Abstract
PURPOSE: The slow flash multifocal electroretinogram (SF-mfERG) stimulates with focal flashes separated by extended dark periods. Compared to standard multifocal stimulation, this permits more complete response development before the next flash and facilitates interpretation of late response components. We wished to determine whether these characteristics enable the SF-mfERG to detect and map diabetic retinal dysfunction. METHODS: Eighteen controls, 12 diabetics without retinopathy (D), and 17 diabetics with mild retinopathy (D+R; predominantly microaneurisms) were tested using VERIS 4. 103 areas of the central 45 deg were stimulated pseudorandomly by flashes (100 cd/m2) separated by at least 53 ms. First order SF-mfERGs were examined in two ways. First, mean responses were computed for each subject in 6 groups concentric with fixation and N1, P1 and N2 implicit times (ITs) and N1, N1-P1, and N2 amplitudes were measured. Second, each subject's N1, P1 and N2 ITs and scalar product amplitudes (SPAs) were measured at all 103 locations. Abnormalities were defined as p<=0.05 and as local values >= 2 Z-scores from controls. RESULTS: [1. Concentric response groups] Abnormal ITs occurred in D+R subjects in N1 (2/6 groups), P1 (4/6) and N2 (5/6). Abnormal amplitudes were obtained for the D+R subjects in 17 of 18 measures. For D subjects, the primary abnormality was P1 IT (4/6 groups); amplitudes were normal. [2. Individual SF-mfERGs] In D+R subjects, N2 ITs were most frequently abnormal. IT abnormalities occurred most frequently beyond 7 deg eccentricity. SPA abnormalities occurred most frequently for D and D+R subjects in the inferior nasal retina. CONCLUSIONS: The SF-mfERG is sensitive to mild diabetic retinopathy. It also detects abnormalities in diabetics without retinopathy. The implicit time of N2 may be the most sensitive parameter for mapping diabetic retinal dysfunction.