Recent rapid progresses of computational and experimental neuroscience are starting to uncover key principles by which neuronal systems realize robust information processing that can often outperform conventional approaches especially for ambiguous real-world tasks. Computation is an inseparable pa…
Recent rapid progresses of computational and experimental neuroscience are starting to uncover key principles by which neuronal systems realize robust information processing that can often outperform conventional approaches especially for ambiguous real-world tasks. Computation is an inseparable part of the communication among neurons and it is realized by the continuous transmission of spikes among neurons in extremely large-scale brain networks. Despite the rather unreliable nature of neurons and synapses, complex interaction between them finally forms robust and reliable communication, for which the topology of large-scale functional networks and noise-induced phenomena in the brain can be utilized. These remarkable features of communication in the brain seem similar with the requirements of currently considered future information networks that should be highly scalable to number of nodes, robust against perturbations, adaptive to environmental changes, and tolerant of noise and diversity.
