Time-Frequency Analysis of ERG With Discrete Wavelet Transform and Matching Pursuits for Glaucoma

Area of Science:

• Ophthalmology
• Biomedical Engineering
• Signal Processing

Background:

• Glaucoma is a leading cause of irreversible blindness.
• Accurate prediction of glaucoma severity is crucial for effective patient management.
• Current diagnostic methods may not fully capture the nuances of disease progression.

Purpose of the Study:

• To evaluate two time-frequency feature extraction techniques for predicting glaucoma severity.
• To compare the performance of discrete wavelet transform (DWT) and matching pursuit (MP) in analyzing electroretinograms (ERGs).
• To determine if time-frequency features improve upon traditional amplitude markers for glaucoma assessment.

Main Methods:

• Electroretinograms (ERGs) targeting the photopic negative response were recorded from 103 eyes of 55 glaucoma patients.
• Time-frequency features were extracted using DWT and MP decomposition.
• Linear and multivariate adaptive regression spline (MARS) models were used to predict retinal ganglion cell counts, a glaucoma severity measure.

Main Results:

• Models incorporating DWT and MP time-frequency features, along with amplitude markers, significantly outperformed models using only amplitude markers (P = 0.001 for linear, P ≤ 0.011 for MARS).
• MARS models with DWT and MP features achieved higher proportions of variance explained (R2 = 0.53 and 0.63, respectively) compared to amplitude markers alone (R2 = 0.34).
• These findings demonstrate the added predictive value of novel time-frequency features.

Conclusions:

• Novel time-frequency features derived from photopic ERGs substantially enhance the prediction of glaucoma severity.
• These advanced features offer significant advantages over traditional time-domain amplitude markers alone.
• The study highlights the potential of ERG time-frequency analysis in glaucoma management.