Predictions penetrate perception: Converging insights from brain,behaviour and disorder

It is argued that during ongoing visual perception, the brain is generating top-down predictions to facilitate, guide and constrain the processing of incoming sensory input. Here we demonstrate that these predictions are drawn from a diverse range of cognitive processes, in order to generate the richest and most informative prediction signals. This is consistent with a central role for cognitive penetrability in visual perception. We review behavioural and mechanistic evidence that indicate a wide spectrum of domains—including object recognition, contextual associations, cognitive biases and affective state—that can directly influence visual perception. We combine these insights from the healthy brain with novel observations from neuropsychiatric disorders involving visual hallucinations, which highlight the consequences of imbalance between top-down signals and incoming sensory information. Together, these lines of evidence converge to indicate that predictive penetration, be it cognitive, social or emotional, should be considered a fundamental framework that supports visual perception.     

Mapping emotions through time: how affective trajectories inform the language of emotion

The words used to describe emotions can provide insight into the basic processes that contribute to emotional experience. We propose that emotions arise partly from interacting evaluations of one's current affective state, previous affective state, predictions for how these may change in the future, and the experienced outcomes following these predictions. These states can be represented and inferred from neural systems that encode shifts in outcomes and make predictions. In two studies, we demonstrate that emotion labels are reliably differentiated from one another using only simple cues about these affective trajectories through time. For example, when a worse-than-expected outcome follows the prediction that something good will happen, that situation is labeled as causing anger, whereas when a worse-than-expected outcome follows the prediction that something bad will happen, that situation is labeled as causing sadness. Emotion categories are more differentiated when participants are required to think categorically than when participants have the option to consider multiple emotions and degrees of emotions. This work indicates that information about affective movement through time and changes in affective trajectory may be a fundamental aspect of emotion categories. Future studies of emotion must account for the dynamic way that we absorb and process information.     

Stronger effects of self-generated than cue-induced expectations when verifying predictions in data graphs

Expectations can come in different forms when analyzing and presenting data. Prior studies have documented stronger effects (behavioral and electrophysiological) of self-generated predictions as compared to cues. While participants presumably cannot help but use their own predictions, they might sometimes ignore cues (of low validity). In two experiments we compared the impact of cues (verbal and visual) and self-generated predictions on the performance of participants checking their current prediction against a presented data graph (linear upward or downward trend). Different from prior studies, the setup allowed for within-experiment comparison of different cue formats and ensured that cues could not be ignored. Nevertheless we found that self-generated predictions had a stronger impact than cues. Verbal cues had a stronger effect than visual cues without verbalization. Responses to graphs with a linear upward trend were faster and were influenced more strongly by predictions, than the response to graphs with a downward trend.     

What will Phil do next? : Spontaneously inferred traits influence predictions of behavior

Past research has shown that perceivers intentionally may make trait inferences about others and use this information to make predictions about these others' future behaviors. Other research has also shown that people can make trait inferences without intent—that is, spontaneously. However, one unexplored avenue is whether spontaneous trait inferences (STI), affect how perceivers predict others' will behavior. Three studies explored this issue. Results from Studies 1 and 2 showed that: (1) exposure to trait-implicative behaviors describing an actor influences subsequent behavior predictions made about the actor in a trait-consistent manner, and (2) predictions occurred regardless of behavior recall, implying that the behavior predictions were derived from prior trait inferences and not from behavior recall. Results from Study 3 bolstered this conclusion by showing that behavior predictions were similar regardless of whether subjects were explicitly instructed to make inferences or not, but that a manipulation known to interfere with inference generation (lie detection instructions) muted behavior predictions. Results from Study 3 also suggested that the prediction effects had both automatic and controlled components, and that reductions observed in the lie detection condition of Study 3 were caused by alterations in the automatic influence of trait knowledge to the behavior predictions. These results suggest that STI may be causal inferences about the actors' dispositions.     

The difference engine: A model of diversity in speeded cognition

A theory of diversity in speeded cognition, the difference engine, is proposed, in which information processing is represented as a series of generic computational steps. Some individuals tend to perform all of these computations relatively quickly and other individuals tend to perform them all relatively slowly, reflecting the existence of a general cognitive speed factor, but the time required for response selection and execution is assumed to be independent of cognitive speed. The difference engine correctly predicts the positively accelerated form of the relation between diversity of performance, as measured by the standard deviation for the group, and task difficulty, as indexed by the mean response time (RT) for the group. In addition, the difference engine correctly predicts approximately linear relations between the RTs of any individual and average performance for the group, with the regression lines for fast individuals having slopes less than 1.0 (and positive intercepts) and the regression lines for slow individuals having slopes greater than 1.0 (and negative intercepts). Similar predictions are made for comparisons of slow, average, and fast subgroups, regardless of whether those subgroups are formed on the basis of differences in ability, age, or health status. These predictions are consistent with evidence from studies of healthy young and older adults as well as from studies of depressed and age-matched control groups.     

Frequency, probability, and prediction: easy solutions to cognitive illusions?

Many errors in probabilistic judgment have been attributed to people's inability to think in statistical terms when faced with information about a single case. Prior theoretical analyses and empirical results imply that the errors associated with case-specific reasoning may be reduced when people make frequentistic predictions about a set of cases. In studies of three previously identified cognitive biases, we find that frequency-based predictions are different from-but no better than-case-specific judgments of probability. First, in studies of the "planning fallacy, " we compare the accuracy of aggregate frequency and case-specific probability judgments in predictions of students' real-life projects. When aggregate and single-case predictions are collected from different respondents, there is little difference between the two: Both are overly optimistic and show little predictive validity. However, in within-subject comparisons, the aggregate judgments are significantly more conservative than the single-case predictions, though still optimistically biased. Results from studies of overconfidence in general knowledge and base rate neglect in categorical prediction underline a general conclusion. Frequentistic predictions made for sets of events are no more statistically sophisticated, nor more accurate, than predictions made for individual events using subjective probability.     

From origami to software development: a review of studies on judgment-based predictions of performance time

This article provides an integrative review of the literature on judgment-based predictions of performance time, often described as task duration predictions in psychology and as expert-based effort estimation in engineering and management science. We summarize results on the characteristics of performance time predictions, processes and strategies, the influence of task characteristics and contextual factors, and the relations between estimates and characteristics of the estimator. Although dependent on the type of study and the level of analysis, underestimation was more frequently reported than overestimation in studies from the engineering and management literature. However, this was not the case in studies from the psychology literature. Our summaries challenge earlier results regarding the effects of factors such as complexity/difficulty and experience. We also question the recurrent finding that small tasks are overestimated and large tasks are underestimated, as this to some extent can be a statistical artifact caused by random error. Several other influences on predictions are identified and discussed. These include various types of anchoring effects, performance and accuracy incentives, task decomposition, request formats, group estimation, revisions of initial ideal or incomplete estimates, level of abstraction, and superficial cues. We summarize similarities and differences between performance time predictions (e.g., number of work hours) and completion time predictions (e.g., delivery dates) because many studies fail to distinguish between these 2 types of predictions. Finally, we discuss methodological issues in time prediction research and implications for research and application.     

Investigation of Perceptual-Motor Behavior Across the Expert Athlete to Disabled Patient Skill Continuum can Advance Theory and Practical Application

A framework is presented of how theoretical predictions can be tested across the expert athlete to disabled patient skill continuum. Common-coding theory is used as the exemplar to discuss sensory and motor system contributions to perceptual-motor behavior. Behavioral and neural studies investigating expert athletes and patients recovering from cerebral stroke are reviewed. They provide evidence of bi-directional contributions of visual and motor systems to perceptual-motor behavior. Majority of this research is focused on perceptual-motor performance or learning, with less on transfer. The field is ripe for research designed to test theoretical predictions across the expert athlete to disabled patient skill continuum. Our view has implications for theory and practice in sports science, physical education, and rehabilitation.     

Tied to expectations: Predicting features speeds processing even under adverse circumstances

When participants predict the upcoming stimulus in a randomized choice reaction task, a match between prediction and stimulus increases processing speed at a level similar to that observed in cueing studies with highly valid cues. This might be taken to suggest that people cannot help but fully use their self-generated, verbalized predictions for preparing task processing. Thus, we tested how flexibly participants can control formation and implementation of predictions. In Experiment 1, we varied validity and response-relevance of predictions. We observed that prediction effects on RT can be boosted by increasing validity, but prevail under adverse circumstances. This was not the case in a control group who read rather than predicted the feature words, suggesting that the effect was specific to predictions as such. Experiment 2 provided further evidence for limited control of participants over implementing and forming predictions. Participants were provided with practice on stimuli occurring with varying frequency, but neither learned to strategically choose predictions to maximize the number of match trials, nor did they reduce the amount of prediction-based preparation when predicting an infrequent stimulus. As sequential aftereffects of prediction match did not vary with validity, they were identified as an independent effect of verbalizing a response-relevant stimulus feature. The results are consistent with the view that the predicted stimulus feature is represented in the focus of attention in working memory and that the amount of implementation can be subject to weighting.     

Predicting the intensity of losses vs. non-gains and non-losses vs. gains in judging fairness and value: A test of the loss aversion explanation

Three studies examined the predictions that in the context of evaluation of fairness and concessions in negotiations, losses would be perceived as more intensely negative than non-gains, and that non-losses would be perceived as more positive than gains. Extant studies tested only the first of these predictions. These predictions derive from the principle of loss aversion (LA), according to which losses are experienced more intensely than gains of similar objective magnitude. In this view, losses and non-losses are measured against the steep loss part of the value curve, whereas gains and non-gains are measured against the shallow part of the value curve. Our studies replicated extant studies in confirming the first prediction but failed to confirm the second prediction. Specifically, opposite to the prediction of LA, gains were perceived as more intensely positive than non-losses. It seems, therefore, that LA is not a sufficient explanation of why losses are perceived as more averse than gains. Feature positive and regulatory focus effects are discussed as additional potential contributors to the phenomenon.     

"Object categorization: reversals and explanations of the basic-level advantage" (Rogers & Patterson, 2007): a simplicity account

T. T. Rogers and K. Patterson (2007), in their article "Object Categorization: Reversals and Explanations of the Basic-Level Advantage" (Journal of Experimental Psychology: General, 136, 451-469), reported an impressive set of results demonstrating a reversal of the highly robust basic-level advantage both in patients with semantic dementia and in healthy individuals engaged in a speeded categorization task. To explain their results, as well as the usual basic-level advantage seen in healthy individuals, the authors employed a parallel distributed processing theory of conceptual knowledge. In this paper, we introduce an alternative way of explaining the results of Rogers and Patterson, which is premised on a more restricted set of assumptions born from standard categorization theory. Specifically, we provide evidence that their results can be accounted for based on the predictions of the simplicity model of unsupervised categorization.     

On bias in magnitude scaling and some conjectures of Stevens

Bias in magnitude scaling can be viewed as involving deviations of judgments from proportionality. A model of bias is shown to provide a theoretical basis for Stevens's conjecture about geometrically averaging magnitude estimation and magnitude production exponents in order to obtain an estimate of the psychophysical exponent. An overlooked result is that one can also obtain an estimate of the magnitude of the bias. Examples from several well-known studies are presented. The bias is also shown to vary in response to experimental manipulation of the stimulus range. Aspects of predicting exponents across experiments are clarified, and a new prediction is examined. The model of bias fills some theoretical gaps in magnitude scaling and clarifies underlying assumptions and predictions.     

Bayes and blickets: effects of knowledge on causal induction in children and adults

People are adept at inferring novel causal relations, even from only a few observations. Prior knowledge about the probability of encountering causal relations of various types and the nature of the mechanisms relating causes and effects plays a crucial role in these inferences. We test a formal account of how this knowledge can be used and acquired, based on analyzing causal induction as Bayesian inference. Five studies explored the predictions of this account with adults and 4-year-olds, using tasks in which participants learned about the causal properties of a set of objects. The studies varied the two factors that our Bayesian approach predicted should be relevant to causal induction: the prior probability with which causal relations exist, and the assumption of a deterministic or a probabilistic relation between cause and effect. Adults' judgments (Experiments 1, 2, and 4) were in close correspondence with the quantitative predictions of the model, and children's judgments (Experiments 3 and 5) agreed qualitatively with this account.     

When distance pays off: The role of construal level in spending predictions

Self predictions are often optimistically biased, even for recurrent events. People could generate more realistic predictions by using information about past experiences, however they tend to disregard this cognitive approach. Drawing on Construal Level Theory, we propose that increases in construal level facilitate the use of information from past experience, and thereby increase prediction accuracy. This proposal was tested in two studies examining predictions of personal spending. Consistent with the hypotheses, individuals induced to construe the prediction target at a higher level of abstraction generated more accurate predictions (Study 1) and the effect of increased construal level on prediction was attributable to a greater reliance on past experience (Studies 1 and 2). The findings indicate that high-level construal can sometimes benefit prediction accuracy.     

“Object Categorization: Reversals and Explanations of the Basic-Level Advantage” (Rogers & Patterson, 2007): A simplicity account

T. T. Rogers and K. Patterson (2007), in their article “Object Categorization: Reversals and Explanations of the Basic-Level Advantage” (Journal of Experimental Psychology: General, 136, 451–469), reported an impressive set of results demonstrating a reversal of the highly robust basic-level advantage both in patients with semantic dementia and in healthy individuals engaged in a speeded categorization task. To explain their results, as well as the usual basic-level advantage seen in healthy individuals, the authors employed a parallel distributed processing theory of conceptual knowledge. In this paper, we introduce an alternative way of explaining the results of Rogers and Patterson, which is premised on a more restricted set of assumptions born from standard categorization theory. Specifically, we provide evidence that their results can be accounted for based on the predictions of the simplicity model of unsupervised categorization.     

Sentential synonymity judgment

Despite the acknowledged theoretical importance of sentential synonymity, there has been little interest in the psychological mechanisms responsible for recognizing it. Here an attempt is made to explore ways in which specific models originally devised for other particular psycholinguistic tasks (e.g., sentence-picture verification, three-term series problems, syllogistic reasoning) can be generalized to synonymity judgment and ways in which recent psychosemantic theories can be made more specific to deal with this same paradigm. Differential predictions from these models are derived for a set of spatiotemporal bidimensional comparative synonym pairs (e.g., High before Low, First above Second), and an experiment to test these predictions is reported. It is argued that, although the results are consonant with predictions from one type of model, they do not favor it definitively. This is because models of this type do not detail how bidimensional comparatives can be parsed into the same propositional representation and because other models can be subjected to further post hoc modifications to account for the data. Finally, general problems arising from attempts to adapt existing models to new tasks are discussed.     

Context processing in older adults: evidence for a theory relating cognitive control to neurobiology in healthy aging.

A theory of cognitive aging is presented in which healthy older adults are hypothesized to suffer from disturbances in the processing of context that impair cognitive control function across multiple domains, including attention, inhibition, and working memory. These cognitive disturbances are postulated to be directly related to age-related decline in the function of the dopamine (DA) system in the prefrontal cortex (PFC). A connectionist computational model is described that implements specific mechanisms for the role of DA and PFC in context processing. The behavioral predictions of the model were tested in a large sample of older (N = 81) and young (N = 175) adults performing variants of a simple cognitive control task that placed differential demands on context processing. Older adults exhibited both performance decrements and, counterintuitively, performance improvements that are in close agreement with model predictions.     

Metacognition in the classroom: The association between students’ exam predictions and their desired grades

Students are overconfident when making grade predictions, and worse, the lowest-performing students are generally the most overconfident. Because metacognitive accuracy is associated with academic performance, multiple studies have attempted to improve metacognitive accuracy with mixed results. However, these studies may be of limited use because we do not understand the types of information university students use to make performance predictions. The current studies examined the possibility that university students’ predictions are associated with their desires—the grade they want to receive. Studies 1–4 demonstrated that students’ desired grades were strongly associated with their grade predictions across different courses, universities, and measurement strategies. Study 4 also showed that, if warned about the previous results, students could reduce their reliance on their desired grades and improve the accuracy of their predictions relative to control. Together, results demonstrated that students’ exam predictions are associated with their desired grades.     

Crossmodal integration in the identification of consonant segments

Although speechreading can be facilitated by auditory or tactile supplements, the process that integrates cues across modalities is not well understood. This paper describes two “optimal processing” models for the types of integration that can be used in speechreading consonant segments and compares their predictions with those of the Fuzzy Logical Model of Perception (FLMP, Massaro, 1987). In “pre-labelling” integration, continuous sensory data is combined across modalities before response labels are assigned. In “post-labelling” integration, the responses that would be made under unimodal conditions are combined, and a joint response is derived from the pair. To describe pre-labelling integration, confusion matrices are characterized by a multidimensional decision model that allows performance to be described by a subject's sensitivity and bias in using continuous-valued cues. The cue space is characterized by the locations of stimulus and response centres. The distance between a pair of stimulus centres determines how well two stimuli can be distinguished in a given experiment. In the multimodal case, the cue space is assumed to be the product space of the cue spaces corresponding to the stimulation modes. Measurements of multimodal accuracy in five modern studies of consonant identification are more consistent with the predictions of the pre-labelling integration model than the FLMP or the post-labelling model.