MENSTRUAL PHASE-DRIVEN FRONTAL WHITE MATTER MICROSTRUCTURAL DYNAMICS TRACK PROGESTERONE AND CRAVING VARIABILITY IN WOMEN WITH SUBSTANCE USE DISORDERS

Substance use disorders (SUDs) constitute a chronic and relapsing condition characterized by persistent cognitive impairments and craving, accompanied by neurological adaptations often involving the prefrontal cortex. Previous studies highlight critical contributions of white matter adaptations underlying addiction-related functional deficits, employing diffusion magnetic resonance imaging to estimate microstructural features of the brain’s anatomical connections. While dynamic cytoarchitectural and volumetric changes in white matter are known to be mediated by sex steroid hormones, possible relationships between cyclical hormonal fluctuations and white matter pathology in addiction are unknown. Here, we used noninvasive diffusion MRI to analyze white matter microstructure in naturally cycling women with SUDs as a function of menstrual phase, hormones, and drug craving. Twenty women with a diagnosed SUD, 17 cocaine primary (CUD)/3 heroin primary (HUD), were recruited for a study examining menstrual cycle effects on the neurocognitive correlates of addiction. Diffusion-weighted images (DWI) and blood concentrations of estradiol and progesterone were acquired from participants twice, during the late follicular and mid luteal phases of their menstrual cycle. Scans were preprocessed using the MRtrix3 FSL pipeline, and diffusion tensor imaging maps of axial, radial, and mean diffusivities (AD, RD, MD) and fractional anisotropy (FA) were computed with tract-based spatial statistics. Menstrual phase effects were analyzed voxelwise in whole-brain white matter using a withinsubject design in FSL Randomise with 5000 permutations per comparison. Due to the small number of HUD participants, DWI from an additional 13 HUD primary women without menstrual data were included to assess SUD-specific effects. Self-reported drug cue-induced craving scores were obtained from participants’ average ratings of 36 picture cues on a 5point scale. Voxelwise comparisons identified clusters of significantly increased AD in the luteal vs. follicular phase, located bilaterally in the anterior corona radiata, genu and body of the corpus callosum, and anterior and posterior limbs of the internal capsule (mean effect: paired t=6.71, p < .001). Tests for FA and other diffusivity metrics were not significant (p’s > .05). Across all scans and controlling for phase, AD from the significant voxels was positively correlated with participants’ progesterone levels (r=0.51, p=.007), even when controlling for estradiol (progesterone/estradiol ratio: r=0.52, p=.002). In CUD subjects, follicular AD correlated negatively with cue-induced craving scores (r=-0.64, p=.008); similar correlations were not found in the luteal phase or in HUD participants (p’s > .126), for whom menstrual phases were mostly unknown. In general, the HUD group exhibited lower AD compared to CUD luteal scans (t=3.48, p=.002), with no significant difference compared to CUD follicular scans (t=1.39, p=.173). Results indicate menstrual cycle-related white matter microstructural dynamics characterized by higher AD in the luteal vs. follicular phase, driven by typical progesterone fluctuations, which negatively tracked cocaine craving in the follicular phase. These findings advance a growing body of evidence for white matter microstructural alterations in the pathology of SUD, highlighting the menstrual cycle as a potential source of sex-specific individual differences in addiction. Further work is needed to better understand effects of sex hormones across SUDs, including dense sampling over the full menstrual cycle, and in postmenopausal and non-naturally cycling populations.