Empirical Mode Decomposition-Based Filter Applied to Multifocal Electroretinograms in Multiple Sclerosis Diagnosis

Area of Science:

• Ophthalmology
• Neuroscience
• Biomedical Engineering

Background:

• Multiple sclerosis (MS) frequently impacts the visual pathway, making visual electrophysiological tests valuable for diagnosis.
• Early detection of MS is crucial for effective management and treatment.

Purpose of the Study:

• To investigate novel processing methods for multifocal electroretinogram (mfERG) recordings to enhance the diagnostic capability for MS.
• To assess the efficacy of empirical mode decomposition (EMD) filtering combined with normative database correlation for MS detection.

Main Methods:

• mfERG recordings were obtained from early-stage MS patients and healthy controls.
• A normative database was established from control subject signals.
• Empirical mode decomposition (EMD) was employed for filtering mfERG signals, with correlation to the normative database serving as a classification feature.

Main Results:

• The combined EMD-based filtering and correlation method achieved a mean area under the curve (AUC) of 0.90.
• High discriminant capacity was observed in specific regions, with AUC values of 0.96 for ring 4 and 0.94 for the inferior nasal quadrant.
• The approach demonstrated significant potential for classifying early-stage MS patients.

Conclusions:

• The integration of EMD filtering and normative database correlation offers a robust method for analyzing mfERG waveforms.
• This technique shows applicability for assessing multiple sclerosis in early-stage patients, potentially improving diagnostic accuracy.
• Further research can validate this approach for broader clinical use in MS diagnostics.