Dynamics of hidden brain states in post-stroke cognitive impairment: Linking neural activity to neurotransmitter
Changjun Ma1, Xiulin Wang2, Jiahui Liu2
1School of Biomedical Engineering, Faculty of Medicine, Dalian University of Technology, Dalian, China; Stem Cell Clinical Research Center, National Joint Engineering Laboratory, Regenerative Medicine, The First Affiliated Hospital of Dalian Medical University, Dalian, Liaoning Province, China; Dalian Innovation Institute of Stem Cell and Precision Medicine, Dalian, Liaoning Province, China.
Background:
Post-stroke cognitive impairment (PSCI) is associated with abnormal dynamic functional connectivity, yet the temporal dynamic of brain activity and their underlying molecular mechanisms remain unclear.
Methods:
Participants were classified into two groups based on neuropsychological assessments: PSCI group (N = 67) and post-stroke with no cognitive impairment (NPSCI) group (N = 65), alongside 47 healthy controls (HCs). Dynamic brain states were analyzed using a Hidden Markov Model (HMM) with the Brainnetome Atlas, yielding metrics like fractional occupancy (FO), mean dwell time (MDT), switching rate (SR) and transition probabilities (TP) based on resting-state functional magnetic resonance imaging (rs-fMRI). Finally, we further assessed the spatial correlations between the mean activation of HMM state and neurotransmitter receptors/transporters distribution, cognitive relative term, and gene expression profiles.
Results:
Five HMM states were identified. Compared with HCs and NPSCI group, patients with PSCI group exhibited different dynamics, including FO, MDT, SR, and TP. Additionally, we found that the mean activation maps of HMM state were associated with the neurotransmitter receptors/transporters distribution and cognitive relative term. Furthermore, our results demonstrated a spatial correlation between the mean activation maps of state 5 and gene expression patterns. Finally, enrichment analysis indicated that PLS-positive genes were enriched in pathways related to DNA/RNA metabolism, signal transduction and regulation, and immune-disease associations, whereas, PLS-negative genes were mainly enriched in lipid metabolism and insulin response, virus-cytokine interactions, and influenza response pathways.
Conclusions:
This study provides new insights into characterizing dynamic neural activity in PSCI. The brain network dynamics defined by HMM analysis may deepen our understanding of the neurobiological underpinnings of PSCI, indicating a linkage between neural configuration and gene expression in PSCI.
