Abstract
Vision prostheses aim to restore, at least partially, vision that leads to restore visual perception we assume and know to be available in (partially) sighted people. Clinical and psychophysical vision tests are rarely designed to determine the ability of prosthetic device to restore vision; they were designed to measure the limits of a working, functioning visual system that serves our perceptual needs effectively. Yet they are almost exclusively used for such testing. The testing is usually applied using multiple alternative forced choice (MAFC) procedures. While these are perfectly suitable for many studies, they are frequently problematic in prosthetic vision evaluation. There are two such cases; one is where nuisance variables provide spurious cues which can be learnt in training, defeating the purpose of the test. The other is when the MAFC test is properly designed and performed, but the outcome does not actually measure what the researchers believe; the interpretation is wrong. The ability to perform on the clinical tests is necessary, but it is insufficient to prove vision (restoration), therefore additional tests covering the latter are needed. Examples for both types of errors will be presented, covering vision restoration evaluations, clinical and other studies. Scanning head movements, forced by the limited field of view or imposed by the need to increase resolution, are frequently used to trace the contour of test objects. The limitation of this method of use and of evaluation, prostheses, with or without MAFC, is also discussed.