Developmental Sex Differences in White Matter using Advanced and Conventional dMRI Models

1263 

Abstract Submission 

Sebastian Benavidez¹, Katherine Lawrence¹, Gaon Kim¹, Zvart Abaryan², Emily Laltoo¹, James McCracken³, Paul Thompson¹

¹Imaging Genetics Center, Mark and Mary Stevens Neuroimaging & Informatics Institute, USC, Marina del Rey, CA, ²Children’s Hospital of Los Angeles, Los Angeles, CA, ³Department of Psychiatry, University of California San Francisco, San Francisco, CA

Sebastian Benavidez  
Imaging Genetics Center, Mark and Mary Stevens Neuroimaging & Informatics Institute, USC
Marina del Rey, CA

Katherine Lawrence  
Imaging Genetics Center, Mark and Mary Stevens Neuroimaging & Informatics Institute, USC
Marina del Rey, CA

Gaon Kim  
Imaging Genetics Center, Mark and Mary Stevens Neuroimaging & Informatics Institute, USC
Marina del Rey, CA

Zvart Abaryan  
Children’s Hospital of Los Angeles
Los Angeles, CA

Emily Laltoo  
Imaging Genetics Center, Mark and Mary Stevens Neuroimaging & Informatics Institute, USC
Marina del Rey, CA

James McCracken  
Department of Psychiatry, University of California San Francisco
San Francisco, CA

Paul Thompson, PhD  
Imaging Genetics Center, Mark and Mary Stevens Neuroimaging & Informatics Institute, USC
Marina del Rey, CA

Childhood and adolescence are periods of substantial white matter (WM) maturation. WM abnormalities have also been implicated in adolescent-onset psychiatric disorders. Understanding sex differences in the typically developing brain may offer insight into sex-specific variation in these brain-based disorders (Boyd et al., 2015). Although prior research has extensively characterized WM sex differences in adults (Salminen et al., 2022), WM sex differences during development are less well understood. Here we aimed to characterize sex differences in WM microstructure during typical development – from early childhood through emerging adulthood – using diffusion-weighted MRI (dMRI) data modeled with conventional diffusion tensor imaging (DTI) (Basser et al., 1994) and the advanced tensor distribution function (TDF) (Leow et al., 2009).

We analyzed dMRI data sourced from the Healthy Brain Network (Alexander et al., 2017), from 239 typically developing participants aged 5-22 years (46.0% female). Participants were included in our sample if they received a full clinical evaluation and were not given a clinical diagnosis. dMRI scans were acquired across four scanners: three 3T scanners (1.8 mm isotropic voxel size, 72 slices), and one 1.5T scanner (2.0 mm isotropic voxel size, 72 slices). All dMRI scans had one b = 0 s/mm² volume and 64 directions at b = 1000 s/mm². All dMRI scans were preprocessed using the ENIGMA-DTI protocol (Jahanshad et al., 2013). Measures of fractional anisotropy (FA_DTI), mean diffusivity (MD), axial diffusivity (AD), and radial diffusivity (RD) were generated using DTI. The TDF model was used to generate FA_TDF, which is similar to FA_DTI but more accurately describes voxels with crossing fibers. These measures were then skeletonized using tract-based spatial statistics. Using the JHU WM brain atlas (Mori et al., 2008), the average value along the WM skeleton was extracted for each metric in 25 regions of interest (ROI). We harmonized the ROI data using ComBat within each metric (Radua et al., 2020; Fortin et al., 2018). A fixed-effects linear regression was used to assess WM sex differences, adjusting for age, age-squared, the interaction between sex and age, and the interaction between sex and age-squared. The false discovery rate (Benjamini & Hochberg, 1995) was used to correct for multiple comparisons across the 25 ROIs, and significance was determined by a corrected p-value of 0.05.

The effect of sex was regionally consistent, as boys generally exhibited lower FA_DTI and FA_TDF and higher MD, AD, and RD than girls. For the DTI metrics, boys displayed significantly lower FA_DTI in the sagittal stratum (SS); higher MD across the whole brain, in the corona radiata (CR), posterior corona radiata (PCR), anterior corona radiata (ACR), external capsule (EC), uncinate fasciculus (UNC), posterior thalamic radiation (PTR), superior longitudinal fasciculus (SLF), and SS; higher AD in the PCR, SLF, and SS; and higher RD in the SS compared to girls. For FA_TDF, boys exhibited significantly lower values than girls in the CR, superior corona radiata (SCR), PCR, EC, fornix/stria terminalis (FXST), PTR, retrolenticular limb of the internal capsule (RLIC), SLF, SS, and tapetum (TAP). The TDF model was more sensitive to sex differences than DTI, detecting more significant ROIs than the DTI model (Fig. 1).

We found widespread sex differences in WM microstructure during typical development. The observed sex differences across WM regions suggest greater FA and lower diffusivity for females compared to males, which may have implications for understanding sex-specific vulnerabilities in psychiatric disorders. As a whole, these findings highlight the need to consider sex in neurodevelopmental research and underscore the value of the advanced TDF model.

Early life, Adolescence, Aging ¹

Diffusion MRI Modeling and Analysis ²

Development

Sexual Dimorphism

WHITE MATTER IMAGING - DTI, HARDI, DSI, ETC