2020/02/24 by Anar Amgalan, Patrick Taylor, Amgalan, Anar +5
Computer Science · Neuroscience · #FOS: Biological sciences #Functional Brain Connectivity Studies #Neural Networks and Applications #Neural dynamics and brain function #Neurons and Cognition (q-bio.NC)
paper · pdf · doi:10.48550/arxiv.2002.10568
openalex publication_date 2020/02/24 · openalex created_date 2022/07/26 · openalex updated_date 2026/08/04
Identifying the brain's neuronal cluster size to be presented as nodes in a\nnetwork computation is critical to both neuroscience and artificial\nintelligence, as these define the cognitive blocks required for building\nintelligent computation. Experiments support many forms and sizes of neural\nclustering, while neural mass models (NMM) assume scale-invariant\nfunctionality. Here, we use computational simulations with brain-derived fMRI\nnetwork to show that not only brain network stays structurally self-similar\ncontinuously across scales, but also neuron-like signal integration\nfunctionality is preserved at particular scales. As such, we propose a\ncoarse-graining of network of neurons to ensemble-nodes, with multiple spikes\nmaking up its ensemble-spike, and time re-scaling factor defining its\nensemble-time step. The fractal-like spatiotemporal structure and function that\nemerge permit strategic choice in bridging across experimental scales for\ncomputational modeling, while also suggesting regulatory constraints on\ndevelopmental and/or evolutionary "growth spurts" in brain size, as per\npunctuated equilibrium theories in evolutionary biology.\n