A novel method to reduce noise in electroretinography using skin electrodes: a study of noise level, inter-session variability, and reproducibility

To determine the feasibility of recording reproducible electroretinograms (ERGs) with skin electrodes using a new ERG system. Seventeen healthy volunteers were studied. The dark-adapted, bright-flash ERGs were recorded with a new ERG recording system (LE-4000, Tomey, Nagoya, Japan) in which the stimulus alternated between the eyes every 15 s, and each eye was stimulated eight times. The active skin electrode was placed on the lower eyelids of both eyes. The voltage changes of the non-stimulated eye were subtracted from that of the stimulated eye to try to increase the signal-to-noise ratio for eight stimulus cycles. The noise levels were measured from 12 subjects with and without the subtraction steps. ERGs were also recorded on five different days from five subjects, and the coefficient of variation (CV) and the intra-class correlation coefficients (ICCs) were calculated. The noise level without the subtraction step was 18.4 ± 8.4 μV, and it was significantly reduced to 13.8 ± 4.0 μV with the subtraction step (P = 0.001). Reproducible ERGs were obtained from each subject, and the average CV for the five subjects was 6.1 % for the a-wave amplitude, 7.7 % for the b-wave amplitude, and 7.7 % for the sum of the oscillatory potential (OP) amplitude. The ICC was 0.76 for the a-wave amplitude, 0.68 for the b-wave amplitude, and 0.72 for the sum of the OPs amplitude. These findings indicate that our new ERG recording methods shows noise reduction and good reproducibility with low inter-session variability even with skin electrodes.