Serial position curves in rats: automatic versus effortful information processing.

Rats were trained to remember lists of five arms on an Olton eight-arm radial maze. A forced-choice memory recognition test procedure was used which required the animal to choose between an arm previously visited during the study phase of a trial and an arm not visited. To receive additional food reinforcement, 10 animals were required to return to the previously visited arm (win-stay) and 10 animals were required to choose the novel, unvisited arm (win-shift). In this way, a direct comparison was made between the serial position curves (SPCs) generated by win-stay trained and win-shift trained animals. The results indicated that only win-stay trained animals produced the classical U-shaped SPC, which included significant primacy and recency effects. Win-shift subjects showed only recency effects. These findings are discussed in terms of differential processing requirements for the two procedures. It is suggested that the win-stay rule necessitates relatively more effortful, elaborative processing than does the win-shift rule, which is used automatically.