Increasing evidence suggests that endogenous sex steroid changes affect human brain functional connectivity, which could be obtained by resting-state fMRI (RS-fMRI). Nevertheless, RS studies on the menstrual cycle (MC) are underrepresented and yield inconsistent results. We attribute these inconsistencies to the use of various methods in exploratory approaches and small sample sizes. Hormonal fluctuations along the MC likely elicit subtle changes that, however, may still have profound impact on network dynamics when affecting key brain nodes. To address these issues, we propose a ROI-based multimodal analysis approach focusing on areas of high functional relevance to adequately capture these changes. To that end, sixty naturally cycling women underwent RS-fMRI in three different cycle phases and we performed the following analyses: (1) group-independent component analyses to identify intrinsic connectivity networks, (2) eigenvector centrality (EC) as a measure of centrality in the global connectivity hierarchy, (3) amplitude of low-frequency fluctuations (ALFF) as a measure of oscillatory activity and (4) seed-based analyses to investigate functional connectivity from the ROIs. For (2)-(4), we applied a hypothesis-driven ROI approach in the hippocampus, caudate and putamen. In the luteal phase, we found (1) decreased intrinsic connectivity of the right angular gyrus with the default mode network, (2) heightened EC for the hippocampus, and (3) increased ALFF for the caudate. Furthermore, we observed (4) stronger putamen-thalamic connectivity during the luteal phase and stronger fronto-striatal connectivity during the pre-ovulatory phase. This hormonal modulation of connectivity dynamics may underlie behavioural, emotional and sensorimotor changes along the MC.

Sex differences in cognitive functions are heavily debated. Recent work suggests that sex differences do stem from different processing strategies utilized by men and women. While these processing strategies are likely reflected in different brain networks, so far the link between brain networks and processing strategies remains speculative. In the present study we seek for the first time to link sex differences in brain activation patterns to sex differences in processing strategies utilizing a semantic verbal fluency task in a large sample of 35 men and 35 women, all scanned thrice. For verbal fluency, strategies of clustering and switching have been described. Our results show that men show higher activation in the brain network supporting clustering, while women show higher activation in the brain network supporting switching. Furthermore, converging evidence from activation results, lateralization indices and connectivity analyses suggests that men recruit the right hemisphere more strongly during clustering, but women during switching. These results may explain findings of differential performance and strategy-use in previous behavioral studies.

Number processing and visual hierarchical processing (global-local processing) have much in common. However, the shared aspects of number processing and global-local processing have not been investigated so far. Most visual stimuli are hierarchical with global structures made up of local parts. Processing of global and local aspects occurs in parallel and the global advantage effect indicates faster reactions to global than local aspects. Likewise, multidigit numbers can be represented holistically (whole number magnitudes) or in a decomposed fashion (single digit magnitudes). During comparison of 2-digit numbers, the unit-decade compatibility effect indicates slower responses when the larger number contains the smaller unit digit and has been suggested as a measure of how strongly participants rely on decomposed number representations. However, this interpretation of the compatibility effect is still controversial and a link between global-local processing and the individual tendency to rely on decomposed representations of multidigit numbers remains to be established. To that end we assessed whether the compatibility effect during number comparison was related to various measures of global advantage. To answer this question we drew upon existing data from participants who had completed both, the number comparison task and 2 global-local tasks. Results show that the compatibility effect is indeed negatively related to several measures of global advantage in women, with no evidence for such a relationship in men. These results demonstrate that global-local processing transcends into the numerical domain but also suggest that the compatibility effect reflects different mechanisms in men and women.