Summary: Researchers have identified stable, recurring patterns of brain activity across 337 healthy adults by simplifying complex fMRI data. These reproducible activity motifs appear across individuals and correlate with cognitive performance, emotion regulation, and substance use, suggesting they could become useful biomarkers for psychiatric conditions and improve our understanding of individual behavioral differences.
By reducing the dimensionality of resting‑state fMRI recordings, the team revealed shared dynamic brain states that are consistent across participants yet show measurable individual variation in how long people occupy each state and how they transition between them. The investigators plan to apply the same approach to clinical populations to detect patterns specific to psychiatric disorders.
Key Facts:
- Analysis of resting‑state fMRI from 337 young adults revealed several recurring, reproducible brain activity patterns.
- These shared patterns relate to cognitive functioning, emotion regulation, and alcohol and substance use measures.
- The method may help identify neuroimaging biomarkers for psychiatric diagnoses and enhance clinical assessment tools.
Source: Yale
Challenge and approach: Functional magnetic resonance imaging (fMRI) captures very complex, high‑dimensional signals reflecting moment‑to‑moment brain activity. That complexity can obscure reproducible patterns tied to behavior. To address this, researchers at Yale used data dimension reduction techniques to simplify brain activity into a smaller set of meaningful features, making it possible to detect common dynamic states shared across many people.
The study used resting‑state fMRI data collected from 337 healthy young adults, each scanned in four 15‑minute sessions. These scans produced time‑series “snapshots” of whole‑brain activity, capturing how networks interact and fluctuate at each moment. Reducing the dimensionality of these data allowed the investigators to extract dominant co‑activation patterns — compact descriptors of complex, time‑varying brain activity.
Applying this dimensionality reduction revealed three principal brain states that recurred consistently across participants and across scanning sessions. While the same states were present in nearly everyone, individuals differed in how often they occupied each state, how long they remained in a given state, and the sequences of state transitions. Those individual differences related to behavioral measures, suggesting the states reflect both shared neural motifs and person‑specific traits.
The results, published Sept. 24 in PLOS Biology, indicate that combining state‑level (within‑session) and trait‑level (between‑individual) analyses with feature reduction can uncover robust, reproducible neural motifs. These motifs co‑vary with principal behavioral variations — notably measures of cognitive ability, emotion regulation, and alcohol and substance use — reinforcing the link between dynamic neural states and meaningful aspects of behavior.
“Human brain activity is extremely complex, which can make reproducibility challenging,” said Kangjoo Lee, lead author and associate research scientist in Yale’s Department of Psychiatry. “We focused on features of brain dynamics that are both tied to behavior and consistent across people.”
Co‑senior author John Murray, who led the physics and psychiatry efforts at Yale and is now at Dartmouth College, noted the clinical potential: “If the same analysis is applied to clinical populations, we may discover recurring brain states that are common in a patient group but rare in healthy controls. Those shared patterns could serve as neuroimaging biomarkers useful in diagnosis and treatment planning.”
The team plans to extend this work to psychiatric cohorts to test whether specific recurrent brain states correspond to particular symptoms or diagnoses and to explore how these patterns vary across individuals and over time. Ultimately, reproducible state‑trait markers derived from fMRI could inform personalized approaches to mental health care and provide objective neural targets for interventions.
About this behavior and neuroscience research news
Author: Mallory Locklear
Source: Yale
Contact: Mallory Locklear – Yale
Image: The image is credited to Neuroscience News
Original Research: Open access. “Human brain state dynamics are highly reproducible and associated with neural and behavioral features” by Kangjoo Lee et al., PLOS Biology
Abstract
Human brain state dynamics are highly reproducible and associated with neural and behavioral features
Neural activity and behavior both vary within individuals (states) and between individuals (traits). How these state‑trait neural variations map onto behavior is not fully understood.
To address this gap, the study quantifies moment‑to‑moment changes in brain‑wide co‑activation patterns extracted from resting‑state fMRI. In healthy young adults, the authors identify reproducible spatiotemporal features of these co‑activation patterns at the single‑subject level.
A joint analysis that combines state‑level and trait‑level neural variation with dimensionality reduction reveals general motifs of individual differences. These motifs include both state‑specific features and more general neural characteristics that exhibit day‑to‑day variability.
Principal neural variations co‑vary with principal behavioral dimensions, notably cognitive function, emotion regulation, and alcohol and substance use. Individual probabilities of occupying particular co‑activation patterns are reproducible and associated with neural and behavioral features.
This combined state‑trait approach offers a promising path toward developing reproducible neuroimaging markers that relate to individual life outcomes and may ultimately inform clinical assessment and treatment strategies.