Image Acquisition and Processing

Neuroimaging data were acquired using an 8-channel, receive-only head coil on a 3-T GE Discovery MR750W MRI system (GE Healthcare).²² Adolescents could participate in a mock scanning session before scanning to become familiar with the procedure and scanning environment. After a 3-plane localizer scan, structural MRI was acquired using a high-resolution T1-weighted coronal inversion recovery–fast spoiled gradient recalled sequence (GE option BRAVO; repetition time, 8.77 ms; time to echo, 3.4 ms; inversion time, 600 ms; flip angle, 10°; matrix size, 220 × 220; field of view, 220 × 220 mm; slice thickness, 1 mm; and autocalibrating reconstruction for Cartesian imaging acceleration factor, 2).²⁷

Structural MRI data were processed with the FreeSurfer analysis suite, version 6.0 (FreeSurfer).²⁸ After converting the DICOM (Digital Imaging and Communications in Medicine) data to the MGZ file format using the FreeSurfer mri_convert tool, cortical reconstruction and volumetric segmentation was conducted. Specifically, nonbrain tissue was removed, voxel intensities were normalized for the B1 field in homogeneities, and voxels were segmented into gray matter, white matter, and cerebral spinal fluid. Automatic subcortical segmentation was also performed, and volumes in cubic millimeters were extracted for the hippocampus and amygdala—the subcortical structures of interest in this study. Our group has developed a metric of image quality that automatically characterizes the amount of motion and/or artifact based on signal intensities outside of the brain.²⁹ We additionally controlled for this metric that quantifies motion artifacts and quality. Global metrics of volume were extracted (total brain volume and subcortical volume), and a number of subcortical and cortical structures (eg, amygdala, hippocampus) were automatically labeled.