On the topochronic map of the human brain dynamics

Abstract

Large-scale brain activity evolves dynamically over time across multiple time-scales. The structural connectome imposes a spatial network constraint since two structurally connected brain regions are more likely to coordinate their activity. It also imposes a temporal network constraint by virtue of time delays via signal transmission, which has modulatory effects on oscillatory signals. Specifically, the lengths of the structural bundles, their widths, myelination, and the topology of the structural connectome influence the timing of the interactions across the brain network. Together, they define a space-time structure (topochronic map) spanned by connection strengths (space) and signal transmission delays (time), which together establish the deterministic scaffold underlying the evolution of brain dynamics. We introduce a novel in vivo approach for directly measuring functional delays across the whole brain using magneto/electroencephalography and integrating them with the structural connectome derived from magnetic resonance imaging. We developed a map of the functional delays characterizing the connections across the human brain and a map of the corresponding functional velocities. This yields a topochronic map of the human brain dynamics. The functional delays are tightly regulated, with a trend showing, as expected, that larger structural bundles had faster velocities, with the delays varying much less than expected if they only depended upon distance. Then, we estimated the delays from magnetoencephalography (MEG) data in a cohort of multiple sclerosis patients, who have damaged myelin sheaths, and demonstrated that patients showed greater delays across the whole network than a matched control group. Furthermore, within each patient, individual lesioned connections were slowed down more than unaffected ones. Our technique provides a practical approach for estimating functional transmission delays in vivo at the single-subject and single-fiber level and, thus, opens the possibility for novel diagnostic and curative interventions as well as providing empirical, subject-specific constraints to tailor brain models.

Efficient Inference on a Network of Spiking Neurons using Deep Learning

“…On the other hand, SBI offers efficient Bayesian estimation, even without detailed knowledge of the system's state-space representation. This aligns with findings from recent studies that highlight the efficiency of SBI across various challenging inverse problems (Gonçalves et al, 2020;Deistler et al, 2022;Boelts et al, 2022Boelts et al, , 2023Hashemi et al, 2023;Lavanga et al, 2023;Yalccinkaya et al, 2023;Rabuffo et al, 2023;Sorrentino et al, 2023). By benchmarking and addressing questions such as computational cost, uncertainty quantification, inter-dependency exploration, and data availability, we conclude that SBI is more efficient than alternatives in making informed choices from microscopic states to emergent dynamics at the macro scale.…”

Neuronal avalanches in temporal lobe epilepsy as a diagnostic tool: a noninvasive investigation of intrinsic resting state dynamics

“…Using source-reconstructed EEG signals, we individuated the presence of NA and used them to calculate the probability of consecutive activations between brain regions (i.e., the topographical spreading of the avalanches). This information was stored within an adjacency matrix named avalanche transition matrix (ATM) (Sorrentino et al, 2022) which was then used as a feature for a support vector machine (SVM) classifier.…”