Refractive error is associated with intracranial volume

Myopia is part of the spectrum of refractive error. Myopia is associated with psychometric intelligence and, the link between brain anatomy and myopia has been hypothesized. Here we aimed to identify the associations between brain structures and refractive error in developed young adults. In a study cohort of 1,319 normal educated young adults, the refractive error showed a significant negative correlation with total intracranial volume and total cerebrospinal fluid (CSF) volume but not with total gray matter volume (GMV) or total white matter volume (WMV). Time spent studying was associated with refractive error but could not explain the aforementioned associations with brain volume parameters. The R2 values of the simple regression between spherical equivalent and outcome variables for each sex in non-whole brain imaging analyses were less than 0.05 in all cases and thus were weak. Psychometric intelligence was not associated with refractive error or total CSF volume, but it weakly positively correlated with total GMV and total WMV in this study population. Thus, refractive error appears to be primarily (weakly) associated with the volume of the cranium, whereas psychometric intelligence was associated with the volume of the brain.


Supplemental online material Supplemental Methods
Subjects. The present study, which is a part of an ongoing project to investigate the association between brain imaging, cognitive function, and aging, included spherical equivalent and structural data from 1,319 healthy, right-handed individuals (763 men and 556 women). The mean age of the subjects was 20.8 years [standard deviation (SD), 1.8; age range: 18-27 years old]. The following descriptions were mostly reproduced from another study of ours from the same project using the exactly same methods regarding these issues 1 . Some of the subjects who took part in this study also became subjects of our intervention studies (psychological data and imaging data recorded before the intervention were used in this study) 2 . Psychological tests and MRI scans not described in this study were performed together with those described in this study. All subjects were university students, postgraduates, or university graduates of less than one year's standing. All subjects had normal vision and none had a history of neurological or psychiatric illness. Handedness was evaluated using the Edinburgh Handedness Inventory 3 . Written informed consent was obtained from each subject. For nonadult subjects, written informed consent was obtained from their parents (guardians). This study was approved by the Ethics Committee of Tohoku University.
Subjects were instructed to get sufficient sleep, maintain their conditions, eat sufficient breakfast, and to consume their normal amounts of caffeinated foods and drinks in the day of cognitive tests and MRI scans. In addition, subjects were instructed to avoid alcohol the night before the assessment.
Pre-processing of structural data. The methods for the preprocessing of T1WIs were described in our previous study and reproduced below 4  tissues. In this new segmentation process, default parameters were used, except that the Thorough Clean option was used to eliminate any odd voxel, affine regularization was performed with the International Consortium for Brain Mapping template for East Asian brains, and the sampling distance was set at 1 mm. We then proceeded to the diffeomorphic anatomical registration through exponentiated lie algebra (DARTEL) registration process implemented in SPM12. We used DARTEL import images of the 2 TPMs from the abovementioned new segmentation process. First, the template for the DARTEL procedures was created using imaging data from 800 participants (400 males and 400 females). The following methods were the same as in our previous study and descriptions were reproduced from our previous study 5 . Next, using this existing template, the DARTEL procedures were performed for all of the subjects in the present study. In these procedures, default parameter settings were used. The resulting images were spatially normalized to the Montreal Neurological Institute (MNI) space to give images with 1.5 1.5  1.5 mm 3 voxels. In addition, we performed a volume change correction (modulation) by modulating each voxel with the Jacobian determinants derived from spatial normalization, which allowed us to determine regional differences in the absolute amount of brain tissue 6 . Subsequently, normalized rGMV, rWMV, and rCSF volume images were smoothed by convolving them with an isotropic Gaussian 6 kernel of 8 mm full width at half maximum (FWHM).

Pre-processing of diffusion data
The methods for the preprocessing of diffusion data were described in our previous study and reproduced below 7 . Preprocessing and analysis of diffusion data were performed using Statistical Parametric Mapping (SPM) 8 implemented in Matlab.
Using a previously validated two-step new segmentation algorithm of diffusion images and the previously validated diffeomorphic anatomical registration through exponentiated lie algebra (DARTEL)-based registration process 8 , all images, including gray matter segment [regional gray matter density (rGMD) map], white matter segment [regional white matter density (rWMD) map], cerebrospinal fluid (CSF) segments [regional CSF density (rCSFD) map] of diffusion images, were normalized. The voxel size of normalized FA images and MD images and segmented images, was 1.5  1.5  1.5 mm 3 . See our previous work 9 for the details of these procedures including the information of the template.
Next, we created average images of normalized rGMD and rWMD images from the normalized rGMD and rWMD images from the subset of the entire sample (63 subjects) 8 . Subsequently, for the analyses of MD images from the normalized images of the (a) MD, (b) rGMD, and (c) rCSFD maps, we created images where areas that were not strongly likely to be gray or white matter in our averaged normalized rGMD and rWMD images (defined by "gray matter tissue probability + white matter tissue probability < 0.99") were removed (to exclude the strong effects of CSF on MD throughout analyses). These images were then smoothed (8 mm full-width halfmaximum) and carried through to the second-level analyses of MD. We did not use T1 weighted structural images for these preprocessing procedures. As described previously "This is because T1 weighted structural images and EPI images have apparent differences due to the distortion caused by 3T MRI and simply it is apparently not suited for the accurate and precise segmentation and normalization images of MD maps." 11 .