Electrically evoked compound action potentials artefact rejection by independent component analysis: procedure...

Area of Science:

• Biomedical Engineering
• Neuroscience
• Signal Processing

Background:

• Independent Component Analysis (ICA) has proven effective in separating electrically-evoked compound action potentials (ECAPs) from noise and stimulation artefacts.
• Previous methods relied on manual artefact rejection, introducing potential biases.

Purpose of the Study:

• To develop and validate an automated artefact cancellation process for ECAP analysis using ICA.
• To provide a less biased alternative to traditional artefact rejection methods in cochlear implant research.

Main Methods:

• Raw ECAP data were recorded from intra-cochlear electrodes without initial artefact rejection.
• Amplifier-saturated recordings were discarded, and data from various stimulus conditions were concatenated.
• Independent Component Analysis (ICA) was applied with a restriction to 4 sources, enabling automatic classification of artefact, noise, and ECAP components.

Main Results:

• The automated ECAP-ICA procedure was tested on 8 cochlear implant users.
• Artefact and noise components were successfully identified and discarded, isolating the ECAP.
• The automated method accurately extracted ECAPs in 22 out of 26 cases when compared to the standard clinical forward masking paradigm.

Conclusions:

• The automated ECAP-ICA procedure offers a robust and unbiased method for ECAP extraction.
• This technique eliminates the need for complex buffer combinations required by traditional methods.
• ECAP-ICA holds significant clinical relevance, particularly as a component in automated ECAP-threshold detection for cochlear implant fitting software.