Continuous-wavelet-transform analysis of the multifocal ERG waveform in glaucoma diagnosis

Area of Science:

• Ophthalmology
• Biomedical Engineering
• Signal Processing

Background:

• Current multifocal electroretinogram (mfERG) analyses for glaucoma primarily focus on signal amplitudes and latencies.
• A need exists for advanced signal processing techniques to improve glaucoma detection accuracy.

Purpose of the Study:

• To investigate the application of wavelet analysis for diagnosing glaucoma using mfERG signals.
• To evaluate the efficacy of continuous wavelet transform (CWT) combined with neural network classification for glaucoma detection.

Main Methods:

• Continuous Wavelet Transform (CWT) using the real Morlet wavelet was applied to mfERG signals.
• CWT coefficients at the scale of maximum correlation were used as input features for a neural network classifier.
• mfERG data were collected from 47 chronic open-angle glaucoma patients and 24 healthy controls.

Main Results:

• The wavelet analysis achieved a high classification sensitivity of 0.894 for detecting glaucomatous sectors.
• A specificity of 0.844 was obtained, indicating accurate detection of healthy sectors.
• These results demonstrate improved diagnostic performance compared to previous analyses on the same dataset.

Conclusions:

• Wavelet analysis of mfERG signals is a promising tool for glaucoma diagnosis.
• The combination of CWT and neural networks offers a reliable method for identifying glaucomatous changes.
• This approach enhances the diagnostic accuracy of mfERG in detecting glaucoma.