Position of delay and delay-associated cues in maze learning

Abstract

Forty-eight rats were run in a multiple T maze, and either immediately reinforced, delayed in the middle of the maze, or delayed at the end of the maze. Cues associated with the position of delay were manipulated. Both delay groups exhibited pre-delay decrements and post-delay increments in performance. Ss delayed at the end of the maze required fewer trials to criterion than Ss delayed in the middle of the maze. The data were interpreted as consistent with Amsel's ( 1958) hypothesis. Amsel (1958) hypothesized that delay produces frustration, an aversive motivational state, and causes an increment in drive. A secondary form of frustration, fractional anticipatory frustration (rf), develops over trials through classical conditioning and serves as an inhibitory mechanism. Thus, pre-delay performance suffers a decrement because of the anticipation of frustration, and postdelay performance is facilitated because the delay-produced frustration adds an increment to drive. Wist (1962) proposed that the effects of delay are largely local; that is, the effects of delay are exerted on those portions of the behavior chain that immediately precede and immediately follow the point of delay.Delay in the middle of a response chain should adversely affect the pre-delay portion of the chain whereas the post-delay portion should be incrementally affected. Delay at the end of a response chain should adversely affect only that portion of the chain immediately preceding the point of delay. Because the consummatory response occurs at the end of the chain, the cues located at that point should become conditioned to the consummatory response. These cues would provide sources of secondary reinforcement and facilitate the learning of the responses that lead to delay. Because the cues in a delay section in the middle of the response chain are further removed, both temporally and spatially, from the consummatory response than are the cues at the end of the chain, the cues at the end should acquire greater secondary reinforcing properties than those in the middle. The effect should be to enhance the conditioning of rf to the reward and to the responses that lead to reward. Therefore, delay imposed at the end of a response chain should result in faster acquisition of a response chain than delay enforced in the middle of a response chain.Subjects. The Ss were 48 experimentally naive, male, Sprague-Dawley rats, 70 days old at the start of the experiment. Apparatus. The apparatus used was an eight-unit multiple T maze. The pattern was LRRLRLLR. Each arm of each Twas 15 in. long and 4 in. wide; the height of the walls was 12 in. A I4-in. x 4-in. x IS-in. cul-de-sac was attached to the end of each incorrect arm of each section, with the exception of sections fOUT and eight. The stem of section one constituted the start box; the right arm of section eight constituted the goal box. The left arm of section four was constructed so that delay could be enforced in that section. The start box, delay section, and goal box were separated from the rest of ...