Electrocorticographic interictal spike removal via denoising source separation for improved neuroprosthesis control

Area of Science:

• Neuroscience
• Biomedical Engineering
• Signal Processing

Background:

• Electrocorticographic (ECoG) neuroprostheses offer communication and control for individuals with motor impairments.
• Clinical use of subdural electrodes is limited to epilepsy diagnostics, resulting in interictal activity in sensorimotor cortex recordings.
• Interictal spikes can obscure amplitude-modulated features crucial for modeling hand trajectories.

Purpose of the Study:

• To investigate the impact of removing interictal activity on the linear prediction of hand trajectories using ECoG data.
• To assess the efficacy of denoising source separation techniques in improving ECoG-based motor control models.

Main Methods:

• Utilized denoising source separation framework to process ECoG recordings.
• Exploited the quasiperiodic nature of interictal spikes for targeted removal.
• Applied linear prediction models to hand trajectory data before and after spike removal.

Main Results:

• Removal of interictal activity significantly improved the linear prediction of hand trajectories.
• Denoising source separation effectively isolated and removed disruptive interictal spikes.
• Enhanced signal clarity facilitated more accurate modeling of motor function.

Conclusions:

• Interictal spikes present a significant challenge in ECoG-based neuroprosthetics.
• Denoising source separation is a viable method for mitigating the effects of interictal activity.
• Improving signal quality through artifact removal enhances the potential of ECoG neuroprostheses for restoring motor control.