Several studies show that adaptation to visual distortion depends on activity. I start with an old but illustrative study, and then present newer evidence. To find an answer to the question why we see the world upright despite retinal inversion, Stratton (1897,1896) did experiments with goggles that revolve sight by 180o. In the main experiment (Stratton, 1897), he wore the goggles over a period of eight days. In the beginning, everything looked upside down and the eye-hand coordination was strongly handicapped (manual operations were easier if done blindly). At the end (the last two days), the conflict between the operation of the hands and the visual impression vanished. Stratton could even have the impression that everything looked the right way up. For him, a new representation was learned next to the old one. In general, active operations enhanced the switch to the new representation; he notes, ``In rapid, complicated, yet practiced movements, the harmony of the localization by sight and that by touch or motor perception--the actual identity of the positions reported in these various ways--came out with much greater force than when I sat down and passively observed the scene'' (p. 356). His findings therefore suggest that the interaction with the world leads to our impression of upright vision. From a study with monkeys that wore inverting goggles for months, Sugita (1996) reported a reorganization of the visual cortex. But, the effect of the adaptation on perception is still debated, and some see Stratton's description of the inversion of vision as exaggerated (Linden et al., 1999).
The importance of active movement in visual adaptation has been further observed in experiments with wedge-prism goggles and with underwater vision. Wedge-prism goggles displace sight by a couple of degrees to one side. In the experiment by Held and Freedman (1963), subjects reached to target points either actively or passively (external force), while wearing these goggles. After the removal of the goggles, the subjects showed after-effects, resulting from an adaptation to the changed visual projection, only in trials following active movements. In the same direction goes a study with divers (Luria and Kinney, 1970). When wearing diving goggles under water, objects appear closer to the untrained eye; this also leads to pointing errors. Here, adaptation was faster if the divers were engaged in activities like placing a weight on a checkerboard grid.
Given these two studies, it may be still argued that the adaption is on a low sensorimotor level, and that it does not influence perception. However, a study with left-hemispatial-neglect patients (which have a neurological deficit of attention, perception, and doing actions within their left-sided space) shows that prism adaptation also involves higher-level space perception (Rossetti et al., 1998). After doing pointing tasks with prisms, the patients--after removal of the goggles--could turn their awareness toward the neglected side (to a degree corresponding to the visual shift resulting from the prism). This awareness shift was demonstrated with tests on reading and on drawing. A following study could induce neglect in healthy humans (Colent et al., 2000). The adaptation to left-deviating prisms resulted in a rightward bias for perceptual and motor line-bisection tasks.
Also a treadmill (a conveyor-belt for indoor exercises) can alter your perception. After being exposed to a static visual input while running, a runner observes an after-effect when standing still. The surroundings appear (erroneously) to be moving toward the runner (Pelah and Barlow, 1996). If our interpretation of vision would depend solely on the retinal image, such a finding could not be explained.