We are grateful for the comments on our recent article about learning and fatigue effects in flicker defined form perimetry. Bagnis and associates point out that fatigue effects have been shown for several perimetry techniques such as standard automated perimetry (SAP), short-wavelength automated perimetry (SWAP), frequency doubling technology, Rarebit perimetry, and flicker defined form perimetry.
They mention that learning effects occur for up to 5 sessions in SWAP, but only between the first and the second or third session in achromatic perimetry, and they therefore ask whether chromatic discrimination may be an additional factor influencing the perimetric learning effect. This is an interesting suggestion. A prolonged learning curve for SWAP also has be shown by Gardiner and associates, who examined the same patient group with both SAP and SWAP to investigate the evidence for continued learning in perimetry over several years. To imitate a realistic clinical setting, SAP and SWAP were performed annually over a period of 8 years. Interestingly, mean sensitivity for SAP increased over the first year to stay stable until after year 5, when it started to decline. In contrast to these findings, mean sensitivity continued to improve until year 6 for SWAP.
It is somewhat unlikely that object identification and form play a primer role, because although different pathways may be stimulated selectively using each of these tests, the task for the patient remains the same for both SAP and SWAP. The patient must report the presence of a luminance increment on a background. However from our understanding, adaptation phenomenons may play an important role in explaining the prolonged learning curve of SWAP. We know that the density of small bistratified cells mediating the blue cone response is much smaller than for ganglion cells mediating the achromatic response and that the blue cone pathway is much more sluggish, and therefore adaptation takes longer. Published literature suggests that SWAP is more difficult to undertake, because patients who underwent SAP, frequency doubling technology, and SWAP ranked SWAP as the worst test. Given that SWAP may be more difficult to undertake than SAP, as well as acknowledging the critical adaptation requirements for S-cone pathway isolation, it may be that on the initial test, sensitivity readings for SWAP were of a much lower value than the real sensitivity value. Whereas in SAP, where adaptation may be more stable, a sensitivity value closer to the real value may be obtained sooner. Consideration also should be given to the respective dynamic ranges of the instruments with regard to normal sensitivity of each pathway before true comparisons can be made.
In conclusion, several factors can be responsible for learning effects in perimetry. We agree that this issue should be clarified and understood properly because this is very important for future development of perimetric techniques. Future studies on this topic should be performed with patients naïve to any perimetric test.