Human Predictive Learning Research Articles

Increasing previous but not concurrent extinction attenuates the "extinction makes acquisition context specific" effect in human predictive learning.

Four experiments in human predictive learning evaluated whether the extinction makes the acquisition context specific (EMACS) effect is attenuated when the increase in prediction error that extinction produces disappears. Participants had to evaluate the relationship between a given food (cue) that was ingested by an imaginary client of a given restaurant (context) and a potential gastric illness (outcome). The task was implemented using Gorilla online software. All participants received the relevant training in context A, and equivalent exposure to context B. Cue E was presented paired with the outcome in all groups. Cue E was then either extinguished (group E) or not extinguished (group NE), either previously or concurrently to training of the target cue (P). P was then tested in contexts A and B. When extinction was conducted concurrently, performance to P became context-dependent regardless of the number of extinction trials (12 or 24)-the EMACS effect. The EMACS effect disappeared when extinction was elongated to 24 trials, and conducted before acquisition of P. Implications of these results for attentional explanations of context processing are discussed. (PsycInfo Database Record (c) 2024 APA, all rights reserved).

"Uncertainty and predictiveness modulate attention in human predictive learning": Correction to Chao, McGregor, and Sanderson (2020).

Reports an error in "Uncertainty and predictiveness modulate attention in human predictive learning" by Chang-Mao Chao, Anthony McGregor and David J. Sanderson (Journal of Experimental Psychology: General, Advanced Online Publication, Nov 30, 2020, np). In the article, formatting for UK Research Councils funding was omitted. The author note and copyright line now reflect the standard acknowledgment of and formatting for the funding received for this article. All versions of this article have been corrected. (The following abstract of the original article appeared in record 2020-88205-001.) Attention determines which cues receive processing and are learned about. Learning, however, leads to attentional biases. In the study of animal learning, in some circumstances, cues that have been previously predictive of their consequences are subsequently learned about more than are nonpredictive cues, suggesting that they receive more attention. In other circumstances, cues that have previously led to uncertain consequences are learned about more than are predictive cues. In human learning, there is a clear role for predictiveness, but a role for uncertainty has been less clear. Here, in a human learning task, we show that cues that led to uncertain outcomes were subsequently learned about more than were cues that were previously predictive of their outcomes. This effect occurred when there were few uncertain cues. When the number of uncertain cues was increased, attention switched to predictive cues. This pattern of results was found for cues (1) that were uncertain because they led to 2 different outcomes equally often in a nonpredictable manner and (2) that were used in a nonlinear discrimination and were not predictive individually but were predictive in combination with other cues. This suggests that both the opposing predictiveness and uncertainty effects were determined by the relationship between individual cues and outcomes rather than the predictive strength of combined cues. These results demonstrate that learning affects attention; however, the precise nature of the effect on attention depends on the level of task complexity, which reflects a potential switch between exploration and exploitation of cues. (PsycInfo Database Record (c) 2021 APA, all rights reserved).

Category relevance attenuates overshadowing in human predictive learning.

In situations in which multiple predictors anticipate the presence or absence of an outcome, cues compete to anticipate the outcome, resulting in a loss of associative strength compared to control conditions without additional cues. Critically, there are multiple factors modulating the magnitude and direction of such competition, although in some scenarios the effect of these factors remains unexplored. We sought to assess whether the relative salience of the elements in a compound of cues modulates the magnitude of the overshadowing effect in human predictive learning. Two separable categories (i.e., colors and symbols) were used in a predictive learning task. In Experiment 1, different groups of participants were granted with different time of exposure to a compound of cues belonging to different categories (color and symbol) to evaluate potential differences in the magnitude of overshadowing. Furthermore, we used posttest questionnaires to assess whether participants used either only one or both categories during training, and assessed if this impacted the magnitude of overshadowing. In general, overshadowing was not modulated by the time of exposition, except in the case of very short time of exposition with prominent learning about the most salient category. In Experiment 2, the relative salience of a category was biased via prior experience either with a biconditional discrimination or attending only the relevant category (either color or symbol). The previously relevant category was less prone to overshadowing, but not the alternative one. Results are discussed in light of attentional and configural theories of associative learning. (PsycInfo Database Record (c) 2023 APA, all rights reserved).

Subsampling of cues in associative learning.

Theories of learning distinguish between elemental and configural stimulus processing depending on whether stimuli are processed independently or as whole configurations. Evidence for elemental processing comes from findings of summation in animals where a compound of two dissimilar stimuli is deemed to be more predictive than each stimulus alone, whereas configural processing is supported by experiments using similar stimuli in which summation is not found. However, in humans the summation effect is robust and impervious to similarity manipulations. In three experiments in human predictive learning, we show that summation can be obliterated when partially reinforced cues are added to the summands in training and tests. This lack of summation only holds when the partially reinforced cues are similar to the reinforced cues (experiment 1) and seems to depend on participants sampling only the most salient cue in each trial (experiments 2a and 2b) in a sequential visual search process. Instead of attributing our and others' instances of lack of summation to the customary idea of configural processing, we offer a formal subsampling rule that might be applied to situations in which the stimuli are hard to parse from each other.

Reversing the relationship between a nontarget cue and the outcome facilitates subsequent human predictive learning

Two experiments were conducted to test the effect of experiencing associative interference on later learning. A predictive learning task was used in which human participants had to evaluate whether plants would grow or not (Outcome) after being watered with different fertilizers (Cues). Experiment 1 found that the increase in the prediction error produced by following a pre-exposed nontarget cue by the outcome, facilitated subsequent acquisition of the relationship between the pre-exposed target cue and the outcome. Experiment 2 compared whether learning about the target cue was differentially affected by experiencing two types of associative interference with the nontarget cue: Pairing the pre-exposed cue with the outcome and presenting the cue without outcome after being paired with it. The experience of associative interference with nontarget cues similarly facilitated subsequent learning about the target cue, regardless of the direction of the change in the nontarget cue-outcome relationship. It is suggested that the increase in prediction error produced by the experience of associative interference may lead to a general increase in attention that facilitates subsequent learning.

Uncertainty and predictiveness modulate attention in human predictive learning.

Attention determines which cues receive processing and are learned about. Learning, however, leads to attentional biases. In the study of animal learning, in some circumstances, cues that have been previously predictive of their consequences are subsequently learned about more than are nonpredictive cues, suggesting that they receive more attention. In other circumstances, cues that have previously led to uncertain consequences are learned about more than are predictive cues. In human learning, there is a clear role for predictiveness, but a role for uncertainty has been less clear. Here, in a human learning task, we show that cues that led to uncertain outcomes were subsequently learned about more than were cues that were previously predictive of their outcomes. This effect occurred when there were few uncertain cues. When the number of uncertain cues was increased, attention switched to predictive cues. This pattern of results was found for cues (1) that were uncertain because they led to 2 different outcomes equally often in a nonpredictable manner and (2) that were used in a nonlinear discrimination and were not predictive individually but were predictive in combination with other cues. This suggests that both the opposing predictiveness and uncertainty effects were determined by the relationship between individual cues and outcomes rather than the predictive strength of combined cues. These results demonstrate that learning affects attention; however, the precise nature of the effect on attention depends on the level of task complexity, which reflects a potential switch between exploration and exploitation of cues.

Concurrent evidence of extinction making acquisition context specific and ABA and ABC renewal effects in human predictive learning.

Two experiments evaluated whether the experience of extinction makes acquisition context specific (EMACS) while the extinction learning itself also becomes context dependent under ABA and ABC renewal designs in a human predictive learning situation. Two groups of participants received X-Outcome pairings in context A followed by P-Outcome pairings in context B. For participants in group E, cue X was then extinguished in context B while cue P was trained. Participants in group NE were trained with P, but they did not have the extinction experience. Testing target cues outside the context B (i.e. the context in which P was trained and X was extinguished) in group E led to an increase in responding to cue X (Renewal effect) and a decrease in responding to cue P (EMACS effect) regardless of whether the test was conducted in context A (Experiment 1) or in an alternative context C (Experiment 2). Combined results suggest that Renewal and EMACS effects may be based on the same underlying mechanism. (PsycInfo Database Record (c) 2021 APA, all rights reserved).

Supplemental Material for Concurrent Evidence of Extinction Making Acquisition Context Specific and ABA and ABC Renewal Effects in Human Predictive Learning

Supplemental Material for Concurrent Evidence of Extinction Making Acquisition Context Specific and ABA and ABC Renewal Effects in Human Predictive Learning

The Redundancy Effect on Human Predictive Learning: Evidence against a Propositional Interpretation

The redundancy effect is the finding of greater learning when an X stimulus is trained in an A+ AX+ blocking procedure, than when a Y stimulus is trained in a BY+ CY- discrimination procedure. These findings are new and theoretically challenging for all conditioning theories that calculate learning based on a common error. For this reason, we alternatively examined the possibility that the phenomenon is the result of a propositional reasoning. In an experiment, we replicated the basic effect and we found out that the addition of instructions on the occurrence of the consequences at a submaximal level does not have a significant impact on the redundancy effect. These findings are discussed with regard to a propositional and associative approach based on the assumption that the experimental stimuli share a common feature.

Does learning history shape the associability of outcomes? Further tests of the outcome predictability effect.

In recent years, several studies of human predictive learning demonstrated better learning about outcomes that have previously been experienced as consistently predictable compared to outcomes previously experienced as less predictable, namely the outcome predictability effect. As this effect may have wide-reaching implications for current theories of associative learning, the present study aimed to examine the generality of the effect with a human goal-tracking paradigm, employing three different designs to manipulate the predictability of outcomes in an initial training phase. In contrast to the previous studies, learning in a subsequent phase, when every outcome was equally predictable by novel cues, was not reliably affected by the outcomes’ predictability in the first phase. This lack of an outcome predictability effect provides insights into the parameters of the effect and its underlying mechanisms.

Supplemental Material for Uncertainty and Predictiveness Modulate Attention in Human Predictive Learning

Supplemental Material for Uncertainty and Predictiveness Modulate Attention in Human Predictive Learning

Adding or removing context components equally disrupts extinction in human predictive learning

In two human predictive learning experiments, we investigated the impact of adding or removing context components on extinction performance toward a cue. In each experiment, participants initially received repeated pairings of a cue and an outcome in a context composed of two distinctive components (context AB). Initial training was followed by a series of trials in which the cue was no longer followed by the outcome. This extinction treatment was conducted in the presence of a different pair of distinctive context components (context CD). During a final test, we observed that changing the extinction context CD disrupted extinction performance toward the cue regardless of whether the context was changed by adding or removing context components. We discuss implications of our results for theories of associative learning.

Test-retest reliability of response recovery after discrimination reversal learning

In one human predictive learning experiment, we demonstrated that an individual’s propensity for response recovery following discrimination reversal learning is stable over time. Participants received four sessions of training with the first three sessions being separated by one week each, while the last session was conducted after a delay of four weeks. During each session, participants initially received discrimination training (E+, F-) in one context, followed by discrimination reversal training (E-, F+) in another context. Sessions each completed with a test, in which the stimuli were presented in the context of initial acquisition. Each test revealed response recovery according to the initially acquired stimulus-outcome contingencies. Furthermore, the strength of response recovery was correlated across sessions that were separated by one week (Sessions 1 and 2), and across sessions separated by four weeks (Sessions 3 and 4). Overall, intra-individual test behavior was stable in 87 % of participants across two sessions, and in 79 % of participants across four sessions. Our results indicate that inter-individual differences in response recovery are a reliable phenomenon, which is a finding that is not accounted by current theories of context-dependent learning.

Prior extinction increases acquisition context specificity in human predictive learning

One experiment evaluated the effect of extinction on the context dependence of non-extinguished information in a situation in which physical (images), rather than predominantly verbal, contexts were used in human predictive learning. Participants received training in which different foods (Cues) were associated with the presence or the absence of gastric illness (outcome) in customers of different restaurants (contexts). One cue was associated with the gastric illness while a different cue was either extinguished or not between groups. A change in the context at test led to a general decrease in both predictive judgments and the speed of responding to the non-extinguished cue. However, these decreases were greater when training was conducted during extinction of the different cue demonstrating the extinction makes acquisition context-specific (EMACS) effect. Results are contrasted with failures to find the effect in other reports and discussed in terms of extinction leading to an allocation of attentional resources to the context, facilitating the context dependence of information.

Attentional instructions modulate differential context-switch effects after short and long training in human predictive learning.

Three experiments in human predictive learning assessed the modulating role of instructions on context-switch effects on performance after different levels of training. Cue X (a food name) was paired with an outcome (gastric malaise) in Context A (a specific restaurant), whereas another cue, Y, was presented in the absence of outcome in Context B. The series manipulated the testing context (same or different from the acquisition context), the length of training (short vs. long), and the instructions participants received before testing (attentional or neutral). Attentional instructions intended to either focusing participants' attention on the context (Experiments 1 and 2) or to take attention away from the context (Experiment 3). In agreement with the predictions of the attentional theory of context processing, instructions that focused participants' attention on the context made retrieval of information after long training context specific, something that did not occur in the absence of attentional instructions. Conversely, instructions that took participants' attention away from the context (by focusing their attention on the cue) attenuated context-switch effects that otherwise appear after short training. (PsycINFO Database Record (c) 2019 APA, all rights reserved).

Experiencing extinction with a non-target cue facilitates reversal of a target conditioned inhibitor in human predictive learning

Three experiments tested the effect of experiencing extinction on learning about a differential conditioned inhibitor that was trained as an excitor. A human predictive learning task was used in which participants had to evaluate the probability of different colored fertilizers (Cues) leading plants to flourish or not (Outcome). Experiment 1 found that presenting the target cue without outcome while other cues were followed by the outcome made the target cue a conditioned inhibitor, passing both, retardation (Experiment 1a) and summation (Experiment 1b) tests of conditioned inhibition. Subsequent extinction of a different cue facilitated reversing the relationship between the conditioned inhibitor and the outcome regardless of whether the situation could be solved by using simple rules (Experiment 2) or not (Experiment 3). Results are discussed in terms of attentional theories that suggest extinction produces a nonspecific increase in attention that facilitates learning.

The fate of redundant cues in human predictive learning: The outcome ratio effect.

In four experiments, participants were shown a sequence of pairs of pictures of food and asked to predict whether each pair signalled an allergic reaction in a hypothetical patient. The pairs of pictures were used to present two simple discriminations that differed in their outcome ratio. A rich discrimination, 3AX+ BX-, involved three trials in which the compound of two foods, AX, was followed by a reaction, for every trial in which the compound BX was not followed by the outcome. A lean discrimination, CY+ 3DY- was based on the opposite outcome ratio. Upon the completion of this training, participants were asked to rate how likely an individual food would be followed by the allergic reaction. In each experiment, the rating for X was stronger than for Y. This outcome ratio effect poses a challenge for theories of learning that assume changes in associative strength are governed by a common error term, based on the significance of all the cues present on a trial. Instead, the results are consistent with the assumption that changes in associative strength are governed by an individual error term, based on the significance of a single cue.

Attention toward contexts modulates context-specificity of behavior in human predictive learning: Evidence from the n-back task.

According to the attentional theory of context processing (ATCP), learning becomes context specific when acquired under conditions that promote attention toward contextual stimuli regardless of whether attention deployment is guided by learning experience or by other factors unrelated to learning. In one experiment with humans, we investigated whether performance in a predictive learning task can be brought under contextual control by means of a secondary task that was unrelated to predictive learning, but supposed to modulate participants' attention toward contexts. Initially, participants acquired cue-outcome relationships presented in contexts that were each composed of two elements from two dimensions. Acquisition training in the predictive learning task was combined with a one-back task that required participants to match across consecutive trials context elements belonging to one of the two dimensions. During a subsequent test, we observed that acquisition behavior in the predictive learning task was disrupted by changing the acquisition context along the dimension that was relevant for the one-back task, while there was no evidence for context specificity of predictive learning when the acquisition context was changed along the dimension that was irrelevant for the one-back task. Our results support the generality of the principles advocated by ATCP.

A test for a difference in the associability of blocked and uninformative cues in human predictive learning.

In human predictive learning, blocking, A+ AB+, and a simple discrimination, UX+ VX-, result in a stronger response to the blocked, B, than the uninformative cue, X (where letters represent cues and + and - represent different outcomes). To assess whether these different treatments result in more attention being paid to blocked than uninformative cues, Stage 1 in each of three experiments generated two blocked cues, B and E, and two uninformative cues, X and Y. In Stage 2, participants received two simple discriminations: either BX+ EX- and BY+ EY-, or BX+ BY- and EX+ EY-. If more attention is paid to blocked than uninformative cues, then the first pair of discriminations will be solved more readily than the second pair. In contrast to this prediction, both discriminations were acquired at the same rate. These results are explained by the theory of Mackintosh, by virtue of the assumption that learning is governed by an individual rather than a common error term.

The Reinstatement Effect in Human Predictive Learning: Contextual Modulation and the Impact of Extinction Reminders.

One of the most relevant phenomena both from a theoretical and clinical perspective is extinction. In particular, several researchers are interested in the response recovery effects from extinction. Reinstatement is an effect that has been proposed as a laboratory model to study relapse from extinction-based therapeutic treatments. We designed two experiments with humans to evaluate the reinstatement effect in a predictive learning task. In both experiments, participants learned a specific relationship between two cues (X and Y) and two outcomes (O1 and O2) during the first phase. Throughout extinction, both cues were presented without outcomes. After an exposure to the original outcomes, reinstatement of the first-learned information was observed during testing in both experiments. However, we found that the reinstatement effect was contextual modulated (Experiment 1; ηp2 = .78, 90% CI [.48, .86], p < .0001). Furthermore, in Experiment 2 we showed a reduction of reinstatement when an extinction reminder was used ηp2 = .45, 90% CI [.07, .65], p = .012. Theoretical implications are discussed, and some potential uses are mentioned.