Jamie G.E. Cochrane¹, Allison B. Sekuler^1,2,3, Patrick J. Bennett¹; ¹Department of Psychology, Neuroscience and Behaviour, McMaster University, ²Rotman Research Institute, Baycrest Academy for Research and Education, ³Department of Psychology, University of Toronto

In the double training paradigm, two tasks are trained at different visual field positions to promote transfer of perceptual learning between them. Prior studies using this method have focused on learning in low-level feature detection and discrimination tasks. We investigated whether double training increases transfer in identification tasks that rely on inferotemporal processing. Previous studies found that perceptual learning in texture and face identification is stimulus- and position-specific. We tested whether double training increases transfer across positions in these tasks. Participants completed two interleaved 1-of-5 identification tasks with two sets of faces or textures. During training, one stimulus set was always shown above fixation, and the other was presented below. At test, positions of the two stimulus sets were reversed. For the texture task, double training successfully induced full transfer across positions: accuracy at the new location matched accuracy at the end of training. However, transfer was asymmetric for the face task. Full transfer occurred only for faces presented above fixation during training, perhaps because identity information is concentrated near the eyes and is therefore closer to fixation for faces tested below. We further investigated the effects of task relevancy by training a new set of participants with intermixed face and texture identification tasks. During training, the positions of the textures and faces were fixed (above or below fixation) and the positions were switched during testing. Only partial transfer was found in both tasks. Compared to the first experiment, learning in the texture identification task exhibited less transfer to a new stimulus location, but learning in the face identification task, when trained above, exhibited greater transfer to the test location. These results show a potential benefit of double training for inducing position transfer in higher-level tasks; however, factors such as task relevance appear to undermine its efficacy.

Acknowledgements: NSERC