Edinburgh Napier University

  Recent research shows that activity patterns play an important role in neural information processing. This seems to be the case both at micro- and macro-level processes within the nervous system. An important question is how to extract such activity pattern languages from neural data. Here we present a methodology that we use to extract an activity pattern language from high-resolution EEG data. First we determine typical activity patterns in the data, and then we establish probabilistic continuation rules describing which typical activity patterns follow earlier activity patterns present in the data. The accuracy of the established rules is tested using previously unseen data.