Trial-by-trial detection of cognitive events in neural time-series

Abstract

Measuring the time-course of neural events that make up cognitive processing is crucial to understand the relation between brain and behavior. To this aim, we formulated a method to discover a trial-wise sequence of events in multivariate neural signals such as electro- or magneto-encephalograpic (E/MEG) recordings. This sequence of events is assumed to be represented by multivariate patterns in neural time-series, with inter-event durations following probability distributions. By estimating event-specific multivariate patterns, and between-event duration distributions, the method allows to recover the by-trial onsets of brain responses. We demonstrate the properties and robustness of this hidden multivariate pattern (HMP) method through simulations, including robustness to low signal-to-noise ratio, as typically observed in EEG recordings. The applicability of HMP is illustrated using previously published data from a speed-accuracy trade-off task. We show how HMP provides, for any experiment or condition, an estimate of the number of events, the sensors contributing to each event (e.g. EEG scalp topography), and the durations between each event. Traditional exploration of tasks' cognitive architectures can thus be enhanced by HMP estimates.

Bayesian estimation of group event-related potential components (BEGEP): testing a model for synthetic and real datasets

“…However, in this case, it will be impossible to perform quantitative comparisons of signal estimates and other parameters of model ( 1), obtained for different subjects, due to the inherent uncertainty in the scales and signal shifts along the time axis for this model. Perhaps for the analysis of ERP signals in single trials, it would be better to use specially developed methods, for example HsMM-MVPA [53,54], for which the above uncertainty is not inherent.…”