Early visual deprivation changes cortical anatomical covariance in dorsal-stream structures

NeuroImage 108 (2015) 194–202

Contents lists available at ScienceDirect

NeuroImage journal homepage: www.elsevier.com/locate/ynimg

Early visual deprivation changes cortical anatomical covariance in dorsal-stream structures Patrice Voss a,b,⁎, Robert J. Zatorre a,b a b

Montreal Neurological Institute, McGill University, Montreal, Canada International laboratory for Brain, Music and Sound research (BRAMS), Montreal, Canada

a r t i c l e

i n f o

Article history: Accepted 24 December 2014 Available online 3 January 2015 Keywords: Blindness Cortical Thickness Covariation Auditory

a b s t r a c t Early blind individuals possess thicker occipital cortex compared to sighted ones. Occipital cortical thickness is also predictive of performance on several auditory discrimination tasks in the blind, which suggests that it can serve as a neuroanatomical marker of auditory behavioural abilities. In light of this atypical relationship between occipital thickness and auditory function, we sought to investigate here the covariation of occipital cortical morphology in occipital areas with that of all other areas across the cortical surface, to assess whether the anatomical covariance with the occipital cortex differs between early blind and sighted individuals. We observed a reduction in anatomical covariance between the right occipital cortex and several areas of the visual dorsal stream in a group of early blind individuals relative to sighted controls. In a separate analysis, we show that the performance of the early blind in a transposed melody discrimination task was strongly predicted by the strength of the cortical covariance between the occipital cortex and intraparietal sulcus, a region for which cortical thickness in the sighted was previously shown to predict performance in the same task. These findings therefore constitute the first evidence linking altered anatomical covariance to early sensory deprivation. Moreover, since covariation of cortical morphology could potentially be related to anatomical connectivity or driven by experiencedependent plasticity, it could consequently help guide future functional connectivity and diffusion tractography studies. © 2015 Elsevier Inc. All rights reserved.

Introduction The scientific literature is quite rich with research documenting the striking crossmodal plasticity that takes place in the brain of congenitally and early blind (EB) individuals (see Voss et al., 2010; Merabet and Pascual-Leone, 2010 for reviews), where deafferented visual brain areas are recruited to carry out non-visual processing. More recently, several groups have started investigating the potential neuroanatomical markers of crossmodal plasticity in the blind as there is ample evidence that experience can shape structural features within the normal brain (see Zatorre et al., 2012). One important neuroplastic change observed in early and congenitally blind individuals relates to increased occipital cortical thickness (CT) relative to sighted controls (Park et al., 2009; Jiang et al., 2009; Bridge et al., 2009; Anurova et al., 2014). Importantly, it has been shown that occipital CT is strongly predictive of auditory abilities in the blind (Voss and Zatorre, 2012), thus confirming that these neuroanatomical changes are behaviourally relevant, and likely result from some form of adaptive compensatory plasticity. However,

⁎ Corresponding author at: Cognitive Neuroscience Unit, Montreal Neurological Institute, 3801 rue University, Montréal, Québec H3A 2B4, Canada. E-mail address: [email protected] (P. Voss).

http://dx.doi.org/10.1016/j.neuroimage.2014.12.063 1053-8119/© 2015 Elsevier Inc. All rights reserved.

it remains unknown whether these changes in cortical structure reflect only local alterations or may also reflect network-level modulations. Neuroanatomical measurements have been proposed as means to characterize brain connectivity through covariation analyses (Bullmore et al., 1998; Mechelli et al., 2005; Lerch et al., 2006). Indeed, brain areas that are highly correlated in size are often part of systems that are known to subserve particular behavioural or cognitive functions (see Alexander-Bloch et al., 2013). For instance, posterior (Wernicke) and anterior language (Broca) areas co-vary strongly in the cortical thickness (Lerch et al., 2006). Similarly, the grey matter volume of the hippocampus co-varies most strongly with other regions known to be part of the memory system (Bohbot et al., 2007). CT in particular has been proposed as a valid measure of cortical covariation since previous morphometric correlations of cortical thickness data have successfully produced structural networks of covariance that resemble tract-tracing data (Mitelman et al., 2005; Lerch et al., 2006; Bernhardt et al., 2008; Gong et al., 2012), while animal tracer studies have additionally shown that cortical thickness can predict anatomical connectivity (Barbas, 1986; Barbas and Rempel-Clower, 1997; Dombrowski et al., 2001). Here, in light of the above findings, we sought to investigate the covariation (i.e. anatomical covariance) of occipital cortical morphology with that of all other points (i.e. vertices (see methods)) across the cortical surface, following the method proposed by Lerch et al. (2006),

P. Voss, R.J. Zatorre / NeuroImage 108 (2015) 194–202

in order to assess how early visual deprivation may alter the corticocortical connectivity patterns. We used a seed-based approach to carry out the covariation analyses. The choice of an occipital seed was based on our previous findings relating auditory behaviour to occipital CT (Voss and Zatorre, 2012). We hypothesized that the seed region (located in the right cuneus) in the EB might show reduced covariance with other non-occipital visual areas, now deafferented from their usual input, whereas enhanced covariance with other primary sensory areas might be observed given the abundant crossmodal recruitment observed in the occipital cortex of the EB. Furthermore, previous work has shown that examining anatomical covariance can be an effective way to characterize individual differences. For instance, Lee et al. (2013) showed that greater anatomical covariance between several linguistically relevant areas was associated with better vocabulary abilities in a sample of developing children. Similarly, here we asked whether better auditory abilities in the EB, could be linked to greater cross-cortical covariance between the seed region and task-relevant brain areas. Finally, although the primary goal here was to investigate the effects of early visual deprivation specifically, we also present some data obtained from late-onset blind (LB) individuals. This was done to assess whether an onset of blindness occurring after the visual system had fully developed with proper input would lead to differential patterns of covariance compared to early blindness. Methods Subjects Data from fourteen early-blind (EB) individuals (38.2 ± 13.8 years (y) of age; 10 males and 4 females), thirteen late-blind (LB) (46.6 ± 8.5 y; 5 males and 8 females), and nineteen sighted subjects (37.6 ± 12.0 y; 8 males and 11 females) were included in the study. EB individuals all lost their sight no later than the age of 4 (average age of blindness onset was of 0.5 ± 1.2 y and average duration of blindness was of 37.7 ± 14.3 y), whereas the LB all lost their sight after the age of 7 (average age of blindness onset was of 29.4 ± 15.4 y and average duration of blindness was of 17.5 ± 10.6 y). Complete demographic data and causes of blindness are displayed in Table 1. All subjects gave written informed consent in accordance with the guidelines approved by the Montreal Neurological Institute and the Nazareth and Louis-Braille Institute (NLBI) for the blind. The research protocols were approved by the ethics committees of the Centre de Recherche Interdisciplinaire en Réadaptation (CRIR), which coordinate research with blind subjects sponsored by the Nazareth and Louis-Braille Institute (NLBI) for the blind, and by the research ethics board of the Montreal Neurological Institute, where the scanning procedures were carried out. Behavioural tasks The behavioural data used in the current study are taken from two auditory tasks that are described in greater detail elsewhere (see Voss and Zatorre, 2012). The first, a pitch discrimination task, required subjects to specify which of the two sequentially presented tones was higher in pitch. The reference tone was a 500 Hz pure tone and the task followed a 2-down/1-up staircase procedure, which produces runs of increasing and decreasing the difference between stimuli whose endpoints (reversal points) bracket the 71% threshold (Levitt, 1970). One staircase run was completed after 15 reversals and the geometric mean of the value of the last 8 reversals was taken as the threshold. The threshold is therefore unaffected by the choice of starting difference, because the first seven endpoints are not entered into the calculation. Four separate runs were conducted for each subject and averaged to produce the final value (i.e. the individual thresholds). The second task, first described by Foster and Zatorre (2010b), was a transposed-melody discrimination task. Here, subjects were required

195

Table 1 The ‘onset’ column refers to the age at which the subjects lost their sight. The ‘duration’ column refers to the number of years that the subjects have been blind. The ‘LP’ column indicates whether subjects still had any residual light perception. Subject demographic information Subjects Age Gender Onset Duration LP

Cause of blindness

EB1 EB2 EB3 EB4 EB5 EB6 EB7 EB8

32 27 37 57 21 26 40 52

F F M F M M M M

4 0 0 0 0 1 0 0

28 27 37 57 21 26 40 52

No No No No No No No No

EB9

20

M

0

20

Yes

EB10 EB11 EB12 EB13 EB14 LB1 LB2 LB3 LB4 LB5 LB6 LB7 LB8 LB9 LB10 LB11 LB12 LB13

59 24 45 39 56 44 48 48 53 53 60 52 59 46 48 29 37 42

M M M M F M F F F F F M F F M F M M

0 2 0 0 0 21 28 24 30 19 48 50 54 43 32 17 9 7

59 21 45 39 56 23 20 24 23 34 12 2 5 3 16 13 18 35

Yes no no Yes Yes No No No Yes No No Yes No Yes Yes Yes No No

Retinoblastoma Retinal detachment Congenital glaucoma Retinopathy of prematurity Retinopathy of prematurity Congenital cataracts Retinopathy of prematurity Medical accident (retina damage) Congenital malformation (no cristallin) Congenital cataracts Retinoblastoma Congenital glaucoma Retinopathy of prematurity Retinal detachment Congenital glaucoma Diabetic retinopathy Ischemic retinopathy Retinal degeneration Glaucoma Failed cornea transplant Dehydration of the optic nerve Retinitis pigmentosa Glaucoma + retinal detachment Retinal detachment Congenital Glaucoma Congenital Glaucoma Lenticular fibroplasia

to determine whether two sequential melodies (unfamiliar melodies in the Western major scale) were identical or different. The difficulty of this task lies in the fact that the all the notes of the second stimulus pattern were transposed 4 semitones higher in pitch (in both the “same” and “different” trials). In “different” trials, one note was altered by 1 semitone to a pitch outside the pattern's new key, maintaining the melodic contour. This task therefore requires that the listener compares the pattern of pitch intervals (frequency ratios) between each successive tone, and as such requires a more abstract, relational type of processing. Image acquisition and cortical thickness measurements T1-weighted magnetization-prepared rapid gradient-echo images (time echo = 2.98 ms, time repetition = 2300 ms, matrix size: 256 × 256, FOV = 256 mm, slice thickness = 1 mm, flip angle = 9°, voxel size 1 mm3) were acquired on a Siemens 3 T MRI scanner. All T1 images were then submitted to the CIVET pipeline (version 1.1.9; Ad-Dab'bagh et al., 2006; Zijdenbos et al., 2002). T1 images were registered to the ICBM152 nonlinear sixth generation template with a 12-parameter linear transformation (Collins et al., 1994), RF inhomogeneity-corrected (Sled et al., 1998) and tissue-classified (Tohka et al., 2004; Zijdenbos et al., 1998). Deformable models were then used to create the white/grey matter and grey matter/cerebrospinal fluid interfaces for each hemisphere separately (MacDonald et al., 2000; Kim et al., 2005), resulting in four polygonal mesh surfaces of 40,962 vertices each. Both surfaces were non-linearly aligned to a surface template (Lyttelton et al., 2007) using a 2D registration procedure that improves the anatomical correspondence of vertices in all subjects (Robbins, 2004). From these surfaces, the t-Laplace metric was derived by using the Laplacian method for determining the distance between the white and grey surfaces (Haidar and Soul, 2006; Lerch and Evans, 2005). The thickness data were subsequently blurred using a 20-mm surface-based diffusion blurring kernel in preparation for statistical

196

P. Voss, R.J. Zatorre / NeuroImage 108 (2015) 194–202

analyses (Chung and Taylor, 2004). Diffusion smoothing, unlike the volumetric blurring used in VBM, follows the curvature of the surface and thus respects anatomical boundaries. Twenty millimeters was chosen as the kernel size in order to maximize statistical power while still minimizing false positives (Lerch et al., 2005). Unnormalized, native-space thickness values were used in all analyses owing to the poor correlation between cortical thickness and brain volume (Ad-Dab'bagh et al., 2005; Sowell et al., 2007). This is because normalizing for global brain size, when it has little pertinence to cortical thickness, risks introducing noise and reducing power. Nonetheless, we compared the groups on overall brain and cortical volume to rule out any systematic group differences and did not find any significant differences for brain volume (all p N 0.26) (EB:1259 ± 152 cm3; LB:1239 ± 109 cm3; SC:1332 ± 117 cm3) or for cortical volume (all p N 0.19) (EB:485 ± 77 cm3; LB:470 ± 49 cm3; SC:510 ± 59 cm3). Statistical analysis All statistical analyses were performed using in-house software (Surfstat Matlab toolbox; http://www.math.mcgill.ca/keith/surfstat/) developed at the Montreal Neurological Institute (Worsley et al., 2009). Analyses of anatomical covariance were performed by first selecting a seed vertex of interest from the cortical surface and testing the correlational strength between this and every other vertex along the cortical surface (Lerch et al., 2006). The chosen seed vertex was from the right cuneus (x = 8, y = −73, z = 21; MNI coordinates), an area whose CT is strongly linked to performance in both auditory tasks (Voss and Zatorre, 2012). The resulting statistic gives an indication of the degree to which cortical thickness throughout the brain covaries with that of the seed region across subjects. We used linear interaction models to compare the covariance strength between the EB and SC. These models contained simple terms for seed thickness and group, together with a seed thickness × group interaction term. We also performed the above-mentioned steps with an additional control seed vertex located within the right auditory cortex and whose coordinates were chosen arbitrarily (superior temporal gyrus (STG); (x = 50, y = − 20, z = 5)) to ascertain the specificity of the covariation measures to the occipital cortex. Furthermore, we used additional linear interaction models where we included the auditory behavioural scores in the regression model to ascertain how the covariation between the seed region CT and CT across the rest of the cortex varied as a function of performance in the tasks in the EB. These models contained simple terms for seed thickness and score as well as a parametric interaction

term for seed thickness × score. While all results were thresholded at p b 0.0005 (uncorrected), we also report which of these are significant when controlling for the false discovery rate (FDR) at p b 0.05. Results Cortical covariation analyses The results of the vertex-wise CT covariation analyses with the CT of the seed vertex are displayed in Fig. 1. It is readily visible that the area of significant correlation in the sighted was much more expansive than in the EB, with several foci in frontal and parietal areas (Fig. 1a). In contrast, significant areas in the EB were more constrained and mostly located in occipital and cingulate areas (Fig. 1b). Fig. 2 displays the only area in the EB that showed significantly more covariance with the seed vertex, the left inferior occipital gyrus, than in the sighted (though the threshold in this case had to be lowered to p b 0.001, as no differences were observable at an uncorrected threshold of p b 0.0005). In contrast, several regions showed the reverse trend (and were significant when controlling for FDR at p b 0.05), being more strongly coupled in the sighted than in the EB (Fig. 3; see also Table 2 for a complete listing of areas). Notably, several areas, that appear to overlap with regions known to be part of the visual dorsal stream, were found to be more strongly coupled with the occipital seed in the sighted. These include an area of the middle frontal gyrus overlapping with the frontal eye field (FEF) (see Lobel et al., 2001; see also discussion), an area of the left superior frontal gyrus overlapping with the left supplementary eye field (SEF) (see Sharika et al., 2013; Neggers et al., 2012), and the left caudal inferior parietal lobule (IPL). Moreover, the area of the primary motor cortex linked with eye muscle proprioception and motor representations control (see Balslev et al., 2011) also showed higher covariance in the sighted. Fig. 4 plots the relationship between the CT of the FEF, SEF and IPL with that of the occipital seed and shows that while there are strong positive correlations in the sighted, the correlation coefficients are near zero in the EB. Moreover, for comparison purposes, we have also plotted the relationship for the LB group for whom the correlation resembles that of the sighted, which suggests that despite being blind for many years, the presence of eyesight during early brain development is likely the key factor underlying the strong covariations in CT between these areas. Finally, when the vertex-wise CT covariation analyses were performed with a control seed vertex located in the right STG (x = 50, y = −20, z = 5), no group significant group differences were observed in either direction (all p N 0.001).

Fig. 1. Group covariation analyses. Illustrated here are the brain where CT was shown to significantly covary with the CT of the seed region in the right cuneus for both the EB (left panel) and SC (right panel).

P. Voss, R.J. Zatorre / NeuroImage 108 (2015) 194–202

197

Fig. 2. Group contrast (EB N SC). Illustrated here is the only brain region whose CT significantly covaried more with the seed region (illustrated with a red X) in the EB than in the sighted. IOG = inferior occipital gyrus. Images were thresholded at p b 0.001 (uncorrected).

Cortical covariation analyses with behaviour The results obtained from the behavioural tasks (group contrasts and regression analyses between CT and behaviour) have been previously reported elsewhere (Voss and Zatorre, 2012), and showed that the EB were significantly better at performing both tasks than the sighted. The behavioural data of the EB were used in the current study to understand the relationship between cortical anatomical covariance and behavioural abilities. Fig. 5 highlights the areas in the EB where CT significantly co-varied with that of the occipital seed as a function of behavioural performance on the two auditory tasks (left: pitch; right: transposed melody task; a complete list of foci can be found in Table 2). Only one significant focus (p b 0.0005 uncorrected) was found for the pitch task and was located in the right intraparietal sulcus (IPS); this finding indicates that greater anatomical covariance between the IPS and the occipital seed is associated with better performance. The

analysis using the transposed melody performance scores yielded several significant foci (p b 0.0005 uncorrected), including the left intraparietal sulcus (IPS) and cingulate areas (see Table 2).

Discussion The primary purpose of the present study was to investigate how early visual deprivation affects the cortical anatomical covariance between occipital cortex and the rest of the brain, revealing changes to structural networks. To achieve this goal, we used an interregional correlational analysis technique that reveals how the thickness of a region of interest covaries with thicknesses across the entire cortex (Lerch et al., 2006). The result allows one to inspect the degree to which areas covary in thickness across subjects and, thereby, serves as an indication of possible structural and functional interdependence.

Fig. 3. Group contrast (SC N EB). Illustrated here are the brain regions whose CT significantly covaried more with the seed region in the sighted than in the EB. Note that the brain surfaces are tilted to better highlight the significant foci. FEF = frontal eye field; SEF = supplementary eye field; IPL = inferior parietal lobule; PoG = postcentral gyrus; PrG = precentral gyrus. Images were thresholded at p b 0.0005 (uncorrected).

198

P. Voss, R.J. Zatorre / NeuroImage 108 (2015) 194–202

Table 2 List of significant foci identified (p b 0.0005, uncorrected) for each contrast or regression analysis. EB: early blind. SC: sighted controls. FEF: frontal eye fields. SEF: supplementary eye fields. + indicates below the pre-established uncorrected threshold of p = 0.005. * indicates significant when corrected using FDR (p = 0.05). List of significant foci

EB N SC SC N EB

Cortical covariation with pitch task (EB) Cortical covariation with transposed melody task (EB)

Region

Coordinates

LH inferior occipital gyrus/fusiform gyrus LH inferior parietal lobule LH middle frontal gyrus (FEF) RH postcentral gyrus RH postcentral gyrus LH superior frontal gyrus (SEF) RH precentral gyrus RH inferior frontal gyrus LH precentral gyrus RH inferior parietal lobule LH postcentral gyrus RH intraparietal sulcus LH cingulate gyrus LH cingulate gyrus LH intraparietal sulcus RH inferior parietal gyrus RH middle frontal gyrus LH inferior frontal gyrus

We selected a region of interest in the right cuneus because we had previously shown that its cortical thickness correlated with auditory task behaviour in the EB (Voss and Zatorre, 2012). Reduced covariance with the seed region was observed in the EB compared to sighted controls (SC) in several parietal and frontal regions (Figs. 1 and 3) that are normally associated with the dorsal visual stream and the oculomotor system (see Kravitz et al., 2011). The specificity of the covariation findings was ascertained by showing that no significant group differences were found when a control seed vertex located in the auditory cortex was selected. The location of the foci within the left primary motor cortex is highly consistent with the location of the motor and proprioceptive neural representations of the eye muscles (Balslev et al., 2011). The location of the focus lying on the superior frontal gyrus is consistent with the location of the SEF previously reported in the literature (Lobel et al., 2001; Neggers et al., 2012; Sharika et al., 2013), though positioned slightly anterior to other reports (Grosbras et al., 1999). Similarly, the location of the area labelled FEF here (along the middle frontal gyrus) is consistent with some previous functional neuroimaging studies (Lobel et al., 2001) but positioned slightly anterior to other reports (Neggers et al., 2012; Chica et al., 2013). Several reasons could account for this slight discrepancy along the rostral–caudal axis. First, several meta-analyses have revealed the extensive variability that exists in the reported coordinates for both regions between studies (SEF: Grosbras et al., 1999: FEF: Paus, 1996), and their exact location still remains rather controversial (see Amiez and Petrides, 2009), possibly a reflection of significant inter-individual variability as to their specific anatomical locations. Moreover, it should be stated that SEF and FEF are functionally defined regions, and are difficult to reliably identify through anatomical landmarks. Nonetheless, the identified frontal foci remain overall quite consistent with the functionally defined anatomical locations of SEF and FEF, and their relative cortical decoupling with the occipital cortex in the EB is in line with our initial predictions within the context of blindness. Finally, the left IPL was also shown to have higher covariance with the occipital seed region in the controls compared to the EB. The IPL has been suggested to be part of a network supporting a general preparatory state for the perception of visual targets (Chica et al., 2013). It also has extensive connections with both the occipital cortex, the FEF and the posterior parietal cortex; the latter in turn projects extensively to the SEF (Kravitz et al., 2011). Interestingly, the scatterplots presented in Fig. 4 illustrate that the LB do not show a reduction in covariance between the occipital seed and the frontal oculomotor structures. On the contrary, the correlation coefficients observed in the LB are highly similar to those observed in the SC,

x

y

z

−36 −37 −37 19 46 −15 60 45 −48 40 −39 33 −5 −9 −31 49 42 −34

−73 −68 9 −31 −16 19 3 31 −6 −71 −29 −59 −16 43 −58 −62 1 28

−13 45 56 71 51 61 22 −2 48 41 64 50 41 −1 46 37 54 5

T score

p-Value

3.36 4.31 4.23 4.13 4.11 4.05 4.02 3.88 3.82 3.88 3.71 4.46 5.11 4.98 4.89 4.82 4.70 4.70

b0.001+ b0.00005* b0.0001* b0.0001* b0.0001* b0.0005* b0.0005* b0.0005* b0.0005* b0.0005* b0.0005* b0.0005 b0.0005 b0.0005 b0.0005 b0.0005 b0.0005 b0.0005

indicating that the decoupling observed in the EB is likely the result of early visual deprivation and not simply of a prolonged period of blindness, as both blind groups had long durations of blindness (38 years for the EB and 29 for the LB). Finally, the opposite contrast (EB N SC), revealed that stronger covariance with the seed region (right cuneus) in the EB was only observed in a region located near the boundary of the left inferior occipital gyrus and fusiform gyrus (Fig. 2). While the functional significance of this increased covariation is unclear at this time, it is consistent with a previous finding showing that both regions (right cuneus and left fusiform gyrus) were the two throughout the entire brain where CT is most strongly correlated with duration of blindness (Voss and Zatorre, 2012), and thus might simply be the result of parallel increases in CT in both regions as function of time spent without visual input. Is the decreased covariation with occipital cortex specific to the left fronto-parietal network? An interesting and unexpected finding was that the decoupling between the seed region and areas within the fronto-parietal stream in the EB was only observed contralaterally to the seed region (i.e. in the left hemisphere). To ascertain if this effect might be specific to a rightsided seed region, we repeated the original covariation analysis using a seed in the left cuneus (identical y and z coordinates with a mirrored x coordinate). The pattern of results was nearly identical to those found for the original analysis, though the strength of the covariations was slightly weaker (see Fig. 6), which further confirms that the cortical decoupling with the deafferented occipital cortex occurs only with the fronto-parietal network in the left hemisphere. This finding raises an obviously important question as to why this is the case. Part of the answer may come from the rich line of research investigating auditory spatial processing in the EB. Many studies show that the right dorsal stream is heavily reorganized and recruited to carry out spatial auditory tasks (see Collignon et al., 2009), and that activity from within the occipital cortex is primarily functionally connected with regions of the right frontal and parietal cortex that are typically involved in spatial awareness and attention (Collignon et al., 2011; Voss et al., 2011). In contrast, components of the left visual dorsal stream seem to be involved in language processing (Burton et al., 2002; Bedny et al., 2011), and congenitally blind individuals have increased functional connectivity between the occipital cortex and left frontal language areas (Bedny et al., 2011), as opposed to left frontal areas that are part of the dorsal visual stream.

P. Voss, R.J. Zatorre / NeuroImage 108 (2015) 194–202

199

Fig. 4. Covariation plots. Here is plotted the relationship between the CT of the occipital and the CT of the FEF (top), SEF (middle), and IPL (bottom). In addition to the EB (red) and the SC (blue), we have plotted the same relationship obtained from our LB sample (green) for comparison purposes. Interestingly, the LB show a pattern of results quite similar to the SC, indicating that the decoupling observed in the EB is a likely the result of early visual deprivation and not simply of a prolonged period of blindness. Images were thresholded at p b 0.0005 (uncorrected).

Does covariation mean connectivity? Whether cortical covariance is more akin to functional connectivity than to anatomical connectivity is not without debate, though there is precedent to believe that the latter scenario could be true (Bullmore and Sporns, 2009; Sporns et al., 2005). Brain areas that are highly correlated in size are often part of systems that are known to subserve particular behavioural or cognitive functions (see Alexander-Bloch et al., 2013). With regards to structural covariance analyses, CT measures in particular have been proposed as a valid covariate because they provide a metric that reflects the size, density and arrangement of neurons in a biological and topological meaningful way (Parent and

Carpenter, 1995). Importantly, previous morphometric correlations of cortical thickness data have successfully reproduced structural networks of covariance that resemble tract-tracing data (Mitelman et al., 2005; Lerch et al., 2006; Bernhardt et al., 2008; Gong et al., 2012). Furthermore animal tracer studies have additionally shown that cortical thickness can predict anatomical connectivity (Barbas, 1986; Barbas and Rempel-Clower, 1997; Dombrowski et al., 2001). Should CT covariance indeed characterize the level of connectivity (whether it been anatomical or functional in nature), such connectivity could be induced by mutually trophic, developmental, and maturational influences. For instance, synapses between neurons can have a mutually trophic and protective effect on subsequent neuronal development, possibly via

200

P. Voss, R.J. Zatorre / NeuroImage 108 (2015) 194–202

Fig. 5. Covariation as a function of behaviour. Illustrated here are the results of the covariation analyses as a function of the behavioural performance of the EB on the two auditory tasks. The image on the left illustrates that the strength of the covariation between right IPS' CT and the occipital seed's CT was highly predictive of the performance on the pitch task (i.e. the stronger the cortical coupling, the better the performance). A similar finding was found for the transposed melody task (right image), though the influential coupling was between the left IPS and the occipital seed. Note that the brain surface on the right was tilted to better highlight the area of significant coupling. Images were thresholded at p b 0.0005 (uncorrected).

glutamatergic NMDA pathways (Burgoyne et al., 1993), with large numbers of such connections possibly leading to co-variance at the macroanatomic level (Bullmore et al., 1997). Synchronous firing can induce synaptogenesis between neurons (Katz and Shatz, 1996; Bi and Poo, 1999), which suggests the possibility of use-dependent coordinated growth. Within the context of blindness, crossmodal experience-dependant plasticity could constitute the driving force behind corticocortical connectivity between particular brain regions and associated covariance measures. Cortical covariation as a behaviourally relevant metric Another important goal of the present study was to use the CT covariation analyses to assess how cortical covariance relates to behaviour. Performance in the transposed melody task was shown to vary as a function of the covariance strength between the occipital seed and the left IPS (Fig. 5). This result is of particular significance, as the CT of the IPS has been previously shown to be highly predictive of performance in the same task in sighted individuals (Foster and Zatorre, 2010a), suggesting that this region has likely maintained a functional role in performing the task in the blind, albeit dependent on its interactions with the right cuneus, rather than auditory cortices. In addition, the activity of the IPS has also previously been linked to performance

on the transposed melody task, providing a high level of consistency across studies (Foster and Zatorre, 2010b; Foster et al., 2013). The IPS is a multimodal region with a role in performing systematic transformations on sensory representations, (Grefkes and Fink, 2005; Zacks, 2008) and abstract operations like arithmetic (Kong et al., 2005), thus providing an ideal neural apparatus for transforming melodic information. Our similar finding with the right IPS and performance on the basic pitch discrimination task (Fig. 5) was somewhat more surprising. However, although the IPS is not typically associated with basic processing of pitch, it is not undocumented (Zatorre et al., 1994; Merrill et al., 2012). In addition, previous work has shown that the IPS' close anatomical neighbours, the inferior and superior parietal lobules, to be involved in pitch processing (see Koelsch and Siebel, 2005), and in particular the supramarginal gyrus (Gaab et al., 2003). Furthermore, the IPS receives converging anatomical inputs from visual, auditory, and tactile sensory cortices (Frey et al., 2008 and Schroeder and Foxe, 2002), suggesting that it might also just act as a relay point that feeds auditory input into occipital cortex. Summary and conclusion Overall, the present findings provide compelling evidence that investigating anatomical covariance can be an effective way to characterize

Fig. 6. Group contrast (SC N EB) with left occipital seed. Illustrated here is the result of the same analysis depicted in Fig. 3 but with a different occipital seed. To verify if the decoupling effect in the EB was specific to the left hemisphere or is contralateral to the seed region, we used this time a seed in the left cuneus (identical y and z coordinates of the original seed with a mirrored x coordinate). This new seed produced an extremely similar pattern of decoupling in the EB (relative to the sighted) as did the original seed in the right hemisphere, albeit the decoupling was slightly weaker here. Images were thresholded at p b 0.001 (uncorrected).

P. Voss, R.J. Zatorre / NeuroImage 108 (2015) 194–202

individual (behavioural) differences and plasticity. We have shown that the degree to which the CT of two regions covary with one another can be a marker of crossmodal reorganization, and highly predictive of a behavioural outcome. Future work will be required to ascertain how this cortical covariance relates to functional connectivity between the regions and to determine if the actual wiring between them corresponds to the cortical covariance findings. Acknowledgments We thank all the individuals that volunteered to participate in this study as well as the INLB for its assistance in recruiting blind participants. We thank Alan Evans for making the CIVET pipeline available to us. We thank Marc Bouffard and Patrick Bermudez for their assistance in running many of the image pre-processing steps. We also would like to thank the staff at the McConnell Brain Imaging Centre for their assistance during the scanning procedures. This research was funded by the Canadian Institutes of Health Research, the Canada Fund for Innovation, and the Fonds de Recherche Nature et Technologies/Société et Culture via its funding of the Centre for Research in Brain, Language and Music. P.V. was supported by a Banting Postdoctoral Fellowship, administered by the Natural Science and Engineering Research Council (NSERC) of Canada. Conflict of interest None. References Ad-Dab'bagh, Y., et al., 2005. Native space cortical thickness measurement and the absence of correlation to cerebral volume. In: Zilles, K. (Ed.), Proceedings of the 11th Annual Meeting of the Organization for Human Brain Mapping. NeuroImage, Toronto. Ad-Dab'bagh, Y., et al., 2006. The CIVET image-processing environment: a fully automated comprehensive pipeline for anatomical neuroimaging research. In: Corbetta, M. (Ed.), Proceedings of the 12th Annual Meeting of the Organization for Human Brain Mapping. NeuroImage, Florence, Italy. Alexander-Bloch, A., Giedd, J.N., Bullmore, E., 2013. Imaging structural covariance between human brain regions. Nat. Rev. Neurosci. 14, 322–336. Amiez, C., Petrides, M., 2009. Anatomical organization of the eye fields in the human and non-human primate frontal cortex. Prog. Neurobiol. 89, 220–230. Anurova, I., Renier, L.A., De Volder, A.G., Carlson, S., Rauschecker, J.P., 2014. Relationship between cortical thickness and functional activation in the early blind. Cereb. Cortex (Epub ahead of print). Balslev, D., Albert, N.B., Miall, C., 2011. Eye muscle proprioception is represented bilaterally in the sensorimotor cortex. Hum. Brain Mapp. 32, 624–631. Barbas, H., 1986. Pattern in the laminar origin of corticocortical connections. J. Comp. Neurol. 252, 415–422. Barbas, H., Rempel-Clower, N., 1997. Cortical structure predicts the pattern of corticocortical connections. Cereb. Cortex 7, 635–646. Bedny, M., Pascual-Leone, A., Dodell-Feder, D., Fedorenko, E., Saxe, R., 2011. Language processing in the occipital cortex of congenitally blind adults. Proc. Natl. Acad. Sci. U. S. A. 108, 4429–4434. Bernhardt, B.C., Worsley, K.J., Besson, P., Concha, L., Lerch, J.P., Evans, A.C., Bernasconi, N., 2008. Mapping limbic network organization in temporal lobe epilepsy using morphometric correlations: insights on the relation between mesiotemporal connectivity and cortical atrophy. Neuroimage 42, 515–524. Bi, G., Poo, M., 1999. Distributed synaptic modification in neural networks induced by patterned stimulation. Nature 401, 792–796. Bohbot, V.D., Lerch, J., Thorndycraft, B., Iaria, G., Zijenbos, A.P., 2007. Gray matter differences correlate with spontaneous strategies in a human virtual navigation task. J. Neurosci. 27, 10078–10083. Bridge, H., Cowey, A., Ragge, N., Watkins, K., 2009. Imaging studies in congenital anophthalmia reveal preservation of brain architecture in ‘visual’ cortex. Brain 132, 3467–3480. Bullmore, E.T., Frangou, S., Murray, R.M., 1997. The dysplastic net hypothesis: an integration of developmental and dysconnectivity theories of schizophrenia. Schizophr. Res. 28, 143–156. Bullmore, E.T., Woodruff, P.W., Wright, I.C., Rabe-Hesketh, S., Howard, R.J., Shuriquie, N., Murray, R.M., 1998. Does dysplasia cause anatomical dysconnectivity in schizophrenia? Schizophr. Res. 30, 127–135. Bullmore, E., Sporns, O., 2009. Complex brain networks: graph theoretical analysis of structural and functional systems. Nat Rev Neurosci. 10 (3), 186–198. Burgoyne, R.D., Graham, M.E., Cambray-Deakin, M., 1993. Neurotrophic effects of NMDA receptor activation on developing cerebellar granule cells. J. Neurocytol. 22, 689–695.

201

Burton, H., Snyder, A.Z., Diamond, J.B., Raichle, M.E., 2002. Adaptive changes in early and late blind: a FMRI study of verb generation to heard nouns. J Neurophysiol. 88 (6), 3359–3371. Chica, A.B., Paz-Alonso, P.M., Valero-Cabré, A., Bartolomeo, P., 2013. Neural bases of the interactions between spatial attention and conscious perception. Cereb. Cortex 23, 1269–1279. Chung, M.K., Taylor, J., 2004. Diffusion smoothing on brain surface via finite element method. 2004 2nd IEEE Int Symp Biomed Imaging Nano to Macro (IEEE Cat No 04EX821). Collignon, O., Voss, P., Lassonde, M., Lepore, F., 2009. Cross-modal plasticity for the spatial processing of sounds in visually deprived subjects. Exp. Brain Res. 192, 343–358. Collignon, O., Vandewalle, G., Voss, P., Albouy, G., Charbonneau, G., Lassonde, M., Lepore, F., 2011. Functional specialization for auditory-spatial processing in the occipital cortex of congenitally blind humans. Proc. Natl. Acad. Sci. U. S. A. 108, 4435–4440. Collins, D.L., Neelin, P., Peters, T.M., Evans, A.C., 1994. Automatic 3d inter-subject registration of mr volumetric data in standardized Talairach space. J. Comput. Assist. Tomogr. 18, 192–205. Dombrowski, S.M., Hilgetag, C.C., Barbas, H., 2001. Quantitative architecture distinguishes prefrontal cortical systems in the rhesus monkey. Cereb. Cortex 11, 975–988. Foster, N.E.V., Zatorre, R.J., 2010a. Cortical structure predicts success in performing musical transformation judgments. NeuroImage 53, 26–36. Foster, N.E.V., Zatorre, R.J., 2010b. A role for the intraparietal sulcus in transforming musical pitch information. Cereb. Cortex 20, 1350–1359. Foster, N.E.V., Halpern, A., Zatorre, R.J., 2013. Common parietal activations in musical mental transformations across pitch and time. NeuroImage 75, 27–35. Frey, S., Campbell, J.S.W., Pike, G.B., Petrides, M., 2008. Dissociating the human language pathways with high angular resolution diffusion fiber tractography. J. Neurosci. 28, 11435–11444. Gaab, N., Gaser, C., Zaehle, T., Jancke, L., Schlaug, G., 2003. Functional anatomy of pitch memory — an fMRI study with sparse temporal sampling. Neuroimage 19, 1417–1426. Gong, G., He, Y., Chen, Z.J., Evans, A.C., 2012. Convergence and divergence of thickness correlations with diffusion connections across the human cerebral cortex. NeuroImage 59, 1239–1248. Grefkes, C., Fink, G.R., 2005. The functional organization of the intraparietal sulcus in humans and monkeys. J. Anat. 207, 3–17. Grosbras, M.H., Lobel, E., Van de Moortele, P.F., LeBihan, D., Berthoz, A., 1999. An anatomical landmark for the supplementary eye fields in human revealed with functional magnetic resonance imaging. Cereb. Cortex 9, 705–711. Haidar, H., Soul, J.S., 2006. Measurement of cortical thickness in 3D brain MRI data: validation of the Laplacian method. J. Neuroimaging 16, 146–153. Jiang, J., Zhu, W., Shi, F., Liu, Y., Li, J., Qin, W., Li, K., Yu, C., Jiang, T., 2009. Thick visual cortex in the early blind. J. Neurosci. 29, 2205–2211. Katz, L.C., Shatz, C.J., 1996. Synaptic activity and the construction of cortical circuits. Science 274, 1133–1138. Kim, J.S., Singh, V., Lee, J.K., Jerch, J., Ad-Dab'bagh, Y., MacDonald, D., Lee, J.M., Kim, S.I., Evans, A.C., 2005. Automated 3-D extraction and evaluation of the inner and outer cortical surfaces using a Laplacian map and partial volume effect classification. Neuroimage 27, 210–221. Koelsch, S., Siebel, W.A., 2005. Towards a neural basis of music perception. Trends Cogn. Sci. 9, 578–584. Kong, J., Wang, C., Kwong, K., Vangel, M., Chua, E., Gollub, R., 2005. The neural substrate of arithmetic operations and procedure complexity. Cogn. Brain Res. 22, 397–405. Kravitz, D.J., Saleem, K.S., Baker, C.I., Mishkin, M., 2011. A new neural framework for visuospatial processing. Nat. Rev. Neurosci. 12, 217–230. Lee, N.R., Raznahan, A., Wallace, G.L., Alexander-Bloch, A., Clasen, L.S., Lerch, J.P., Giedd, J.N., 2013. Anatomical coupling among distributed cortical regions in youth varies as a function of individual differences in vocabulary abilities. Hum. Brain Mapp. http://dx.doi.org/10.1002/hbm.22299. Lerch, J.P., Evans, A.C., 2005. Cortical thickness analysis examined through power analysis and a population simulation. NeuroImage 24, 163–173. Lerch, J.P., Pruessner, J.C., Zijdenbos, A., Hampel, H., Teipel, S.J., Evans, A.C., 2005. Focal decline of cortical thickness in Alzheimer's disease identified by computational neuroanatomy. Cereb. Cortex 15, 995–1001. Lerch, J.P., Worsley, K., Shaw, W.P., Greenstein, D.K., Lenroot, R.K., Giedd, J., Evans, A.C., 2006. Mapping anatomical correlations across cerebral cortex (MACACC) using cortical thickness from MRI. Neuroimage 31, 993–1003. Levitt, H., 1970. Transformed up–down methods in psychoacoustics. J. Acoust. Soc. Am. 49, 467–477. Lobel, E., Kahane, P., Leonards, U., Grosbras, M., Lehéricy, S., Le Bihan, D., Berthoz, A., 2001. Localization of human frontal eye fields: anatomical and functional findings of functional magnetic resonance imaging and intracerebral electrical stimulation. J. Neurosurg. 95, 804–815. Lyttelton, O., Boucher, M., Robbins, S., Evans, A., 2007. An unbiased iterative group registration template for cortical surface analysis. NeuroImage 3, 1535–1544. MacDonald, D., Kabani, N., Avis, D., Evans, A.C., 2000. Automated 3-D extraction of inner and outer surfaces of cerebral cortex from MRI. NeuroImage 12, 340–356. Mechelli, A., Friston, K.J., Frackowiak, R.S., Price, C.J., 2005. Structural covariance in the human cortex. J. Neurosci. 25, 8303–8310. Merabet, L.B., Pascual-Leone, A., 2010. Neural reorganization following sensory loss: the opportunity of change. Nat. Rev. Neurosci. 11, 44–52. Merrill, J., Sammler, D., Bangert, M., Goldhahn, D., Lohmann, G., Turner, R., Friederici, A.D., 2012. Perception of words and pitch patterns in song and speech. Front. Psychol. 3. Mitelman, S.A., Buchsbaum, M.S., Brickman, A.M., Shihabuddin, L., 2005. Cortical intercorrelations of frontal area volumes in schizophrenia. Neuroimage 27, 753–770.

202

P. Voss, R.J. Zatorre / NeuroImage 108 (2015) 194–202

Neggers, S.F.W., van Diepen, R.M., Zandbelt, B.B., Vink, M., Mandl, R.C.W., Gutteling, T.P., 2012. A functional and structural investigation of the human fronto-basal volitional saccade network. PLoS ONE 7, e29517. Parent, A., Carpenter, M.B., 1995. Human neuroanatomy. Williams & Wilkins, Baltimore (MD). Park, H.J., Lee, J.D., Kim, E.Y., Park, B., Oh, M.K., Lee, S.C., Kim, J.J., 2009. Morphological alterations in the congenital blind based on the analysis of cortical thickness and surface area. NeuroImage 47, 98–106. Paus, T., 1996. Location and function of the human frontal eye-field: a selective review. Neuropsychologia 34, 475–483. Robbins, S.M., 2004. Anatomical Standardization of the Human Brain in Euclidean 3-Space and on the Cortical 2-Manifold. Ph.D. Thesis. School of Computer Science. McGill University, Montreal. Schroeder, C.E., Foxe, J.J., 2002. The timing and laminar profile of converging inputs to multisensory areas of the macaque neocortex. Cognitive Brain Research, pp. 187–198. Sharika, K.M., Neggers, S.F.W., Gutteling, T.P., Van der Stigchel, S., Dijkerman, H.C., Murthy, A., 2013. Proactive control of sequential saccades in the human supplementary eye field. Proc. Natl. Acad. Sci. U. S. A. 110, E1311–E1320. Sled, J.G., Zijdenbos, A.P., Evans, A.C., 1998. A non-parametric method for automatic correction of intensity non-uniformity in MRI data. IEEE Trans. Med. Imaging 17, 87–97. Sowell, E.R., Peterson, B.S., Kan, E., Woods, R.P., Yoshii, J., Bansal, R., Xu, D., Zhu, H., Thompson, P.M., Toga, A.W., 2007. Sex differences in cortical thickness mapped in 176 healthy individuals between 7 and 87 years of age. Cereb. Cortex 17, 1550–1560. Sporns, O., Tononi, G., Kötter, R., 2005. The human connectome: a structural description of the human brain. PLoS Comput. Biol. 1, 0245–0251. Tohka, J., Zijdenbos, A., Evans, A.C., 2004. Fast and robust parameter estimation for statistical partial volume models in brain MRI. NeuroImage 23, 84–97.

Voss, P., Zatorre, R.J., 2012. Occipital cortical thickness predicts performance on pitch and musical tasks in blind individuals. Cereb. Cortex 22, 2455–2465. Voss, P., Collignon, O., Lassonde, M., Lepore, F., 2010. Adaptation to sensory loss. Wiley Interdiscip. Rev. Cogn. Sci. 1, 308–328. Voss, P., Lepore, F., Gougoux, F., Zatorre, R.J., 2011. Relevance of spectral cues for auditory spatial processing in the occipital cortex of the blind. Front. Psychol. 2, 48. Worsley, K.J., Taylor, J.E., Carbonell, F., Chung, M.K., Duerden, E., Bernhardt, B., Lyttelton, O., Boucher, M., Evans, A.C., 2009. SurfStat: a Matlab toolbox for the statistical analysis of univariate and multivariate surface and volumetric data using linear mixed effects models and random field theory. NeuroImage 47, S102. Zacks, J.M., 2008. Neuroimaging studies of mental rotation: a meta-analysis and review. J. Cogn. Neurosci. 20, 1–19. Zatorre, R.J., Evans, A.C., Meyer, E., 1994. Neural mechanisms underlying melodic perception and memory for pitch. J. Neurosci. 14, 1908–1919. Zatorre, R.J., Fields, R.D., Johansen-Berg, H., 2012. Plasticity in gray and white: neuroimaging changes in brain structure during learning. Nat. Neurosci. 15, 528–536. Zijdenbos, A., Forghani, R., Evans, A., 1998. Automatic quantification of MS lesions in 3D MRI brain data sets: validation of INSECT. In: Wells, W.M., Colchester, A., Delp, S. (Eds.), Proceeding of the Medical Image Computing and Computer-Assisted Interventation MICCAI98. Springer-Verlag Berlin Heidelberg, Cambridge, MA, pp. 439–448. Zijdenbos, A., Forghani, R., Evans, A., 2002. Automatic pipeline analysis of 3-D MRI data for clinical trials: application to multiple sclerosis. IEEE Trans. Med. Imaging 21, 1280–1291.