Signal processing techniques for oscillatory potential extraction in the electroretinogram: automated highpass cutoff...

Area of Science:

• Ophthalmology
• Neuroscience
• Biomedical Engineering

Background:

• Oscillatory potentials (OPs) are crucial components of the electroretinogram (ERG), reflecting retinal function.
• Accurate isolation of OPs requires precise bandpass filtering, specifically the highpass cutoff frequency.
• Improper highpass cutoff frequencies can lead to contamination or loss of significant OP signal energy.

Purpose of the Study:

• To develop and evaluate automated methods for estimating the optimal highpass cutoff frequency for OP isolation from ERG signals.
• To compare the performance of automated methods against expert-selected cutoff frequencies across a range of luminances.
• To assess the influence of luminance on the critical selection of highpass cutoff frequency.

Main Methods:

• Two automated methods were developed: OP amplitude analysis and time-varying exponential model fitting to the b-wave.
• The OP amplitude analysis method utilized trends in maximum OP amplitude variation with cutoff frequency.
• The exponential fitting method minimized errors between the residual ERG signal and the fitted model.

Main Results:

• Automated cutoff frequency selection was less critical at low luminances but significantly impacted OP signal shape at higher luminances.
• The OP amplitude analysis method yielded estimations statistically indistinguishable from expert selections.
• Both automated methods provided excellent fits to manual selections at the four highest stimulus luminance values.

Conclusions:

• Automated methods for highpass cutoff frequency estimation in ERG analysis are effective and reliable.
• The OP amplitude analysis method demonstrates particular promise for accurate and objective OP isolation.
• These automated approaches can improve the consistency and efficiency of ERG analysis in research and clinical settings.