Neuronal ensembles have been shown to exhibit synchronized activity that is thought to be related to behavioral tasks and learning. We develop an algorithm that adaptively identifies significant synchronization of neurons based on a multivariate model of simultaneous spiking processes. We extend the statistical inference framework of adaptive Granger causal analysis to this setting in order to quantify the strength and dynamics of synchronization. We demonstrate the utility of our proposed method on simulated ensemble spiking data.
展开▼