Modulation of proximity to criticality enhances slow activity fluctuations during free recall

Abstract

AbstractCategorically bounded free recall allows generating perceptual and cognitive contents within specific categories while avoiding unrelated intrusions. Previous research suggested that this is implemented via amplification of ultra-slow spontaneous activity fluctuations, initiating a spontaneous recall event. However, the underlying amplification mechanism remains unclear. Here, we demonstrate, using a simulation of a simple random recurrent neuronal network operating near a critical point, that such selective amplification can be generated by a small shift towards this critical point, resulting in a dynamical phenomenon termed “critical slowing down”. By fitting physiological parameters and applying stochastic white noise input, we simulated ultra-slow fluctuations observed during rest and categorically bounded visual recall in the human cortex. Our findings suggest that modulation of spontaneous fluctuations linked to free recall can be explained by a stochastically driven recurrent network near a critical point, providing insight into the rapid and flexible formation of categorical boundaries in human cognition.