Abstract
While perceptual learning (PL) is facilitated by feedback such as the correctness of a subject's response (response feedback), it can occur even without any feedback (e.g. Karni & Sagi, 1991; Poggio et al., 1992; Watanbe et al., 2001). Thus, response feedback has been considered to play a superfluous role in PL. However, recently it has been found that mere top-down processing, such as imagination, could induce PL (Dupuis-Roy & Gosselin, 2007; Tartaglia et al., 2009). This finding implies that response feedback could play a more important role than has been thought. To test this hypothesis, we conducted a seven-day experiment that employed an oriented sinusoidal grating (signal) embedded in spatial noise (noise). During five days of training, subjects were asked to identify the orientation of a grating patch by choosing one of two given orientations (2AFC). Each orientation was exclusively associated with one of the two conditions, one with the signal condition wherein an actual stimulus was presented, and the other with the noise condition wherein only noise was presented without any actual stimulus. Response feedback was provided after each trial in both conditions. Note that, in the noise condition, response feedback was fake feedback because no orientation was actually presented. Since the grating patch in the signal condition had 5% signal to noise ratio, with which subjects could hardly identify the grating, subjects could not discriminate between the two conditions. After training, subjects' sensitivity improved not only towards the orientation in the signal condition, but also towards the orientation indicated by fake feedback in the noise condition. These results indicate that response feedback itself can solely induce PL. They further imply that there are at least two types of PL, one by feedback and the other by exposure to a feature, with mechanisms distinguishable from each other.
NIH R01 (EY015980, EY019466, AG031941, MH091801).