Reduced Inhibitory Lateral Interactions by Perceptual Learning in Normal Subjects

In psychology of visual perception for a long time has been considered that presentation of a single stimulus is enough. However, works of Polat and Sagi (1993) demonstrated that stimulus perception depends from the contextual surroundings. Moreover, there were investigations with a goal to understand whether perceptual learning affect these contextual influences (Polat and Sagi, 1994), as well as investigation how these this modulation of contextual influences by learning could help people with visual developmental problems (amblyopia) to improve their vision (Polat et. al, 2004).

The learning procedure included training on contrast detection of a target stimulus flanked by masks at a range of distances (0-12 lambda) during each daily session, and this was repeated several times per week. Polat and Sagi (1994) trained two observers on sets of stimuli having relatively large separations between target and masks.
Learning occurred in conditions where target and masks were aligned along their local orientation (ororthogonal to it) but not when local orientation was 45° from the global orientation. Results show an increase in interaction range up to a factor of two for small signal wavelengths (lambda): 0.075°, 0.15° (fig. 14).

These correspond to high spatial frequency values: 13 and 6.66 c/deg. No consistent learning effect was found at lower spatial frequencies. A similar result was found by Adini, Sagi and Tsodyks (1997). They found that practice had an interesting influence on the observers’ performance. At the early stage of practice, the chain depression effect (indicating facilitation) was more pronounced under the parallel configuration condition, but practice had the effect of increasing depression on the collinear configuration condition, thus making the difference between the two conditions insignificant.