Fig. 1: Illustration of time-delay embedded Hidden Markov Model (TDE-HMM) approach applied on MEG data. | Communications Biology

Fig. 1: Illustration of time-delay embedded Hidden Markov Model (TDE-HMM) approach applied on MEG data.

From: A Hidden Markov Model reveals magnetoencephalography spectral frequency-specific abnormalities of brain state power and phase-coupling in neuropathic pain

Fig. 1

MEG data preprocessing (Independent component analysis (ICA) for artifact removal, source reconstruction, parcellation, leakage correction, and sign disambiguation—see Methods). We used a linearly constrained minimum variance beamformer to extract a continuous time series for nodes of the dynamic pain connectome based on anatomical 3D T1 scans of each participant. Spontaneous cortical activity transiently organizes into frequency-specific phase-coupling networks. Principal component analysis was applied for dimensionality reduction, and HMM inference was then used to identify the state time-courses (state probability) and the state parameters in HCs and those with neuropathic pain. Each state was characterized as having its own distinct spatial, temporal, and spectral properties. Individual power maps and phase-coupling patterns (networks) were estimated for active states and illustrated by their temporal features (i.e., fractional occupancy, interval times, switching rates, and lifetimes). States were spatially defined and spectrally resolved according to the main frequency bands (i.e., delta/theta, alpha, beta, and low gamma).

Back to article page