Anatomical Harmonics Basis based Brain Source Localization with Application to Epilepsy

Brain Source Localization (BSL) using Electroencephalogram (EEG) has been a useful noninvasive modality for the diagnosis of epileptogenic zones, study of evoked related potentials, and brain disorders. The inverse solution of BSL is limited by high computational cost and localization error. The performance is additionally limited by head shape assumption and the corresponding harmonics basis function. In this work, an anatomical harmonics basis (Spherical Harmonics (SH), and more particularly Head Harmonics (H2)) based BSL is presented. The spatio-temporal four shell head model is formulated in SH domain. The performance of spatial subspace based Multiple Signal Classification (MUSIC) and Recursively Applied and Projected (RAP)-MUSIC method is compared with the proposed SH and H2 counterparts on simulated data. SH and H2 domain processing effectively resolves the problem of high computational cost without sacrificing the inverse source localization accuracy. The proposed H2 MUSIC was additionally validated for epileptogenic zone localization on clinical EEG data. The proposed framework offers an effective solution to clinicians in automated and time efficient seizure localization.

Comparison of Machine Learning Techniques Based Brain Source Localization Using EEG Signals

Background: The Brain is the most complex organ of human body with millions of connections and activations. The electromagnetic signals are generated inside the brain due to a mental or physical task performed. These signals excite a bunch of neurons within a particular lobe depending upon nature of task performed. To localize this activity, certain machine learning (ML) techniques in conjunction with a neuroimaging technique (M/EEG, fMRI, PET) are developed. Different ML techniques are provided in literature for brain source localization. Among them, the most common are: minimum norm estimation (MNE), low resolution brain electromagnetic tomography (LORETA) and Bayesian framework based multiple sparse priors (MSP). Aims: In this research work, EEG is used as a neuroimaging technique. Methods: EEG data is synthetically generated at SNR=5dB. Afterwards, ML techniques are applied to estimate the active sources. Each dataset is run for multiple trials (>40). The performance is analyzed using free energy and localization error as performance indicators. Furthermore, MSP is applied with variant number of patches to observe the impact of patches on source localization. Results: It is observed that with increased number of patches, the sources are localized with more precision and accuracy as expressed in terms of free energy and localization error respectively. Conclusion: The patches optimization within Bayesian Framework produces improved results in terms of free energy and localization error.