Statistical Inference of Functional Connectivity in Neuronal Networks using Frequent Episodes

AbstractIdentifying the spatio-temporal network structure of brain activity from multi-neuronal data streams is one ofthe biggest challenges in neuroscience. Repeating patterns of precisely timed activity across a group of neurons ispotentially indicative of a microcircuit in the underlying neural tissue. Frequent episode discovery, a temporal datamining framework, has recently been shown to be a computationally ecient method of counting the occurrences ofsuch patterns. In this paper, we propose a framework to determine when the counts are statistically signi cant bymodeling the counting process. Our model allows direct estimation of the strengths of functional connections be-tween neurons with improved resolution over previously published methods. It can also be used to rank the patternsdiscovered in a network of neurons according to their strengths and begin to reconstruct the graph structure of thenetwork that produced the spike data. We validate our methods on simulated data and present analysis of patternsdiscovered in data from cultures of cortical neurons.keywords: event sequences, spike trains, multi-electrode array, microcircuits, temporal data mining, frequentepisodes, non-overlapped occurrences, statistical inferences