Brain morphometry in blind and sighted subjects

Journal of Clinical Neuroscience xxx (2016) xxx–xxx

Contents lists available at ScienceDirect

Journal of Clinical Neuroscience journal homepage: www.elsevier.com/locate/jocn

Clinical Study

Brain morphometry in blind and sighted subjects Jerome J. Maller a,⇑, Richard H. Thomson a, Amanda Ng b, Collette Mann c, Michael Eager d, Helen Ackland e, Paul B. Fitzgerald a, Gary Egan f,g, Jeffrey V. Rosenfeld h,i,j,k a

Monash Alfred Psychiatry Research Centre, The Alfred & Monash University Central Clinical School, 607 St Kilda Rd, Melbourne, VIC 3181, Australia Howard Florey Institute, University of Melbourne, VIC, Australia c Monash Vision Group, Department of Electrical and Computer Systems Engineering, Monash University, Melbourne, VIC, Australia d Monash Biomedical Imaging and Monash e-Research, Monash University, Melbourne, VIC, Australia e National Trauma Research Institute, The Alfred hospital and Department of Epidemiology and Preventive Medicine, Monash University, Melbourne, VIC, Australia f Monash Biomedical Imaging, Monash University, Clayton, VIC, Australia g School of Psychology & Psychiatry, Monash University, Melbourne, VIC, Australia h Division of Clinical Sciences and Department of Surgery, Central Clinical School, Monash University, VIC, Australia i Department of Neurosurgery, Alfred Hospital, Melbourne, VIC, Australia j Monash Institute of Medical Engineering, Melbourne, VIC, Australia k F. Edward Hebert School of Medicine, Uniformed Services University of the Health Sciences, Bethesda, MD, USA b

a r t i c l e

i n f o

Article history: Received 19 October 2015 Accepted 17 January 2016 Available online xxxx Keywords: Blindness Magnetic resonance imaging Morphometry Visual cortex

a b s t r a c t Previous neuroimaging studies have demonstrated structural brain alterations in blind subjects, but most have focused on primary open angle glaucoma or retinopathy of prematurity, used low-field scanners, a limited number of receive channels, or have presented uncorrected results. We recruited 10 blind and 10 age and sex-matched controls to undergo high-resolution MRI using a 3T scanner and a 32-channel receive coil. We evaluated whole-brain morphological differences between the groups as well as manual segmentation of regional hippocampal volumes. There were no hippocampal volume differences between the groups. Whole-brain morphometry showed white matter volume differences between blind and sighted groups including localised larger regions in the visual cortex (occipital gyral volume and thickness) among those with blindness early in life compared to those with blindness later in life. Hence, in our patients, blindness resulted in brain volumetric differences that depend upon duration of blindness. Ó 2016 Elsevier Ltd. All rights reserved.

1. Introduction Whilst brain structure has been extensively examined in normal subjects, relatively few studies have investigated patients with blindness (Table 1). Many of these studies have reported regions of increased size or thickness in blind subjects, whilst others have found the opposite. For example, using a whole brain voxelbased morphometry approach, Chen and colleagues [1] reported that a bilateral advanced glaucoma study group (N = 15) compared to a control group (N = 15) showed significantly decreased gray matter volume in widespread regions such as the lingual gyrus, calcarine gyrus, postcentral gyrus, and an array of frontal regions; gray matter volume was significantly larger in the study group than in the control group in occipital and parietal gyri, and left occipital gyri. Zikou and colleagues [2] found reduction in the left visual cortex volume, left LGN, and chiasm in those with POAG

(N = 18). Similarly, Yu and colleagues [3] found POAG patients (N = 36) to show significant bilateral cortical thinning in the anterior half of the visual cortex around the calcarine sulci (left BA17 and BA18, right BA17) and in some smaller regions located in the left middle temporal gyrus (BA37) and fusiform gyrus (BA19) when compared to 40 matched controls. In a smaller sample, Bridge and colleagues [4] reported the banks of the calcarine sulci to be significantly thicker in six anophthalmic subjects compared with controls. The authors hypothesized that a lack of visual input in these patients from an early age may have prevented axonal and synaptic pruning in this region hence resulting in thicker cortex. Other groups have also found increased cortical thickness in these regions in early/congenital blind subjects [5–8]. As Table 1 indicates, field strength, number of receive channels, and voxel dimensions have been inconsistent in morphometric MRI studies of blindness, as have the analysis methods employed. Cause of blindness has also been heterogeneous. All of these factors prevent strong conclusions from being formed.

⇑ Corresponding author. Tel.: +61 3 9076 2404; fax: +61 3 9076 8545. E-mail address: [email protected] (J.J. Maller). http://dx.doi.org/10.1016/j.jocn.2016.01.040 0967-5868/Ó 2016 Elsevier Ltd. All rights reserved.

Please cite this article in press as: Maller JJ et al. Brain morphometry in blind and sighted subjects. J Clin Neurosci (2016), http://dx.doi.org/10.1016/j. jocn.2016.01.040

10 (5:5), 23–48, X = 38.6 15 (9:6), 40–50, 43.3 ± 4.1

26 (21:5), 25–58, X = 35.4

12 EB (9:3), 19–55, X = 33.8 7 LB (4:3), 22–57, X = 33.9 10 (3:7), 52–76, X = 63.1 ± 7.7

Chebat (2007)

Dai (2011)

Fortin (2008)

11 (6:5), 6–45, X = 23

14 (7:7), 38–58, X = 47.1 ± 5.8

21 EB (16:5), 17–36, X = 27.1 ± 5.4 12 LB (7:5), 18–38, X = 28.8 ± 6.6 11 (6:5), 22–68, X = 33.0

Manara (2014)

Pan (2007)

Park (2009)

5 (3:2), 27–54, X = 40.4

13 (6:7), 21–70, X = 46

14 EB (10:4), 20–59, X = 38.2 ± 13.8 13 LB (5:8), 29–60, X = 46.6 ± 8.5

Shimony (2006)

Trampel (2011)

Voss (2011)

Ptito (2008)

16 EB (10:6), 19–55, X = 36.2 ± 9.8 16 LB (10:6), 22–56, X = 38.2 ± 10.2)

Lepore (2010)

Lepore (2009)

Jiang (2009)

Gupta (2009)

Chen (2013)

17 EB (9:8), 15–31, X = 22.6 ± 4.0 19 LB (12:7), 18–35), X = 24.2 ± 5.0 22 (14:8), X = 36.9 ± 11.0

12 (5:7), 34–58, X = 49 ± 9.2 6 (4:2), 18–31, X = 24.2 ± 5.2

Anurova (2015)

Bridge (2009)

Clinical N (M:F), age (years)

Study

19 (8:11), 37.6 ± 12.0

15 (8:7), 20–31, X = 25

7 (4:3), 20–56, X = 34.7

21 (11:10), 20–54, X = 35.6

35 (21:14), 22–37, 26.7 ± 4.1

16 (8:8), 42–55, 49.8 ± 4.1

19 (6:5), X = 23 (6–43)

32 (20:12), divided into: 16 (10:6) 22–44, 35.3 ± 9.5 (matched for EB) 16 (10:6), 22–57, X = 38.2 ± 10.3 (matched for LB)

28 (17:11), X = 34.1 ± 10.6

29 (15:14), X = 22.6 ± 3.1

8 (3:5), 46–71, X58.6 ± 10.0

19 (13:6), 19–56, X = 36.0

26 (21:5), 21–57, X = 35.4

15 (9:6), X = 43.9 ± 3.8

10, 24–49, X = 34.0

6 (3:3), X = 27.0 ± 6.0 and 20 (10:10), 20–35, X = 28.0 ± 4.6

12 (5:7), X = 45 ± 10.9

Control N (M:F), age (years)

Table 1 Previous structural brain MRI studies in blind patients

EB and LB

CB

EB

CB

CB and LB

EB

Alstrom Syndrome

EB and LB

NS

EB and LB

POAG

EB and LB

POAG

POAG

EB

Anophthalmia

EB

Cause of blindness

SPM; whole brain

In-house, 3D Slicer, and Analyze; LGN, V1/V2 WM and GM

1.5T, 1 mm  0.94  0.94 mm

1.5T, 48d, Sl = 1  1  1.25 mm reformatted to 2.5 mm isotropic

CIVET

Signal intensity of Stria of Gennari

Freesurfer; whole brain

3T, 1.2  0.98  0.98 mm

7T, 24 channel head coil, 0.5 mm isotropic 3T, 1 mm isotropic

SPM; whole brain

SPM; whole brain

ANIMAL, Multitracer (for CC), BrainSuite (whole brain)

Display (only hippocampus)

FreeSurfer; whole brain

Magicweb (only LGN)

Display (only hippocampus)

Advantage workstation (only LGN)

1.5T, 32d, 1 mm isotropic (reconstructed to 1  0.66  0.66 mm) 1.5T, 1.2  0.94  0.94 mm, Av = 2.

1.5T, 0.98 mm (note: this is NS, but matrix = 256  256 and FOV = 250 mm) 1.5T, 1 mm isotropic

1.5T, 8 channel head coil, 4 mm gap 10 mm, PD: 2 mm and Av = 5 (for LGN measurements) 3T, 1 mm isotropic

3T, 8 channel head coil, 1 mm isotropic, PD: 1.8  0.64  0.64 (for LGN measurements) 1.5T, 1 mm isotropic

SPM

Display (only hippocampus)

1.5T, 1 mm  0.94  0.94 mm 3T, 8 channel head coil, 1 mm isotropic

BrainVoyager QX Cortical thickness; preselected ROIs Whole brain; VBM FSL and FS

Analysis technique

3T, 12 channel head coil, 1 mm isotropic 3T, 12 channel head coil, 1 mm isotropic. PD (for LGN): 2  0.75  0.75 mm

Scanner and T1-weighted parameters

Analyses between EB and LB not corrected for multiple comparisons

; L temporal lobe " EB primary visual cortex (compared with controls or LB) " EB L fusiform (compared to LB) " R hippocampal head ; hippocampal tail

Stria of Gennari is detectable in CB subjects EB " L lingual and R lateral occipital (compared with controls) LB ; lingual, cuneus, L inferior/ middle occipital (compared with controls) LB ; lingual, cuneus, fusiform, lateral occipital (compared with EB)

; GM V1 and cuneus/lingual gyrus ; WM bilateral optic radiation and R anterior temporal lobe CB: " pericalcarine, lateral occipital, lingual ; L somatosensory R auditory cortex (compared to LB and controls) ; whole brain GM and WM ; GM: LGN and R pulvinars, bilateral BA17/18/19, MTG, caudate, posterior Hippocampus, R SFG, R ITG, R lateral orbital, R insular ; WM: splenium CC, ILF, OR, fornix " WM: FOF, SLF, genu CC ;WM V1/V2 (but not GM)

; (EB and LB, but greater in EB) V1/ V2, cingulate, L SMA, PMA, L entorhinal " prefrontal, parietal subcortical WM, cerebellum ; (only EB) CC (splenium, isthmus) ;GM anterior part calcarine cortex and fusiform gyri

1.5T, 8 channel head coil

;L LGN height

Small sample size; 1.5T; T1weighted data reformatted to low resolution 7T

1.5T

1.5T

1.5T

1.5T

1.5T

1.5T

8 channel head coil

1.5T, low intra-class reliabilities (0.730.98) 8 channel head coil; first reduced the number of voxels entering the statistical computation

" hippocampal head (in both EB and LB compared with controls)

; GM lingual, calcarine, postcentral, SFG, IFG, rolandic operculum, inferior occipital, paracentral, supramarginal, cuneus " MTG, IPG, angular, parietal, precuneus, middle occipital ;LGN height and volume

12 channel head coil

" Occipital, frontal STG, superior parietal, anterior cingulate VBM: Nothing when corrected; (uncorrected: " Primary visual cortex) FS: " Thickness and GM volume (corrected) calcarine sulcus, MTG, putamen. WM: ;thalamus, internal capsule, occipital. Smaller LGNs. ; R hippocampal tail 12 channel head coil; very small sample

Comments

Findings

2 J.J. Maller et al. / Journal of Clinical Neuroscience xxx (2016) xxx–xxx

Please cite this article in press as: Maller JJ et al. Brain morphometry in blind and sighted subjects. J Clin Neurosci (2016), http://dx.doi.org/10.1016/j. jocn.2016.01.040

AB = acquired blindness; Av = Averages; BA = Brodmann area; CB = congenitally blind; CC = corpus callosum; EB = early blindness; FOF = fronto-occipital fasciculus; GM = gray matter; IFG = inferior frontal gyrus; ILF = inferior longitudinal fasciculus; ITG = inferior temporal gyrus; L = left; LB = late blindness; LGN-lateral geniculate nucleus; mm = millimetres; MTG = middle temporal gyrus; MTR = magnetization transfer resonance; NS = not stated; OR = optic radiation; PD = proton density; POAG = primary open angle glaucoma; R = right; ROI = Region of interest; SFG = superior frontal gyrus; Sl = slice thickness; SPM = statistical parametric mapping; STG = superior temporal gyrus; T = tesla; VBM = Voxel-based morphometry; WM = white matter; X = mean age; ; = reduced; " = increased. Bolded words represent results which were consistent with those of the current study.

INSECT

POAG (8.30 ± 5.14 years of blindness) 18 (age- and sex-matched; statistics NS)

SPM, whole brain VBM

3T, 8 channel head coil, Sl-1 mm isotropic 1.5T, 1 mm isotropic POAG 40 (29:11), 22–62, X = 46.5

36 (27:9), 18–72, X = 46.5 18 (NS) X = 57.05 ± 11.38 Zikou (2012)

Yu (2013)

3T, 1 mm isotropic EB 30 (stated as 8:7), 17–28, X = 22.5 15 (8:7), 17–30, X = 23.2 Yang (2012)

Premature Wan (2013)

24 (13:11), 21–63, X = 42

16 (NS), 22–53, X = 37

BrainSuite and HAMMER

8 channel head coil; uncorrected results 1.5T

SPM; whole brain 1.5T, 1  0.5  0.5 mm

Uncorrected results

Resampled to 2 mm isotropic; not all results corrected for multiple comparisons

" GM (fusiform, middle occipital, cuneus, lingual) ; GM and WM in EB (compared to LB) No MTR differences after multiplecorrections ; total WM ; GM in cuneus, lingual, middle occipital, precuneus, parietal, thalamus ; WM in cuneus, lingual, cingulate " globus pallidus ; left visual cortex (BA17, 18) " L associated visual cortex (BA19), posterior cingulate (BA23/31), cerebellum ; GM in L calcarine, cuneus, MTG, fusiform, R calcarine ; GM in L calcarine, pulvinar, putamen, precuenus, R caudate, superior parietal CIVET (T1-weighted); Surfstat (MTR) 3T, 1 mm isotropic MTR: 3 mm  1  1 mm, all data resampled to 2 mm isotropic 19 (8:11), 37.6 ± 12.0 Voss (2014)

14 EB (10:4), 20–59, X = 38.2 ± 13.8 13 LB (5:8), 29–60, X = 48.4 ± 8.7

EB and LB

1.5T

Comments Findings Analysis technique Scanner and T1-weighted parameters Cause of blindness Control N (M:F), age (years) Clinical N (M:F), age (years) Study

Table 1 (continued)

J.J. Maller et al. / Journal of Clinical Neuroscience xxx (2016) xxx–xxx

3

There are also some reports of altered regional hippocampus size in studies of blindness. For example, reduced right hippocampal tail in blindness was found by Chebat and colleagues [9] and enlarged right hippocampal head was reported by Fortin and colleagues [10]. Similarly, Lepore and colleagues [11] found larger right hippocampal head and smaller hippocampal tail in blind subjects; the authors suggested that these findings may represent increased functional demands on memory systems and/or adaptive responses to sensory deprivation. The aim of the current study was to statistically determine whether there were regional morphometric differences in highresolution brain MRI scans from 10 blind subjects and 10 sighted subjects. We hypothesized on the basis of previous studies that there would be reduced volume and thickness in the blind group in the regions related to primary visual cortex, and that those with early blindness would have greater volume and thickness than those with late blindness and control subjects. Based upon previous findings of differences in hippocampal volumes in blind subjects we also hypothesized that there would be regional hippocampal volumetric differences. 2. Materials and methods 2.1. Participants The subjects comprised two groups. The first was a group of 10 patients who had acquired blindness as a result of retinoblastoma (N = 2), retinitis pigmentosa (N = 2), glaucoma (N = 2), aniridia (N = 1), eye accident (N = 3), mean age = 49.5 ± 15.6 (six men, four women; eight right-handed). These clinically blind subjects (Snellen’s test Type 6/60 or less) were recruited from the ophthalmology clinic at The Alfred hospital, and via an information session at Vision Australia. Six subjects were classified as early blindness (blindness before the age of 6) and four as late blindness (blindness after the age of 20). The blind patients were age, sex and handedness matched with a group of 10 sighted subjects (Table 2) with stable vision of at least Snellen’s Test Type 6/12 recruited from the community. No subjects had a history of psychiatric disorder or neurological condition apart from blindness. The study received Alfred hospital and Monash University Ethics approval and was carried out in accordance with the Code of Ethics of the World Medical Association (Declaration of Helsinki). Informed consent was obtained from all subjects. 2.2. MRI protocol A Skyra 3T MRI scanner (Siemens Medical, Erlangen, Germany) and a 32-channel head coil was used to acquire the following sequences after a scout scan for AC-PC alignment: T1-weighted MPRAGE (TR = 1540, TE = 2.55, TI = 900, flip angle = 9, slice thickness = 1.00 mm, in-plane = 1.00 mm  1.00 mm, matrix = 256  256), T1-weighted MP2RAGE (TR = 5000, TE = 2.98, TI = 0, 700, 2500, flip angle = 4, slice thickness = 1 mm, in-plane = 1 mm  1 mm, matrix = 240  256), FLAIR SPACE (TR = 6000, TE = 402, TI = 2100, flip angle = 120, slice thickness = 0.9 mm, inplane = 0.9 mm  0.9 mm, matrix = 256  256. A fish oil capsule was placed on the right forehead of each subject before scanning so that the right and left sides would always be identifiable in each software package used. 2.3. Volume and cortical thickness To extract reliable volume and thickness estimates, images were automatically processed with the longitudinal stream in FreeSurfer (FS) [12] using the MPRAGE data. Each T1-weighted

Please cite this article in press as: Maller JJ et al. Brain morphometry in blind and sighted subjects. J Clin Neurosci (2016), http://dx.doi.org/10.1016/j. jocn.2016.01.040

4

J.J. Maller et al. / Journal of Clinical Neuroscience xxx (2016) xxx–xxx

Table 2 Descriptive statistics for blind and control subjects Variable

Group

Number (M:F) Age mean (SD)

Years of blindness Early or late Age of blindness Total GM (L) Total WM (L) CSF (L) TBV (L) ICV (L) TBV/ICV

Blind

Control

10 (6:4) 49.50 ± 15.59 Early (N = 4) 42.00 ± 20.74 (Range:20–63) Left eye 21.22 ± 19.99 (Range:2–55) Early = 4 2.25 ± 2.50 (Range:1–6) 0.732 ± 0.078 0.394 ± 0.046 0.270 ± 0.055 1.127 ± 0.119 1.397 ± 0.118 0.806 ± 0.039

10 (6:4) 49.60 ± 15.10 Late (N = 6) 54.50 ± 10.21 (Range:40–65) Left and Right 23.70 ± 19.21 (Range:2–55) Late = 6 39.50 (Range:21–63)

Right eye 25.10 ± 20.82 (Range:2–63)

–

0.717 ± 0.085 0.432 ± 0.057 0.278 ± 0.045 1.15 ± 0.099 1.428 ± 0.072 0.804 ± 0.037

CSF = Cerebrospinal fluid; F = Females; GM = Gray matter; ICV = Intracranial volume; L = Liters; M = Males; SD = Standard deviation; TBV = Total brain volume; WM = White matter.

scan was segmented into 32 cortical regions in each hemisphere and the corpus callosum [13]. FS also segmented subcortical structures that were not analysed in Qdec (a Graphical User Interface tool for FS) but were exported into SPSS for Windows version 22.0 (IBM, New York, NY, USA) for analysis of the volumes for each region. All FS output was inspected for segmentation accuracy and scatterplots of FS volumetric variables were inspected to check whether any outliers reflected misclassification. After all of the individuals’ images were reconstructed using the above processes, cortical volume group analysis was undertaken within Qdec. BA17 thickness measures were extracted following the method by Hinds and colleagues [14]. Qdec was used to model the data based on the General Linear Model as well as for permutation testing (10,000 iterations) to correct for multiple comparisons [15]. Group comparisons were then conducted between the two groups. The subsequent parameters within Qdec were used for all of the analyses: Measure = Volume (and then thickness), Smoothing = 10, false discovery rate (FDR) and Monte Carlo simulation was used to correct for multiple comparisons using a threshold of 1.3 (p < 0.05) and the appropriate one-tailed test (either positive or negative) was applied. As FS does not calculate regional hippocampal volumes, we segmented the hippocampal heads, bodies and tails of all 20 subjects using Analyze 12.0 (Brain Imaging Resource, Mayo Clinic, MN). Regions were manually outlined on consecutive coronal slices and verified from axial and sagittal orientations. The hippocampi were manually outlined by a single experienced tracer (JJM) from coronal orientated MR images in an anterior–posterior direction as described in Boccardi et al. [16]. Volumes were measured twice for each subject’s scan and then averaged and the intra-rater kappa was 0.98. As FS does not calculate extra-ventricular CSF volume, SPM8 which includes extra-ventricular CSF volume [17], was used to

Fig. 1. Significant clusters of white matter volume differences between sighted and blind subjects using FreeSurfer. Upper row: Parietal. Lower row: Occipital.

calculate gross volumetrics (total GM, WM, CSF volume, and intracranial volume (ICV; which is the sum of GM, WM and CSF volumes)). Total brain volume (TBV) was the sum of GM and WM.

2.4. Statistical analysis Demographic, SPM gross volumetrics, FS output and regional hippocampal volumes were statistically analysed using SPSS for Windows version 22.0. Analyses were two-tailed and evaluated for significance at the 0.05 alpha level. Simple t-tests and analysis of variance (ANOVA) were used to compare demographics between

Table 3 Significant clusters detected in the right and left hemispheres between controls and blind subjects after multiple-corrections Group

Hem

Cluster number

Morphology

Max

VtxMax

Size (mm2)

TalX

TAlY

TAlZ

CWP

CWPLow

CWPHi

Annotation

Controls vs blind

R

1

WM

3.341

33742

1667.75

15.9

83.0

34.0

0.03700

0.03460

0.03940

Superiorparietal

L

1

WM

3.520

1526

900.79

40.6

82.3

4.8

0.04450

0.04190

0.04720

Lateraloccipital

CWP = p-value of the cluster, CWPLow = lower 90% confidence interval CWPHi = higher 90% confidence interval for CWP, Annotation = annotation of segmented region as defined by Freesurfer, Hem = Hemisphere, L = Left, Max = maximum  log10(p value) in the cluster, R = Right, Raw = Raw volumes, TalX = Talairach (MNI305) coordinate of the maximum for  direction, TalY = Talairach (MNI305) coordinate of the maximum for y direction, TalZ = Talairach (MNI305) coordinate of the maximum for z direction, TBV = Controlling for TBV, Th = Thickness, V = Volume, VtxMax = vertex number at the maximum, Size (mm2) = surface area (mm2) of cluster.

Please cite this article in press as: Maller JJ et al. Brain morphometry in blind and sighted subjects. J Clin Neurosci (2016), http://dx.doi.org/10.1016/j. jocn.2016.01.040

5

J.J. Maller et al. / Journal of Clinical Neuroscience xxx (2016) xxx–xxx Table 4 Significant clusters between the blind and control groups in SPSS. Hemisphere

Variable

Significance

Greater

Correlation with blindness duration

Left

Thalamus proper volume Banks STS volume Banks STS area Lateral occipital area Postcentral area Pericallosal thickness G_and_S occipital inferior area G_and_S occipital superior area WM lateral occipital volume WM lingual volume WM pericalcarine volume WM postcentral volume WM precuneus volume WM rostral anterior cingulate volume BA3a thickness (primary somatosensory)

0.024 0.029 0.016 0.012 0.018 0.009 0.029 0.021 0.001 0.024 0.055y 0.002 0.012 0.028 NS

Sighted Sighted Sighted Sighted Sighted Blind Sighted Sighted Sighted Sighted Sighted Sighted Sighted Sighted

NS NS NS NS NS NS NS NS NS NS 0.652 (p = 0.080y)à NS NS NS 0.760 (p = 0.028)à

Right

Entorhinal volume

0.041

Blind

Entorhinal area Lingual area Superior parietal area G_and_S occipital inferior volume G_and_S occipital inferior area G_and_S paracentral volume G_and_S paracentral thickness G_and_S paracentral area G occipital superior volume G subcallosal thickness G subcallosal area G occipital superior area G occipito-temporal medial Lingual area WM lateral occipital volume WM lingual volume WM pericalcarine volume WM postcentral volume WM precuneus volume WM superior parietal volume

0.031 0.040 0.030 0.034 0.028 0.002 0.025 0.009 0.013 0.033 0.044 0.008 0.028 0.002 0.021 0.027 0.060y 0.043 0.002

Blind Sighted Sighted Sighted Sighted Blind Blind Blind Sighted Sighted Sighted Sighted Sighted Sighted Sighted Sighted Sighted Sighted Sighted

parahippocampal thickness: 0.800 (p = 0.010)¥, 0.793 (p = 0.019)à parahippocampal volume: 0.856 (p = 0.003)¥, 0.859 (p = 0.006)à NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS

y = statistical trend, à = covarying for age and TBV, ¥ = covarying for age, Bolded = statistical significance less than 0.01, G = Gyrus, NS = Not significant, S = Sulcus, STS = Superior temporal sulcus, WM = White matter.

groups. Pearson correlations were used to investigate any relationship to duration of blindness. 3. Results 3.1. Demographics There were no significant differences (p > 0.05) between the proportion of males and females in the groups or between their mean ages. Mean duration of blindness (average of right and left blindness) was 23.70 ± 19.21 years (Table 2). 3.2. Neuroimaging 3.2.1. Volumetrics and cortical thickness There were no significant whole brain differences in GM, WM, CSF, or ICV between the blind and sighted groups (Table 2). FS demonstrated no whole brain cortical volumetric or thickness differences between the groups after correction for multiplecomparisons, but there was a significant WM cluster in the left lateral occipital lobe and right superior parietal region after Monte Carlo simulation but not when FDR was applied (Table 3, Fig. 1). There were localized volumetric, thickness and area differences when data was exported to SPSS (Table 4). These regions included

the occipital lobe (e.g. pericalcarine), both WM and GM. Using a stricter significance of p < 0.01, there were still some differences between the groups, including occipital regions (Table 4, bolded), and with the strongest significances in right and left WM of the occipital lobes (p = 0.002 and p = 0.001, respectively). There were no significant differences in total or regional hippocampal volumes between the groups. 3.2.2. Duration of blindness and morphometry Duration of blindness did not significantly correlate with age or TBV. No morphometric clusters demonstrated significant correlations with duration of blindness in Qdec after correction for multiple-comparisons. Covarying for age and/or TBV made no difference to the results. In SPSS, there were significant correlations between years of right eye blindness and left sulcal occipito-temporal lateral thickness (r2 = 0.800, p = 0.005), and between years of right eye and left eye blindness and right WM temporal pole (right eye: r2 = 0.772, p = 0.009; left eye: r2 = 0.756, p = 0.018), and when covarying for age (right eye: r2 = 0.-770, p = 0.015; left eye: r2 = 0.741, p = 0.035). There were also significant positive correlations between duration of blindness and left BA3a thickness and right parahippocampal thickness and volume (Table 4). There were no significant differences in age or TBV between the early and late blindness subjects (all p > 0.05). There were no

Please cite this article in press as: Maller JJ et al. Brain morphometry in blind and sighted subjects. J Clin Neurosci (2016), http://dx.doi.org/10.1016/j. jocn.2016.01.040

6

J.J. Maller et al. / Journal of Clinical Neuroscience xxx (2016) xxx–xxx

significant differences in regional volume or thickness in Qdec between the groups, with or without covarying for age and/or TBV, after applying multiple corrections. In SPSS, those with early blindness had significantly larger left BA3a thickness (F(1,8) = 8.907, p = 0.017; after covarying for TBV: F(1,7) = 8.860, p = 0.021), left middle occipital gyrus thickness (F(1,8) = 5.137, p = 0.053; after covarying for TBV: F(1,7) = 5.224, p = 0.056), left superior temporal gyrus thickness (F(1,8) = 6.355, p = 0.036; after covarying for TBV: F(1,7) = 1.630, p = 0.242), right medial lingual gyrus thickness (F(1,8) = 6.104, p = 0.039; after covarying for TBV: F(1,7) = 1.200, p = 0.309), right lateral occipital volume (F(1,8) = 5.410, p = 0.048; after covarying for TBV: F(1,7) = 5.829, p = 0.046), and right fusiform gyrus volume (F(1,8) = 6.255, p = 0.037; after covarying for TBV: F(1,7) = 28.928, p = 0.001). 4. Discussion The current study demonstrated that blind subjects have localized reduced volume and thickness in visual regions including those representing BA17 that varied with the duration of blindness. We did not find regional hippocampal volumetric differences between the groups. 4.1. Volume and thickness Left hemisphere postcentral gyral WM (and cortical area) was lower in the blind group, as was the right paracentral region. This is consistent with Bedny and colleagues [18] who reported changes in resting-state correlations in the congenitally blind relative to the sighted group included mostly occipital and postcentral gyrus. Whilst Chen and colleagues [1] found cortical differences after multiple-comparisons between POAG and sighted subjects, they first reduced the number of voxels entering the statistical computation. Yu and colleagues presented uncorrected results, and multiple-comparisons corrected data of Bridge and colleagues [4] were not significant. Results presented as significant by Zikou and colleagues [2] (reduction in the left visual cortex volume, left lateral geniculate nucleus, and chiasm in those with POAG) were corrected for multiple-comparisons. We found reduced regional volumes (specifically in the occipital lobe) before and after correction for multiple-comparisons. Left BA3 (somatosensory, postcentral gyrus) thickness (in SPSS) was positively correlated with duration of blindness, and of the subcortical structures, the left thalamus was significantly smaller in the blind group but the right entorhinal cortex (and parahippocampal gyrus) was larger and thicker and positively correlated with blindness duration. These results are consistent with Bhattacharjee and colleagues [19] who reported accelerated somatosensory processing in congenitally blind Braille readers, and with Debowska and colleagues [20] who found significant activations in several cortical areas, including bilateral primary and secondary somatosensory cortices during a Braille Character Stimulator task. Jahn and colleagues [21] reported imagined stance and locomotion to be associated with activation in the occipital visual areas, thalamus, parahippocampal gyrus, and somatosensory cortex. Furthermore, Jahn and colleagues [22] have shown that the parahippocampal gyrus (particularly the right side) is connected to visual cortical areas, and is important for visually guided locomotion and landmark recognition during navigation. We used two different statistical analysis techniques (FS Qdec and SPSS based on FreeSurfer output) in order to statistically investigate whether there were group differences based on whether the data were voxel-based or based on brain regions such as gyri. Furthermore, Qdec does not analyse subcortical structures,

whereas exporting the data into SPSS allowed for statistical comparisons for those structures. Qdec displays statistical results based on correction for multiple-comparisons, but does not provide for results without correction for multiple-comparisons. Hence, in order to statistically investigate whether there were differences based on results without correction for multiple-comparisons, we exported the data to SPSS and conducted ANOVAs and Pearson correlational analysis to investigate correlations of structural brain changes with blindness duration. There is a difference between the FS method voxelwise method for computing differences and the region-wide tests in SPSS. Although SPSS showed significant differences between the groups that were not shown with Qdec, these SPSS statistics were not corrected for multiple-comparisons. Only some of the differences were P < 0.01, hence, results must be interpreted with caution. Cortical thickness was reduced among the blind subjects in diffuse regions, and of the variables that showed the highest significance were the right and left WM of the occipital lobes. 4.2. Hippocampal volumetrics We did not find significant differences between the groups in terms of total or regional hippocampal volumes. This supports previous studies of whole brain analyses but is in contrast to studies of regional hippocampal volumetrics [9–11]. This could possibly be due to our small sample, or, differences in protocols. For example, Chebat and colleagues [9] had lower intra-class reliabilities (0.73 to 0.98) and measured data acquired from lower field strength (1.5T), as did Fortin and colleagues [10] and Lepore et al. [11]. 4.3. Blindness duration and cause of blindness We divided our clinical sample into early and late blindness based on whether they had been without sight since before or after 5 years of age. This was in order to divide the group into near-equal halves. However, previous studies on blindness duration have usually classified early blindness as being without sight since the first or second year of life [8,9,23–25]. Hence, our categorization of blindness duration was different (although the subjects with early blindness in the study by Pan and colleagues [26] had blindness onset up until 6 years of age – their results were similar to ours i.e. reduction in volume in visual cortices). In previous studies, most of our patients would therefore have been classified as late blindness due to their blindness onset being after 2 years of age. Additionally, the cause of blindness in many subjects in previous studies was retinopathy of prematurity. Their analyses would thus be clouded by factors related to development of the premature brain. Furthermore, many previous studies have solely focused on subjects who had non-rapid degenerative blindness (e.g. POAG), whereas most of our subjects had relatively rapid loss of vision (e.g. accident, retinoblastoma, aniridia). Hence, the underlying pathophysiology and neurophysiological mechanisms would differ between our subjects and those studied in previous studies of blindness. Those with early blindness had significantly increased left BA3a thickness compared with late blindness. This is consistent with our finding of a positive correlation of left BA3a thickness with duration of blindness. Furthermore, those with early blindness had significantly greater occipital gyrus thickness, lateral occipital volume, and fusiform gyral volume, compared to those with late blindness. This supports previous research demonstrating that those with early blindness may benefit more from the plasticity or compensatory adaptations due to a larger potential of plasticity [27], particularly in the occipital lobe [28–31]. It is also consistent

Please cite this article in press as: Maller JJ et al. Brain morphometry in blind and sighted subjects. J Clin Neurosci (2016), http://dx.doi.org/10.1016/j. jocn.2016.01.040

J.J. Maller et al. / Journal of Clinical Neuroscience xxx (2016) xxx–xxx

with previous morphometric studies (Table 1) comparing early to late blindness [5–7]. 4.4. Imaging technique The 3T MRI protocol used in this study provided improved imaging data for morphometric analyses compared to previous morphometric studies in blind subjects. We acquired data that utilised a 32-channel head coil which yielded higher signal and therefore greater image contrast and generated more accurate segmentations. Around half of the previous structural brain studies had used either lower field scanners (1.5T), and data were acquired with fewer receive coils, although some studies did not state this latter variable. Future research could utilise recently developed techniques (e.g. multiband, compressed sensing) to acquire more averages yet still within clinically acceptable times. For example, with a multiband factor of 3 and compressed sensing factor of 2, scan time will be reduced to one sixth of the original protocol. 4.5. Limitations We had a small sample size compared to some previous structural MRI studies of blindness, although our groups were well ageand sex-matched. Additionally, the mixed population of early and late blind subjects considered as a single group renders any anatomical comparisons difficult to interpret as the developmental processes between congenital, early and late blind people may differ considerably. We also did not administer comprehensive psychiatric assessments that may have impacted upon the results, as previous research has demonstrated reduction in regional brain volumes in psychiatric illness.

5. Conclusions The current study demonstrated that patients with blindness have regional volumetric or thickness differences and that these differences may vary with blindness duration. Whilst sample size was limited, we used an improved imaging protocol that can be applied clinically and our results are consistent with the literature.

Funding This work was supported by funding received from the Monash Vision Group, an Australian Research Council Special Research Initiative in Bionic Vision Science and Technology (Grant Number SR1000006).

Conflicts of Interest/Disclosures The authors declare that they have no financial or other conflicts of interest in relation to this research and its publication. Study performed on behalf of the Monash Vision Group MVG (Bionic Vision Project). Acknowledgements The authors are grateful to the radiographers at Monash Biomedical Imaging and the subjects who were MRI brain scanned as part of the Monash Vision Group Bionic Vision project.

7

References [1] Chen WW, Wang N, Cai S, et al. Structural brain abnormalities in patients with primary open-angle glaucoma: a study with 3T MR imaging. Invest Ophthalmol Vis Sci 2013;54:545–54. [2] Zikou AK, Kitsos G, Tzarouchi LC, et al. Voxel-based morphometry and diffusion tensor imaging of the optic pathway in primary open-angle glaucoma: a preliminary study. AJNR Am J Neuroradiol 2012;33:128–34. [3] Yu L, Xie B, Yin X, et al. Reduced cortical thickness in primary open-angle glaucoma and its relationship to the retinal nerve fiber layer thickness. PLoS One 2013;8:e73208. [4] Bridge H, Cowey A, Ragge N, et al. Imaging studies in congenital anophthalmia reveal preservation of brain architecture in ‘visual’ cortex. Brain 2009;132:3467–80. [5] Park HJ, Lee JD, Kim EY, et al. Morphological alterations in the congenital blind based on the analysis of cortical thickness and surface area. NeuroImage 2009;47:98–106. [6] Jiang J, Zhu W, Shi F, et al. Thick visual cortex in the early blind. J Neurosci 2009;29:2205–11. [7] Voss P, Zatorre RJ. Occipital cortical thickness predicts performance on pitch and musical tasks in blind individuals. Cereb Cortex 2012;22:2455–65. [8] Anurova I, Renier LA, De Volder AG, et al. Relationship between cortical thickness and functional activation in the early blind. Cereb Cortex 2014. [9] Chebat DR, Chen JK, Schneider F, et al. Alterations in right posterior hippocampus in early blind individuals. NeuroReport 2007;18:329–33. [10] Fortin M, Voss P, Lord C, et al. Wayfinding in the blind: larger hippocampal volume and supranormal spatial navigation. Brain 2008;131:2995–3005. [11] Lepore N, Shi Y, Lepore F, et al. Pattern of hippocampal shape and volume differences in blind subjects. NeuroImage 2009;46:949–57. [12] Reuter M, Schmansky NJ, Rosas HD, et al. Within-subject template estimation for unbiased longitudinal image analysis. NeuroImage 2012;61:1402–18. [13] Desikan RS, Segonne F, Fischl B, et al. An automated labeling system for subdividing the human cerebral cortex on MRI scans into gyral based regions of interest. NeuroImage 2006;31:968–80. [14] Hinds OP, Rajendran N, Polimeni JR, et al. Accurate prediction of V1 location from cortical folds in a surface coordinate system. NeuroImage 2008;39:1585–99. [15] Golland P, Fischl B. Permutation tests for classification: towards statistical significance in image-based studies, Information processing in medical imaging: proceedings of the conference, vol. 18. p. 330–41. [16] Boccardi M, Ganzola R, Bocchetta M, et al. Survey of protocols for the manual segmentation of the hippocampus: preparatory steps towards a joint EADCADNI harmonized protocol. J Alzheimer’s Dis 2011;26:61–75. [17] Valverde S, Oliver A, Cabezas M, et al. Comparison of 10 brain tissue segmentation methods using revisited IBSR annotations. MRI 2014. [18] Bedny M, Pascual-Leone A, Dodell-Feder D, et al. Language processing in the occipital cortex of congenitally blind adults. Proc Natl Acad Sci USA 2011;108:4429–34. [19] Bhattacharjee A, Ye AJ, Lisak JA, et al. Vibrotactile masking experiments reveal accelerated somatosensory processing in congenitally blind braille readers. J Neurosci 2010;30:14288–98. [20] Debowska W, Wolak T, Soluch P, et al. Design and evaluation of an innovative MRI-compatible Braille stimulator with high spatial and temporal resolution. J Neurosci Methods 2013;213:32–8. [21] Jahn K, Deutschländer A, Stephan T, et al. Brain activation patterns during imagined stance and locomotion in functional magnetic resonance imaging. NeuroImage 2004;22:1722–31. [22] Jahn K, Wagner J, Deutschländer A, et al. Human hippocampal activation during stance and locomotion: fMRI study on healthy, blind, and vestibularloss subjects. Ann NY Acad Sci 2009;1164:229–35. [23] Shimony JS, Burton H, Epstein AA, et al. Diffusion tensor imaging reveals white matter reorganization in early blind humans. Cereb Cortex 2006;16:1653–61. [24] Ptito M, Schneider FC, Paulson OB, et al. Alterations of the visual pathways in congenital blindness. Exp Brain Res 2008;187:41–9. [25] Yang C, Wu S, Lu W, et al. Anatomic differences in early blindness: a deformation-based morphometry MRI study. J Neuroimag 2014;24:68–73. [26] Pan WJ, Wu G, Li CX, et al. Progressive atrophy in the optic pathway and visual cortex of early blind Chinese adults: a voxel-based morphometry magnetic resonance imaging study. NeuroImage 2007;37:212–20. [27] Wang D, Qin W, Liu Y, et al. Altered white matter integrity in the congenital and late blind people. Neural Plast 2013;2013:128236. [28] Lazzouni L, Lepore F. Compensatory plasticity: time matters. Front Hum Neurosci 2014;8:340. [29] Striem-Amit E, Amedi A. Visual cortex extrastriate body-selective area activation in congenitally blind people ‘‘seeing” by using sounds. Curr Biol 2014;24:687–92. [30] Striem-Amit E, Bubic A, Amedi A. Neurophysiological mechanisms underlying plastic changes and rehabilitation following sensory loss in blindness and deafness. In: Murray MM, Wallace MT, editors. The neural bases of multisensory processes. Boca Raton (FL)2012. [31] Buchel C, Price C, Frackowiak RS, et al. Different activation patterns in the visual cortex of late and congenitally blind subjects. Brain 1998;121:409–19.

Please cite this article in press as: Maller JJ et al. Brain morphometry in blind and sighted subjects. J Clin Neurosci (2016), http://dx.doi.org/10.1016/j. jocn.2016.01.040