Overshadowing/generalization vs. preexperimental bias in utilization of motivation-specific odors in rats

Abstract

The notion that rats emit odors which can serve as sources of behavioral control for conspecifics has received substantial support in recent years. However, several investigators have reported that there may exist certain limits on the discriminative use of such odors. In experiments using a straight alleyway, test rats have shown an impaired ability to utilize, as discriminative cues, odors from startbox-placed donors when the deprivation states and reinforcers of donors and test subjects were different. Earlier experiments have not permitted an assessment of why such a learning difficulty might exist. Since odors emitted by test subjects were allowed to accumulate inside the goalbox, it is possible that goalbox odors overshadowed control by "incongruent" donor odors or interfered with donor cues via generalization decrement. On the other hand, it is possible that a preexperimental bias or constraint exists against utilization of incongruent donor cues. The present experiment attempted to determine whether test rats could learn discriminations based on incongruent donor odors when donors were placed inside the goalbox and emissions from test subjects were not allowed to accumulate and, thus, interfere with learning about incongruent odors. Additionally, the contribution of odors or remnants from donors' reinforcers to test subjects' discriminations was assessed, as was the importance of visible urine left by donors to those discriminations. All subjects were either food or water deprived. Results of a 5-phase experiment revealed that test subjects did not readily acquire discriminations based on incongruent donor cues. Discriminations based on congruent odors (i.e., when both donors and test subjects were maintained in the same deprivation state) were readily acquired. Thus, it appeared that a preexperimental bias or constraint against use of incongruent odors exists. Such a constraint is not absolute since test subjects provided with incongruent cues did eventually acquire discriminations. Also, shifting a test subject's deprivation state from congruent to incongruent did not destroy discriminations once acquired. Evaluation of the importance of reinforcer-generated cues and visible urine to test subjects' discriminations revealed that neither was necessary for maintaining discriminations.