Central perimetric sensitivity estimates are directly influenced by the fixation target

Purpose: Perimetry is increasingly being used to measure sensitivity at central visual field locations. For many tasks, the central (0 degrees, 0 degrees) location is functionally the most important, however threshold estimates at this location may be affected by masking by the nearby spatial structure of the fixation target. We investigated this effect. Methods: First we retrospectively analysed microperimetry (MAIA-2; CenterVue, Padova, Italy) data from 60 healthy subjects, tested on a custom grid with 1 degrees central spacing. We compared sensitivity at (0 degrees, 0 degrees) to the mean sensitivity at the eight adjacent locations. We then prospectively tested 15 further healthy subjects on the same instrument using a cross-shaped test pattern with 1 degrees spacing. Testing was carried out with and without the central fixation target, and sensitivity estimates at (0 degrees, 0 degrees) were compared. We also compared sensitivity at (0 degrees, 0 degrees) to the mean of the adjacent four locations in each condition. Three subjects undertook 10 repeated tests with the fixation target in place to assess within-subject variability of the effect. Results: In the retrospective analysis, central sensitivity was median 2.8 dB lower (95% range 0.1-8.8 dB lower, p < 0.0001) than the mean of the adjacent locations. In the prospective study, central sensitivity was median 2.0 dB lower with the fixation target vs without (95% range 0.4-4.7 dB lower, p = 0.0011). With the fixation target in place central sensitivity was median 2.5 dB lower than mean sensitivity of adjacent locations (95% range 0.8-4.2 dB lower, p = 0.0007), whilst without the fixation target there was no difference (mean 0.4 dB lower, S.D. 0.9 dB, p = 0.15). These differences could not be explained by reduced fixation stability. Mean within subject standard deviation in the difference between central and adjacent locations' sensitivity was 1.84 dB for the repeated tests. Conclusions: Perimetric sensitivity estimates from the central (0 degrees, 0 degrees) location are, on-average, reduced by 2 to 3 dB, corresponding to a 60-100% increase in stimulus luminance at threshold. This effect can be explained by masking by the nearby fixation target. The considerable within-and between-subject variability in magnitude, and the unknown effects of disease may hamper attempts to compensate threshold estimates for this effect. Clinicians should interpret central perimetric sensitivity estimates with caution, especially in patients with reduced sensitivity due to disease.