The brain’s internal echo: Longer timescales, stronger recurrent connections and higher neural excitation in self regions
Makaleler > Tam Makale · 2025
Yazarlar: Kaan KESKİN, Yasir Catal, Angelika Wolman, Mehmet Çağdaş EKER, Ali Saffet Gonul, Georg Northoff
NeuroImage, Uluslararası, Wos, 2025, cilt 312, ss 121221-121221
DOI: 10.1016/j.neuroimage.2025.121221
Özet
• Healthy subjects show lower myelin content, longer ACW, and decreased GSCORR in self regions compared to non-self regions. • Multi-level analysis revealed ACW as a mediator between myelin and GSCORR, with a hierarchical model showing higher predictive accuracy than direct linear models, emphasizing the role of intrinsic neural timescales in shaping global brain activity. • Neural mass modeling indicated that myelin-related differences in recurrent excitatory connections and excitatory tone approximate distinctions between self/non-self and unimodal/transmodal regions. Understanding the brain's intrinsic architecture has long been a central focus of neuroscience, with recent advances shedding light on its topographic organization along uni and transmodal regions. How the brain’s global uni-transmodal topography relates to psychological features like our sense of self remains yet unclear, though. We here combine fMRI brain imaging with computational modeling (Wilson Cowan model) to better understand the temporal, spatial and physiological features underlying the distinction of self and non-self regions within the brain’s global topography. fMRI resting state shows lower myelin content, longer timescales (measured by the autocorrelation window/ACW), and lower global functional connectivity/synchronization (measured by global signal correlation/GSCORR) in self regions (based on the three-layer self topography; Qin et al. 2020) compared to non-self regions. Next, we fit the fMRI data with a neural mass model, the Wilson-Cowan model, which is enriched by structural and functional connectivity data from human MRI/fMRI. We first replicate the empirical data with longer ACW and lower GSCORR in self regions. Next, we demonstrate that self and non-self regions can, based on the same measures in the model, not only be distinguished within the brain’s global topography but also within the unimodal and transmodal regions themselves, respectively. Finally, the neural mass model shows that such topographic differentiation relates to two physiological features: self regions exhibit higher intra-regional excitatory recurrent connection and higher levels in their basal neural excitation than non-self regions. Our findings demonstrate the intrinsic nature of the distinction of self and non-self regions within the brain’s global uni-transmodal topography as well as their underlying physiological differences with higher levels in both recurrent connections and neural excitation in self regions. The increased recurrent connections in self regions, together with their higher levels of neural excitation and the longer autocorrelation window, may be ideally suited to mediate their self-referential processing: this can thus be seen as a form of ‘psychological recurrence’ where one and the same input/stimulus is processed in a prolonged echo-chamber like way, that is, an internal echo within the self regions themselves.