Abstract
Specificity is a key property of perceptual learning. Improvements in one task after extensive training may not transfer to different stimuli or tasks. We argued that specificity or transfer depend critically on the precision of the tasks. Jeter et al. (2006) used the simultaneous training paradigms of Liu et al. (1998) and found full specificity for a high precision task that differed in both orientation and position. The current experiment extended this investigation to test tasks in simultaneous training that differed only in orientation, only in position, or in both position and orientation in a 2AFC high precision (+/−5° from a reference angle) orientation discrimination task using Gabor patches in external noise at a fixed contrast. Trials were interleaved in uneven proportions in A-A-B-A-A-B… order. ‘A’ trials appeared on one position diagonal (e.g., NW/SE) with one orientation reference, while ‘B’ trials shared positions, reference orientations, or neither. Accuracy on the ‘B’ trials should equal the first half of all ‘A’ trials for independent training; if transfer occurs, the ‘B’ accuracy should be better. We generally found independence of learning ‘A’ and ‘B’ tasks, whether they differ in one or both features. Training the same orientations in different positions yielded slight positive transfer, while different orientations in the same locations yielded slight negative transfer for certain layout configurations. Overall, these results support the previous claims of Jeter et al. of independence of high precision tasks in simultaneous learning protocols.