Enhancing Electrical Cochlear Response Quality in Cochlear Implant Apical Electrodes Using Empirical Mode Decomposition

The auditory evoked potential, recorded on the scalp while an implanted subject uses their device in everyday mode, is called the Electrical Cochlear Response (ECR). The waveform and amplitude of this potential vary with sound stimulus intensity and frequency, thus containing useful information about hearing provided by Cochlear Implant (CI). Depending on the CI design criterion, the electrical current stimulation rate in apical electrodes may be higher than in middle and basal electrodes. A high stimulation rate produces an electrical artifact interfering with ECR detection and measuring in apical electrodes. In this paper, we compare two methods based on Empirical Mode Decomposition (EMD) to select and remove the electrical artifact produced in the apical electrodes due to a high current stimulation rate. The performance of both methods was evaluated using ECR signal-to-noise ratio (SNR) and waveform quality after artifact removal. Both methods generally had similar performance with a significant SNR improvement. However, only one of them preserves ECR waveform across electrodes. Clinical Relevance—Clinical judgment about hearing provided by Cochlear Implants to the patient is based on ECR detection and latency and peak amplitude measure in each stimulation electrode; hence, SNR and waveform quality improvement of ECR allows clinicians to determine these features better