2013/11/27 by Guillaume Lajoie, Jean-Philippe Thivierge, Lajoie, Guillaume +5
Biochemistry, Genetics and Molecular Biology · Computer Science · Mathematics · Neuroscience · Physics and Astronomy · #37H99 #92B20 #Artificial intelligence #Artificial neural network #Biological Physics (physics.bio-ph) #Biology #Chaotic #Chaotic Dynamics (nlin.CD) #Computer science #Dynamical Systems (math.DS) #ENCODE #Entropy (arrow of time) #FOS: Biological sciences #FOS: Mathematics #FOS: Physical sciences #Mathematics #Network dynamics #Neural Networks and Applications #Neural dynamics and brain function #Neurons and Cognition (q-bio.NC) #Physics #Spike (software development) #Spike train #Statistical physics #Stimulus (psychology) #math.DS #msc:37H99 #msc:92B20 #nlin.CD #physics.bio-ph #q-bio.NC #stochastic dynamics and bifurcation
paper · pdf · doi:10.48550/arxiv.1311.7128
9 pages, 5 figures
openalex publication_date 2013/11/27 · arxiv created 2014/02/22 · arxiv updated 2014/02/25 · openalex created_date 2022/10/02 · openalex updated_date 2026/08/06
Large networks of sparsely coupled, excitatory and inhibitory cells occur\nthroughout the brain. A striking feature of these networks is that they are\nchaotic. How does this chaos manifest in the neural code? Specifically, how\nvariable are the spike patterns that such a network produces in response to an\ninput signal? To answer this, we derive a bound for the entropy of multi-cell\nspike pattern distributions in large recurrent networks of spiking neurons\nresponding to fluctuating inputs. The analysis is based on results from random\ndynamical systems theory and is complimented by detailed numerical simulations.\nWe find that the spike pattern entropy is an order of magnitude lower than what\nwould be extrapolated from single cells. This holds despite the fact that\nnetwork coupling becomes vanishingly sparse as network size grows -- a\nphenomenon that depends on ``extensive chaos," as previously discovered for\nbalanced networks without stimulus drive. Moreover, we show how spike pattern\nentropy is controlled by temporal features of the inputs. Our findings provide\ninsight into how neural networks may encode stimuli in the presence of\ninherently chaotic dynamics.\n