network): the shortest path length between a tree node I and any other
303
node from the tree. Eccentricity decreases when nodes become more central in the
304
tree.
305 306
v)
Radius (nodal): the smallest node eccentricity in the tree. The lower the eccentricity, the more central a node in a tree.
13
ACCEPTED MANUSCRIPT 307
vi)
weights (Hagmann et al., 2010; Rubinov and Sporns, 2010).
308 309
Strength (nodal): the tree node strength is a summation of all nodal connection
vii)
Maximum betweenness centrality (BCmax): a network hub metric which relies on the identification of the number of shortest paths that pass through a node (Rubinov and
311
Sporns, 2010). The more the passages, the higher the betweenness-centrality (i.e.
312
hubness), which is defined by bc i =
1 ( n − 1)( n − 2)
n
RI PT
310
∑
j ≠ k ,k ≠ i , j ≠ i
g jk (i ) g jk
where gjk is the
shortest path between two nodes and gjk(i) is the number of node paths that actually
314
pass through i. We summarized the tree by taking the maximum betweenness
315
centrality.
317
viii)
M AN U
316
SC
313
Closeness centrality (nodal): the inverse of the sum of all distances to other nodes (Sabidussi, 1966).
318
2.8 Statistical analyses
320
All statistical analyses were performed within a Bayesian framework. Differences between
321
groups, eye conditions and interactions between groups and eye conditions were evaluated
322
with Bayes factors. Bayes factors were extracted from Bayesian model comparisons. This was
323
done for all frequency bands separately. We determined the model likelihood of a null model
324
without an interaction effect of group and condition as well as the likelihood of an alternative
325
model with an interaction effect of group and condition for functional connectivity strength.
326
This was done for all MST metrics and for their potential relationship with sign language
327
experience. Bayes factors give the ratio of model likelihoods, indicating which model is
328
supported by the data. For example, if – given the data – a null model (M0) without an effect
329
of condition (i.e. eyes-open versus eyes-closed) on functional connectivity strength has a very
330
low probability, whereas an alternative model (M1) with an effect of condition on functional
331
connectivity strength has a high probability, this would yield a high Bayes factor (e.g. 50).
AC C
EP
TE D
319
14
ACCEPTED MANUSCRIPT This may be interpreted as very strong evidence for M1 as M1 is fifty times more likely than
333
M0 in explaining the data. Table 2 gives an overview of Bayes factors and their interpretation
334
(Raftery, 1995). Since sex (Boersma et al., 2011) and age (Smit et al., 2012) influence
335
functional network topologies, we tested whether models with strong evidence were affected
336
by sex- and age.
RI PT
332
All network analyses, statistical modeling and visualizations were performed in R
338
(http://www.r-project.org/) using the packages igraph, BayesFactor, reshape2, dplyr and
339
ggplot2. Epoch data and scripts are freely available at the Open Science Framework in
340
anonymized form (Otte et al., 2018b) and the GitHub repository (Otte et al., 2018a).
SC
337
M AN U
341
AC C
EP
TE D
342
15
ACCEPTED MANUSCRIPT 3. Results
343 344
3.1 Functional connectivity strength between auditory, visual and somatosensory cortices
346
Figure 4 and Suppl. Figure 6 show the functional connectivity strength between the occipital
347
cortex and the parietal cortex for the different frequency bands. Occipital-parietal functional
348
connectivity strength was lower in the eyes-open condition compared to the eyes-closed
349
condition in the theta, alpha and beta frequency bands. After transition from eyes-closed to
350
eyes-open, differences in functional connectivity strength were most pronounced in the alpha
351
frequency band with a significant reduction of 54.9% (95% confidence interval (CI): -68.2%
352
to -41.6%) in controls and an even larger reduction of 88.0% (95% CI: -112% to -63.6% in
353
deaf (Figure 4B). A decrease in functional connectivity strength from eyes-closed to eyes-
354
open was also present in the beta frequency band, with a reduction of 27.7% (95% CI: -37.3
355
to -18.0) in controls and 36.1% (95% CI: -49.2 to -23.0) in deaf. These reductions in
356
functional connectivity strength in the alpha and beta band were supported by Bayes factors
357
of respectively >100 (i.e. labeled as ‘extreme evidence’) and 2.6 (‘moderate evidence’)
358
(Table 3).
EP
TE D
M AN U
SC
RI PT
345
Similar trends were seen for functional connectivity strength between the parietal
360
cortex and temporal cortex (Suppl. Figure 4 and Suppl. Figure 7) as well as between the
361
occipital cortex and temporal cortex (Suppl. Figure 5 and Suppl. Figure 8). However, for
362
those connections no significant differences were found between controls and deaf in
363
functional connectivity reduction from eyes-open to eyes-closed, as indicated by Bayes
364
factors of <1 (Suppl. Table 1 and Suppl. Table 2). The parietal-temporal delta-band
365
connectivity strength was significantly more reduced from eyes-closed to eyes-open in deaf,
366
while no reduction was found in controls (Suppl. Figure 5), as indicated by the high Bayes
AC C
359
16
ACCEPTED MANUSCRIPT 367
factor (Suppl. Table 2). All evidence was still present if age and sex were added to the
368
models.
369
3.2 Functional backbone differences between eyes-open and eyes-closed conditions
371
Transition from eyes-closed to eyes-open initiated visible changes in functional network
372
topology, as indicated by changes in backbone metrics. The most notable effects were found
373
in the alpha and beta bands (Table 4), which are shown in Figure 5. Overall the backbone
374
leaf number, the average and maximum strength as well as the kappa were lower in the eyes-
375
open condition as compared to the eyes-closed condition. In contrast, the diameter,
376
eccentricity, radius as well as the median and maximum closeness centrality were higher in
377
the eyes-open condition compared to the eyes-closed condition. Interestingly, in contrast to
378
deaf, controls did not show a transition effect on some network metrics in the alpha band (i.e.
379
leaf number, eccentricity, radius and diameter).
M AN U
SC
RI PT
370
TE D
380
3.3 Larger functional backbone modifications in deaf
382
Some functional backbone characteristics were different between deaf and hearing controls
383
(Figure 5). In both the alpha and beta band, functional backbone strength was stronger in the
384
eyes-closed condition and weaker in the eyes-open condition in deaf as compared to hearing
385
controls. In the alpha band the leaf number was lower in deaf, whereas the diameter and
386
radius were larger in the eyes-open condition.
AC C
387
EP
381
Several functional backbone characteristics showed larger shifts in deaf than controls
388
when comparing eyes-closed to eyes-open (Figure 5). For both the alpha and beta band,
389
evidence was found for a larger decrease in the average and maximum backbone connectivity
390
strength in deaf (as indicated by the Bayes factors in Table 4). Furthermore, for the alpha
391
band there was moderate to strong evidence for a larger increase in both betweenness and
392
closeness centrality in deaf. Anecdotal evidence was found for larger shifts in leaf number and 17
ACCEPTED MANUSCRIPT 393
diameter – in the alpha band – as well as in closeness centrality – in the beta band – in deaf
394
(Table 4). All evidence was still present if age and sex were added to the models. We found that the decrease in average connectivity strength (Suppl. Figure 9) and
396
increase in closeness centrality (Suppl. Figure 10) from eyes-closed to eyes-open was
397
different across the congenital, pre-lingual and post-lingual deafness types. This is supported
398
by the Bayes factors in Suppl. Table 3.
RI PT
395
399
3.4 Relation between backbone characteristics and American Sign Language
401
We investigated the relationship between sign language experience and functional backbone
402
characteristics. Initially, with no distinction made between deaf and controls. Sign language
403
experience was related to altered backbone characteristics in the delta, theta and alpha band.
404
The most pronounced effects were found in the theta band (Table 5), which are visualized in
405
Figures 6 and 7. More specifically, for the theta band an increase in ASL experience was
406
related to a higher average backbone connectivity strength and a lower closeness centrality for
407
the eyes-closed condition (Figure 6). Furthermore, strong evidence was found for a positive
408
relationship between sign language experience and higher backbone connectivity strength for
409
the delta band (Table 5). Figure 7 shows the relationship between sign language experience
410
and functional backbone characteristics for deaf subjects only. Again most pronounced effects
411
were found for the theta band (Table 6) with a positive relation between sign language
412
experience and higher average backbone connectivity strength. All associations were still
413
present if age and sex were added to the models.
AC C
EP
TE D
M AN U
SC
400
414
18
ACCEPTED MANUSCRIPT 4. Discussion
416
The present study used resting-state EEG to map functional network backbone differences
417
between deaf and hearing people. We showed that transition from eyes-closed to eyes-open
418
was associated with changes in functional connectivity strength between occipital and parietal
419
lobes as well as changes in functional backbone topology. In both deaf and hearing, these
420
changes occurred especially in the alpha and beta frequency bands. Moreover, the difference
421
in functional connectivity strength as well as in functional backbone characteristics between
422
eyes-closed to eyes-open tended to be larger in deaf as compared to hearing controls.
SC
RI PT
415
423
4.1 Functional cortical remapping
425
We examined functional connectivity strength between sensory regions involved in auditory,
426
visual and somatosensory information processing. With some evidence we can state that
427
functional connectivity appears stronger in deaf, specifically in the alpha and beta bands. This
428
might indicate increased integration between the different sensory cortical regions in deaf.
429
Our results are in agreement with previous fMRI studies that also found increased audiovisual
430
connectivity in deaf (Bola et al., 2017; Li et al., 2016, 2013; Shiell et al., 2014). With the
431
shifts from an eyes-closed to an eyes-open condition we found a reduction in functional
432
connectivity strength. This is expected given the shift from more ‘interoceptive’ awareness,
433
characterized by integrative multisensory activity, towards more ‘exteroceptive’ awareness
434
focused on attention and oculomotor activity (Marx et al., 2004, 2003; Xu et al., 2014).
435
Specifically for the alpha band, this change may be related to ‘alpha desynchronization’
436
during the eyes-open state (Barry et al., 2009; Gómez-Ramírez et al., 2017; Xu et al., 2014).
437
Moreover, we found this reduction in functional connectivity strength from eyes-closed to
438
eyes-open to be larger in deaf than hearing controls. This may be explained by long-term
439
auditory deprivation and related cross-modal plasticity mechanisms (Bavelier and Neville,
440
2002; Merabet and Pascual-Leone, 2010). It may be explained more specifically by the
AC C
EP
TE D
M AN U
424
19
ACCEPTED MANUSCRIPT enhanced sensitivity of the auditory cortex to non-auditory stimuli (e.g. Almeida et al., 2015;
442
Auer et al., 2010; Brozinsky and Bavelier, 2004; Dye et al., 2007; Finney and Dobkins, 2001;
443
Levänen and Hamdorf, 2001; Neville and Lawson, 1987a, 1987b) and increased audiovisual
444
connectivity in deaf (Bola et al., 2017; Li et al., 2016, 2013; Shiell et al., 2014) together with
445
the reduced integrative multi-sensory activity in the eyes-open state (Gómez-Ramírez et al.,
446
2017; Xu et al., 2014).
RI PT
441
447
4.2 Functional backbone differences between eyes-open and eyes-closed conditions
449
We found functional backbone differences between the eyes-open and eyes-closed conditions
450
in both deaf and hearing subjects, mainly in the alpha and beta bands. During the eyes-open
451
condition the functional network backbone showed an increased diameter and eccentricity
452
combined with a decreased leaf number and connectivity strength, although hearing controls
453
did not show differences in diameter and leaf number in the alpha band. These network
454
metrics indicate that the functional backbone topology in the eyes-open condition was more
455
chainlike (i.e. less integrated and with reduced global efficiency), whereas during the eyes-
456
closed condition the topology was more star-like (i.e. more functional integration and
457
increased global efficiency) (Stam et al., 2014). These results also nicely fit with the reduced
458
connectivity and integration in the ‘exteroceptive’ state, as well as the alpha
459
desynchronization, during the eyes-open condition (Barry et al., 2007; Chen et al., 2008;
460
Gómez-Ramírez et al., 2017; Marx et al., 2004; Xu et al., 2014). Other studies reported
461
increased global efficiency and decreased clustering of functional networks in the alpha band
462
(Miraglia et al., 2016; Tan et al., 2013) and beta-band (Gaál et al., 2010; Knyazev et al., 2015)
463
in the eyes-open condition as compared to the eyes-closed condition, which seem to partly
464
contradict our findings. This could potentially be explained by the distinct age groups. One
465
study showed similar effects as our study in participants between 18-35 years, but the
AC C
EP
TE D
M AN U
SC
448
20
ACCEPTED MANUSCRIPT opposite effects in participants between 51-80 years (Knyazev et al., 2015). This pattern is in
467
line with (Miraglia et al., 2016). However, two other studies also showed contrasting results
468
in younger participants (Gaál et al., 2010; Tan et al., 2013). These discrepancies might be due
469
to methodological differences as two studies (i.e. Miraglia et al., 2016; Tan et al., 2013) used
470
a large number of EEG electrodes or regions of interest (i.e. respectively 128 and 84)
471
compared to our study. Increased network sizes, with potential higher densities, may arguably
472
affect alteration of topological network characteristics if analyzed with classical network
473
metrics (van Wijk et al., 2010). Network size also affects backbone metrics (Tewarie et al.,
474
2014), which may explain the discrepancy between our study and Gaál et al. (2010).
SC
RI PT
466
M AN U
475
4.3 Larger functional backbone modifications in deaf
477
We found some backbone differences between deaf and hearing subjects. Functional
478
backbones in deaf showed higher strength in the eyes-closed, while lower strength during the
479
eyes-open condition, in both the alpha and beta band. Also in the eyes-open condition, the leaf
480
number was lower, while the radius and diameter were larger, which means that the functional
481
network topology in this state is different between deaf and hearing. More specifically, the
482
functional backbone of deaf shows less integration (i.e. reduced global efficiency) and
483
increased clustering. Previous network studies in deaf also showed increased clustering and
484
local efficiency, although the small-worldness was preserved (Kim et al., 2014; Li et al.,
485
2016).
EP
AC C
486
TE D
476
Furthermore, we found larger functional backbone shifts in these functional backbone
487
characteristics, from the eyes-open to the eyes-closed condition in deaf people. The patterns in
488
functional backbone strength correspond with changes we found in functional connectivity
489
strength between the different sensory cortical regions. They additionally indicate large-scale
490
functional connectivity changes, i.e. beyond the sensory cortices, since the functional
21
ACCEPTED MANUSCRIPT backbone topology is more altered in deaf when going from eyes-closed to eyes-open.
492
Accordingly, these findings may also be explained by transition from ‘exteroceptive’
493
awareness to ‘interoceptive’ awareness (Marx et al., 2004, 2003; Xu et al., 2014) together
494
with alterations in inter-regional connectivity – partly due to alpha desynchronization (Barry
495
et al., 2009; Gómez-Ramírez et al., 2017; Xu et al., 2014) – combined with auditory
496
deprivation and cross-modal plasticity (Bavelier and Neville, 2002; Merabet and Pascual-
497
Leone, 2010).
RI PT
491
We also found that the functional backbone connectivity strength differed between the
499
different forms of deafness. Functional backbone connectivity might depend on whether
500
people are born deaf or acquired deafness later in life. Both congenital and pre-lingual
501
deafness showed similar patterns as hearing controls, whereas post-lingual deaf showed the
502
largest deviation, which is in agreement with previous findings. It has been shown that cross-
503
modal plasticity is dependent on the age of onset and duration of deafness (Brotherton et al.,
504
2016; Li et al., 2016, 2013; Sadato et al., 2004). However, given the limited number of post-
505
lingual deaf in our study as well as the complex relationship of cross-modal plasticity with
506
other factors, such as age of onset, duration of deafness and sign-language experience, it is
507
impossible to be conclusive.
M AN U
TE D
EP
508
SC
498
4.4 Relation between backbone characteristics and sign language experience
510
Sign language comprehension have shown to be related to activation of brain regions which
511
are normally considered to be involved in unimodal (e.g. speech or sound) auditory
512
processing, such as frontal and temporal regions (Li et al., 2016; Liu et al., 2017; Malaia et
513
al., 2014; Meyer et al., 2007; Neville et al., 1998; Nishimura et al., 1999; Petitto, 2000; Sadato
514
et al., 2005). Language comprehension requires higher cognitive functions. Increased use and
515
experience in sign language may therefore arguably enhance functional integration between
AC C
509
22
ACCEPTED MANUSCRIPT different brain regions as well as cross-modal plasticity (Nishimura et al., 1999; Petitto, 2000;
517
Sadato et al., 2005). Accordingly, we found a relation between years of sign language
518
experience and functional backbone characteristics. This was found most prominently in the
519
theta band, but also in the delta and alpha bands. In both deaf and hearing, increasing
520
experience in sign language was related to a higher average backbone strength, suggesting
521
increased global efficiency and integration. The decreased closeness centrality suggests more
522
segregation and local clustering. Our reported relations with sign language experience are in
523
line with previous examinations of functional network differences between hearing signers
524
and non-signers. That study found sign-language comprehension to be related to increased
525
local efficiency, small-worldness and modularity (i.e. segregation) (Liu et al., 2017). To our
526
knowledge no network studies were performed on sign-language comprehension in deaf.
527
However, other connectivity studies in deaf showed enhanced functional connectivity
528
between brain regions which are specifically recruited for higher cognitive functions, such as
529
comprehension of sign language (Li et al., 2016; Malaia et al., 2014). Further research is
530
needed to elucidate the mechanisms of reorganization in functional brain networks in relation
531
to sign language acquisition and experience.
SC
M AN U
TE D
EP
532
RI PT
516
4.5. Differences between frequency bands
534
Our study shows that the effects of deafness, eyes-condition or sign-language experience are
535
limited to specific frequency bands, which suggests that auditory deprivation does not alter
536
functional networks as much in all frequency bands. Different brain networks and behavioral
537
functionalities are related to distinct frequency bands (Başar et al., 2000; Buschman and
538
Miller, 2007; Klimesch, 1999; Wróbel, 2000; Wrobel et al., 2007). Cognitive tasks that
539
involve working memory, such as sign language, are mostly related to alpha and theta activity
540
(Klimesch, 1999; Stam and van Straaten, 2012). This may explain the relationship we found
AC C
533
23
ACCEPTED MANUSCRIPT between sign language experience and functional backbone topology in those frequency
542
bands. Furthermore, both the alpha and beta band are related to attention, including visual
543
attention (Stam and van Straaten, 2012; Wróbel, 2000; Wrobel et al., 2007). These bands alter
544
by transition from eyes-closed (i.e. ‘interoceptive’ awareness) to eyes-open (i.e.
545
‘exteroceptive’ awareness) (Barry et al., 2009, 2007; Gómez-Ramírez et al., 2017). Overall, it
546
seems therefore that reorganization in topology of functional brain networks in deaf is
547
strongly associated with cognitive functioning as well as with attentional state. Further
548
research is nonetheless needed to investigate this into more detail.
SC
549
RI PT
541
4.6 Advantages of study design and tools
551
Our study used the unbiased MST approach (Tewarie et al., 2015) for network analyses. The
552
methodology illustrates that acquired backbone metrics are highly suitable in exploring the
553
topology and connectivity of brain networks and cross modal neuroplasticity (Engels et al.,
554
2015; Tewarie et al., 2014; van Diessen et al., 2016). Our study yielded a unique dataset of
555
subjects with intermediate to long periods of auditory deprivation combined with different
556
levels of sign language experience. Such a dataset would be difficult to acquire in Western
557
countries with well-established health-care systems and many people with hearing disabilities
558
equipped with a cochlear implant. Cochlear implants have shown to be related to functional
559
cortical reorganization (Strelnikov et al., 2010) and almost normal developing auditory
560
language processing (Hammes et al., 2002). In our study population, none of the participants
561
had a cochlear implant. Lastly, our study shows the usefulness of a portable EEG device.
562
These mobile devices are invaluable tools for use in rural or resource-limited settings. It
563
enabled us to acquire recordings from participants in a country where auditory deprivation –
564
and neurological disorders in general – are more prevalent than in Western countries, but
AC C
EP
TE D
M AN U
550
24
ACCEPTED MANUSCRIPT 565
where neuroimaging research is often impossible or burdensome due to geographic
566
conditions, poor health-care infrastructure and high levels of poverty.
567
4.7 Study limitations and future directions
569
Our study has limitations. The neural brain activity was measured at a limited amount of scalp
570
locations. EEG also lacks information from deeper brain structures. The EEG signals are
571
linear combinations of the neural generators they project to the scalp location of the
572
electrodes. The precision of functional network mapping is consequently reduced. Other
573
neuroimaging techniques, such as fMRI, might therefore be more capable to capture different
574
activation patterns of the whole brain at both cortical and sub-cortical level. The use of
575
wireless headsets for EEG recording may have increased the noise in the EEG signal, and
576
affected the backbone computations. However, recent studies have shown similar
577
performance of wireless headsets compared to standard EEG hardware (Badcock et al., 2013;
578
David Hairston et al., 2014; Schiatti et al., 2016).
TE D
M AN U
SC
RI PT
568
History taking in our study might be affected by a recall bias. Patient records (i.e.
580
dates of births) in Nigeria are not stored like they are in modern Western countries. Many
581
people living in rural Nigeria do not exactly know their birthday. This bias may have
582
increased the noise-level in the regression analysis. Apart from backbone analysis future studies may try alternative promising network
AC C
583
EP
579
584
analysis techniques such as dynamic functional connectivity modeling (Avena-Koenigsberger
585
et al., 2017; Breakspear, 2017), Bayesian exponential random graph models (Caimo and Friel,
586
2011; Sinke et al., 2016), mixed-effect models (Simpson and Laurienti, 2015) and Gibbs
587
distribution models (La Rosa et al., 2016). These techniques also enable unbiased comparison
588
of networks differing in size and density but may capture more subtle differences between
589
groups. Combining techniques might further elucidate the role of specific brain areas in
25
ACCEPTED MANUSCRIPT 590
functional network alterations in normal and sensory lacking conditions. This may ultimately
591
improve our understanding of neuroplasticity occurring after auditory and other types of
592
sensory deprivation.
593
5. Conclusion
595
In conclusion, we were able to detect functional backbone differences between eyes-closed
596
and eyes-open conditions as well as larger shifts in functional backbone characteristics in deaf
597
as compared to controls. Effects that are presumably a consequence of auditory deprivation
598
and cross-modal plasticity. Subtle differences were seen between different forms of deafness.
599
Our study demonstrated functional network backbone characteristics to be related with
600
increasing experience of sign language. Our study provide original insights into the
601
organization and reorganization of functional brain networks derived from EEG data, both in
602
deaf and healthy people. Our results further underpin the notion of brain-wide neuroplasticity
603
mechanisms and global network reorganization in the cortex of deaf, which emphasize the
604
importance to study the brain – as well as cross-modal plasticity – from a network
605
perspective. The link between the functional network backbone characteristics and acquired
606
sign language experience reflects ongoing brain adaptation in both hearing and deaf people.
607
EP
TE D
M AN U
SC
RI PT
594
Contributors
609
MRTS, JWB, EvD and WMO were involved in the study design. MRTS, JWB, FvdM, JN and
610
WMO were involved in the data acquisition. MRTS, JWB, and WMO performed the data
611
analysis. MRTS, JWB, FvdM, JN, RMD, EvD and WMO interpreted the data and wrote the
612
manuscript.
AC C
608
613 614
Funding
26
ACCEPTED MANUSCRIPT 615
This work was supported by the Netherlands Organization for Scientific Research (NWO-
616
VENI 016.168.038), and the Dutch Brain Foundation [F2014(1)-06].
617
Disclosure/Conflict of Interest
619
None of the authors has any conflict of interest to disclose in relation to this work.
RI PT
618
620
AC C
EP
TE D
M AN U
SC
621
27
ACCEPTED MANUSCRIPT Tables
622 623 Table 1. Participant demographics. Deaf pre-lingual
post-lingual
sign language
no sign
RI PT
congenital
Hearing controls
experience
language
experience
7
65
57
13 males / 29
16 males / 6
7 males
31 males / 34
36 males / 21
females
females
females
females
18 ± 5
19.6 ± 4.4
16.6 ± 4.1
17.3 ± 5.6
20 ± 7.6
(range 10-30)
(range 10-26)
(range 10-22)
(range 12-39)
(range 12-43)
First eyes
25 open /
11 open /
2 open / 5
30 open / 35
28 open / 29
condition
17 closed
11 closed
closed
closed
closed
Sign
8.9 ± 3.5
8.9 ± 3.4
8.1 ± 3.7
2.3 ± 2.6
-
(range 3-15)
(range 1-17)
Sex
Age (years)
language
624
625
AC C
(years)
(range 1-16)
EP
experience
(range 2-16)
SC
22
M AN U
42
TE D
N total
28
ACCEPTED MANUSCRIPT Table 2. Bayes factors and their interpretations, based on (Raftery, 1995). M0: the baseline model without interaction term, M1: the alternative model with interaction term. In the subsequent tables the Bayes factors for M1 are reported only. Interpretation 100
Extreme evidence for M1
30
–
100
Very strong evidence for M1
10
–
30
Strong evidence for M1
3
–
10
Moderate evidence for M1
1
–
3
Anecdotal evidence for M1
M AN U
SC
>
1
628
629
630
–
1
Anecdotal evidence for M0
1/10
–
1/3
Moderate evidence for M0
1/30
–
1/10
Strong evidence for M0
1/100
–
1/30
Very strong evidence for M0
<
1/100
TE D
1/3
Extreme evidence for M0
EP
627
No evidence
AC C
626
RI PT
Bayes factor
631
632
633
29
ACCEPTED MANUSCRIPT Table 3. Bayes factors of the interaction effect between condition (i.e. eyes-closed and eyes-open)
635
and group (i.e. deaf and controls) in relation to occipital-parietal functional connectivity
636
strength (per frequency band).
Bayes factor
Delta (0.5 – 4 Hz)
0.04
Theta (4 – 8 Hz)
0.07
Alpha (8 – 16 Hz)
>100 V
Beta (16 – 32 Hz)
2.39 I
637
I
0.04
M AN U
Gamma (32 – 64 Hz) Anecdotal evidence, V Extreme evidence for interaction
638
643 644 645 646 647 648 649
EP
642
AC C
641
TE D
639 640
SC
Frequency
RI PT
634
650 Table 4. Bayes factors of the interaction effect between condition (i.e. eyes-closed and eyesopen) and group (i.e. deaf and controls) in relation to the network backbone metrics (shown per frequency band).
30
ACCEPTED MANUSCRIPT Gamma Alpha band
Beta band
band (32-
(0.5-4 Hz)
(4-8 Hz)
(8-16 Hz)
(16-32 Hz)
64 Hz)
Strength (max)
0.09
0.05
>100 V
1.37 I
0.06
Strength (mean)
0.03
0.19
>100 V
57.82 IV
0.07
Degree (max)
0.04
0.06
0.27
BC (max)
0.02
0.06
0.08
BC (median)
0.06
0.06
4.63 II
CC (max)
0.04
0.27
26.64 III
CC (median)
0.04
0.26
93.37 IV
Leaf
0.07
0.06
Diameter
0.08
0.06
Eccentricity
0.07
0.06
Radius
0.08
0.05
Tree-hierarchy
0.06
0.05
Kappa
0.10
0.10
0.06
0.09
0.06
0.06
0.06
0.14
1.59 I
0.09
2.61 I
0.10
2.31 I
0.08
0.09
2.42 I
0.08
0.10
0.39
0.07
0.08
0.16
0.06
0.06
0.73
0.14
0.11
0.62
0.06
0.12
SC
MST-metric
RI PT
band
TE D
Delta band
M AN U
Theta
BC = betweenness centrality, CC = closeness centrality, I Anecdotal evidence, II Moderate
EP
evidence, III Strong evidence, IV Very strong evidence V Extreme evidence for interaction effect
651 652
AC C
between group and condition.
31
ACCEPTED MANUSCRIPT Table 5. Bayes factors of the relationship between the network backbone metrics and sign language experience in all participants. Gamma Theta band
Alpha band
Beta band
band (32-64
(0.5-4 Hz)
(4-8 Hz)
(8-16 Hz)
(16-32 Hz)
Hz)
Strength (max)
20.92 III
0.73
0.07
Strength (mean)
>100 V
>100 V
66.84 IV
0.22
0.06
0.37
3.94 II
0.05
0.06
BC (median)
0.07
0.05
0.09
CC (max)
0.91
2.37 I
11.80 III
Leaf
Degree (max)
0.04
0.05
0.09
0.09
0.08
0.04
SC
MST-metric
RI PT
Delta band
0.05
0.04
0.05
62.60 IV
0.05
0.18
52.06 IV
46.38 IV
0.04
0.25
0.04
0.04
0.21,
0.05
0.05
Diameter
0.04
0.04
0.21
0.04
0.05
Eccentricity
0.84
0.06
0.13
0.05
0.05
Radius
2.40 I
0.05
0.09
0.05
0.05
Tree-hierarchy
0.04
0.04
0.11
0.04
0.04
Kappa
0.17
0.05
0.86
0.08
0.04
TE D
EP
CC (median)
M AN U
0.06
BC (max)
BC = betweenness centrality, CC = closeness centrality, I Anecdotal evidence, II Moderate
AC C
evidence, III Strong evidence, IV Very strong evidence, V Extreme evidence for relation of MST metric with sign language experience.
653 654 655 656 657
32
ACCEPTED MANUSCRIPT Table 6. Bayes factors of the relationship between the network backbone metric and ASLexperience in deaf people only. MST-metric
Delta band
Theta band
Alpha band
Beta band
Gamma
(0.5-4 Hz)
(4-8 Hz)
(8-16 Hz)
(16-32 Hz)
band (32-64
1.56 I
3.37 II
0.05
Strength (mean)
0.18
43.35 IV
0.71
Degree (max)
8.14 II
0.22
0.51
BC (max)
1.97 I
0.06
0.07
BC (median)
0.61
0.45
0.09
CC (max)
0.04
0.22
CC (median)
0.07
Leaf
0.12
0.10
0.11
0.09
0.10
0.16
SC
Strength (max)
RI PT
Hz)
0.07
0.06
0.07
2.08 I
0.06
0.07
0.99
1.75 I
0.07
0.07
0.36
0.46
0.15
0.10
0.08
Diameter
0.38
0.46
0.16
0.10
0.08
Eccentricity
0.69
0.10
0.10
0.06
0.05
Radius
2.44 I
0.10
0.08
0.05
0.07
Tree-hierarchy
0.04
0.38
0.11
0.09
0.06
5.00 II
0.44
0.50
0.21
0.09
TE D
EP
Kappa
M AN U
0.06
BC = betweenness centrality, CC = closeness centrality, I Anecdotal evidence, II Moderate
658 659 660
AC C
evidence, IV Very strong evidence for relation of MST metric with sign language experience.
661 662
33
ACCEPTED MANUSCRIPT References
663 664
Almeida, J., He, D., Chen, Q., Mahon, B.Z., Zhang, F., Gonçalves, Ó.F., Fang, F., Bi, Y., 2015.
665
Decoding visual location from neural patterns in the auditory cortex of the congenitally deaf.
666
Psychol. Sci. 26, 1771–1782. doi:10.1177/0956797615598970
668 669
Aspinall, P., Mavros, P., Coyne, R., Roe, J., 2013. The urban brain: analysing outdoor physical
activity with mobile EEG. Br. J. Sports Med. online, 1–6. doi:10.1136/bjsports-2012-091877 Auer, E., Bernstein, L., Sungkarat, W., Singh, M., 2010. Vibrotactile Activation of the Auditory Cortices in Deaf versus Hearing Adults. Changes 38, 319–335.
671
doi:10.1146/annurev.neuro.31.060407.125627.Brain
674
Avena-Koenigsberger, A., Misic, B., Sporns, O., 2017. Communication dynamics in complex brain
M AN U
673
SC
670
672
RI PT
667
networks. Nat. Rev. Neurosci. 19, 17–33. doi:10.1038/nrn.2017.149 Badcock, N.A., Mousikou, P., Mahajan, Y., de Lissa, P., Thie, J., McArthur, G., 2013. Validation of
675
the Emotiv EPOC ® EEG gaming system for measuring research quality auditory ERPs. PeerJ 1,
676
e38. doi:10.7717/peerj.38
Badcock, N.A., Preece, K.A., de Wit, B., Glenn, K., Fieder, N., Thie, J., McArthur, G., 2015.
678
Validation of the Emotiv EPOC EEG system for research quality auditory event-related
679
potentials in children. PeerJ 3, e907. doi:10.7717/peerj.907 Barry, R.J., Clarke, A.R., Johnstone, S.J., Brown, C.R., 2009. EEG differences in children between
EP
680
TE D
677
eyes-closed and eyes-open resting conditions. Clin. Neurophysiol. 120, 1806–1811.
682
doi:10.1016/j.clinph.2009.08.006
683
AC C
681
Barry, R.J., Clarke, A.R., Johnstone, S.J., Magee, C.A., Rushby, J.A., 2007. EEG differences between
684
eyes-closed and eyes-open resting conditions. Clin. Neurophysiol. 118, 2765–2773.
685
doi:10.1016/j.clinph.2007.07.028
686
Başar, E., Başar-Eroglu, C., Karakaş, S., Schürmann, M., 2000. Gamma, alpha, delta, and theta
687
oscillations govern cognitive processes. Int. J. Psychophysiol. 39, 241–248. doi:10.1016/S0167-
688
8760(00)00145-8
689
Bassett, D.S., Bullmore, E.T., 2009. Human brain networks in health and disease. Curr Opin Neurol.
34
ACCEPTED MANUSCRIPT
692 693 694 695 696 697 698 699
Bassett, D.S., Sporns, O., 2017. Network Neuroscience. Nat. Neurosci. 20, 353–364. doi:10.1038/nn.4502.Network Bassett, D.S., Zurn, P., Gold, J.I., 2018. On the nature and use of models in network neuroscience. Nat. Rev. Neurosci. 19, 566–578. doi:10.1038/s41583-018-0038-8
RI PT
691
22, 340–347. doi:10.1097/WCO.0b013e32832d93dd.Human
Bavelier, D., Dye, M.W.G., Hauser, P.C., 2006. Do deaf individuals see better? Trends Cogn. Sci. 10, 512–518. doi:10.1016/j.tics.2006.09.006
Bavelier, D., Neville, H.J., 2002. Cross-modal plasticity: where and how? Nat. Rev. Neurosci. 3, 443– 52. doi:10.1038/nrn848
SC
690
Bavelier, D., Tomann, A., Hutton, C., Mitchell, T., Corina, D., Liu, G., Neville, H., 2000. Visual attention to the periphery is enhanced in congenitally deaf individuals. J. Neurosci. 20, RC93--
701
RC93.
702
M AN U
700
Boersma, M., Smit, D.J.A., Boomsma, D.I., De Geus, E.J.C., Delemarre-van de Waal, H.A., Stam, C.J., 2013. Growing trees in child brains: graph theoretical analysis of electroencephalography-
704
derived minimum spanning tree in 5- and 7-year-old children reflects brain maturation. Brain
705
Connect. 3, 50–60. doi:10.1089/brain.2012.0106
706
TE D
703
Boersma, M., Smit, D.J.A., De Bie, H.M.A., Van Baal, G.C.M., Boomsma, D.I., De Geus, E.J.C., Delemarre-Van De Waal, H.A., Stam, C.J., 2011. Network analysis of resting state EEG in the
708
developing young brain: Structure comes with maturation. Hum. Brain Mapp. 32, 413–425.
709
doi:10.1002/hbm.21030
AC C
EP
707
710
Bola, Ł., Zimmermann, M., Mostowski, P., Jednoróg, K., Marchewka, A., Rutkowski, P., Szwed, M.,
711
2017. Task-specific reorganization of the auditory cortex in deaf humans. Proc. Natl. Acad. Sci.
712
114, E600–E609. doi:10.1073/pnas.1609000114
713
Bosworth, R.G., Dobkins, K.R., 2002. The effects of spatial attention on motion processing in deaf
714
signers, hearing signers, and hearing nonsigners. Brain Cogn. 49, 152–69.
715
doi:10.1006/brcg.2001.1497
716 717
Breakspear, M., 2017. Dynamic models of large-scale brain activity. Nat. Neurosci. 20, 340–352. doi:10.1038/nn.4497
35
ACCEPTED MANUSCRIPT 718
Brotherton, H., Plack, C.J., Schaette, R., Munro, K.J., 2016. Time course and frequency specificity of
719
sub-cortical plasticity in adults following acute unilateral deprivation. Hear. Res. 341, 210–219.
720
doi:10.1016/j.heares.2016.09.003
721
Brozinsky, C.J., Bavelier, D., 2004. Motion velocity thresholds in deaf signers: Changes in lateralization but not in overall sensitivity. Cogn. Brain Res. 21, 1–10.
723
doi:10.1016/j.cogbrainres.2004.05.002
724
RI PT
722
Buckley, K.A., Tobey, E.A., 2011. Cross-modal plasticity and speech perception in pre- and postlingually deaf cochlear implant users. Ear Hear. 32, 2–15.
726
doi:10.1097/AUD.0b013e3181e8534c
SC
725
Bullmore, E., Barnes, A., Bassett, D.S., Fornito, A., Kitzbichler, M., Meunier, D., Suckling, J., 2009.
728
Generic aspects of complexity in brain imaging data and other biological systems. Neuroimage
729
47, 1125–1134. doi:10.1016/j.neuroimage.2009.05.032
731 732 733 734
Bullmore, E., Sporns, O., 2012. The economy of brain network organization. Nat. Rev. Neurosci. 13, 336–349. doi:10.1038/nrn3214
Bullmore, E., Sporns, O., 2009. Complex brain networks: graph theoretical analysis of structural and
TE D
730
M AN U
727
functional systems. Nat Rev Neurosci 10, 186–198. doi:10.1038/nrn2575 Buschman, T.J., Miller, E.K., 2007. Top-Down Versus Bottom-Up Control of Attention in the Prefrontal and Posterior Parietal Cortices. Science (80-. ). 315, 1860–1862.
736
doi:10.1126/science.1138071
738 739 740 741 742 743
Caimo, A., Friel, N., 2011. Bayesian inference for exponential random graph models. Soc. Networks
AC C
737
EP
735
33, 41–55. doi:10.1016/j.socnet.2010.09.004 Campbell, J., Sharma, A., 2014. Cross-modal re-organization in adults with early stage hearing loss. PLoS One 9. doi:10.1371/journal.pone.0090594 Campbell, J., Sharma, A., 2013. Compensatory changes in cortical resource allocation in adults with hearing loss. Front. Syst. Neurosci. 7, 1–9. doi:10.3389/fnsys.2013.00071 Chen, A.C.N., Feng, W., Zhao, H., Yin, Y., Wang, P., 2008. EEG default mode network in the human
744
brain: Spectral regional field powers. Neuroimage 41, 561–574.
745
doi:10.1016/j.neuroimage.2007.12.064
36
ACCEPTED MANUSCRIPT 746
David Hairston, W., Whitaker, K.W., Ries, A.J., Vettel, J.M., Cortney Bradford, J., Kerick, S.E.,
747
McDowell, K., 2014. Usability of four commercially-oriented EEG systems. J. Neural Eng. 11,
748
046018. doi:10.1088/1741-2560/11/4/046018
749
Dewey, R.S., Hartley, D.E.H., 2015. Cortical cross-modal plasticity following deafness measured using functional near-infrared spectroscopy. Hear. Res. 325, 55–63.
751
doi:10.1016/j.heares.2015.03.007
752
RI PT
750
Ding, H., Qin, W., Liang, M., Ming, D., Wan, B., Li, Q., Yu, C., 2015. Cross-modal activation of auditory regions during visuo-spatial working memory in early deafness. Brain 138, 2750–2765.
754
doi:10.1093/brain/awv165
755
SC
753
Doucet, M.E., Bergeron, F., Lassonde, M., Ferron, P., Lepore, F., 2006. Cross-modal reorganization and speech perception in cochlear implant users. Brain 129, 3376–3383.
757
doi:10.1093/brain/awl264
759 760
Durkin, M., 2002. The epidemiology of developmental disabilities in low-income countries. Ment. Retard. Dev. Disabil. Res. Rev. 8, 206–211. doi:10.1002/mrdd.10039 Dye, M.W.G., Baril, D.E., Bavelier, D., 2007. Which aspects of visual attention are changed by
TE D
758
M AN U
756
761
deafness? The case of the Attentional Network Test. Neuropsychologia 45, 1801–1811.
762
doi:10.1016/j.neuropsychologia.2006.12.019
765
EP
764
Eleweke, C.J., Rodda, M., 2002. The challenge of enhancing inclusive education in developing countries. Int. J. Incl. Educ. 6, 113–126. doi:10.1080/13603110110067190 Engels, M.M.A., Stam, C.J., van der Flier, W.M., Scheltens, P., de Waal, H., van Straaten, E.C.W.,
AC C
763
766
2015. Declining functional connectivity and changing hub locations in Alzheimer’s disease: an
767
EEG study. BMC Neurol. 15, 145. doi:10.1186/s12883-015-0400-7
768
Finney, E.M., Dobkins, K.R., 2001. Visual contrast sensitivity in deaf versus hearing populations:
769
Exploring the perceptual consequences of auditory deprivation and experience with a visual
770
language. Cogn. Brain Res. 11, 171–183. doi:10.1016/S0926-6410(00)00082-3
771 772 773
Finney, E.M., Fine, I., Dobkins, K.R., 2001. Visual stimuli activate auditory cortex in the deaf. Nat. Neurosci. 4, 1171–1173. doi:10.1038/nn763 Fraschini, M., Demuru, M., Crobe, A., Marrosu, F., Stam, C.J., Hillebrand, A., 2016. The effect of
37
ACCEPTED MANUSCRIPT 774
epoch length on estimated EEG functional connectivity and brain network organisation. J. Neural
775
Eng. 13, 036015. doi:10.1088/1741-2560/13/3/036015
776
Gaál, Z.A., Boha, R., Stam, C.J., Molnár, M., 2010. Age-dependent features of EEG-reactivitySpectral, complexity, and network characteristics. Neurosci. Lett. 479, 79–84.
778
doi:10.1016/j.neulet.2010.05.037
RI PT
777
779
Gómez-Ramírez, J., Freedman, S., Mateos, D., Pérez-Velázquez, J.L., Valiante, T., 2017. Eyes closed
780
or Eyes open? Exploring the alpha desynchronization hypothesis in resting state functional
781
connectivity networks with intracranial EEG 1–26. doi:10.1101/118174
783
Hackett, T.A., 2012. Information flow in the auditory cortical network. Hear. Res. 271, 133–146. doi:10.1016/j.heares.2010.01.011.Information
SC
782
Hagmann, P., Sporns, O., Madan, N., Cammoun, L., Pienaar, R., Wedeen, V.J., Meuli, R., Thiran, J.-
785
P., Grant, P.E., 2010. White matter maturation reshapes structural connectivity in the late
786
developing human brain. Proc. Natl. Acad. Sci. U. S. A. 107, 19067–19072.
787
doi:10.1073/pnas.1009073107
Hammes, D.M., Novak, M.A., Rotz, L.A., Willis, M., Edmondson, D.M., Thomas, J.F., 2002. Early
TE D
788
M AN U
784
789
identification and cochlear implantation: Critical factors for spoken language development. Ann.
790
Otol. Rhinol. Laryngol. Suppl. 189, 74–78.
Hauser, P.C., Dye, M.W.G., Boutla, M., Green, C.S., Bavelier, D., 2007. Deafness and visual
EP
791
enumeration: Not all aspects of attention are modified by deafness. Brain Res. 1153, 178–187.
793
doi:10.1016/j.brainres.2007.03.065
794
AC C
792
Heuvel, M.P. Van Den, Kahn, R.S., Goñi, J., Sporns, O., 2012. Brain Communication. Proc. Natl.
795
Acad. Sci. U. S. A. 109, 11372–77. doi:10.1073/pnas.1203593109/-
796
/DCSupplemental.www.pnas.org/cgi/doi/10.1073/pnas.1203593109
797 798 799
Hidalgo, C.A., Klinger, B., Barabasi, A.-L., Hausmann, R., 2007. The product space conditions the development of nations. Science (80-. ). 317, 482–487. doi:10.1126/science.1144581 Hüfner, K., Stephan, T., Flanagin, V.L., Deutschländer, A., Stein, A., Kalla, R., Dera, T., Fesl, G.,
800
Jahn, K., Strupp, M., Brandt, T., 2009. Differential effects of eyes open or closed in darkness on
801
brain activation patterns in blind subjects. Neurosci. Lett. 466, 30–34.
38
ACCEPTED MANUSCRIPT 802
doi:10.1016/j.neulet.2009.09.010 Hüfner, K., Stephan, T., Glasauer, S., Kalla, R., Riedel, E., Deutschländer, A., Dera, T., Wiesmann,
804
M., Strupp, M., Brandt, T., 2008. Differences in saccade-evoked brain activation patterns with
805
eyes open or eyes closed in complete darkness. Exp. Brain Res. 186, 419–430.
806
doi:10.1007/s00221-007-1247-y
RI PT
803
807
Jackson, T.S., Read, N., 2010. Theory of minimum spanning trees. I. Mean-field theory and strongly
808
disordered spin-glass model. Phys. Rev. E - Stat. Nonlinear, Soft Matter Phys. 81, 1–18.
Karns, C.M., Dow, M.W., Neville, H.J., 2012. Altered cross-modal processing in the primary auditory
810
cortex of congenitally deaf adults: A visual-somatosensory fMRI study with a double-flash
811
illusion. J. Neurosci. 32, 9626–9638. doi:10.1523/JNEUROSCI.6488-11.2012 Kim, E., Kang, H., Lee, H., Lee, H.-J., Suh, M.-W., Song, J.-J., Oh, S.-H., Lee, D.S., 2014.
M AN U
812
SC
809
813
Morphological brain network assessed using graph theory and network filtration in deaf adults.
814
Hear. Res. 315, 88–98. doi:10.1016/j.heares.2014.06.007
Klimesch, W., 1999. EEG alpha and theta oscillations reflect cognitive and memory performance: a
816
review and analysis. Brain Res. Rev. 29, 169–195. doi:10.1016/S0165-0173(98)00056-3
817
TE D
815
Knyazev, G.G., Volf, N. V., Belousova, L. V., 2015. Age-related differences in electroencephalogram connectivity and network topology. Neurobiol. Aging 36, 1849–1859.
819
doi:10.1016/j.neurobiolaging.2015.02.007
821 822
Kruskal, J.B., 1956. On the shortest spanning subtree of a graph and the traveling salesman problem. Proc. Am. Math. Soc. 7, 48–50. doi:10.2307/2033241
AC C
820
EP
818
La Rosa, P.S., Brooks, T.L., Deych, E., Shands, B., Prior, F., Larson-Prior, L.J., Shannon, W.D., 2016.
823
Gibbs distribution for statistical analysis of graphical data with a sample application to fcMRI
824
brain images. Stat. Med. 35, 566–580. doi:10.1002/sim.6757
825 826 827 828 829
Levänen, S., Hamdorf, D., 2001. Feeling vibrations: Enhanced tactile sensitivity in congenitally deaf humans. Neurosci. Lett. 301, 75–77. doi:10.1016/S0304-3940(01)01597-X Li, W., Li, J., Wang, J., Zhou, P., Wang, Z., Xian, J., He, H., 2016. Functional reorganizations of brain network in prelingually deaf adolescents. Neural Plast. 2016, 1–10. doi:10.1155/2016/9849087 Li, W., Li, J., Wang, Z., Li, Y., Liu, Z., Yan, F., Xian, J., He, H., 2015. Grey matter connectivity
39
ACCEPTED MANUSCRIPT 830
within and between auditory, language and visual systems in prelingually deaf adolescents.
831
Restor. Neurol. Neurosci. 33, 279–290. doi:10.3233/RNN-140437
832
Li, Y., Booth, J.R., Peng, D., Zang, Y., Li, J., Yan, C., Ding, G., 2013. Altered intra- and interregional synchronization of superior temporal cortex in deaf people. Cereb. Cortex 23, 1988–
834
1996. doi:10.1093/cercor/bhs185
835
RI PT
833
Liu, L., Yan, X., Liu, J., Xia, M., Lu, C., Emmorey, K., Chu, M., Ding, G., 2017. Graph theoretical
836
analysis of functional network for comprehension of sign language. Brain Res. 1671, 55–66.
837
doi:10.1016/j.brainres.2017.06.031
840
SC
839
Malaia, E., Talavage, T.M., Wilbur, R.B., 2014. Functional connectivity in task-negative network of the Deaf: effects of sign language experience. PeerJ 2, e446. doi:10.7717/peerj.446 Marx, E., Deutschländer, A., Stephan, T., Dieterich, M., Wiesmann, M., Brandt, T., 2004. Eyes open
M AN U
838
841
and eyes closed as rest conditions: Impact on brain activation patterns. Neuroimage 21, 1818–
842
1824. doi:10.1016/j.neuroimage.2003.12.026
843
Marx, E., Stephan, T., Nolte, A., Deutschländer, A., Seelos, K.C., Dieterich, M., Brandt, T., 2003. Eye closure in darkness animates sensory systems. Neuroimage 19, 924–934. doi:10.1016/S1053-
845
8119(03)00150-2
TE D
844
McMahan, T., Parberry, I., Parsons, T.D., 2015. Modality specific assessment of video game player’s
847
experience using the Emotiv. Entertain. Comput. 7, 1–6. doi:10.1016/j.entcom.2015.03.001
848
Merabet, L.B., Pascual-Leone, A., 2010. Neural reorganization following sensory loss: the opportunity of change. Nat. Rev. Neurosci. 11, 44–52. doi:10.1038/nrn2758
AC C
849
EP
846
850
Meredith, M.A., Lomber, S.G., 2011. Somatosensory and visual crossmodal plasticity in the anterior
851
auditory field of early-deaf cats. Hear. Res. 280, 38–47. doi:10.1016/j.heares.2011.02.004
852
Meyer, M., Toepel, U., Keller, J., Nussbaumer, D., Zysset, S., Friederici, A.D., 2007. Neuroplasticity
853
of sign language: implications from structural and functional brain imaging. Restor. Neurol.
854
Neurosci. 25, 335–351.
855
Miraglia, F., Vecchio, F., Bramanti, P., Rossini, P.M., 2016. EEG characteristics in “eyes-open”
856
versus “eyes-closed” conditions: Small-world network architecture in healthy aging and age-
857
related brain degeneration. Clin. Neurophysiol. 127, 1261–1268.
40
ACCEPTED MANUSCRIPT 858 859 860 861
doi:10.1016/j.clinph.2015.07.040 Mulwafu, W., Kuper, H., Ensink, R.J.H., 2016. Prevalence and causes of hearing impairment in Africa. Trop. Med. Int. Heal. 21, 158–165. doi:10.1111/tmi.12640 Neville, H.J., Bavelier, D., Corina, D., Rauschecker, J., Karni, A., Lalwani, A., Braun, A., Clark, V., Jezzard, P., Turner, R., 1998. Cerebral organization for language in deaf and hearing subjects:
863
biological constraints and effects of experience. Proc. Natl. Acad. Sci. U. S. A. 95, 922–9.
864
doi:10.1073/pnas.95.3.922
867 868 869
detection task: II. Brain Res. 405, 268–283.
SC
866
Neville, H.J., Lawson, D., 1987a. Attention to central and peripheral visual space in a movement
Neville, H.J., Lawson, D., 1987b. Attention to central and peripheral visual space in a movement detection task: III. Brain Res. 405, 284–294.
M AN U
865
RI PT
862
Nishimura, H., Hashikawa, K., Doi, K., Iwaki, T., Watanabe, Y., Kusuoka, H., Nishimura, T., Kubo,
870
T., 1999. Sign language “heard” in the auditory cortex. Nature 397, 116. doi:10.1038/16376
871
Otte, W.M., Sinke, M.R.T., Van Diessen, E., 2018a. Functional brain network analysis with minimum
873 874
spanning trees. https://zenodo.org/record/2066281#.XAudsRNKhDU.
TE D
872
Otte, W.M., Sinke, M.R.T., van Diessen, E., Dijkhuizen, R.M., 2018b. Functional Connectivity and Brain Networks in Deaf and Hearing. Open Sci. Framew. Otte, W.M., van Diessen, E., Paul, S., Ramaswamy, R., Subramanyam Rallabandi, V.P., Stam, C.J.,
876
Roy, P.K., 2015. Aging alterations in whole-brain networks during adulthood mapped with the
877
minimum spanning tree indices: The interplay of density, connectivity cost and life-time
878
trajectory. Neuroimage 109, 171–189. doi:10.1016/j.neuroimage.2015.01.011
AC C
EP
875
879
Pénicaud, S., Klein, D., Zatorre, R.J., Chen, J.K., Witcher, P., Hyde, K., Mayberry, R.I., 2013.
880
Structural brain changes linked to delayed first language acquisition in congenitally deaf
881
individuals. Neuroimage 66, 42–49. doi:10.1016/j.neuroimage.2012.09.076
882
Petitto, L.A., 2000. Speech-like cerebral activity in profoundly deaf people processing signed
883
languages: implications for the neural basis of human language. Proc. Natl Acad. Sci. USA 97,
884
13961–13966.
885
Pförtner, K., 2014. Community-based inclusive education: Best practices from Nicaragua, El
41
ACCEPTED MANUSCRIPT 886
Salvador, Guatemala and Honduras. Disabil. CBR Incl. Dev. 25, 72–81.
887
doi:10.5463/DCID.v25i1.321
889 890
Pillai, J., Sperling, M.R., 2006. Interictal EEG and the diagnosis of epilepsy. Epilepsia 47, 14–22. doi:10.1111/j.1528-1167.2006.00654.x Prause, N., Siegle, G.J., Deblieck, C., Wu, A., Iacoboni, M., 2016. EEG to primary rewards: Predictive
RI PT
888
891
utility and malleability by brain stimulation. PLoS One 11. doi:10.1371/journal.pone.0165646
892
Ptito, M., Giguere, J.F., Boire, D., Frost, D.O., Casanova, C., 2001. When the auditory cortex turns
895 896 897
Raftery, A.E., 1995. Bayesian model selection in social research. Sociol. Methodol.
SC
894
visual. Prog. Brain Res. 134, 447–458.
doi:10.2307/271063
Rubinov, M., Sporns, O., 2010. Complex network measures of brain connectivity: Uses and
M AN U
893
interpretations. Neuroimage 52, 1059–1069. doi:10.1016/j.neuroimage.2009.10.003
898
Sabidussi, G., 1966. The centrality index of a graph. Psychometrika 31, 581–603.
899
Sadato, N., Okada, T., Honda, M., Matsuki, K.-I., Yoshida, M., Kashikura, K.-I., Takei, W., Sato, T., Kochiyama, T., Yonekura, Y., 2005. Cross-modal integration and plastic changes revealed by lip
901
movement, random-dot motion and sign languages in the hearing and deaf. Cereb. Cortex 15,
902
1113–1122. doi:10.1093/cercor/bhh210
Sadato, N., Yamada, H., Okada, T., Yoshida, M., Hasegawa, T., Matsuki, K.-I., Yonekura, Y., Itoh,
EP
903
TE D
900
H., 2004. Age-dependent plasticity in the superior temporal sulcus in deaf humans: a functional
905
MRI study. BMC Neurosci. 5, 1–6. doi:10.1186/1471-2202-5-56
AC C
904
906
Schiatti, L., Faes, L., Tessadori, J., Barresi, G., Mattos, L., 2016. Mutual information-based feature
907
selection for low-cost BCIs based on motor imagery. Proc. Annu. Int. Conf. IEEE Eng. Med.
908
Biol. Soc. EMBS 2016–Octob, 2772–2775. doi:10.1109/EMBC.2016.7591305
909
Scott, G.D., Karns, C.M., Dow, M.W., Stevens, C., Neville, H.J., 2014. Enhanced peripheral visual
910
processing in congenitally deaf humans is supported by multiple brain regions, including primary
911
auditory cortex. Front. Hum. Neurosci. 8, 1–9. doi:10.3389/fnhum.2014.00177
912 913
Shiell, M.M., Champoux, F., Zatorre, R.J., 2016. The right hemisphere planum temporale supports enhanced visual motion detection ability in deaf people: Evidence from cortical thickness. Neural
42
ACCEPTED MANUSCRIPT 914 915
Plast. 1–10. doi:10.1155/2016/7217630 Shiell, M.M., Champoux, F., Zatorre, R.J., 2014. Reorganization of auditory cortex in early-deaf
916
people: Functional connectivity and relationship to hearing aid use. J. Cogn. Neurosci. 27, 150–
917
163. doi:10.1162/jocn
919 920
Simpson, S.L., Laurienti, P.J., 2015. A two-part mixed-effects modeling framework for analyzing
RI PT
918
whole- brain network data. Neuroimage 113, 310–319.
Sinke, M.R.T., Dijkhuizen, R.M., Caimo, A., Stam, C.J., Otte, W.M., 2016. Bayesian exponential random graph modeling of whole-brain structural networks across lifespan. Neuroimage 135,
922
79–91. doi:10.1016/j.neuroimage.2016.04.066
SC
921
Smit, D.J.A., Boersma, M., Schnack, H.G., Micheloyannis, S., Boomsma, D.I., Hulshoff Pol, H.E.,
924
Stam, C.J., de Geus, E.J.., 2012. The brain matures with stronger functional connectivity and
925
decreased randomness of its network. PLoS One 7, e36896. doi:10.1371/journal.pone.0036896
M AN U
923
Smit, D.J.A., Geus, E.J.C. de, Boersma, M., Boomsma, D.I., Stam, C.J., 2016. Life-span development
927
of brain network integration assessed with Phase Lag Index connectivity and minimum spanning
928
tree graphs. Brain Connect. 8, 1–38. doi:10.1089/brain.2015.0359
TE D
926
Smith, K.M., Mecoli, M.D., Altaye, M., Komlos, M., Maitra, R., Eaton, K.P., Egelhoff, J.C., Holland,
930
S.K., 2011. Morphometric differences in the heschl’s gyrus of hearing impaired and normal
931
hearing infants. Cereb. Cortex 21, 991–998. doi:10.1093/cercor/bhq164
933
Stam, C.J., 2014. Modern network science of neurological disorders. Nat. Rev. Neurosci. 15, 683–695. doi:10.1038/nrn3801
AC C
932
EP
929
934
Stam, C.J., Nolte, G., Daffertshofer, A., 2007. Phase lag index: Assessment of functional connectivity
935
from multi channel EEG and MEG with diminished bias from common sources. Hum. Brain
936
Mapp. 28, 1178–1193. doi:10.1002/hbm.20346
937
Stam, C.J., Tewarie, P., Van Dellen, E., van Straaten, E.C.W., Hillebrand, a, Van Mieghem, P., 2014.
938
The trees and the forest: Characterization of complex brain networks with minimum spanning
939
trees. Int. J. Psychophysiol. 92, 129–138. doi:10.1016/j.ijpsycho.2014.04.001
940 941
Stam, C.J., van Straaten, E.C.W., 2012. The organization of physiological brain networks. Clin. Neurophysiol. 123, 1067–1087. doi:10.1016/j.clinph.2012.01.011
43
ACCEPTED MANUSCRIPT 942
Stevens, G., Flaxman, S., Brunskill, E., Mascarenhas, M., Mathers, C.D., Finucane, M., 2013. Global
943
and regional hearing impairment prevalence: An analysis of 42 studies in 29 countries. Eur. J.
944
Public Health 23, 146–152. doi:10.1093/eurpub/ckr176
945
Strelnikov, K., Rouger, J., Demonet, J.F., Lagleyre, S., Fraysse, B., Deguine, O., Barone, P., 2010. Does brain activity at rest reflect adaptive strategies? evidence from speech processing after
947
cochlear implantation. Cereb. Cortex 20, 1217–1222. doi:10.1093/cercor/bhp183
948
RI PT
946
Striem-Amit, E., Almeida, J., Belledonne, M., Chen, Q., Fang, Y., Han, Z., Caramazza, A., Bi, Y., 2016. Topographical functional connectivity patterns exist in the congenitally, prelingually deaf.
950
Sci. Rep. 6, 29375. doi:10.1038/srep29375
951
SC
949
Tan, B., Kong, X., Yang, P., Jin, Z., Li, L., 2013. The difference of brain functional connectivity between eyes-closed and eyes-open using graph theoretical analysis. Comput Math Methods Med
953
1–15. doi:10.1155/2013/976365
M AN U
952
Tewarie, P., Hillebrand, A., Schoonheim, M.M., van Dijk, B.W., Geurts, J.J.G., Barkhof, F., Polman,
955
C.H., Stam, C.J., 2014. Functional brain network analysis using minimum spanning trees in
956
Multiple Sclerosis: An MEG source-space study. Neuroimage 88, 308–318.
957
doi:10.1016/j.neuroimage.2013.10.022
958
TE D
954
Tewarie, P., van Dellen, E., Hillebrand, A., Stam, C.J., 2015. The minimum spanning tree: An unbiased method for brain network analysis. Neuroimage 104, 177–188.
960
doi:10.1016/j.neuroimage.2014.10.015 van den Heuvel, M.P., Hulshoff Pol, H.E., 2010. Exploring the brain network: A review on resting-
AC C
961
EP
959
962
state fMRI functional connectivity. Eur. Neuropsychopharmacol. 20, 519–534.
963
doi:10.1016/j.euroneuro.2010.03.008
964
van Diessen, E., Numan, T., van Dellen, E., van der Kooi, A.W., Boersma, M., Hofman, D., van
965
Lutterveld, R., van Dijk, B.W., van Straaten, E.C.W., Hillebrand, A., Stam, C.J., 2015.
966
Opportunities and methodological challenges in EEG and MEG resting state functional brain
967
network research. Clin. Neurophysiol. 126, 1468–1481. doi:10.1016/j.clinph.2014.11.018
968 969
van Diessen, E., Otte, W.M., Braun, K.P.J., Stam, C.J., Jansen, F.E., 2014. Does sleep deprivation alter functional EEG networks in children with focal epilepsy? Front. Syst. Neurosci. 8, 67.
44
ACCEPTED MANUSCRIPT 970
doi:10.3389/fnsys.2014.00067
971
van Diessen, E., Otte, W.M., Stam, C.J., Braun, K.P.J., Jansen, F.E., 2016. Electroencephalography
972
based functional networks in newly diagnosed childhood epilepsies. Clin. Neurophysiol. 127,
973
2325–2332. doi:10.1016/j.clinph.2016.03.015 van Wijk, B.C.M., Stam, C.J., Daffertshofer, A., 2010. Comparing brain networks of different size and
975
connectivity density using graph theory. PLoS One 5, e13701. doi:10.1371/journal.pone.0013701
976 977
RI PT
974
Vohr, B., 2003. Infants and children with hearing loss - part 1. Ment. Retard. Dev. Disabil. Res. Rev. 9, 218–9. doi:10.1002/mrdd.10082
Vourkas, M., Karakonstantaki, E., Simos, P.G., Tsirka, V., Antonakakis, M., Vamvoukas, M., Stam,
979
C.J., Dimitriadis, S., Micheloyannis, S., 2014. Simple and difficult mathematics in children: A
980
minimum spanning tree EEG network analysis. Neurosci. Lett. 576, 28–33.
981
doi:10.1016/j.neulet.2014.05.048
985 986 987 988 989 990
M AN U
WHO, 2014. WHO | Deafness and hearing loss. World Heal. Organ.
TE D
984
442.
doi:/entity/mediacentre/factsheets/fs300/en/index.html WHO, 2012. WHO global estimates on prevalence of hearing loss. World Heal. Organ. 1–15. doi:10.1002/2014GB005021
EP
983
Watts, D.J., Strogatz, S.H., 1998. Collective dynamics of `small-world’ networks. Nature 393, 440–
Wróbel, A., 2000. Beta activity: a carrier for visual attention Andrzej. Acta Neurobiol Exp 60, 247–60. doi:10.1016/j.procs.2015.07.351
AC C
982
SC
978
Wrobel, A., Ghazaryan, A., Bekisz, M., Bogdan, W., Kaminski, J., 2007. Two Streams ofAttention-
991
Dependent B Activity in the Striate Recipient Zone of Cat’s Lateral Posterior–Pulvinar Complex.
992
J. Neurosci. 27, 2230–2240. doi:10.1080/00268976.2013.805848
993
Xu, P., Huang, R., Wang, J., Van Dam, N.T., Xie, T., Dong, Z., Chen, C., Gu, R., Zang, Y.F., He, Y.,
994
Fan, J., Luo, Y. jia, 2014. Different topological organization of human brain functional networks
995
with eyes open versus eyes closed. Neuroimage 90, 246–255.
996
doi:10.1016/j.neuroimage.2013.12.060
997
Yoshida, H., Kanda, Y., Takahashi, H., Miyamoto, I., Chiba, K., 2011. Observation of cortical activity
45
ACCEPTED MANUSCRIPT 998
during speech stimulation in prelingually deafened adults with cochlear implantation by positron
999
emission tomography-computed tomography. Ann. Otol. Rhinol. Laryngol. 120, 499–504.
1000 1001
doi:10.1177/000348941112000802 Yu, J.-H., Sim, K.-B., 2016. Classification of color imagination using Emotiv EPOC and event-related potential in electroencephalogram. Opt. - Int. J. Light Electron Opt. 127, 9711–9718.
1003
doi:10.1016/j.ijleo.2016.07.074
1004
RI PT
1002
Zalesky, A., Fornito, A., Bullmore, E.T., 2010. Network-based statistic: Identifying differences in brain networks. Neuroimage 53, 1197–1207. doi:10.1016/j.neuroimage.2010.06.041
1006
Zhang, D., Liang, B., Wu, X., Wang, Z., Xu, P., Chang, S., Liu, B., Liu, M., Huang, R., 2015.
SC
1005
Directionality of large-scale resting-state brain networks during eyes open and eyes closed
1008
conditions. Front. Hum. Neurosci. 9, 1–10. doi:10.3389/fnhum.2015.00081
M AN U
1007
AC C
EP
TE D
1009
46
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
ACCEPTED MANUSCRIPT Figure 1. Schematic overview of the study pipeline. Resting-state electroencephalography (EEG) was acquired for five minutes with wireless headsets in deaf and controls. The sensor locations corresponding to the fourteen channels are shown in orange; the reference sensors are shown in blue. The first minute of data acquisition, required for acclimatization, was
RI PT
discarded, yielding two blocks of recordings – two minutes with eyes-closed and two minutes with eyes-open. The order of the eyes condition (‘open’ and ‘closed’) was alternated. Subsequently, functional networks were constructed from ten-second epochs from distinct
SC
frequency bands (i.e. delta, theta, alpha, beta and gamma), and functional network backbone metrics were determined for all bands. These network backbone metrics were related to years
M AN U
of sign language experience. Main effects between groups (i.e. deaf and control) and conditions (i.e. eyes-open and eyes-closed) as well as interaction effects between group and condition were assessed using model selection within a Bayesian framework.
TE D
Figure 2. Raw EEG time-series of one participant. The visualized time-series were acquired from a deaf male (22 years old). Vertical grey lines indicate the initial acclimatization period, the first two minutes (i.e. eyes-closed in this subject) and the second
EP
two minutes (i.e. eyes-open). The y-scaling is arbitrary. Labels of the fourteen channels are
AC C
shown on top of the time-series. Abbreviations of these channel labels are explained in the main text.
Figure 3. Schematic overview of functional network and minimum spanning tree (MST) construction. (A) EEG epochs were measured from 14 electrodes, represented as network nodes. The electrodes were placed across the left (L, yellow) and right (R, orange) hemisphere (top), two electrodes served as references (blue). The recorded time-series (bottom) were used to determine functional connectivity between brain regions. (B) The determined functional
ACCEPTED MANUSCRIPT connections can either be depicted as a weighted functional network graph (top) or an adjacency matrix (bottom) where columns and rows represent nodes and colored squares indicate functional connectivity between nodes. The color indicate functional connectivity strength (i.e. darker = stronger). (C) The weighted functional network can be used to
RI PT
determine the minimal spanning tree (MST), which only includes the strongest connections, forming a functional network backbone. This determined MST can also be depicted as a network graph (top) or an adjacency matrix (bottom). Subsequently, MST metrics, such as
SC
leaf fraction and diameter, can be determined. Abbreviations of channel labels are explained
M AN U
in the main text.
Figure 4. Functional connectivity strength between the occipital cortex and the parietal cortex. (A) The average functional connectivity, quantified with the phase lag index, between O1/O2 and P7/P8 (y-axis) is shown for eyes-open and eyes-closed conditions (x-axis) and all
TE D
frequency bands (top) in controls and deaf people. (B) The delta functional connectivity between eyes-open en eyes-closed is plotted as percentage change relative to the eyes-open functional connectivity values for each frequency band, based on the data shown in the left
AC C
EP
panel: ∆ in % = 100 × ( closed – open ) / open.
Figure 5. Network backbone comparisons between eyes-open and eyes-closed conditions for both deaf and controls. Functional network backbone characteristics in the alpha band (8-16Hz) and beta band (16-32 Hz) (top), are shown for deaf (yellow) and controls (blue) for both the eyes-open and eyes-closed condition (x-axis) and indicated by the following minimum spanning tree metrics (y-axis), (from top-left to bottom-right): diameter, maximum closeness centrality, median closeness centrality, leaf number, maximum strength, mean strength, eccentricity, radius and kappa. Error bars represent the 95% confidence intervals.
ACCEPTED MANUSCRIPT Figure 6. The relation between American Sign Language (ASL) and functional backbone characteristics. The sign language experience in years (x-axis) is related to functional backbone characteristics (y-axis) in the theta band (4-8 Hz) for both the eyes-open
RI PT
(yellow) and eyes-closed (blue) condition, as indicated by the following minimum spanning tree metrics (from top-left to bottom-right): mean strength, maximum degree, median betweenness centrality (BC), median closeness centrality (CC), leaf number, diameter, radius
M AN U
SC
and kappa. Shaded areas: 95% confidence intervals.
Figure 7. The relation between American Sign Language (ASL) and functional backbone characteristics in deaf subject only. The sign language experience in years (xaxis) related to functional backbone characteristics (y-axis) in the theta band (4–8 Hz) for both the eyes-open (yellow) and eyes-closed (blue) condition, as indicated by the following
TE D
minimum spanning tree metrics (from top-left to bottom-right): mean strength, maximum degree, median betweenness centrality (BC), median closeness centrality (CC), leaf number,
AC C
EP
diameter, radius and kappa. Shaded areas: 95% confidence intervals.
ACCEPTED MANUSCRIPT Highlights
EP
TE D
M AN U
SC
RI PT
Functional brain mapping in homogeneous population in rural Nigeria. Increased functional synchronization in deaf people in alpha en beta frequency. Synchronization effects are present in brain-wide network backbone structures. Topological network reorganization is associated with sign language acquisition.
AC C
• • • •