Related Experiment Video
Updated: Apr 21, 2026
08:44
Dynamic Clamp Methods to Investigate Impaired Neuronal Excitability Associated with Autism
Published on: October 17, 2025
839
Characterizing and Differentiating Brain State Dynamics via Hidden Markov Models
Jinli Ou1, Li Xie1, Changfeng Jin2
1School of Biomedical Engineering & Instrument Science, Zhejiang University, Hangzhou, China.
Brain Topography
|October 22, 2014
Summary
This study introduces a new computational framework using hidden Markov models (HMMs) to analyze brain state dynamics from resting-state fMRI (R-fMRI) data. The research found distinct brain state transitions in post-traumatic stress disorder (PTSD) patients, aiding in classification.
Area of Science:
- Neuroscience
- Computational Psychiatry
- Medical Imaging Analysis
Background:
- Resting-state functional magnetic resonance imaging (R-fMRI) is crucial for understanding brain function and identifying neurological conditions.
- While functional connectivity is well-studied, the dynamic transitions between brain functional states remain underexplored.
- Characterizing dynamic brain states is essential for a comprehensive understanding of brain activity and disorders.
Purpose of the Study:
- To develop and validate a novel computational framework for quantitatively characterizing brain state dynamics using hidden Markov models (HMMs).
- To investigate the temporal dynamics of functional connectomics and identify distinct functional connectome states.
- To differentiate between post-traumatic stress disorder (PTSD) patients and normal controls (NC) based on their brain state dynamics.
Main Methods:
- Application of a novel computational framework utilizing hidden Markov models (HMMs) to analyze R-fMRI data.
- Learning HMMs from temporally dynamic functional connectomics to define functional connectome states.
- Analysis of an R-fMRI dataset comprising 44 PTSD patients and 51 NC subjects.
Main Results:
- Both PTSD and NC subjects exhibit significant dynamic changes in resting-state brain activity, transitioning among several functional states.
- HMM analysis revealed that PTSD patients tend to enter, but struggle to disengage from, a negative mood state.
- A classification accuracy of 84% for PTSD patients and 86% for NC subjects was achieved using multiple HMMs with majority voting.
Conclusions:
- The proposed HMM-based framework effectively characterizes brain state dynamics and functional connectome states from R-fMRI data.
- Dynamic brain state analysis provides valuable insights into the neural underpinnings of PTSD, particularly regarding mood regulation.
- This computational approach demonstrates high potential for classifying neurological conditions like PTSD based on functional brain dynamics.
