Automated spatiotemporal response identification and separation for averaged and single-trial EEG and MEG data

Area of Science:

• Neuroscience
• Biomedical Engineering
• Signal Processing

Background:

• Conventional electroencephalography (EEG) and magnetoencephalography (MEG) analysis methods rely on assumptions of independence and fixed parameters, which do not reflect the complex, overlapping nature of cortical activity.
• These limitations hinder accurate interpretation of neural responses and limit the validity of research and clinical inferences.
• Existing methods often fail to account for spatial and temporal overlaps in neural activity and individual variability.

Purpose of the Study:

• To develop a novel algorithm for reliably isolating spatiotemporally localized neural evoked responses.
• To overcome the limitations of conventional analysis approaches that assume independence or uncorrelatedness of neural signals.
• To enhance the accuracy and interpretability of EEG and MEG data analysis for improved research and clinical applications.

Main Methods:

• Developed Spatiotemporal Event Response ENcoding (SEREN), an algorithm utilizing Gaussian kernels in time and space to leverage spatiotemporal density properties of post-synaptic currents.
• SEREN automatically identifies and extracts spatiotemporally localized evoked response components.
• The algorithm operates in both sensor and source space and can process both averaged and single-trial data.

Main Results:

• Demonstrated SEREN's effectiveness on auditory and visual-evoked MEG data and simulated datasets.
• Showcased SEREN's capability for robust single-trial monitoring in noisy EEG systems using transcranial magnetic stimulation-evoked potentials.
• Validated SEREN's performance in simulated real-time applications.

Conclusions:

• SEREN reliably isolates cortical evoked responses, addressing limitations of conventional methods that overlook inter-response overlaps and individualization.
• The algorithm offers improved precision in neural response extraction, advancing the analysis of neural dynamics.
• SEREN provides a powerful tool for enhancing the validity of research and clinical applications in neuroscience.