How to build a brain: from function to implementation

To have a fully integrated understanding of neurobiological systems, we must address two fundamental questions: 1. What do brains do (what is their function)? and 2. How do brains do whatever it is that they do (how is that function implemented)? I begin by arguing that these questions are necessarily inter-related. Thus, addressing one without consideration of an answer to the other, as is often done, is a mistake. I then describe what I take to be the best available approach to addressing both questions. Specifically, to address 2, I adopt the Neural Engineering Framework (NEF) of Eliasmith &; Anderson [Neural engineering: Computation representation and dynamics in neurobiological systems. Cambridge, MA: MIT Press, 2003] which identifies implementational principles for neural models. To address 1, I suggest that adopting statistical modeling methods for perception and action will be functionally sufficient for capturing biological behavior. I show how these two answers will be mutually constraining, since the process of model selection for the statistical method in this approach can be informed by known anatomical and physiological properties of the brain, captured by the NEF. Similarly, the application of the NEF must be informed by functional hypotheses, captured by the statistical modeling approach.     

How to build a connectionist idiot (savant)

This paper describes the development of a connectionist model of an idiot savant date calculator using a multi-layer back-propagation network. The model mimics the behaviour of idiot savant date calculators in learning to perform the date calculation task without the need either to develop arithmetical skills or to encode explicit knowledge about regularities in the structure of dates. However, the performance of the model highlights limitations on the use of back-propagation as a general-purpose learning algorithm in complex problem domains. Even with a relatively simple problem like date calculation back-propagation can only succeed when forced to take a highly structured and modular approach to learning.     

The development of numerical estimation: evidence against a representational shift

How do our mental representations of number change over development? The dominant view holds that children (and adults) possess multiple representations of number, and that age and experience lead to a shift from greater reliance upon logarithmically organized number representations to greater reliance upon more accurate, linear representations. Here we present a new theoretically motivated and empirically supported account of the development of numerical estimation, based on the idea that number‐line estimation tasks entail judgments of proportion. We extend existing models of perceptual proportion judgment to the case of abstract numerical magnitude. Two experiments provide support for these models; three likely sources of developmental change in children’s estimation performance are identified and discussed. This work demonstrates that proportion‐judgment models provide a unified account of estimation patterns that have previously been explained in terms of a developmental shift from logarithmic to linear representations of number.     

Bridging the representational gap in the dynamic systems approach to development

We describe the relationship between the dynamic systems approach to development and a recent approach to the dynamics of representational states – the dynamic field approach. Both approaches share an emphasis on the concepts of stability (attractor states), instability (especially bifurcations), soft-assembly and flexibility. But the dynamic field approach adds the concept of ‘activation’ to capture the strength with which behaviorally relevant information is specified. By explicitly linking these dynamic systems approaches, we allow for more direct comparisons between dynamic systems theory and connectionism. We note three current differences between these two approaches to development: (1) the notion of stability is central to how representational states are conceptualized in the dynamic field approach; (2) the dynamic field approach is more directly concerned with the sensorimotor origins of cognition; and (3) the dynamic approach is less advanced with regard to learning. We conclude that proponents of the two approaches can learn from the respective strengths of each approach. We suspect these differences will largely disappear in the next 20 years.     

The development of representational insight: beyond the model/room paradigm

Two experiments were conducted to determine whether very young children have difficulty with dual representations when change in perspective has been eliminated, that is, the task no longer requires transfer between a traversable and surveyable array. In a preliminary experiment children 30 and 36 months of age attempted to transfer an action observed with one of four tools associated with a child-size workbench to a miniature replica of that workbench or vice versa. Children at both ages exhibited limited success on both transfer and memory trials; the demands of selecting a tool and transferring the appropriate imitative response were hypothesized to exceed their processing capacity. A second experiment using an object-retrieval procedure involving the two workbenches revealed successful transfer by 36-month-olds, but not 30-month-olds. The results confirm that limitations in dual representation extend to contexts other than the room and model of the room paradigm where spatial perspective differs. The results are discussed in terms of procedures that support or interfere with children’s ability to demonstrate symbolic insight.     

Police responses to diversity: A social representational study of rural British policing in a changing representational context

In the first decade of the 21st century, British policing faced two new challenges in how it responded to social diversity: As well as instituting reforms in response to a highly publicized report describing the British police as ‘institutionally racist’ (Macpherson, 1999 ), they faced challenges associated with rapid increases in numbers of immigrants into the UK. Studying social representations at such times of change allows access into processes, themes and value systems that may otherwise remain hidden. This paper uses social representations theory to explore interview accounts provided by regular police officers of interactions with members of minority groups. Empirically, we focus on an area of diversity policing that has received relatively little previous attention: Police work in a rural context that has recently played host to large numbers of migrant workers. Our analysis shows that interviews operate as a site of resistance in which respondents attempt to rework hegemonic representations of the police as prejudiced and to re‐present themselves and their work as able to respond appropriately to diversity. Fairness as a defining characteristic of good police practice is a central representational theme that links identity construction processes to police work with minority groups. Recent immigrant groups are represented as both needing and deserving help to assimilate into British culture: Their lack of acculturation and language difficulties provide additional challenges to police work and to how this is judged as ‘fair’. We discuss the broader origins and implications of police officers' understandings of fairness and the use of social representations theory to study representational fields within organizational settings. Copyright © 2010 John Wiley & Sons, Ltd.     

Young children's use of scale models: testing an alternative to representational insight

Using a symbolic object such as a model as a source of information about something else requires some appreciation of the relation between the symbol and what it represents. Representational insight has been proposed as essential to success in a symbolic retrieval task in which children must use information from a hiding event in a scale model to find a toy hidden in a room. The two studies reported here examine and reject a proposed alternative account for success in the model task. The results with 2.5-year-olds and 3-year-olds show that children's successful use of a scale model cannot be attributed to the simple detection of the correspondences between the objects in the two spaces. A higher-level representation of the model-room relation (i.e. representational insight) is required. The results are discussed with respect to the coalescence of multiple factors in determining performance in the model task.     

On the representational systems underlying prospection: Evidence from the event-cueing paradigm

The ability to think about the future-prospection-is central to many aspects of human cognition and behavior, from planning and decision making, to self-control and the construction of a sense of identity. Yet, the exact nature of the representational systems underlying prospection is not fully understood. Recent findings point to the critical role of episodic memory in imagining specific future events, but it is unlikely that prospection depends solely on this system. Using an event-cueing paradigm in two studies, we here show that specific events that people imagine might happen in their personal future are commonly embedded in broader event sequences-termed event clusters-that link a set of envisioned events according to causal and thematic relations. These findings provide novel evidence that prospection relies on multiple representational systems, with general autobiographical knowledge structures providing a frame that organizes imagined events in overarching event sequences. The results further suggest that knowledge about personal goals plays an important role in structuring these event sequences, especially for the distant future.     

Internal representational models of peers: implications for the development of problematic behavior.

The authors investigated the relation between children's knowledge structures for peers and externalizing behavior problems. Initial levels of aggression were evaluated in 135 boys and 124 girls (Grades 1-3; 40% African American, 60% Caucasian) in Year 1 and again in Years 6 and 9. In Year 6, 3 aspects of their social knowledge structures were assessed: quality, density, and appropriateness. Results indicate that knowledge structures are related to children's concurrent levels of externalizing behaviors and that knowledge structures are related to children's concurrent levels of externalizing behaviors and predict externalizing behaviors 3 years later even after controlling for current levels of behavior. In addition, knowledge structures in Year 6 mediate the relation between aggression in Year 1 and externalizing behaviors in Year 9. The role of knowledge structures in the maintenance and growth of children's antisocial behavior is discussed.     

Appraising the threat of illness: a mental representational approach

This article (a) describes a mental model underlying initial evaluations of illness signs and (b) reports an experiment demonstrating the model's utility by showing how the model represents evidence of defensiveness among people who test positively for a sign of illness. The model consists of a set of cognitive elements that people consider to evaluate the threat represented by a sign of possible illness. Seventy-two undergraduates were led to believe that they tested positively or negatively on a saliva test for a fictitious risk factor for a disease. In addition, half the participants were told about the existence of a simple preventive treatment for the disease, whereas the others were not. Subsequently the participants answered questions about elements of the threat-appraisal model. Analysis of their responses reveals evidence of defensiveness on several elements of the model. Those testing positively for the risk factor, especially those uninformed about its treatment, minimized threat by (a) increasing their estimates of the false-positive rate of the test, (b) decreasing their estimates of the seriousness of the risk factor, and (c) decreasing their estimates of the extent to which the disease itself is life-threatening. Applications of the model to actual illness threats and the relation between threat-related judgments and health-related behavior are discussed.     

The representational status of pretence: Evidence from typical development and autism

The question of whether understanding pretend play requires meta‐representational skill was examined among typically developing children and individuals with autism. Participants were presented with closely equated true and false pretence trials in which they had to judge a protagonist's pretend reading of a situation, which either matched or differed from their own. Results showed that individuals' theory of mind abilities determined their performance on false, but not true, pretence trials. These findings imply that meta‐representation is involved when an individual has to make sense of a pretend state of mind that differs from their own, but, crucially, they also show that pretend play can often be understood without meta‐representational competence.     

How to manage the evolution of a computerized information system

Existing computerized fiscal and patient information systems do not seem well suited for the mental health field. To design its own system, an institution would do better to form an information system staff rather than to use an outside software house, because doing it externally is less responsive, slower, and, for some tasks, simply impossible. The in-house staff must be selected to be competent generalists, at ease with computer technology. Working in close and friendly cooperation with various user groups, the systems evolve out of small beginnings. As the users gain working experience with early versions of a system, they are able to refine their goals, with the result that the final system meets their needs well, even though it may seem to be different from what was originally expected.     

On How to Build a Moral Machine

Herein we make a plea to machine ethicists for the inclusion of constraints on their theories consistent with empirical data on human moral cognition. As philosophers, we clearly lack widely accepted solutions to issues regarding the existence of free will, the nature of persons and firm conditions on moral agency/patienthood; all of which are indispensable concepts to be deployed by any machine able to make moral judgments. No agreement seems forthcoming on these matters, and we don’t hold out hope for machines that can both always do the right thing (on some general ethic) and produce explanations for its behavior that would be understandable to a human confederate. Our tentative solution involves understanding the folk concepts associated with our moral intuitions regarding these matters, and how they might be dependent upon the nature of human cognitive architecture. It is in this spirit that we begin to explore the complexities inherent in human moral judgment via computational theories of the human cognitive architecture, rather than under the extreme constraints imposed by rational-actor models assumed throughout much of the literature on philosophical ethics. After discussing the various advantages and challenges of taking this particular perspective on the development of artificial moral agents, we computationally explore a case study of human intuitions about the self and causal responsibility. We hypothesize that a significant portion of the variance in reported intuitions for this case might be explained by appeal to an interplay between the human ability to mindread and to the way that knowledge is organized conceptually in the cognitive system. In the present paper, we build on a pre-existing computational model of mindreading (Bello et al. 2007) by adding constraints related to psychological distance (Trope and Liberman 2010), a well-established psychological theory of conceptual organization. Our initial results suggest that studies of folk concepts involved in moral intuitions lead us to an enriched understanding of cognitive architecture and a more systematic method for interpreting the data generated by such studies.     

Theories of reasoning and the representational level: A reply to Oaksford

Oaksford (2001) considers that the findings of Espino, Santamaría, and García Madruga (2000a) could be explained by the Probability Heuristics Model (PHM) proposed by Chater and Oaksford (1999). He specifically voices three objections, the two main ones being based on the fact that PHM is not a theory about syllogism representation. If this is the case, we consider that PHM cannot explain our data, because most of them were registered before the participants evaluated the conclusion. We argue that only a theory at the representational level can property explain these data.     

On the costs of accessible attitudes: detecting that the attitude object has changed

The present research examined whether individuals with more accessible attitudes have more difficulty detecting that the attitude object has changed. While being repeatedly exposed to photographs of undergraduates, participants either rehearsed their attitudes toward each photo or performed a control task. They then saw these original photos and computer-generated morphs representing varying degrees of change in an original. Participants in the attitude rehearsal condition required more time to correctly identify morphs that were similar to the original as "different" (Experiment 1) and made more errors in response to such morphs (Experiment 2). Experiment 3 revealed that participants with accessible attitudes perceived relatively less change; they were less likely to view a morph as a photo of a novel person and more likely to view it as a different photo of a person seen before. The costs and benefits of accessible attitudes are discussed.