Abstract
Sex hormone fluctuations modulate structural and functional brain dynamics, yet little is known how sex steroid levels map onto the expression of sex differences in the brain. Here, we trained machine learning models for brain sex classification based on anatomical structures in cross-sectional data of N = 1090 individuals (50% females, age matched). Applied to dense sampled data of one male and two females in different hormonal states (naturally cycling, oral contraceptive user, pregnancy), we linked inter- and intra-individual fluctuations in brain sex to neuroendocrine modulation. We found lower variation in brain sex across time in the male compared to the female subjects. Oral contraceptive use was associated with a more female-like brain, while (inverted) U-shaped brain sex trajectories emerged across menstrual cycle phases and pregnancy trimesters. Overall, our findings suggest that changes in brain sex capture hormone-related plasticity over time in dense sampled individuals.