Discrimination learning using different CS-US intervals in classical conditioning of the rabbit’s nictitating membrane response

Abstract

Differential conditioning of the rabbit’s nictitating membrane (NM) response was examined when two CSs each signaled the US, but at different CS-US intervals (e.g., 200 and 600 msec). For each subject, one CS was a 600-Hz tone, and the other CS was either a 2100-, 1000-, or 660-Hz tone. This task took advantage of the distinctive temporal character of the NM conditioned response (CR); namely, the NM closure reaches its peak near the point of US delivery. The distribution of CR peaks around the points of US delivery varied as a function of the difference between the tones. For the differences that were “easy” (600 vs. 2100 Hz) and “medium” (600 vs. 1000 Hz), the distributions overlapped but conformed to their CS-US intervals. For the “hard” difference (600 vs. 660 Hz), the distributions overlapped entirely. In an analysis of individual subjects using signal detection measures, half the subjects showed very high degrees of sensitivity to the easy and medium tone differences. The remaining subjects in those two groups showed high response biases that obscured any sensitivity. For the hard tonal difference, there was no evidence of sensitivity, and variation between subjects reflected only variation in response bias. The results suggest that this task can help illuminate the neural basis of encoding for the sensory and temporal differences among CSs in a useful way.