Skip to main content
Figure 5 | BioMedical Engineering OnLine

Figure 5

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

Figure 5

The transition diagram and the posterior probabilities for HMM opt14D using the original feature space. (a) The state transition diagram for HMMopt14D is given along with a sample LFP recording. For an eight state model, some of the states represent the same electrophysiological dynamics and hence can be consolidated. The HMM states S 1 , S 4 and S 7 represent the late tonic activities. They are interconnected and are always followed by the chronic activity and preceded by early tonic firing states S 2 and S 5 . The early tonic activity of this HMM is also allowed to return to interictal state without generating seizures. (b) The mapping of HMMopt14D to the electrophysiological events is provided. Typically, this model is able to differentiate early and late tonic firing. For this particular example, the model output switches between interictal and early tonic state until about 46 s.

Back to article page