How non-reciprocity, plasticity and feedback determine the energetic cost of collective performance in biological networks
Tuan Pham (DIEP Research Fellow)
Abstract:
Non-reciprocal and fluctuating interactions fundamentally reshape the collective dynamics of biological networks by driving them out of equilibrium and hence incurring an energetic cost for achieving a target performance. In this talk, we present a theoretical framework that connects irreversibility with the emergence of collective behaviors near criticality, including the onset of oscillation and the edge of chaos. Specifically, we use two example systems for illustration: a neural network with synaptic connections evolving on a concurrent timescale with the neuronal dynamics and a cellular network with feedback between its two layers. For the former system, a universal scaling between entropy production rate and the population-level variance of the neural activity is established and tested in Human-Brain data. For the latter system, we found an energy-driven switching from sensing-based to a memory-based behavior, as quantified by the information flow between the two layers.
This seminar will take place in room C3.011