Coupled hidden Markov models for biosignal interaction modelling

Analysis of physiological systems frequently involves studying the interactions between them. In fact, the interactions themselves can give us clues about the state of health of the patient. For instance, interactions between cardio-respiratory variables are known to be abnormal in autonomic neuropathy. Conversely, if the type of interaction is known a departure from this state is, by definition, an anomaly. For instance, in multichannel EEG analysis we would expect coupling between the channels. If the states of the channels are divergent one may suspect some recording error (electrode failure). Traditionally, such interactions would be measured using correlation or, more generally, mutual information measures. There are, however, major limitations associated with correlation, namely, the linearity and stationarity assumptions. Mutual information, on the other hand, can handle nonlinearity. However, it is unable to infer anti-correlation and it also assumes stationarity (for the density estimation). Some recent work has focused on hidden Markov models (HMM) in order to avoid some of the stationarity assumptions (the chain is still assumed to be ergodic) and model the dynamics explicitly. HMMs handle the dynamics of the time series through the use of a discrete state-space. The paper discusses coupled HMM for biosignal interaction modeling