Skip to main content
Figure 9 | BioMedical Engineering OnLine

Figure 9

From: Wavelet-based Gaussian-mixture hidden Markov model for the detection of multistage seizure dynamics: A proof-of-concept study

Figure 9

The transition diagrams and the posterior probabilities of the reduced feature space after mRMR and AICc. (a) The state transition diagram for HMMAIC is given. Similar to the HMMopt14D, the some early tonic activity permissive to seizures is allowed to return to the interictal state. Between the late tonic firing and postictal state, the transition follows a unidirectional progression. (b) The mapping of HMMAIC states to the electrophysiological LFP is shown. In this illustrative example, the marginal probabilities of the model states follow a progression from non-ictal, tonic firing to chronic seizures and postictal state, before returning to non-ictal. The late tonic firing state in this model is typically shorter than the model from Figure 5b because of the significantly smaller transition probability back to itself.

Back to article page