Dynamic Analysis on Simultaneous iEEG-MEG Data via Hidden Markov Model

Background: Intracranial electroencephalography (iEEG) recordings are used for clinical evaluation prior to surgical resection of the focus of epileptic seizures and also provide a window into normal brain function. While these recordings afford detailed information about local brain activity, putting this activity in context and comparing results across patients is challenging. Non-invasive whole-brain Magnetoencephalography (MEG) could help translate iEEG in the context of overall brain activity, and thereby aid group analysis and interpretation. Methods: Simultaneous MEG-iEEG recordings were performed at rest on 11 patients with epilepsy. Pre-processed MEG sensor data was projected to source space. The time delay embedded hidden Markov model (HMM) technique was applied to find recurrent sub-second patterns of network activity in a completely data-driven way. To relate MEG and iEEG results, correlations were computed between HMM state time courses and iEEG power envelopes in equally spaced frequency bins and presented as correlation spectra for the respective states and iEEG channels. Results: Five HMM states were inferred from MEG. Two of them corresponded to the left and right temporal activations and had a spectral signature primarily in the theta/alpha frequency band. The majority of iEEG contacts were also located in left and right temporal areas and the theta/alpha power of the local field potentials (LFP) recorded from these contacts correlated with the time course of the HMM state corresponding to the temporal lobe of the respective hemisphere. Discussion: Our findings are consistent with the fact that most subjects were diagnosed with temporal epilepsy and implanted with temporal electrodes. As the placement of electrodes between patients was inconsistent, their modulation by HMM states could help group the contacts into functional clusters. This is the first time that HMM was applied to simultaneously recorded iEEG-MEG and our pipeline could be used in future similar studies.