Super-latent inhibition with delayed conditioned taste aversion testing

R. E. LUBOW 0 1 2 0 This research was supported by DGES Grant PB95-0529-C02-01 to dressed either to L. G. de la Casa, Department of Experimental Psy chology , Avda. San Francisco Javier, sin, 41005 Seville, Spain (e-rnail: 1 TelAviv University , TelAviv, Israel 2 L. G. DE LA CASA University ofSeville , SeviUe, Spain In three conditioned taste aversion experiments with rats, latent inhibition (U) was examined as a function of the time interval (1 or 21 days) between the conditioning and the test phases. In Experiments 1 and 2, the effects of US intensity on Ll were examined. U increased in the 21-day condition, as compared with the l-day condition, with medium and high US intensity, but not with weak US intensity. Groups not preexposed to the CS flavor had similar aversions when testing was conducted 1 day after conditioning, as compared with 21 days. In Experiment 3A, delay-induced super-Ll was obtained when the delay was spent in the home cage and the experimental stages took place in a different context (as in Experiments 1 and 2). In Experiment 3B, when all the stages, including the delay period, were conducted in the home cage, there was no super-Ll effect. The modulation of delay-induced superU as a function of US intensity and context extinction is discussed in relation to association deficit and retrieval interference theories of U. - Passive exposures to a stimulus, whether it be a light, a tone, or a flavor, have a number of consequences. Of these, the most obvious is the gradual reduction in the magnitude of the unconditioned responses that are elic ited by the repeated stimulus (habituation). However, more subtle effects can be seen when that same stimulus is subsequently paired with an unconditioned stimulus (US). Such pairings of a conditioned stimulus (CS) with a US are less effective in promoting conditioned perfor mance than are those that are not preceded by passive ex posure to the CS. This phenomenon, latent inhibition (LI), has been the subject of extensive research (see Lu bow, 1989, for a review). The deficit in the performance of a learning task that results from such stimulus preex posure is found in many species and test paradigms. LI is most often attributed to some mechanism opera tive during the preexposure period that affects the subse quent associability of the stimulus when it is paired with the US (Lubow, Weiner, & Schnur, 1981; Mackintosh, 1975; Pearce & Hall, 1980; Wagner, 1981). In opposition to the idea that LI represents an acquisition failure, it has been suggested that LI is a result of retrieval failure (e.g., Bouton, 1993; Miller, Kasprow, & Schachtman, 1986). Such a hypothesis proposes that, following stimulus preexposure, the acquisition of the new association to the old stimulus proceeds normally. However, in the test stage, when the animal again encounters the target stimulus, two competing associations are retrieved: the earlier stimulus no-consequence association from the preexposure stage and the stimulus-US association from the acquisition stage. In normal LI, the nonpreexposed group exhibits greater conditioned response strength than does the pre exposed group because there is only the second associa tion to be retrieved, whereas the preexposed group exhibits less conditioned response strength, either because the first association is the one that is retrieved or because the first association interferes with the retrieval of the second one. The primary evidence for the retrieval failure hypoth esis of LI comes from studies that have varied the time between acquisition and test. If LI is found after a short delay between acquisition and test stages, but not after a long delay, and ifthis difference is due to the preexposed delay group's providing more evidence of conditioning than does the preexposed -no-delay group, this is pre sumptive evidence that with the short delay, the CS-US association was present but not manifest and that some thing occurred during the longer delay that allowed the originally encoded CS-US association to be retrieved. However, such an effect of delay has not been obtained consistently. With conditioned taste aversion (CTA), a re instatement of conditioning (i.e., a loss of LI) with long acquisition-test intervals has been reported in several studies (Aguado, Symonds, & Hall, 1994, Experiment I; Bakner, Strohen, Nordeen, & Riccio, 1991, Experiments 1 and 2; De La Casa & Lubow, 1995).\ Other studies have obtained such an effect only when the test was conducted with a flavor that was different from the one in the pre exposure and acquisition phases (Kraemer & Ossen kopp, 1986; Kraemer & Roberts, 1984; Kraemer & Spear, 1992). Alvarez and Lopez (1995, Experiment 2) also failed to find a diminution in LI after a long delay interval when the same flavor was used in preexposure, acquisition, and test. Additional evidence for the retrieval failure hypothe sis comes from CTA experiments that have imposed long delays between preexposure and conditioning stages, re sulting in attenuated LI (Ackil, Carman, Bakner, & Ric cio, 1992; Aguado et aI., 1994; Kraemer & Roberts, 1984). However, as with the experiments manipulating the conditioning-test interval, the results are not com pletely consistent (Rosas & Bouton, 1997). The importance ofLI in contemporary learning theory (Hall, 1991; Mackintosh, 1983) compels one to confront the discrepant data and to uncover mechanisms/rules that would clarify the relationship between the two compet ing explanations of LI. In order to differentiate between acquisition failure and retrieval failure theories, one would need to (I) specify the variables that modulate LI, (2) iden tify the association(s) that are learned during the preex posure phase, and (3) postulate a mechanism by which these associations modulate test performance. Both acquisition failure and retrieval failure theories of LI are based on the premise that something is learned during the preexposure period. They differ in terms of the locus of action of such learning. In the former case, what is learned in preexposure interferes with the devel opment of a new association in the acquisition phase. In the latter case, such acquisition is unaffected by stimulus preexposure. Instead, the association that was learned in the preexposure phase and the association that was learned in the acquisition phase compete for expression in the test phase. The effectiveness of these analyses is, however, de pendent on specifying the associations that are learned in the preexposure and acquisition phases. For the acquisi tion phase, there is agreement that the organism acquires a CS-US association, and perhaps other associations as well, particularly to context. However, no such consen sus exists for the identification of the associations acquired in the preexposure phase. Whether the associa tion is CS-no-consequence, CS-context, context-no consequence, or a higher order conditional association whereby the context be (...truncated)