Is there a learning effect in adaptometry in normal subjects?

Purpose. Test vs. retest changes in the dark adaptation (DA) function of normal subjects were measured to determine if learning effects could be documented. Methods. 16 normal subjects (12 women, 4 men, whose age ranged from 24 to 52 yrs) with no previous adaptometry experience were studied. The average test-retest interval period was 1.35 months. DA was measured with a Goldmann-Weekers adaptometer, modified to use the von Békésy tracking method. The rod and cone limbs of the DA function were fit with the model: thresholds + a+b^-time/c The parameters of the DA function were analyzed with paired t-tests and Wilcoxon signed-rank tests. Results. No clinically significant differences between test and retest were documented in the means of any of the parameters of the DA function (a_cones: -2.44 vs. -2.33 log cd/m², b_cones: 1.21 vs. 1.23 log cd/m², C_cones 0-93 vs. 0.91 min, a_rods: -4.90 vs. -4.86 log cd/m², b_rods: 5.43 vs. 5.23 log cd/m², C_rods: 6.85 vs. 7.12 min, rod-cone break: 5.15 vs, 5.05 min). All these differences were not statistically significant, with the exception of the one for a_cones (p = 0.01 in both tests). Conclusions. These findings suggest that, in normal subjects, there are few, if any, learning effects in adaptometry analogous to those seen in static perimetry. Therefore, any differences between test and retest are probably indicative of actual changes in the DA function.