In this study, the step VEP underestimated psychophysical acuity (GAC) by ∼0.46 logMAR units across all acuity levels. This 0.46 log unit offset may occur for several reasons. First, the skull attenuates all responses uniformly before they are recorded by the scalp electrode, which means that small responses to near-threshold stimuli are very difficult to discern from the background EEG. Second, acuity targets differ substantially: GAC letter charts versus flickering checks. Other studies comparing similar pairs of acuity targets have found similar discrepancies: Sweep VEP underestimated Bailey-Lovie letter acuity by 0.25 log units in a group of normal adults,
19 underestimated Snellen acuity by 0.5 log units in defocused adults,
33 and underestimated preferential-looking acuity by ∼0.37 log units in children with reasonably good acuity.
28 Similarly, tVEPs underestimated behavioral acuity (Keeler or Cardiff cards) by 0.76 logMAR in multiply handicapped children.
12 However, sweep VEP in children with cortical visual impairment gave better acuities than Teller cards by ∼0.3 log units.
24 This finding is in keeping with other studies that show sweep VEPs to provide equal or better acuity estimates than behavioral methods for those with very poor vision.
19 26 27 28 29 33 In contrast, we found no acuity dependence in the relationship between the step VEP and GAC acuity. In experiments where the same acuity target is used for electrophysiological and psychophysical comparisons, VEP thresholds are similar to
45 46 or underestimate psychophysical thresholds by ∼0.08,
33 0.15,
18 or 0.4
47 log units (contrast thresholds), generally smaller discrepancies than those seen for different acuity targets. Third, the oblique effect caused by using a checkerboard whose fundamental frequencies are obliquely oriented may also reduce the measured VEP acuity.
48 Finally, pattern reversal VEP amplitudes drop at midspatial frequencies,
49 and the step VEP may miss VEPs in the notch, although VEPs may be present in response to smaller check sizes. However, the symmetry of the distribution of points in
Figure 2suggests this is not the case. Pattern-reversal VEPs are known to have larger amplitudes than pattern-onset VEPs in those with poor vision,
50 and the notch may be irrelevant, as it occurs below the threshold of many of the subjects when artificially degraded.