Deficits in discrimination and identification characterize a range of visual impairments. Training-based improvements in visual performance provide one possible non-invasive approach for remediation. Ideally, training for remediation will learn and transfer to a range of similar stimuli and tasks. However, improvements in perceptual task performance through perceptual learning or training, and the extent of transfer to related conditions, both depend critically upon the training protocol and the mixture of stimuli and tasks being trained. The current research uses computational models of visual perceptual learning, new and extended training and testing protocols, efficient estimation methods, and empirical tests.
It aims to improve our understanding of the conditions for and the limits of transfer and specificity of perceptual learning. Past research has reported strong modulation of perceptual learning for multiple tasks/stimuli and a range of effects on transfer. Here, we propose new systematic investigations of both phenomena that cry out for an integrated theoretical account of these fundamental issues in perceptual learning.
In Aim 1, we develop and test a theoretical framework to understand how practice on multiple stimuli or tasks interact in perceptual learning, in some cases eliminating learning, and in other cases supporting it.
In Aim 2, we develop and test a theoretical framework to understand and predict the cases in which perceptual training transfers to other stimuli, tasks, and visual locations, and those where it does not. We extend a previous computational model of perceptual learning based on partially-supervised learning algorithms, to incorporate location-independent as well as location-specific visual representations. The goal of this research program is to develop the theories and practical implementation of perceptual learning in normal populations that could contribute to translational applications to developmental learning and to ameliorative training in populations with perceptual deficits.
These aims are consistent with the goals of the NEI's National Plan for Eye and Vision Research.

Public Health Relevance

Perceptual learning through training visual tasks is one approach to remediation of some visual impairment and can contribute to development of visual skills. The current project seeks to understand the conditions for producing the best learning of multiple stimuli and tasks and for improving transfer of training to related stimuli and tasks. Through modeling learning for different training protocols, the proposed research program of model development and empirical testing aims to produce a framework for predicting the value of different training regimens in normal adults, and suggest applications in developmental and rehabilitative training.

Agency
National Institute of Health (NIH)
Institute
National Eye Institute (NEI)
Type
Research Project (R01)
Project #
5R01EY017491-11
Application #
8258711
Study Section
Cognition and Perception Study Section (CP)
Program Officer
Wiggs, Cheri
Project Start
2000-07-15
Project End
2015-04-30
Budget Start
2012-05-01
Budget End
2013-04-30
Support Year
11
Fiscal Year
2012
Total Cost
$387,285
Indirect Cost
$61,885
Name
University of California Irvine
Department
Psychology
Type
Schools of Arts and Sciences
DUNS #
046705849
City
Irvine
State
CA
Country
United States
Zip Code
92697
Xi, Jie; Jia, Wu-Li; Feng, Li-Xia et al. (2014) Perceptual learning improves stereoacuity in amblyopia. Invest Ophthalmol Vis Sci 55:2384-91
Liu, Jiajuan; Dosher, Barbara; Lu, Zhong-Lin (2014) Modeling trial by trial and block feedback in perceptual learning. Vision Res 99:46-56
Dorr, Michael; Lesmes, Luis A; Lu, Zhong-Lin et al. (2013) Rapid and reliable assessment of the contrast sensitivity function on an iPad. Invest Ophthalmol Vis Sci 54:7266-73
Meng, Jianjun; Liu, Ruilong; Wang, Ke et al. (2013) Neural correlates of stimulus spatial frequency-dependent contrast detection. Exp Brain Res 225:377-85
Hou, Fang; Huang, Chang-Bing; Liang, Ju et al. (2013) Contrast gain-control in stereo depth and cyclopean contrast perception. J Vis 13:
Liu, Jiajuan; Lu, Zhong-Lin; Dosher, Barbara Anne (2012) Mixed training at high and low accuracy levels leads to perceptual learning without feedback. Vision Res 61:15-24
Huang, Chang-Bing; Zhou, Jiawei; Lu, Zhong-Lin et al. (2011) Deficient binocular combination reveals mechanisms of anisometropic amblyopia: signal attenuation and interocular inhibition. J Vis 11:
Lu, Zhong-Lin; Hua, Tianmiao; Huang, Chang-Bing et al. (2011) Visual perceptual learning. Neurobiol Learn Mem 95:145-51
Hou, Fang; Huang, Chang-Bing; Tao, Liming et al. (2011) Training in contrast detection improves motion perception of sinewave gratings in amblyopia. Invest Ophthalmol Vis Sci 52:6501-10
Bejjanki, Vikranth R; Beck, Jeffrey M; Lu, Zhong-Lin et al. (2011) Perceptual learning as improved probabilistic inference in early sensory areas. Nat Neurosci 14:642-8

Showing the most recent 10 out of 27 publications