Cochlear Implant Artifacts Removal in EEG-Based Objective Auditory Rehabilitation Assessment

Area of Science:

• Neuroscience
• Biomedical Engineering
• Signal Processing

Background:

• Cochlear implants (CI) restore hearing but introduce electrical artifacts in electroencephalography (EEG) data.
• These artifacts complicate objective assessments of cerebral plasticity and auditory rehabilitation outcomes in CI recipients.
• Existing artifact removal methods struggle with automatic detection and minimizing information loss.

Purpose of the Study:

• To develop an automated method for detecting and removing CI-induced electrical artifacts from EEG signals.
• To improve the reliability of EEG-based assessments of cerebral plasticity in CI users.
• To enable more accurate evaluation of advanced auditory functions post-cochlear implantation.

Main Methods:

• A novel approach combining machine learning (Support Vector Machines - SVM), Independent Component Analysis (ICA), and Ensemble Empirical Mode Decomposition (EEMD).
• Automatic detection of CI artifacts based on the temporal properties of EEG signals.
• Processing and removal of artifacts using EEMD and ICA to refine EEG data.

Main Results:

• The proposed method successfully automates the detection and reduction of CI electrical artifacts in EEG.
• Corrected EEG recordings from CI recipients showed strong alignment with normal-hearing individuals across temporal, frequency, and spatial domains.
• Reliable neural responses across the scalp were restored, free from CI artifacts.

Conclusions:

• The developed approach effectively eliminates CI artifacts from EEG data, restoring reliable neural signal assessment.
• This method offers a significant advancement for objective evaluation of brain function and auditory rehabilitation in CI users.
• Accurate EEG analysis is crucial for understanding cerebral reorganization and optimizing CI outcomes.