Abstract:Perceptual learning becomes completely transferrable with double training when the observers receive additional exposure of the transfer stimulus via an irrelevant task. However, we noticed that Vernier learning transferred to a new location after double training only when the training and transfer stimuli shared the same orientation. A second exposure to a suprathreshold orthogonal Vernier was necessary to enable cross-orientation learning transfer (Wang_et_al._2012). Here we investigated whether orientation learning transfer between orthogonal gratings and symmetric dot patterns also requires such triple training (training plus double exposures). Observers learned orientation discrimination of a grating or symmetric dot pattern, and the learning transfer to the other, orthogonal, stimulus was tested. We found (1) Baseline: Grating orientation learning (35°/125°) had little effect on orthogonal dot-pattern orientation discrimination; (2) Double training: Additional exposure to the orthogonal dot patterns via a brightness discrimination task produced incomplete learning transfer to orthogonal dot patterns, since further dot-pattern orientation training induced extra performance gain; (3) Triple training: Observers received exposure of both new orientation and new stimulus via two separate tasks: grating brightness discrimination at the orthogonal orientation, and dot pattern brightness discrimination at the same or orthogonal orientation. After triple training, dot-pattern orthogonal orientation discrimination was improved, and continued training produced no further gains, indicating complete transfer of grating orientation learning to the orthogonal symmetric dot patterns. (4) Control: The above double exposures by themselves had little impact on dot-pattern orientation discrimination. (5) The same patterns of learning transfer results were replicated from dot-patterns to gratings. These results suggest that the brain performs two remapping to transfer perceptual learning. One remaps learning to an orthogonal orientation of the same stimulus, which is enabled by exposure of the same but orthogonal-oriented stimulus. The other remaps learning to a physically distinct stimulus by exposure of the new stimulus.

Feature Variability Determines Specificity and Transfer in Multiorientation Feature Detection Learning

Can perceptual learning alleviate the global motion direction discrimination deficit in amblyopia

The Transfer of Perceptual Learning to a Physically and Orientation Different Stimulus Requires Triple Training

Under-stimulation at Untrained Orientation May Explain Orientation Specificity in Perceptual Learning

A new format of perceptual learning based on evidence abstraction from multiple stimuli

Transfer of visual perceptual learning over a task-irrelevant feature through feature-invariant representations: Behavioral experiments and model simulations

Enabling Complete Transfer of Perceptual Learning Across Orientations in Foveal Vision Through Double Training

Location and Direction Specificity in Motion Direction Learning Associated with a Single-Level Method of Constant Stimuli

Complete Transfer of Perceptual Learning Across Retinal Locations Enabled by Double Training

Task-irrelevant phase but not contrast variability unlocks generalization in visual perceptual learning

Task Relevancy and Demand Modulate Double-Training Enabled Transfer of Perceptual Learning

BOLD Signal Changes Associated with Orientation Learning and Transfer at Different Brain Areas

Perceptual learning in contour detection transfer across changes in contour path and orientation

Teaching Visual Orientation Discrimination Through Tactile Learning

Generalization in perceptual learning across stimuli and tasks

Under-stimulation at untrained retinal locations may explain location specificity in perceptual learning

Bottom-Up And Top-Down Influences At Untrained Conditions Determine Perceptual Learning Specificity And Transfer

Evidence for Stimulus Abstraction Before Perceptual Learning

Vernier Perceptual Learning Transfers to Completely Untrained Retinal Locations after Double Training: A "piggybacking'' Effect

Orientation Discrimination and Learning May Not Rely on Direct Sensory Inputs from Orientation Detectors

Transfer of Perceptual Learning to Completely Untrained Locations after Double Training