A new method of extrapolating the sweep pattern visual evoked potential acuity

Purpose In order to compare the amplitude-spatial frequency (A-SP) regression method with amplitude-logVA (A-logVA) regression methods in extrapolating the sweep pattern visual evoked potential (SPVEP) acuity. Methods We measured SPVEPs in 21 children and three adults using sinusoidally-modulated horizontal gratings as stimuli. The responses were averaged and displayed through discrete Fourier transformations. SPVER acuity was then estimated by using both the SPVEP amplitude- spatial frequency function (A-SP function regression method) and the SPVEP amplitude-log visual-angle function (A-logVA function regression method). Furthermore, the Bailey Lovie logMAR chart was employed to define visual acuity. Curve estimates were calculated to derive a correlation index (R) for each method. Results There are significant differences (t = 2.71, P < 0.05) between the correlation indices of curves obtained using the A-logVA function (logarithmic model, 0.95 ± 0.01) and that obtained by the A-SP function (inverse model, 0.92 ± 0.02). The overall correlation coefficient (r) between logMAR acuity and acuity calculated by the A-logVA regression method was 0.32 (P < 0.05). The overall correlation coefficient (r) between logMAR acuity and acuity calculated by the A-SP regression method was 0.41 (P < 0.05). Paired t-tests show that SPVEP acuity from the A-logVA function was not significantly different from acuities of the logMAR function (t = 1.77, P = 0.09). The difference in their mean values is 0.14 ± 0.08. However, SPVEP acuity calculated using the A-SP function regression method is significantly different from the acuity calculated from the logMAR function (t = 10.09, P < 0.01). The difference in their mean values is 0.41 ± 0.04. Conclusions The amplitude-logVA function regression method is more accurate in estimating SPVEP acuity in normal subjects with good visual acuity.