We examined neural correlates of perceptual learning in rats receiving visual discrimination training in a Y-shaped water maze. Rats learned to discriminate visual cues to navigate to a hidden escape platform. In trained rats, visual stimuli encountered during training elicited larger evoked potentials in V1 than novel stimuli, regardless of whether stimuli served as CS+ (platform) or CS- (no platform) during training. Training also resulted in facilitation of synaptic plasticity (long-term potentiation) induced by high-frequency stimulation of thalamic afferents to V1. Thus, perceptual learning might involve stimulus-selective facilitation and changes in the plasticity properties (“metaplasticity”) of primary sensory cortices.

“Hebb Award Abstract”
Synaptic plasticity plays a key role in processes of learning and memory formation. Long-term potentiation (LTP) is a relatively stable enhancement of synaptic transmission following specific patterns of electrical stimulation. Some types of learning (e.g. motor learning, fear conditioning) result in LTP-like changes at central synapses. However, no studies have examined whether perceptual learning can result in LTP-like plasticity in the visual cortex.
We used a visual discrimination task to examine changes in LTP in the primary visual cortex (V1) of adult rats. Rats were placed in a Y-shaped water maze and required to swim to one of the choice arms containing a hidden escape platform. Distinct visual cues indicated the presence (CS+) and absence (CS-) of the platform. Rats learned to reliably discriminate the visual cues to successfully navigate to the platform. Control rats received the same procedure, but the two visual stimuli did not have a predictive relation with the escape platform. Following training, trained, control, and task-naïve rats were anesthetized and evoked field potentials (EPs) in V1 were recorded in response to stimuli CS+, CS-, and novel stimuli.
Results indicate that, in both task-naïve and control animals, all visual stimuli (familiar and novel) elicit EPs of similar (p >0.05) amplitude. In contrast, trained animals show significantly larger amplitude EPs to stimuli encountered during training relative to novel stimuli, regardless of whether stimuli act as CS+ or CS-. In addition, trained and control animals also differ in the magnitude of LTP that can be induced by electrical (theta-burst) stimulation of fibers between the lateral geniculate nucleus (LGN) and V1. Trained and control rats showed 71% and 53% potentiation (p < 0.05), respectively, indicating greater plasticity of thalamocortical synapses following visual training. There were no significant differences in LTP between controls and task naïve animals, indicating that swimming and associated stress responses did not exert major influenced on LTP.
These experiments demonstrate that perceptual learning might involve stimulus-selective facilitation of neuronal responses at early stages of visual processing (LGN, V1). The effect requires that stimuli carry some significance (e.g., indicating the platform location) to the animal, while mere passive exposure does not result in the same level of neuronal enhancement. Further, visual experience alters the plasticity properties of V1 (“metaplasticity”) by facilitating LTP along thalamocortical sensory fibers. The later effect may allow past visual experience to influence the efficiency of encoding novel visual features.