A single-system account of the relationship between priming, recognition, and fluency

A single-system computational model of priming and recognition was applied to studies that have looked at the relationship between priming, recognition, and fluency in continuous identification paradigms. The model was applied to 3 findings that have been interpreted as evidence for a multiple-systems account: (a) priming can occur for items not recognized; (b) the pattern of identification reaction times (RTs) to hits, misses, correct rejections, and false alarms can change as a function of recognition performance; and (c) fluency effects (shorter RTs to words judged old vs. judged new) and priming effects (shorter RTs to old vs. new words) can be observed in amnesic patients at levels comparable with healthy adults despite impaired or near-chance recognition. The authors' simulations suggest, contrary to previous interpretations, that these results are consistent with a single-system account.

Challenging the role of implicit processes in probabilistic category learning

Considerable interest in the hypothesis that different cognitive tasks recruit qualitatively distinct processing systems has led to the proposal of separate explicit (declarative) and implicit (procedural) systems. A popular probabilistic category learning task known as the weather prediction task is said to be ideally suited to examine this distinction because its two versions, "observation" and "feedback," are claimed to recruit the declarative and procedural systems, respectively. In two experiments, we found results that were inconsistent with this interpretation. In Experiment 1, a concurrent memory task had a detrimental effect on the implicit (feedback) version of the task. In Experiment 2, participants displayed comparable and accurate insight into the task and their judgment processes in the feedback and observation versions. These findings have important implications for the study of probabilistic category learning in both normal and patient populations.

Disrupted prediction-error signal in psychosis: evidence for an associative account of delusions

Delusions are maladaptive beliefs about the world. Based upon experimental evidence that prediction error-a mismatch between expectancy and outcome--drives belief formation, this study examined the possibility that delusions form because of disrupted prediction--error processing. We used fMRI to determine prediction-error-related brain responses in 12 healthy subjects and 12 individuals (7 males) with delusional beliefs. Frontal cortex responses in the patient group were suggestive of disrupted prediction-error processing. Furthermore, across subjects, the extent of disruption was significantly related to an individual's propensity to delusion formation. Our results support a neurobiological theory of delusion formation that implicates aberrant prediction-error signalling, disrupted attentional allocation and associative learning in the formation of delusional beliefs.

Attention modulates the learning of multiple contingencies

We employed a novel version of the serial reaction time task to test the idea that human implicit learning allows the simultaneous learning of multiple independent contingencies and that this learning may proceed in the absence of attention. Using probabilistic sequences, we showed that both a primary sequence (the focus of the experimental task) and a statistically independent secondary sequence could be learned across 4,800 target localization trials, provided the perceptual load of the primary task was low. However, learning of the secondary sequence was abolished under conditions of high perceptual load. These findings suggest that there are attentional limitations on the learning of multiple contingencies.

Summation in causal learning: elemental processing or configural generalization?

Results from human causal learning tasks that employ multiple cues are often interpreted in terms of the elemental theory of Rescorla and Wagner (1972). However, most results can also be successfully interpreted by the configural model proposed by Pearce (1987, 1994). One method of discriminating between these alternatives is through an investigation of summation and overexpectation. Indeed, demonstrations of these phenomena are fundamental to an elemental approach but are generally incompatible with an account that involves configural processing. Using a procedure in which the magnitude of the outcome varied, evidence for both summation and overexpectation was obtained in two experiments.

Bayesian associative learning

The Bayesian approach to belief updating is par excellence a theory of learning--a theory of how beliefs should be revised in the light of new evidence. Yet despite its considerable utility as a framework for understanding cognition, it has not been prominent in theorizing about elementary learning processes in animals and humans. New research by Kruschke seems set to change this and to offer a bridge between rational and empirically descriptive accounts of learning.

Disruption of sequential priming in organic and pharmacological amnesia: a role for the medial temporal lobes in implicit contextual learning

We examined learning and expression of contextual implicit learning of a sequence of targets in a speeded target-detection task in amnesic and control participants. Amnesia was of organic origin in one participant group and induced psychopharmacologically (diazepam 7.5 or 15 mg) in another. Although the amnesic groups were able to learn the target sequence normally, their expression of sequence knowledge (priming) was attenuated when contextual support was limited. This was evaluated by studying response latencies for targets primed by between 0 and 5 preceding context locations. Whereas control participants showed priming when the current target location was primed by only two previous locations, priming was eliminated with two (but not four) previous locations by a low dose of diazepam and was eliminated even with four elements of context under a high dose of diazepam and in amnesia of organic origin. The results suggest that a function of the hippocampal memory system is to support contextual learning and performance, even when that learning is nondeclarative.

Frontal responses during learning predict vulnerability to the psychotogenic effects of ketamine: linking cognition, brain activity, and psychosis

CONTEXT

Establishing a neurobiological account of delusion formation that links cognitive processes, brain activity, and symptoms is important to furthering our understanding of psychosis.

OBJECTIVE

To explore a theoretical model of delusion formation that implicates prediction error-dependent associative learning processes in a pharmacological functional magnetic resonance imaging study using the psychotomimetic drug ketamine.

DESIGN

Within-subject, randomized, placebo-controlled study.

SETTING

Hospital-based clinical research facility, Addenbrooke's Hospital, Cambridge, England. The work was completed within the Wellcome Trust and Medical Research Council Behavioral and Clinical Neuroscience Institute, Cambridge.

PARTICIPANTS

Fifteen healthy, right-handed volunteers (8 of whom were male) with a mean +/- SD age of 29 +/- 7 years and a mean +/- SD predicted full-scale IQ of 113 +/- 4 were recruited from within the local community by advertisement.

INTERVENTIONS

Subjects were given low-dose ketamine (100 ng/mL of plasma) or placebo while performing a causal associative learning task during functional magnetic resonance imaging. In a separate session outside the scanner, the dose was increased (to 200 ng/mL of plasma) and subjects underwent a structured clinical interview.

MAIN OUTCOME MEASURES

Brain activation, blood plasma levels of ketamine, and scores from psychiatric ratings scales (Brief Psychiatric Ratings Scale, Present State Examination, and Clinician-Administered Dissociative States Scale).

RESULTS

Low-dose ketamine perturbs error-dependent learning activity in the right frontal cortex (P = .03). High-dose ketamine produces perceptual aberrations (P = .01) and delusion-like beliefs (P = .007). Critically, subjects showing the highest degree of frontal activation with placebo show the greatest occurrence of drug-induced perceptual aberrations (P = .03) and ideas or delusions of reference (P = .04).

CONCLUSIONS

These findings relate aberrant prediction error-dependent associative learning to referential ideas and delusions via a perturbation of frontal cortical function. They are consistent with a model of delusion formation positing disruptions in error-dependent learning.

The Comparator Theory Fails to Account for the Selective Role of Within-Compound Associations in Cue-Selection Effects

In a human causal learning experiment, we investigated cue selection effects to test the comparator theory ( Denniston, Savastano, & Miller, 2001 ; Miller & Matzel, 1988 ). The theory predicts that the occurrence of cue selection is independent of whether the relevant learning trials are presented in a standard forward manner or in a backward manner and that within-compound associations are of equal importance in both cases. We found that the strength of the cue-selection effect was positively correlated with knowledge of within-compound associations in the backward condition but not in the forward condition. Furthermore, cue-selection effects were less pronounced in the former than in the latter condition. These results are at variance with the comparator hypothesis but are in agreement with a modified associative theory and with the suggestion that retrospective revaluation might be due to rehearsal processes.

Insight and strategy in multiple-cue learning

In multiple-cue learning (also known as probabilistic category learning) people acquire information about cue-outcome relations and combine these into predictions or judgments. Previous researchers claimed that people can achieve high levels of performance without explicit knowledge of the task structure or insight into their own judgment policies. It has also been argued that people use a variety of suboptimal strategies to solve such tasks. In three experiments the authors reexamined these conclusions by introducing novel measures of task knowledge and self-insight and using "rolling regression" methods to analyze individual learning. Participants successfully learned a four-cue probabilistic environment and showed accurate knowledge of both the task structure and their own judgment processes. Learning analyses suggested that the apparent use of suboptimal strategies emerges from the incremental tracking of statistical contingencies in the environment.

Sequence learning and selection difficulty

The authors studied the role of attention as a selection mechanism in implicit learning by examining the effect on primary sequence learning of performing a demanding target-selection task. Participants were trained on probabilistic sequences in a novel version of the serial reaction time (SRT) task, with dual- and triple-stimulus participants having to ignore irrelevant items in the SRT display. Despite large performance decrements under dual- and triple-stimulus configurations, testing under single-stimulus conditions revealed no impairment to sequence learning. These findings suggest that implicit sequence learning is resistant to disruption of the selection process. Results are discussed in terms of a componential model of attention and in relation to the implicit-explicit distinction.

On the status of unconscious memory: Merikle and Reingold (1991) revisited

Four experiments are reported that reevaluate P. M. Merikle and E. M. Reingold's (1991) demonstration of unconscious memory: the greater sensitivity to familiarity (repetition) of an indirect (implicit) memory task than of a comparable direct (explicit) task. At study, participants named the cued member of a pair of visually presented words. At test, new and uncued study words were presented against a background mask. Participants judged whether each word was old or new (direct task) or whether the contrast between the word and the background was high or low (indirect task). Contrary to the original findings, the sensitivity of the indirect task to familiarity never exceeded that of the direct task. These findings pose a challenge to a key pillar of evidence for unconscious influences of memory.

Individual differences in causal learning and decision making

In judgment and decision making tasks, people tend to neglect the overall frequency of base-rates when they estimate the probability of an event; this is known as the base-rate fallacy. In causal learning, despite people's accuracy at judging causal strength according to one or other normative model (i.e., Power PC, DeltaP), they tend to misperceive base-rate information (e.g., the cause density effect). The present study investigates the relationship between causal learning and decision making by asking whether people weight base-rate information in the same way when estimating causal strength and when making judgments or inferences about the likelihood of an event. The results suggest that people differ according to the weight they place on base-rate information, but the way individuals do this is consistent across causal and decision making tasks. We interpret the results as reflecting a tendency to differentially weight base-rate information which generalizes to a variety of tasks. Additionally, this study provides evidence that causal learning and decision making share some component processes.

Attentional load and implicit sequence learning

A widely employed conceptualization of implicit learning hypothesizes that it makes minimal demands on attentional resources. This conjecture was investigated by comparing learning under single-task and dual-task conditions in the sequential reaction time (SRT) task. Participants learned probabilistic sequences, with dual-task participants additionally having to perform a counting task using stimuli that were targets in the SRT display. Both groups were then tested for sequence knowledge under single-task (Experiments 1 and 2) or dual-task (Experiment 3) conditions. Participants also completed a free generation task (Experiments 2 and 3) under inclusion or exclusion conditions to determine if sequence knowledge was conscious or unconscious in terms of its access to intentional control. The experiments revealed that the secondary task impaired sequence learning and that sequence knowledge was consciously accessible. These findings disconfirm both the notion that implicit learning is able to proceed normally under conditions of divided attention, and that the acquired knowledge is inaccessible to consciousness. A unitary framework for conceptualizing implicit and explicit learning is proposed.

Prediction error during retrospective revaluation of causal associations in humans: fMRI evidence in favor of an associative model of learning

Associative learning theory assumes that prediction error is a driving force in learning. A competing view, probabilistic contrast (PC) theory, is that learning and prediction error are unrelated. We tested a learning phenomenon that has proved troublesome for associative theory--retrospective revaluation--to evaluate these two models. We previously showed that activation in right lateral prefrontal cortex (PFC) provides a reliable signature for the presence of prediction error. Thus, if the associative view is correct, retrospective revaluation should be accompanied by right lateral PFC activation. PC theory would be supported by the absence of this activation. Right PFC and ventral striatal activation occurred during retrospective revaluation, supporting the associative account. Activations appeared to reflect the degree of revaluation, predicting later brain responses to revalued cues. Our results support a modified associative account of retrospective revaluation and demonstrate the potential of functional neuroimaging as a tool for evaluating competing learning models.

Dissociation Between Judgments and Outcome-Expectancy Measures in Covariation Learning: A Signal Detection Theory Approach

A number of studies using trial-by-trial learning tasks have shown that judgments of covariation between a cue c and an outcome o deviate from normative metrics. Parameters based on trial-by-trial predictions were estimated from signal detection theory (SDT) in a standard causal learning task. Results showed that manipulations of P(c) when contingency (deltaP) was held constant did not affect participants' ability to predict the appearance of the outcome (d') but had a significant effect on response criterion (c) and numerical causal judgments. The association between criterion c and judgment was further demonstrated in 2 experiments in which the criterion was directly manipulated by linking payoffs to the predictive responses made by learners. In all cases, the more liberal the criterion c was, the higher judgments were. The results imply that the mechanisms underlying the elaboration of judgments and those involved in the elaboration of predictive responses are partially dissociable.

The role of the lateral frontal cortex in causal associative learning: exploring preventative and super-learning

Prediction error--a mismatch between expected and actual outcome--is critical to associative accounts of inferential learning. However, it has proven difficult to explore the effects of prediction error using functional magnetic resonance imaging (fMRI) while excluding the confounding effects of stimulus novelty and incorrect responses. In this event-related fMRI study we used a three-stage experiment generating preventative- and super-learning conditions. In both cases, it was possible to generate prediction error within a causal associative learning experiment while subtracting the effects of novelty and error. We show that right lateral prefrontal cortex (PFC) activation is sensitive to the magnitude of prediction error. Furthermore, super-learning activation in this region of PFC correlates, across subjects, with the amount learned. We thus provide direct evidence for a brain correlate of the surprise-dependent mechanisms proposed by associative accounts of causal learning. We show that activity in right lateral PFC is sensitive to the magnitude, though not the direction, of the prediction error. Furthermore, its activity is not directly explicable in terms of novelty or response errors and appears directly related to the learning that arises out of prediction error.

On the Role of Recognition in Decision Making

In 2 experiments, the authors sought to distinguish between the claim that recognition of an object is treated simply as a cue among others for the purposes of decision making in a cue-learning task from the claim that recognition is attributed a special status with fundamental, noncompensatory properties. Results of both experiments supported the former interpretation. When recognition had a high predictive validity, it was relied on (solely) by the majority of participants; however, when other cues in the environment had higher validity, recognition was ignored, and these other cues were used. The results provide insight into when, where, and why recognition is used in decision making and also question the elevated status assigned to recognition in some frameworks (e.g., D. G. Goldstein & G. Gigerenzer, 2002).

Intentional control and implicit sequence learning

Sequence knowledge acquired by repeated exposure to targets in a speeded localization task was studied in 3 experiments that sought to test A. Destrebecqz and A. Cleeremans's (2001, 2003) claim that, under certain circumstances, the expression of such sequence knowledge cannot be brought under intentional control. In Experiment 1 participants were trained on either a deterministic or a probabilistic sequence and then performed a free-generation test under either inclusion or exclusion instructions. Participants were found to be capable of both expressing (inclusion) and avoiding expressing (exclusion) sequence knowledge. These results were confirmed in Experiment 2 with a more exact replication of Destrebecqz and Cleeremans's methodology. In Experiment 3 participants performed a trial-by-trial generation test under both inclusion and exclusion conditions after a much longer period of training. All the findings are consistent with the proposal that information acquired during sequence learning is explicit in nature.