BAYESIAN PARALLEL FACTOR ANALYSIS FOR THE STUDY OF EVENT-RELATED POTENTIALS

The aim of this study was to develop a Bayesian probabilistic model for parallel factor analysis of event-related potentials (ERP) of the human brain. Twelve statistical models are proposed that take into account the characteristics of the signals of ERP sources. For these models, Bayesian inference procedures were developed based on Markov chains Monte Carlo sampling. The effectiveness of these procedures was evaluated using both synthetic data with different signal-to-noise ratios and an array of ERPs obtained from 351 subjects in the Go/NoGo task. The procedure for obtaining estimates of the model parameters of the best accuracy was chosen. An analysis of the dependence of the model signals on the type of human activity performed showed that Bayesian parallel factor analysis is able to identify functionally different components of the ERP.