This or not that: select and reject control of relational responding in rats using a blank comparison procedure with odor stimuli

Abstract

The blank comparison (BLC) task was developed to assess stimulus relations in discrimination learning; that is, are subjects learning to "select" the correct stimulus (S+) or "reject" the incorrect stimulus (S-) or both? This task has been used to study exclusion learning, mostly in humans and monkeys, and the present study extends the procedure to rats. The BLC task uses an ambiguous stimulus (BLC+/-) that replaces S+ (in the presence of S-) and replaces S- (in the presence of S+). In the current experiment, four rats were trained to remove session-novel scented lids from sand-filled cups in a two-choice, simultaneous presentation procedure called the Odor Span Task (OST) before being trained on the BLC procedure using odors as the discriminative stimuli. The BLC training procedure utilized simple discrimination training (S+ and S-) and added select (S+ and BLC-) and reject (BLC+ and S-) trial types. All rats demonstrated accurate performance in sessions with both select and reject type trials. Next, BLC probe trials were interspersed in standard OST sessions to assess the form of stimulus control in the OST. Rats performed accurately on select type probe trials (similar to baseline OST performance) and also showed above chance accuracy on reject type trials. Thus, we demonstrated that rats could acquire an odor-based version of the BLC task and that both select and exclusion-based (reject) relations were active in the OST. The finding of exclusion in rats under the rigorous BLC task conditions confirms that exclusion-based responding is not limited to humans and non-human primates.

Keywords: Blank comparison; Exclusion learning; Rats; Reject control; Select control; Stimulus control topography.

Fig. 1

Arena testing apparatus. Note In the testing arena, there are two circular arrays (rings) in which the stimuli can be placed. In the outer ring, there are 12 equidistant holes around the perimeter. Note that a cup with lid is in the 4 o’clock position. There are six equidistant holes in the center ring; there is a cup with lid in the 6 o’clock position

Fig. 2

Percent correct responses by each subject in blocked reject, select, and interspersed select trials in the last five sessions of BLC training. Note Error bars are standard error of the mean. Dashed line indicates chance responding. All means significantly above chance (p < .001, binomial test)

Fig. 3

Percent correct responding by each subject on OST and SD sessions during BLC training. Note Error bars are standard error of the mean. Dashed line indicates chance responding. All means significantly above chance (p < .001, binomial test)

Fig. 4

Precent correct responding by each subject on OST, select, and reject trials across all 25 BLC in OST sessions. Note Error bars are standard error of the mean. Dashed line indicates chance responding. All means significantly above chance (p < .001, binomial test)