A Novel Algorithm to Derive Spread of Excitation Based on Deconvolution.

Area of Science:

• Auditory Neuroscience
• Biomedical Engineering
• Signal Processing

Background:

• The width of the spread of excitation (SOE) curve is traditionally used to estimate neural excitation.
• Previous research has explored correlations between psychophysical parameters and SOE curve width, but definitive relationships remain elusive.
• Current methods may not accurately reflect the true neural SOE.

Purpose of the Study:

• To develop a novel analytical tool to derive SOE by treating SOE curves as a convolution operation.
• To apply deconvolution principles to retrieve excitation areas from measured SOE curves.
• To more accurately reveal the neural spread of excitation (SOE).

Main Methods:

• Analyzed intraoperative SOE curves from 16 subjects with Advanced Bionics implants.
• Recorded evoked compound action potential (ECAP)-based SOE curves using a forward masker paradigm.
• Compared measured SOE curves with predicted SOE curves derived from the convolution of excitation density profiles (EDPs).
• Iteratively adjusted EDPs to fit predicted SOE curves to measured data.

Main Results:

• Achieved a good fit between predicted and measured SOE curves, including asymmetry.
• Smoother EDPs (exponential or Gaussian) provided better fits than rectangular EDPs.
• Found that EDP width generally decreased from the apex to the base of the electrode array.
• EDP widths provided a qualitatively distinct measure of SOE compared to conventional methods.

Conclusions:

• ECAP-based SOE curves measured with forward masking can be modeled as a convolution of EDPs.
• A sloping excitation area is necessary to accurately mimic SOE recordings, as shown by the poor fit with rectangular EDPs.
• The deconvolution method explains the asymmetry of SOE curves as a result of wider apical excitation areas.
• Broader apical EDPs correlate with higher eCAP amplitudes in apical stimulation.