Should physicians be Bayesian agents?

Because physicians use scientific inference for the generalizations of individual observations and the application of general knowledge to particular situations, the Bayesian probability solution to the problem of induction has been proposed and frequently utilized. Several problems with the Bayesian approach are introduced and discussed. These include: subjectivity, the favoring of a weak hypothesis, the problem of the false hypothesis, the old evidence/new theory problem and the observation that physicians are not currently Bayesians. To the complaint that the prior probability is subjective, Bayesians reply that there will be ultimate convergence, but the rebuttal to this is that there will not be uniform convergence. Secondly, since the Bayesian scheme favors a weak hypothesis, theories turn out to be a gratuitous risk. The problem with the false hypothesis comes out in the denominator of the theorem, revealing that a factor which is not a theory at all is being considered in the reasoning. On the old evidence/new theory problem old evidence cannot confirm a new theory so that the posterior probability will equal the prior probability. Finally, empiric studies have shown that current physicians are not Bayesians. But on consideration of Bayesian inference as a system of inference, it can be reasoned that physicians should be Bayesians. However, the problem of physicians' and patients' own subjectivity continue to plague this system of medical decision making.     

Visual variability affects early verb learning

Research demonstrates that within‐category visual variability facilitates noun learning; however, the effect of visual variability on verb learning is unknown. We habituated 24‐month‐old children to a novel verb paired with an animated star‐shaped actor. Across multiple trials, children saw either a single action from an action category (identical actions condition, for example, travelling while repeatedly changing into a circle shape) or multiple actions from that action category (variable actions condition, for example, travelling while changing into a circle shape, then a square shape, then a triangle shape). Four test trials followed habituation. One paired the habituated verb with a new action from the habituated category (e.g., ‘dacking’ + pentagon shape) and one with a completely novel action (e.g., ‘dacking’ + leg movement). The others paired a new verb with a new same‐category action (e.g., ‘keefing’ + pentagon shape), or a completely novel category action (e.g., ‘keefing’ + leg movement). Although all children discriminated novel verb/action pairs, children in the identical actions condition discriminated trials that included the completely novel verb, while children in the variable actions condition discriminated the out‐of‐category action. These data suggest that – as in noun learning – visual variability affects verb learning and children's ability to form action categories.     

Transition structure versus commitment in sequential subjective probability revision

When Ss revise subjective probabilities, in the light of new evidence, a common finding is that they are conservative with respect to Bayes' theorem; revisions are too small. One kind of hypothesis to account for this is ‘model specific’, assuming a breakdown in an otherwise potentially Bayesian process. The other kind assumes that statistically irrelevant, task-specific information is processed. An example of the latter is the commitment hypothesis, assuming a commitment building up to the indications of early evidence, causing Ss to lag behind Bayes' theorem in their later judgements. Evidence is presented suggesting that Ss are not necessarily sensitive to mere sub-sets of a sequence, but that this form of suboptimality may result from overall sequence structure; specifically from a bias against long runs of like evidence. This would fit with findings from other areas of research, and would suggest that there is a general form of suboptimality operating which is relevant to all sequential processing tasks.     

Little Bayesians or little Einsteins? Probability and explanatory virtue in children's inferences

Like scientists, children seek ways to explain causal systems in the world. But are children scientists in the strict Bayesian tradition of maximizing posterior probability? Or do they attend to other explanatory considerations, as laypeople and scientists – such as Einstein – do? Four experiments support the latter possibility. In particular, we demonstrate in four experiments that 4‐ to 8‐year‐old children, like adults, have a robust latent scope bias that leads to inferences that do not maximize posterior probability. When faced with two explanations equally consistent with observed data, where one explanation makes an unverified prediction, children consistently preferred the explanation that does not make this prediction (Experiment 1), even if the prior probabilities are identical (Experiment 3). Additional evidence suggests that this latent scope bias may result from the same explanatory strategies used by adults (Experiments 1 and 2), and can be attenuated by strong prior odds (Experiment 4). We argue that children, like adults, rely on ‘explanatory virtues’ in inference – a strategy that often leads to normative responses, but can also lead to systematic error.     

Categories and causality: the neglected direction

The standard approach guiding research on the relationship between categories and causality views categories as reflecting causal relations in the world. We provide evidence that the opposite direction also holds: categories that have been acquired in previous learning contexts may influence subsequent causal learning. In three experiments we show that identical causal learning input yields different attributions of causal capacity depending on the pre-existing categories to which the learning exemplars are assigned. There is a strong tendency to continue to use old conceptual schemes rather than switch to new ones even when the old categories are not optimal for predicting the new effect, and when they were motivated by goals that differed from the present context of causal discovery. However, we also found that the use of prior categories is dependent on the match between categories and causal effect. Whenever the category labels suggest natural kinds which can be plausibly related to the causal effects, transfer was observed. When the categories were arbitrary, or could not be plausibly related to the causal effect learners abandoned the categories, and used different categories to predict the causal effect.     

DSM-IV criteria for borderline personality disorder: A critical evaluation

The present article critically examines the diagnostic criteria for borderline personality disorder (BPD) recently proposed by the DSM-IV Axis II work group (Gunderson, J. G., et al., Journal of Personality Disorders, 5, 340–352, 1991). It is concluded that some of the revisions are trivial and there is insufficient justification for others. The revisions are likely to increase the overlap between BPD and other disorders and represent an attempt to retain a categorical classification in the face of evidence that a dimensional model is more appropriate. Evidence suggests that many borderline traits are manifestations of extreme neuroticism.     

Learning New Letter-like Writing Patterns Explicitly and Implicitly in Children and Adults

A handwriting task was used to test the assumption that explicit learning is dependent on age and working memory, while implicit learning is not. The effect of age was examined by testing both, typically developing children (5–12 years old, n = 81) and adults (n = 27) in a counterbalanced within-subjects design. Participants were asked to repeatedly write letter-like patterns on a digitizer with a non-inking pen. Reproduction of the pattern was better after explicit learning compared to implicit learning. Age had positive effects on both explicit and implicit learning; working memory did not affect learning in either conditions. These results show that it may be more effective to learn writing new letter-like patterns explicitly and that an explicit teaching method is preferred in mainstream primary education.     

Children learn words easier when they are interested in the category to which the word belongs

The overall pattern of vocabulary development is relatively similar across children learning different languages. However, there are considerable differences in the words known to individual children. Historically, this variability has been explained in terms of differences in the input. Here, we examine the alternate possibility that children's individual interest in specific natural categories shapes the words they are likely to learn – a child who is more interested in animals will learn a new animal name easier relative to a new vehicle name. Two‐year‐old German‐learning children (N = 39) were exposed to four novel word–object associations for objects from four different categories. Prior to the word learning task, we measured their interest in the categories that the objects belonged to. Our measure was pupillary change following exposure to familiar objects from these four categories, with increased pupillary change interpreted as increased interest in that category. Children showed more robust learning of word–object associations from categories they were more interested in relative to categories they were less interested in. We further found that interest in the novel objects themselves influenced learning, with distinct influences of both category interest and object interest on learning. These results suggest that children's interest in different natural categories shapes their word learning. This provides evidence for the strikingly intuitive possibility that a child who is more interested in animals will learn novel animal names easier than a child who is more interested in vehicles.     

Prioritizing selection of new elements: bottom-up versus top-down control

Watson and Humphreys (1997) have proposed that prioritized selection of new over old elements occurs because observers can apply top-down inhibition to the locations of the old elements by a mechanism they refer to as visual marking. However, recent evidence has suggested that the top-down mechanism is questionable (Donk & Theeuwes, 2001). In the present study, we investigated whether prioritized selection of new over old elements occurs in a bottom-up or a top-down fashion. Observers were presented with displays containing one set of elements (old elements) followed, after a certain time interval, by a second set of elements (new elements). The observers were instructed to search for the presence of a target that was presented with equal probability among the old and the new elements (Experiments 1 and 2) or twice as often among the old elements than among the new elements (Experiment 3). The results show that new elements were prioritized for selection over old ones even though the observers had no incentive to do so. The results suggest that prioritized selection of new over old elements is not mediated by a top-down inhibition process, as was proposed by Watson and Humphreys (1997). Instead, prioritization of new elements appears to be a bottom-up process. The implications of these results are discussed in terms of models of attentional control.     

New Foundations for a Relational Theory of Theory-revision

AGM-theory, named after its founders Carlos Alchourrón, Peter Gärdenfors and David Makinson, is the leading contemporary paradigm in the theory of belief-revision. The theory is reformulated here so as to deal with the central relational notions ‘J is a contraction of K with respect to A’ and ‘J is a revision of K with respect to A’. The new theory is based on a principal-case analysis of the domains of definition of the three main kinds of theory-change (expansion, contraction and revision). The new theory is stated by means of introduction and elimination rules for the relational notions. In this new setting one can re-examine the relationship between contraction and revision, using the appropriate versions of the so-called Levi and Harper identities. Among the positive results are the following. One can derive the extensionality of contraction and revision, rather than merely postulating it. Moreover, one can demonstrate the existence of revision-functions satisfying a principle of monotonicity. The full set of AGM-postulates for revision-functions allow for completely bizarre revisions. This motivates a Principle of Minimal Bloating, which needs to be stated as a separate postulate for revision. Moreover, contractions obtained in the usual way from the bizarre revisions, by using the Harper identity, satisfy Recovery. This provides a new reason (in addition to several others already adduced in the literature) for thinking that the contraction postulate of Recovery fails to capture the Principle of Minimal Mutilation. So the search is still on for a proper explication of the notion of minimal mutilation, to do service in both the theory of contraction and the theory of revision. The new relational formulation of AGM-theory, based on principal-case analysis, shares with the original, functional form of AGM-theory the idealizing assumption that the belief-sets of rational agents are to be modelled as consistent, logically closed sets of sentences. The upshot of the results presented here is that the new relational theory does a better job of making important matters clear than does the original functional theory. A new setting has been provided within which one can profitably address two pressing questions for AGM-theory: (1) how is the notion of minimal mutilation (by both contractions and revisions) best analyzed? and (2) how is one to rule out unnecessary bloating by revisions?     

Motivated reasoning and verbal vs. numerical probability assessment: Evidence from an accounting context

A perplexing yet persistent empirical finding is that individuals assess probabilities in words and in numbers nearly equivalently, and theorists have called for future research to search for factors that cause differences. This study uses an accounting context in which individuals are commonly motivated to reach preferred (rather than accurate) conclusions. Within this context, I predict new differences between verbal and numerical probability assessments, as follows: first, individuals will justify an optimistic verbal assessment (e.g., somewhat possible) by retaining the option of re-defining it, in case of negative outcomes, as though the phrase really means something different, and, for that matter, means more things. This re-definition will maintain some connection to the original meaning of the phrase, but de-emphasized relative to the new meaning. Second, based on this behavior, I also predict individuals’ verbal probability assessments to be (1) more biased and yet (2) perceived as more justifiable than their numerical assessments. I find supportive evidence in an experiment designed to test the hypotheses. This study contributes to motivated reasoning and probability assessment theories (1) with new evidence of how individuals can word-smith in multiple attributes of a phrase to justify reaching a preferred conclusion, and (2) with new, reliable differences between verbal and numerical probability assessments. This study has important theoretical and practical implications relevant to organizational contexts in which people assess the likelihoods of uncertainties in words or numbers, and with motivations to reach a preferred conclusion.     

A little labeling goes a long way: Semi‐supervised learning in infancy

There is considerable evidence that labeling supports infants' object categorization. Yet in daily life, most of the category exemplars that infants encounter will remain unlabeled. Inspired by recent evidence from machine learning, we propose that infants successfully exploit this sparsely labeled input through “semi‐supervised learning.” Providing only a few labeled exemplars leads infants to initiate the process of categorization, after which they can integrate all subsequent exemplars, labeled or unlabeled, into their evolving category representations. Using a classic novelty preference task, we introduced 2‐year‐old infants (n = 96) to a novel object category, varying whether and when its exemplars were labeled. Infants were equally successful whether all exemplars were labeled (fully supervised condition) or only the first two exemplars were labeled (semi‐supervised condition), but they failed when no exemplars were labeled (unsupervised condition). Furthermore, the timing of the labeling mattered: when the labeled exemplars were provided at the end, rather than the beginning, of familiarization (reversed semi‐supervised condition), infants failed to learn the category. This provides the first evidence of semi‐supervised learning in infancy, revealing that infants excel at learning from exactly the kind of input that they typically receive in acquiring real‐world categories and their names.     

Acquisition of habitual visual attention and transfer to related tasks

Extensive research has shown that statistical learning affects perception, attention, and action control; however, few studies have directly linked statistical learning with the formation of habits. Evidence that learning can induce a search habit has come from location probability learning, in which people prioritize locations frequently attended to in the past. Here, using an alternating training–testing procedure, we demonstrated that the initial attentional bias arises from short-term intertrial priming, whereas probability learning takes longer to emerge, first reaching significance in covert orienting (measured by reaction times) after about 48 training trials, and in overt orienting (measured by eye movements) after about 96 training trials. We further showed that location probability learning is persistent after training is discontinued, by transferring from a letter search task to a scene search task—emulating another characteristic feature of habits. By identifying the onset of probability learning and investigating its task specificity, this study provides evidence that probability cuing can induce habitual spatial attention.     

Seeing is believing: Priors,trust, and base rate neglect

Kahneman and Tversky (1973) described an effect they called ‘insensitivity to prior probability of outcomes’, later dubbed base rate neglect, which describes people’s tendency to underweight prior information in favor of new data. As probability theory requires that prior probabilities be taken into account, via Bayes’ theorem, the fact that most people fail to do so has been taken as evidence of human irrationality and, by others, of a mismatch between our cognitive processes and the questions being asked (Cosmides & Tooby, 1996). In contrast to both views, we suggest that simplistic Bayesian updating using base rates is not necessarily rational. To that end, we present experiments in which base rate neglect is often the right strategy, and show that people’s base rate usage varies systematically as a function of the extent to which the data that make up a base rate are perceived as trustworthy.     

Spatial probability learning by alcoholic Korsakoff patients.

Performance of 10 alcoholic Korsakoff patients was compared to that of 10 normal and 10 alcoholic control subjects on each of three different schedules of spatial probability learning (50:50, 70:30, and 30:70) using monetary reinforcement and a correction procedure. There was some evidence that the Korsakoff patients were less sensitive than normals to the effects of reward: Although on all three schedules, choice ratios by normal subjects approximated the reinforcement ratios, the choice ratios of Korsakoffs on the second and third schedules remained close to the reinforcement ratio acquired with the first schedule. In addition, the Korsakoffs made an abnormal number of perseverative errors early in training. On most measures, performance by alcoholic controls fell between that of the other two groups.     

Sensory noise explains auditory frequency discrimination learning induced by training with identical stimuli

Thresholds in various visual and auditory perception tasks have been found to improve markedly with practice at intermediate levels of task difficulty. Recently, however, there have been reports that training with identical stimuli, which, by definition, were impossible to discriminate correctly beyond chance, could induce as much discrimination learning as could training with different stimuli. These surprising findings have been interpreted as evidence that discrimination learning can occur in the absence of perceived differences between stimuli and need not involve the fine-tuning of a discrimination mechanism. Here, we show that these counterintuitive findings of discrimination learning without discrimination can be understood simply by considering the effect of internal noise on sensory representations. Because of such noise, physically identical stimuli are unlikely to be perceived as being strictly identical. We show that, given empirically derived levels of sensory noise, perceived differences evoked by identical stimuli are actually not much smaller than those induced by the physical differences typically used in discrimination-learning experiments. We suggest that findings of discrimination learning with identical stimuli can be explained without implicating any fundamentally new learning mechanism.     

Causal learning from probabilistic events in 24‐month‐olds: an action measure

How do young children learn about causal structure in an uncertain and variable world? We tested whether they can use observed probabilistic information to solve causal learning problems. In two experiments, 24‐month‐olds observed an adult produce a probabilistic pattern of causal evidence. The toddlers then were given an opportunity to design their own intervention. In Experiment 1, toddlers saw one object bring about an effect with a higher probability than a second object. In Experiment 2, the frequency of the effect was held constant, though its probability differed. After observing the probabilistic evidence, toddlers in both experiments chose to act on the object that was more likely to produce the effect. The results demonstrate that toddlers can learn about cause and effect without trial‐and‐error or linguistic instruction on the task, simply by observing the probabilistic patterns of evidence resulting from the imperfect actions of other social agents. Such observational causal learning from probabilistic displays supports human children's rapid cultural learning.     

Adaptive spatial working memory assessments for aging pet dogs

Assessments for spatial working memory (SWM) in pet dogs that can detect age-related cognitive deficits in a single session may aid in diagnosing canine dementia and may facilitate translational research on Alzheimer’s disease in humans. Adaptive testing procedures are widely used in single-session assessments for humans with diverse cognitive abilities. In this study, we designed and deployed two up-down staircase assessments for SWM in which 26 pet dogs were required to recall the location of a treat hidden behind one of two identical boxes following delays of variable length. In the first experiment, performance tended to decline with age but few dogs completed the test (n?=?10). However, all of the dogs that participated in the second experiment (n?=?24) completed the assessment and provided reliable evidence of learning and retaining the task. Delay length and age significantly predicted performance supporting the validity of this assessment. The relationships between age and performance were described by inverted U-shaped functions as both old and young dogs displayed deficits in weighted cumulative-scores and trial-by-trial performance. Thus, SWM in pet dogs may develop until midlife and decline thereafter. Exploratory analyses of non-mnemonic fixation strategies, sustained engagement, inhibitory control, and potential improvements for future SWM assessments which adopt this paradigm are also discussed.

     

The development of route learning in Down syndrome,Williams syndrome and typical development: investigations with virtual environments

The ability to navigate new environments has a significant impact on the daily life and independence of people with learning difficulties. The aims of this study were to investigate the development of route learning in Down syndrome (N = 50), Williams syndrome (N = 19), and typically developing children between 5 and 11 years old (N = 108); to investigate use of landmarks; and to relate cognitive functions to route‐learning ability in these groups. Overall, measures of attention and long‐term memory were strongly associated with route learning, even once non‐verbal ability was controlled for. All of the groups, including 5‐ to 6‐year‐old TD children, demonstrated the ability to make use of all landmark types to aid route learning; those near junctions, those further from junctions, and also distant landmarks (e.g. church spire, radio mast). Individuals with WS performed better than a matched subset of TD children on more difficult routes; we suggest that this is supported by relatively strong visual feature recognition in the disorder. Participants with DS who had relatively high levels of non‐verbal ability performed at a similar level to TD participants.     

Connectionism and the learning of probabilistic concepts

Three experiments examine the claim of Gluck and Bower (1986, 1988a, 1988b) that the learning of medical concepts can be simulated by a connectionist network in which the symptoms are the input and the diagnosis is the output. The first experiment replicates the main finding of Gluck and Bower. In this experiment, subjects were required to estimate the probability of each of two diseases, given a particular target symptom. In fact these two probabilities were identical, but because one illness was more common than the other, the target symptom was a better predictor of the rare disease than of the common disease. Contrary to a normative probability judgement account, subjects were biased in that they judged the probability of the rare disease given the target symptom to be greater than the probability of the common disease given the target symptom. Gluck and Bower argued that such a result was predicted by a connectionist network using the Rescorla-Wagner learning rule, but it is argued that Gluck and Bower's network simulation was not appropriate for the experiment they had performed. In fact, it appeared that the connectionist network failed to predict the bias in the subjects' probability estimates.

However, this conclusion rests on an assumption that Gluck and Bower implicitly made. They arranged for P(rare disease/target symptom) and P(com-mon disease/target symptom) to be identical across all trials on which the target symptom occurred, both on its own and with other symptoms present. Gluck and Bower assumed that the subjects were estimating these probabilities. But the results of the second experiment showed instead that the subjects were estimating the probability of each disease given only the target symptom. In the final experiment the design was changed so that this problem might be circumvented. In this experiment, again, the subjects were biased in their probability judgements exactly as the connectionist network predicted. Thus, finally, evidence was found which was compatible with the network model but not with a normative account, but this was true only if the network did not include a layer of hidden units.