Abstract
We can easily attend any selected cues in the enriched environment according to their motivational significance. How does our sensory system give a priority to biologically important signals among abundant, incoming information? To address the question of where such selective function begins in the sensory system, the activity of single neuron in the sensory thalamus (the gateway to the forebrain) was recorded while the rats performed a delayed cue-reward association task. The task required discrimination between cue stimuli that did and did not predict rewards. Most of neurons responded to cue presentation with a short latency (early response component: EC). These neurons were further divided into two types, based on the absence (Type 1) or presence (Type 2) of a late response component (LC), which displayed a gradual buildup with a peak prior to reward delivery. The EC of Type 2 neurons was enhanced by cues associated with reward, while that of Type 1 neurons was not. Altering reward value or timing of reward delivery did not affect the EC but affected the LC. That is, the change in the LC initially appeared around the time of reward delivery, and then expanded retroactively in time toward the cue. These properties resembled the value function in the reinforcement learning theory. In extinction trials applied to Type 2 neurons, the LC disappeared quickly while the EC decreased gradually. Such self-organized modifications in the sensory thalamus might serve as a filter for the forebrain that favors biologically significant stimuli over the others. [J Physiol Sci. 2007;57 Suppl:S57]
