Journal of Experimental Psychology-Animal Learning and Cognition最新文献

In 1988 Robert Rescorla published an article in the Annual Review of Neuroscience that addressed the circumstances under which learning occurs, some key methodological issues, and what constitutes an example of learning. The article has inspired a generation of neuroscientists, opening the door to a wider range of learning phenomena. After reviewing the historical context for his article, its key points are briefly reviewed. The perspective outlined enabled the study of learning in simpler preparations, such as the spinal cord. The period after 1988 revealed that pain (nociceptive) stimuli can induce a lasting sensitization of spinal cord circuits, laying down a kind of memory mediated by signal pathways analogous to those implicated in brain dependent learning and memory. Evidence suggests that the spinal cord is sensitive to instrumental response-outcome (R-O) relations, that learning can induce a peripheral modification (muscle memory) that helps maintain the learned response, and that learning can promote adaptive plasticity (a form of metaplasticity). Conversely, exposure to uncontrollable stimulation disables the capacity to learn. Spinal cord neurons can also abstract that stimuli occur in a regular (predictable) manner, a capacity that appears linked to a neural oscillator (central pattern generator). Disrupting communication with the brain has been shown to transform how GABA affects neuronal function (an example of ionic plasticity), releasing a brake that enables plasticity. We conclude by presenting a framework for understanding these findings and the implications for the broader study of learning. (PsycInfo Database Record (c) 2022 APA, all rights reserved).

Associative accounts of category learning have been, for the most part, abandoned in favor of cognitive explanations (e.g., similarity, explicit rules). In the current work, we implement an Adaptive Linear Filter (ALF) closely related to the Rescorla and Wagner learning rule, and use it to tackle three learning tasks that pose challenges to an associative view of category learning. Across three computational simulations, we show that the ALF is in fact able to make the predictions that seemed problematic. Notably, in our simulations we use exactly the same model and specifications, attesting to the generality of our account. We discuss the consequences of our findings for the category learning literature. (PsycInfo Database Record (c) 2022 APA, all rights reserved).

Six experiments with rats examined the nature of inhibition learned in an operant feature-negative (FN) discrimination. The results of prior experiments that examined instrumental extinction rather than FN learning suggest that inhibition can be very specific to the inhibited response. In Experiment 1, we trained lever-press and chain-pull responses in separate but parallel FN discriminations (AR1+, ABR1-, CR2+, and CDR2-) and then tested both inhibitors (B and D) with both responses. Of primary interest was the extent to which the inhibitors suppressed the response they were trained with (same-response inhibition) versus the other response (cross-response inhibition). We found that cross-response inhibition was robust and essentially equal to same-response inhibition. Experiment 2 replicated this result and confirmed stronger inhibition after FN learning than after a differential inhibition procedure (AR1+, BR1-, CR2+, and DR2-). There was also little evidence that cross-response inhibition was due to a demonstrable competing response. Experiment 3 found that cross-response inhibition did not depend on having the two responses reinforced by a common outcome. Experiments 4 and 5 then found that cross-response inhibition depended substantially (though not completely) on the transfer target response having been trained in its own FN discrimination. However, Experiment 6 found that inhibition after instrumental extinction (as opposed to FN learning) was still highly response-specific when the transferred-to response had been trained in an FN discrimination. The overall results suggest that the characteristics of inhibition in instrumental extinction and FN learning differ and that transfer of FN inhibition across responses depends at least partly on previous "inhibitability" of the target response. (PsycInfo Database Record (c) 2022 APA, all rights reserved).

The present article explored the fate of previously formed response-outcome associations when the relation between R and O was disrupted by arranging for O to occur independently of R. In each of three experiments response independent outcome delivery selectively reduced the R earning that O. Nevertheless, in Experiments 1 and 2, the R continued to show sensitivity to outcome devaluation, suggesting that the strength of the R-O association was undiminished by this treatment. These experiments used a two-lever, two-outcome design introducing the possibility that devaluation reflected the influence of specific Pavlovian lever-outcome associations. In an attempt to nullify the influence of these incidental Pavlovian cues Experiment 3 used a single bidirectional vertical lever that rats could press left or right for different outcomes. Again, response-independent outcome presentations selectively depressed the performance of the R that delivered the response-independent O. However, in this situation, the response independent O also reduced the sensitivity of R to outcome devaluation; whereas the nondegraded R was sensitive to devaluation, the degraded R was not. We conclude that selective degradation of the instrumental contingency can weaken a specific R-O association while leaving other R-O associations intact. Furthermore, the use of a bidirectional vertical lever in Experiment 3 revealed that unidirectional and spatially separated instrumental manipulanda, such as levers or chains, may produce Pavlovian cues capable of forming incidental associations with the instrumental outcome that can obscure the relative influence of R-O associations after various manipulations. (PsycInfo Database Record (c) 2022 APA, all rights reserved).

In this article we investigate how a psychological theory used to model perceptual learning and face recognition can be used to predict that anodal tDCS delivered over the DLPFC at Fp3 site (for 10 mins duration at 1.5 mA intensity) modulates the decision criterion, C, (and not d-prime [d']) in a target detection task. In two between-subjects and double-blind experiments (n = 112) we examined the tDCS effects on C when subjects were engaged in a target detection task, in the first instance involving artificial checkerboard stimuli (Experiment 1a), and subsequently face stimuli (Experiment 1b). The results from both experiments revealed that in the sham/control groups a significantly higher C was used when detecting a target pattern (Experiment 1a) or face (Experiment 1b) presented on a familiar rather than a random background. Importantly, anodal tDCS significantly reduced/reversed this difference between C adopted for familiar and random backgrounds in both Experiment 1a and 1b without affecting d'. These results contribute to advance our understanding of the tDCS-induced effects on stimulus representation and to the literature regarding the modulation of C. (PsycInfo Database Record (c) 2022 APA, all rights reserved).

The nature of the operations that support learning should be evident in the form or shape of the learning curve. For example, models that describe learning as an iterative error-correction process expect that the amount learned on each trial follows a decelerating (negatively inflected) function. That prediction is broadly consistent with the shape of the acquisition and extinction curves derived from mean measures of response strength. However, such evidence can be flawed because group means may not accurately portray the response curves of individual subjects in a conditioning experiment. Moreover, such evidence relies on strong assumptions about the way that what has been learned is expressed in responding. The current work presents a new analytical approach to reveal the rate of change in responding across the course of conditioning in individual animals. When applied to the conditioning and extinction data from a large sample of rats, this analysis confirms that responses are acquired and extinguish gradually and, in both cases, follow a decelerating learning curve. That is, changes in responding are largest at the start of conditioning or extinction and get progressively smaller as responding approaches an asymptote. However, rather than conforming to the specific shape predicted by an error-correction process, the results suggest that the amount learned increases according to a logarithmic function such that responding during conditioning and extinction is proportional to the log of the number of trials. The implications of these findings for models of associative learning are discussed. (PsycInfo Database Record (c) 2022 APA, all rights reserved).

In a signal detection theory approach to associative learning, the perceived (i.e., subjective) contingency between a cue and an outcome is a random variable drawn from a Gaussian distribution. At the end of the sequence, participants report a positive cue-outcome contingency provided the subjective contingency is above some threshold. Some researchers have suggested that the mean of the subjective contingency distributions and the threshold are controlled by different variables. The present data provide empirical support for this claim. In three experiments, participants were exposed to rapid streams of trials at the end of which they had to indicate whether a target outcome O1 was more likely following a target cue X. Interfering treatments were incorporated in some streams to impend participants' ability to identify the objective X-O1 contingency: interference trials (X was paired with an irrelevant outcome O2), nonreinforced trials (X was presented alone), plus control trials (an irrelevant cue W was paired with O2). Overall, both interference and nonreinforced trials impaired participants' sensitivity to the contingencies as measured by signal detection theory's d', but they also enhanced detection of positive contingencies through a cue density effect, with nonreinforced trials being more susceptible to this effect than interference trials. These results are explicable if one assumes interference and nonreinforced trials impact the mean of the associative strength distribution, while the cue density influences the threshold. (PsycInfo Database Record (c) 2022 APA, all rights reserved).

Exposure to two similar stimuli (AX and BX; e.g., two tastes) reduces the extent to which a conditioned response later established to BX generalizes to AX. This example of perceptual learning is more evident when AX and BX are exposed in an alternating manner (AX, BX, AX, BX,…) than when AX and BX occur in separate blocks (e.g., AX, AX,…BX, BX,…). We examined in male rats (N = 126) the impact of rapid alternation to AX and BX on generalization of a taste aversion from BX to AX. Experiment 1 showed that such alternating presentations (with 5-min intervals between AX and BX) reduced generalization relative to blocked exposure; but only as assessed by consumption levels and not by lick cluster size (an index of hedonic reactions). Experiment 1 also showed that the nature of exposure did not affect how A influenced performance to a novel conditioned taste, Y. Experiment 2 replicated the pattern of results involving the different influences of rapidly alternating and blocked exposure on generalization from BX to AX, and showed that this effect was only evident when rats received access to water during the 5-min intervals between AX and BX. These results reinforce parallels between perceptual learning effects in rats and humans, both at empirical and theoretical levels. (PsycInfo Database Record (c) 2022 APA, all rights reserved).

A challenge for generalization models is to specify how excitation generated from a CS+ (i.e., positive evidence) should interact with inhibition from a CS- (i.e., negative evidence) to produce generalized responding. Empirically, many generalization phenomena are consistent with the monotonicity principle, which states that additional positive evidence should increase generalized responding, whereas additional negative evidence should decrease responding. However, a recent study (Lee et al.,, 2019) demonstrated that additional negative evidence can sometimes increase generalization, in direct contrast to animal data and associative accounts of generalization. The current study investigated whether a similar effect could be found in a symmetrical intradimensional discrimination procedure with two sources of negative evidence (CS-s) located on each side of a CS+. In three experiments, we compared generalization along a green-blue dimension between one group of participants who learned that an aqua-colored shape (CS+) predicted an outcome (Single Positive group) with another group who also learned that both a slightly greener and a slightly bluer shape led to no outcome (Double Negative group). Experiments 1A and 1B showed no effect of the additional negative evidence in increasing generalization around the CS+. However, changing a stimulus feature at test (shape) resulted in a higher gradient peak in the Double Negative group relative to the Single Positive group in Experiment 2. Although this result violates the monotonicity principle, an extended version of Blough's (1975) model applying cue competition to multiple stimulus dimensions (i.e., shape and color) successfully replicated the group differences. Our results suggest that associative mechanisms can account for some instances in which negative evidence increases generalization. (PsycInfo Database Record (c) 2022 APA, all rights reserved).

Bridging associative and normative theories of animal learning, I show that an associative system can behave as if performing probabilistic inference by using the function f(V) = 1 - e^-cV to transform associative strengths (V) into response probabilities. For example, using this function, an associative system can respond normatively to a compound stimulus AB, given previous separate experiences with the components A and B. The CR probability formulae that result from the proposed function have a normative interpretation in terms of statistical decision theory. The formulae also suggest a normative interpretation of stimulus generalization as a heuristic to infer whether different stimuli are likely to convey redundant or independent information about reinforcement. (PsycInfo Database Record (c) 2022 APA, all rights reserved).