A Model-Free Method to Quantify Memory Utilization in Neural Point Processes.

Abstract

Objective: Quantifying the predictive capacity of a neural system, intended as the capability to store information and actively utilize it for dynamic system evolution, is a key component of neural information processing. Information storage (IS), the main information-theoretic measure quantifying the active utilization of memory in a dynamic system, is only defined for discrete-time processes, and although recent theoretical work laid the foundations for its continuous-time analysis, a reliable computation method is still needed for broader application to neural data.

Methods: This work introduces a method for the model-free estimation of the so-called memory utilization rate (MUR), the continuous-time counterpart of the IS, specifically designed to quantify the predictive capacity stored in neural point processes. Moreover, a surrogate data-based procedure is used to correct estimation bias and detect significant memory levels in the analyzed point process.

Results: The method is first validated in simulations of Poisson processes, both memoryless and with memory, as well as in realistic models of coupled cortical dynamics and heartbeat dynamics. It is then applied to real spike trains reflecting central and autonomic nervous system activities: in spontaneously growing cortical neuron cultures, the MUR detected increasing levels of memory utilization across maturation stages, linked to the emergence of synchronized bursting; in heartbeat modulation analysis, the MUR reflected sympathetic activation and vagal withdrawal occurring with postural stress, but not with mental stress.

Conclusion and significance: The proposed approach offers a novel, computationally reliable tool for the analysis of spike train data in computational neuroscience and physiology.