Discovering Cognitive Stages in M/EEG Data to Inform Cognitive Models

Computational cognitive models aim to simulate the cognitive processes humans go through when performing a particular task. In this chapter, we discuss a machine learning approach that can discover such cognitive processes in M/EEG data. The method uses a combination of multivariate pattern analysis (MVPA) and hidden semi-Markov models (HsMMs), to take both the spatial extent and the temporal duration of cognitive processes into account. In the first part of this chapter, we will introduce the HsMM-MVPA method and demonstrate its application to an associative recognition dataset. Next, we will use the results of the analysis to inform a high-level cognitive model developed in the ACT-R (adaptive control of thought – rational) architecture. Finally, we will discuss how the HsMM-MVPA method can be extended and how it can inform other modeling paradigms.