High-amplitude network co-fluctuations linked to variation in hormone concentrations over the menstrual cycle

Abstract Many studies have shown that the human endocrine system modulates brain function, reporting associations between fluctuations in hormone concentrations and brain connectivity. However, how hormonal fluctuations impact fast changes in brain network organization over short timescales remains unknown. Here, we leverage a recently proposed framework for modeling co-fluctuations between the activity of pairs of brain regions at a framewise timescale. In previous studies we showed that time points corresponding to high-amplitude co-fluctuations disproportionately contributed to the time-averaged functional connectivity pattern and that these co-fluctuation patterns could be clustered into a low-dimensional set of recurring “states.” Here, we assessed the relationship between these network states and quotidian variation in hormone concentrations. Specifically, we were interested in whether the frequency with which network states occurred was related to hormone concentration. We addressed this question using a dense-sampling dataset (N = 1 brain). In this dataset, a single individual was sampled over the course of two endocrine states: a natural menstrual cycle and while the subject underwent selective progesterone suppression via oral hormonal contraceptives. During each cycle, the subject underwent 30 daily resting-state fMRI scans and blood draws. Our analysis of the imaging data revealed two repeating network states. We found that the frequency with which state 1 occurred in scan sessions was significantly correlated with follicle-stimulating and luteinizing hormone concentrations. We also constructed representative networks for each scan session using only “event frames”—those time points when an event was determined to have occurred. We found that the weights of specific subsets of functional connections were robustly correlated with fluctuations in the concentration of not only luteinizing and follicle-stimulating hormones, but also progesterone and estradiol.

(panels i-l).In each quartet of panels, we highlight the top two communities by frequency.In panel m, we show the similarity of community centroids to one another.In general, we find that when splitting data by experiment, we maintain an excellent correspondence with the original data.We also find a strong correspondence between the original data and data processed without global signal regression.
-3-  .Effect of null model on detected events.In the main text, we analyzed co-fluctuation patterns estimated from resting-state time series by comparing the global RMS time series against a null distribution generated using a circular shift null model.For each scan, we considered a null distribution generated using the empirical data from that scan.Here, we test the effect of using null distributions from all other 59 scans on the event structure.Specifically, we compared the binary mask of statistically significant frames (reported in main text) with masks generated using null models from the remaining 59 scans.
As a measure of similarity we used the Jaccard similarity measure (bounded between 0 and 1, i.e. minimal and maximal overlap between significant frames).
Here, we show that irrespective of the scan session in which the null model was generated, the significant frames vary only slight (note that the colormap varies over a truncated range from 0.85 to 1.00).In the main text we analyzed fMRI data that had been processed using a pipeline that included global signal regression.Here, we report results using the same data but processed without global signal regression.The analysis procedures were performed identically for both datasets.After obtaining consensus clusters, we mapped communities to those obtained following the global signal regression pipeline.Panels a and b show correlations between state frequencies (how often a given community appeared on any one scan session), for communities 1 and 2. We find a positive correspondence in both cases.In the main text we also computed the correlation between state frequency and hormone concentration.We find that without global signal regression, the overall magnitude of correlations is decrease.However, the overall pattern of correlations is largely preserved (see panel c).-8- Material to "High-amplitude network co-fluctuations linked to variation in hormone concentrations over th Authors: Greenwell et al.

Figure S1 .
Figure S1.Centroids for remaining co-fluctuation communities.In the main text we clustered "event" co-fluctuation patterns and analyzed the top two Material to "High-amplitude network co-fluctuations linked to variation in hormone concentrations over th Authors: Greenwell et al.

Figure S2 .
Figure S2.Robustness of communities 1 and 2 to processing decisions.In the main text we clustered "event" co-fluctuation patterns and analyzed the Material to "High-amplitude network co-fluctuations linked to variation in hormone concentrations over th Authors: Greenwell et al.

Figure S3
Figure S3.Effect of null model on detected events.In the main text, we analyzed co-fluctuation patterns estimated from resting-state time series by

Figure S4 .
Figure S4.Robustness of correlation between state frequency and gonadotropins.In the main text we reported correlations between gonadotropin

Figure S5 .
Figure S5.Comparison of hormone concentrations between experiments 1 and 2. All reported tests are two-sample t-tests.

Figure S6 .
Figure S6.Impact of "leave one out" analysis on state frequency and gonadotropin correlations.Distribution of correlation coefficients after removing

Figure S7 .
FigureS7.Edge-and system-level correlations with hormone concentration for community 2. In the main text we calculated the correlation of hormone

Figure S8 .
Figure S8.Correlations of sex hormone concentration with frequency of community 1.In the main text, we reported positive correlations between

Figure S9 .
Figure S9.Effect of global signal regression on state frequency and correlations.In the main text we analyzed fMRI data that had been processed using Material to "High-amplitude network co-fluctuations linked to variation in hormone concentrations over th Authors: Greenwell et al.

Figure S10 .
Figure S10.Similarity of hormone co-fluctuations across experiments.We calculated the (Spearman) correlations between hormone concentrations