Ver. 2 - Time-dependent POMDP model

This release corresponds to the analysis comparing the static POMDP model (previous release) with the time-dependent extension, where the observation matrix contains different observation models for each POMDP time-step in a given trial. This allows the model to leverage more brain data as the trial advances, similar to other dynamic classification approaches used in reactive-BCI.

Ver 1 - Simple POMDP model

First version of the POMDP-BCI that implements a POMDP as an Active Sensing problem to function as the decision step of the BCI process. The same POMDP problem works for three types of BCI: SSVEP, CVEP, MI.

The results are (or will be) published in "POMDP-BCI: A Benchmark of Active BCI Using POMDP to Issue Commands", by J. J. Torre, C. P. C. Chanel and F. Dehais (in press).