Dynamic 3D Clustering of Spatio-temporal Brain Data in the NeuCube Spiking Neural Network Architecture on a Case Study of fMRI and EEG Data

Abstract

The paper presents a novel clustering method for dynamic Spatio-Temporal Brain Data (STBD) on the case study of functional Magnetic Resonance Image (fMRI). The method is based on NeuCube spiking neural network (SNN) architecture, where the spatio-temporal relationships between STBD streams are learned and simultaneously the clusters are created. The clusters are represented as groups of spiking neurons inside the NeuCube’s spiking neural network cube (SNNc). The centroids of the clusters are predefined by spatial location of the brain data sources used as input variables. We illustrate the proposed clustering method on an fMRI case study STBD recorded during a cognitive task. A comparative analysis of the clusters across different mental activities can reveal new findings about the brain processes under study.