抑制性中間神經元的梯度以及大腦動態特徵：基因、腦磁波和全腦模型的洞見

This thesis investigated the relationship between differences in inhibitory interneuron populations across the cortex, as indexed by gene expression patterns, and the dynamic organization of cortical activity, as measured by Magnetoencephalography (MEG). Utilizing data from the Allen Human Brain Atlas (AHBA), we start with a foundational exploration to identify dominant spatial patterns of gene expression similarity across the cortex, focusing on testing if a theoretically informed selection of genes is meaningful. This study uncovered a general pattern of the gene expression similarity across cortex, linking to evolutionary expansion and cortical cytoarchitecture.

Subsequently, we sought to establish a distribution gradient of inhibitory interneuron subtypes through an analysis of marker gene expression. This was then related to various metrics of MEG cortical dynamics, covering different aspects of cortical activity. Our investigation uncovered two primary interneuron subtype gradients that represent different aspects of cortical organization. The first of these gradients correlates with key features derived from MEG signals, suggesting a fundamental role of these neurons in shaping the functional architecture of the human brain. At the same time, the second gradient aligned with independent measures of the functional organization of the brain, emphasizing a unimodal-transmodal axis.

Finally, we explored whole-brain, connectome-based models, specifically the Jansen-Rit neural mass model, to simulate brain dynamics and investigate the role of the spacial distribution of interneurons in cortical function. By adopting MEG data and structural connectivity of Human Connectome Project (HCP), we implemented a biophysically realistic framework that accounts for the complex interactions among neural populations across the cortex. Our findings reveal that integrating our interneuron distribution gradient significantly enhances the model’s ability to replicate features derived from empirical MEG data. In particular, the incorporation of interneuron Gradient 1 causes a strong association with large-scale timeseries features, highlighting the importance of excitation/inhibition balance in cortical dynamics.

These findings highlight the utility of integrating molecular biology with advanced neuroimaging techniques to deepen our understanding of the neural basis underpinning large-scale cortical organization. This thesis further contributes to the broader field of neuroscience by demonstrating the potential of integrating detailed biophysical models with molecular profiles to advance our understanding of brain function and organization.