Automatic positioning of cursors in the transient pattern electroretinogram (PERG) with very poor SNR using an Expert System.

An expert system is described which automatically cursors PERG waveforms of very poor signal-to-noise ratio. The training data-set is derived from shape-perturbated PERG waveforms based on the ISCEV Standard and a series of clinically-normal waveforms. In validation trials, the Expert System is clearly shown to out-perform the Human Expert in the presence of noise. All software was written in MatLab((R)). It is proposed that this system is generic and similarly applicable to other transient recordings. A comprehensive Internet demonstration (SPoC: Smart Positioning of Cursors) is maintained at www.liverpooleye.org.