Advanced Analysis of Electroretinograms Based on Wavelet Scalogram Processing

• Published in: Applied sciences Vol. 12; no. 23; p. 12365
• Main Authors: Zhdanov, Aleksei, Dolganov, Anton, Zanca, Dario, Borisov, Vasilii, Ronkin, Mikhail
• Format: Journal Article
• Language: English
• Published: Basel MDPI AG 01-12-2022
• Subjects: Accuracy; biomedical research; Classification; Decomposition; Digital imaging; Disadvantages; Electrophysiology; electroretinogram; Electroretinograms; Electroretinography; ERG; Evaluation; Fourier transforms; Image processing; International standards; Latency; Learning algorithms; Machine learning; Methods; Ophthalmology; Pediatrics; Physiology; Retina; Signal analysis; Signal processing; Toxicology; Wavelet transforms
• ISSN: 2076-3417; 2076-3417
• DOI: 10.3390/app122312365

The electroretinography (ERG) is a diagnostic test that measures the electrical activity of the retina in response to a light stimulus. The current ERG signal analysis uses four components, namely amplitude, and the latency of a-wave and b-wave. Nowadays, the international electrophysiology community established the standard for electroretinography in 2008. However, in terms of signal analysis, there were no major changes. ERG analysis is still based on a four-component evaluation. The article describes the ERG database, including the classification of signals via the advanced analysis of electroretinograms based on wavelet scalogram processing. To implement an extended analysis of the ERG, the parameters extracted from the wavelet scalogram of the signal were obtained using digital image processing and machine learning methods. Specifically, the study focused on the preprocessing of wavelet scalogram as images, and the extraction of connected components and thier evaluation. As a machine learning method, a decision tree was selected as one that incorporated feature selection. The study results show that the proposed algorithm more accurately implements the classification of adult electroretinogram signals by 19%, and pediatric signals by 20%, in comparison with the classical features of ERG. The promising use of ERG is presented using differential diagnostics, which may also be used in preclinical toxicology and experimental modeling. The problem of developing methods for electrophysiological signals analysis in ophthalmology is associated with the complex morphological structures of electrophysiological signal components.