Hormone-Dependent Microstates and the Reconfiguration of Resting-State Dynamics across the Menstrual Cycle

This study demonstrates that menstrual cycle-induced hormonal fluctuations modulate large-scale brain microstate dynamics, providing a neurophysiological basis for behavioral changes and offering a predictive framework for individual variations in psychological well-being.

The Gist

Imagine your brain isn't a static computer that runs the same program every day. Instead, think of it as a bustling city with traffic patterns that change constantly. Sometimes the traffic flows smoothly through the downtown area; other times, it gets stuck in the suburbs. These "traffic patterns" are what scientists call brain microstates—brief, stable moments where your brain settles into a specific global configuration before switching to the next one.

This study asks a big question: Does the monthly hormonal cycle (the menstrual cycle) act like a traffic controller, changing how these brain patterns flow?

Here is the story of the research, broken down into simple concepts:

1. The Setup: Tracking the "Monthly Weather"

The researchers followed 24 women over the course of their menstrual cycles. They visited them at three specific times:

• Early Follicular: The "fresh start" phase (low hormones).
• Peri-ovulatory: The "peak" phase (high estrogen).
• Mid-Luteal: The "winding down" phase (high progesterone).

At each visit, they took blood samples to measure the "weather" (hormone levels) and used a super-sensitive helmet (MEG) to listen to the electrical "traffic" of the brain.

2. The Discovery: Finding the "Hormone-Sensitive" Patterns

Usually, scientists look for the average brain pattern. But this team wanted to find the specific patterns that change depending on the time of the month.

They used a clever trick: they looked at thousands of brain snapshots and asked, "Which of these snapshots appears more often in the 'Luteal' phase than in the 'Follicular' phase?"

They found two special patterns, which they named Hormone-Dependent Microstates (HDMs).

• The Metaphor: Imagine your brain has a playlist of songs. Most songs play the same amount of time every day. But these two specific songs (HDM 0 and HDM 1) get played much more often when a woman is in the mid-luteal phase (the week before her period) compared to the rest of the month.
• The Location: These patterns light up specific areas of the brain, particularly the left side, involving regions responsible for emotions, memory, and sensory processing (like the amygdala and temporal lobes).

3. The Connection: Hormones Drive the Playlist

The study confirmed that the more the hormones shifted, the more these specific brain patterns appeared.

• The Analogy: Think of the hormones as the DJ and the brain patterns as the music. When the DJ (hormones) changes the mix, the brain (the dance floor) starts dancing to a different beat. The study found a direct link: as the hormonal "DJ" turned up the volume on certain chemicals, the brain's "dance moves" (microstates) changed accordingly.

4. The Human Impact: Does This Affect How We Feel?

This is the most exciting part. The researchers asked: Does this change in brain traffic affect how these women feel about their lives?

They measured "Psychological Well-being" across six areas, such as feeling autonomous, having purpose, and feeling like they are growing as a person.

• The Finding: They found a specific link to "Personal Growth."
• The Story: When the hormonal levels shifted and the brain spent more time in that specific "Hormone-Sensitive" pattern, the women reported subtle changes in how they felt about their personal growth. It's as if the brain's internal traffic pattern acts as a bridge between the body's chemistry and the mind's sense of self.

Why Does This Matter?

For a long time, scientists treated the brain like a machine that works the same way for everyone, regardless of biology. This study suggests that we need to account for the "monthly rhythm" when studying the brain.

• The Big Picture: Just as the ocean has tides, the human brain has monthly tides driven by hormones. Ignoring these tides is like trying to study ocean waves without knowing about the moon.
• The Future: Understanding these "Hormone-Dependent Microstates" could help us better understand why some women feel different (or struggle with mood disorders like PMDD) at certain times of the month. It shows that our biology and our psychology are deeply intertwined, dancing together in a complex, monthly rhythm.

In short: Your brain changes its "traffic patterns" every month based on your hormones, and these changes might be the secret reason why your feelings about your life and growth shift along with your cycle.

Technical Summary

1. Problem Statement

While behavioral fluctuations (mood, cognition) across the natural menstrual cycle (MC) are well-documented, the underlying neural mechanisms remain elusive. Previous research has utilized fMRI to study hemodynamic changes and EEG/MEG for spectral power, but these approaches often rely on time-averaged metrics or assume stationarity.

• The Gap: There is a lack of understanding regarding how sex hormone fluctuations (Estradiol, Progesterone, FSH, LH) modulate large-scale, sub-second brain dynamics in a non-stationary manner.
• The Hypothesis: Cyclic variations in sex hormones modulate large-scale brain activation patterns, creating specific "microstates" (quasi-stable global configurations) that serve as a neurophysiological signature of the MC and predict psychological well-being.

2. Methodology

Participants and Experimental Design

• Cohort: 24 naturally cycling, right-handed women (mean age ~26 years).
• Longitudinal Protocol: Participants underwent MEG recordings and blood sampling across three distinct MC phases:
  1. Early Follicular: Days 1–4 (Low E/P).
  2. Peri-ovulatory: Days 13–15 (High E).
  3. Mid-luteal: Days 21–23 (High E and P).
• Behavioral Assessment: Psychological well-being was measured using the Ryff's Psychological Well-Being Scale (6 dimensions: autonomy, environmental mastery, personal growth, positive relations, purpose in life, self-acceptance).

Data Acquisition and Preprocessing

• MEG: 163-channel system (SQUID-based) with eyes-closed resting-state recordings (3.5 mins per session).
• Source Reconstruction:
  • Preprocessing included PCA for noise attenuation and ICA for artifact removal (cardiac/ocular).
  • Beamforming: Linearly Constrained Minimum Variance (LCMV) beamformer was used to reconstruct source activity.
  • Parcellation: Automated Anatomical Labeling (AAL) atlas with 90 cortical Regions of Interest (ROIs).
  • Filtering: Band-pass filtered between 2–30 Hz.

Analytical Framework: Hormone-Dependent Microstates (HDMs)

The study deviated from traditional microstate analysis (which seeks canonical 4–5 global maps) to identify phase-specific maps.

1. GFP Extraction: Global Field Power (GFP) peaks were extracted from source-reconstructed time series.
2. Clustering Strategy:
   • A pooled dataset of 72,000 GFP peaks (24 subjects × 3 phases) was clustered using k-means.
   • Optimization Criterion: Instead of maximizing Global Explained Variance (GEV), the optimal number of clusters ($k$) was selected to maximize the F-statistic for the difference in "visit" frequency (occurrence) of microstates across the three MC phases.
   • Result: The optimal solution was found at $k=13$.
3. Selection of HDMs: From the 13 clusters, the authors identified specific maps (HDMs) that showed significant phase-dependent modulation.
4. Statistical Modeling:
   • Linear Mixed-Effects Models (LMM): Used to test the relationship between:
     • Hormonal fluctuations (First Principal Component of E, P, FSH, LH) and Microstate occurrences.
     • Microstate occurrences + Hormones and Psychological scores.
   • Validation: Leave-One-Out Cross-Validation (LOOCV) was used to assess predictive power.

3. Key Results

Hormonal Validation

• Hormonal profiles (PC1 of E, P, FSH, LH) showed significant variation across all phase pairs ($p < 0.001$), confirming the physiological validity of the sampling.

Identification of Hormone-Dependent Microstates (HDMs)

• Two specific microstates (HDM 0 and HDM 1) demonstrated significant phase-dependent differences in their occurrence frequency ($p < 0.001$).
• HDM 0 (Primary Focus):
  • Topography: Predominantly left-lateralized, involving the temporal pole, pallidum, insula, amygdala, putamen, inferior frontal orbital gyrus, Rolandic operculum, and thalamus.
  • Dynamics: Occurrence increased progressively from the Early Follicular phase to the Mid-Luteal phase.
  • Robustness: HDM 0 maintained high spatial correlation across different clustering granularities ($k$ values), making it the most robust marker.
• HDM 1: Also left-hemispheric dominant (pre/post-central gyri, parietal, cingulate), but HDM 0 was selected for deeper analysis due to stronger statistical significance.

Hormone-Microstate Coupling

• Hormonal levels (PC1) were a robust positive predictor of HDM 0 occurrences ($\beta = 0.407, p < 0.001$).
• Hormonal shifts accounted for 42.3% of the variance in HDM 0 activity.

Prediction of Psychological Well-being

• Personal Growth: The combination of hormonal levels and HDM 0 occurrences significantly predicted changes in the "Personal Growth" subscale of the Ryff scale.
  • The full model (Hormones + HDM 0) explained ~7% of the variance in personal growth scores ($R^2_{marginal} = 0.07$).
  • Note: While the trend was significant ($p_{uncorrected} = 0.02$), it did not survive strict FDR correction across all 6 subscales, suggesting a specific but subtle effect.
• Other Subscales: No significant predictive relationships were found for autonomy, environmental mastery, etc., or for the aggregate well-being score.

4. Key Contributions

1. Novel Framework (HDMs): Introduced "Hormone-Dependent Microstates," a methodological shift from identifying static canonical brain states to identifying dynamical states specifically sensitive to slow-changing biological variables (hormones) over a month-long timescale.
2. Neurophysiological Substrate: Provided direct evidence that sex hormones modulate large-scale, sub-second cortical dynamics (not just hemodynamic or spectral power), specifically reconfiguring the frequency of visits to left-lateralized limbic and fronto-temporal networks.
3. Tripartite Link: Established a statistical link between Endocrine Fluctuations $\rightarrow$ Neural Dynamics (HDMs) $\rightarrow$ Psychological State (specifically personal growth).
4. Methodological Innovation: Demonstrated that optimizing microstate clustering based on phase-discrimination power (rather than just variance explained) can reveal biologically relevant brain states that traditional methods might miss.

5. Significance and Implications

• Neuroimaging Standards: The study underscores the critical necessity of accounting for the menstrual cycle in neuroimaging studies involving women of fertile age, as hormonal fluctuations induce significant, measurable changes in brain dynamics.
• Clinical Relevance: The findings offer a potential biomarker for conditions like Premenstrual Dysphoric Disorder (PMDD). If healthy women show specific microstate reconfigurations linked to hormones, PMDD patients might exhibit exaggerated or dysregulated versions of these dynamics.
• Multiscale Dynamics: The results support a framework where ultra-slow dynamics (monthly hormonal cycles) constrain and modulate fast dynamics (millisecond-scale neural microstates), challenging the view of the brain as an isolated system.
• Future Directions: Suggests that "Personal Growth" (a dimension of well-being related to self-expansion and potential) may be uniquely sensitive to the neuro-hormonal interplay during the luteal phase, warranting further investigation with larger samples and clinical populations.

Limitations Noted:

• Sample size (N=24) limits power for detecting small effect sizes, particularly in the psychological prediction models.
• Discrete sampling (3 time points) misses intra-cycle variability.
• Findings are specific to naturally cycling women and may not apply to those on hormonal contraceptives.