A three-layer backpropagation neural network is applied to electroretinographical (ERG) waveforms in an attempt to classify eye diseases. Three classes of ERG responses are used: normal, congenital stationary night blindness (CSNB), and rod/cone retinitis pigmentosa (R/C RP). Highly successful classification has been achieved by examining only a single ERG framer. This result demonstrates the potential clinical usefulness of the combination of electroretinograms and neural signal processing in the detection of eye diseases. As the problem is scaled up to include more diseases, the classification problem will become more difficult. Future work needs to include multiple ERG frames in the feature extraction process.

Additional Metadata
Conference Proceedings of the 13th Annual International Conference of the IEEE Engineering in Medicine and Biology Society
Lipoth, Leon L., Hafez, H.M., & Goubran, R. (1991). Electroretinographical (ERG) based classification of eye diseases. In Proceedings of the Annual Conference on Engineering in Medicine and Biology (pp. 1417–1418).