Parameterization of connectionist models

We present a method for estimating parameters of connectionist models that allows the model’s output to fit as closely as possible to empirical data. The method minimizes a cost function that measures the difference between statistics computed from the model’s output and statistics computed from the subjects’ performance. An optimization algorithm finds the values of the parameters that minimize the value of this cost function. The cost function also indicates whether the model’s statistics are significantly different from the data’s. In some cases, the method can find the optimal parameters automatically. In others, the method may facilitate the manual search for optimal parameters. The method has been implemented in Matlab, is fully documented, and is available for free download from the Psychonomic Society Web archive atwww.psychonomic.org/archive/.     

Parameterization of connectionist models.

We present a method for estimating parameters of connectionist models that allows the model's output to fit as closely as possible to empirical data. The method minimizes a cost function that measures the difference between statistics computed from the model's output and statistics computed from the subjects' performance. An optimization algorithm finds the values of the parameters that minimize the value of this cost function. The cost function also indicates whether the model's statistics are significantly different from the data's. In some cases, the method can find the optimal parameters automatically. In others, the method may facilitate the manual search for optimal parameters. The method has been implemented in Matlab, is fully documented, and is available for free download from the Psychonomic Society Web archive at www.psychonomic.org/archive/.     

Internal working models of caregiving and security of attachment at age six

This study describes the relation between internal working models of caregiving, child attachment, and maternal behavior in the home. Thirty-two mothers of 6-year-old children were observed in the home and subsequently interviewed regarding experiential and affective dimensions of parenting. Interviews were examined in order to assess the quality of the mother's thinking regarding two dimensions of caregiving (secure base, competence) which we hypothesized to be related to attachment security. Results indicated a strong correspondence between internal working models of caregiving and child mental representations of attachments as measured from the child's response to a laboratory reunion. The correspondence between mental representations of caregiving and maternal behavior in the home was limited. Representation ratings were most strongly associated with competence-supporting behavior. Implications for infant mental health research and program evaluation are discussed.     

Using connectionist models to guide assessment of psychological disorder

Connectionist models are introduced as promising tools for understanding the nature of psychological disorders and guiding their assessment. Specifically, ways in which connectionist models can guide the following aspects of the assessment process are described: understanding what constructs are relevant to assess, designing approaches to assessing these constructs, and understanding individual differences in data from assessments. Two extended examples are given based on the authors' research on cognitive aspects of depression and Alzheimer's disease.     

Client attachment and the working alliance

The relationship between four dimensions of client attachment (secure, preoccupied, fearful, and dismissing) and client ratings of four dimensions of the working alliance (goals, tasks, bonds, and global) was examined in a sample of first-time clients (n = 63) seeking services through a university-based counselling clinic. Significant relationships were found between secure and fearful attachment and the bond component of the Working Alliance Inventory (WAI). Secure attachment was also significantly associated with the WAI goals subscale as well as global WAI ratings. Implications of these findings for counselling practice and future research are discussed.     

A threshold approach to understanding the origins of attachment disorganization

Disorganized attachment in infancy is known to predict a wide range of maladaptive outcomes, but its origins are poorly understood. Parental lack of resolution concerning loss or trauma has been proposed to result in atypical parenting behaviors, which in turn have a disorganizing effect on the parent-child relationship. The authors review the evidence for this transmission pathway, considering other factors (e.g., social environment, child characteristics) that might enrich understanding of the antecedents of disorganization. A threshold approach is proposed to explain (a) why different parental behaviors are linked to disorganization depending on prevailing social conditions and (b) why certain children appear more vulnerable to forming a disorganized attachment relationship.     

Classical and connectionist models: Levels of description

To begin, I introduce an analysis of interlevel relations that allows us to offer an initial characterization of the debate about the way classical and connectionist models relate. Subsequently, I examine a compatibility thesis and a conditional claim on this issue. With respect to the compatibility thesis, I argue that, even if classical and connectionist models are not necessarily incompatible, the emergence of the latter seems to undermine the best arguments for the Language of Thought Hypothesis, which is essential to the former. I attack the conditional claim of connectionism to eliminativism, presented by Ramsey et al. (1990), by discrediting their discrete characterization of common-sense psychological explanations and pointing to the presence of a moderate holistic constraint. Finally, I conclude that neither of the arguments considered excludes the possibility of viewing connectionist models as forming a part of a representational theory of cognition that dispenses with the Language of Thought Hypothesis.     

OXlearn: A new MATLAB-based simulation tool for connectionist models

OXlearn is a free, platform-independent MATLAB toolbox in which standard connectionist neural network models can be set up, run, and analyzed by means of a user-friendly graphical interface. Due to its seamless integration with the MATLAB programming environment, the inner workings of the simulation tool can be easily inspected and/or extended using native MATLAB commands or components. This combination of usability, transparency, and extendability makes OXlearn an efficient tool for the implementation of basic research projects or the prototyping of more complex research endeavors, as well as for teaching. Both the MATLAB toolbox and a compiled version that does not require access to MATLAB can be downloaded from http://psych.brookes.ac.uk/oxlearn/.     

The continuity principle: A unified approach to disaster and trauma

The continuity principle stipulates that through all stages of disaster, management and treatment should aim at preserving and restoring functional, historical, and interpersonal continuities, at the individual, family, organization, and community levels. Two misconceptions work against this principle and lead to decisional errors: the “abnormalcy bias” which results in underestimating victims' ability to cope with disaster, and the “normalcy bias” which results in underestimating the probability or extent of expected disruption. This article clarifies these biases and details the potential contributions of the continuity principle at the different stages of the disaster. The authors are reserve officers in the Mental Health and Behavioral Sciences Centers of the IDF (Israel Defense Forces) and were, in this capacity, directly involved in the coordination of agencies dealing with Gulf War victims.     

At the intersection of social and cognitive development: internal working models of attachment in infancy

Three visual habituation studies using abstract animations tested the claim that infants' attachment behavior in the Strange Situation procedure corresponds to their expectations about caregiver-infant interactions. Three unique patterns of expectations were revealed. Securely attached infants expected infants to seek comfort from caregivers and expected caregivers to provide comfort. Insecure-resistant infants not only expected infants to seek comfort from caregivers but also expected caregivers to withhold comfort. Insecure-avoidant infants expected infants to avoid seeking comfort from caregivers and expected caregivers to withhold comfort. These data support Bowlby's (1958) original claims-that infants form internal working models of attachment that are expressed in infants' own behavior.     

Patterns of stability in adult attachment: an empirical test of two models of continuity and change

One of the core assumptions of attachment theory is that attachment representations are stable over time. Unfortunately, the data on attachment stability have been ambiguous, and as a result, alternative theoretical perspectives have evolved to explain them. The objective of the present research was to evaluate alternative models of stability by studying adults in 2 intensive longitudinal investigations. Specifically, we assessed attachment representations in 1 sample (N = 203) daily over a 30-day period and in the other sample (N = 388) weekly over a year. Analyses show that the patterns of stability that exist in adult attachment are most consistent with a prototype model--a model assuming that there is a stable factor underlying temporary variations in attachment. Moreover, although the Big Five personality traits exhibited a pattern of stability that was similar to that of attachment, they did not account for the stability observed in attachment.     

The association between working models of attachment and personality: Toward an integrative framework operationalizing global relational models

A theoretical framework is presented which outlines the association between relational models of romantic relationships (romantic attachment anxiety and avoidance) and the conceptually similar dual dimensions of autonomy and sociotropy identified in the literature on personality diatheses of depression. Two studies, along with a meta-analysis of previous research, lend support to the notion that autonomy and avoidance share a common element reflecting variation in the degree to which people seek out or avoid closeness with others, whereas sociotropy and anxiety share a common element reflecting variation in the degree to which people worry about or are content in their social relations. The measures differ in relational specificity, however, as attachment anxiety and avoidance assess models of close and romantic relationships, whereas sociotropy and autonomy reflect more abstracted representations that encompass broader categories of social relation. The proposed framework offers a variety of avenues for research operationalizing and elaborating upon the function of global relational models in the attachment system.     

Positive feedback in hierarchical connectionist models: Applications to language production

Recent connectionist models of the perception and production of words make use of positive feedback from later to earlier levels of processing. This paper focuses on production and identifies several specific effects of phoneme-tomorpheme feedback. In addition, I argue that there is support for the use of this kind of feedback in production from experimental and naturalistic studies of slips of the tongue.     

The relations between maternal behaviors and urban preschool children's internal working models of attachment security

Interviewed 93 African-American, low-income women who had become pregnant as teenagers and their preschool-aged children in their homes. Mothers answered questions regarding their everyday stresses and feelings of depression. Children were assessed for receptive vocabulary ability, then video-taped completing five stories thematically related to attachment experiences with mother and rated on their security of attachment. Mothers and children were also videotaped playing together, and mothers were assessed on their sensitivity to their children's cues. After controlling for children's age and receptive vocabulary ability, mothers' sensitivity significantly predicted children's level of attachment security. The positive association between maternal sensitivity and children's security of attachment, and the strengths and weaknesses of administering the Attachment Story-Completion Task in the home with this population, are discussed. Implications for assessing attachment in the home are considered.     

A multisource approach to self-critical vulnerability to depression: the moderating role of attachment

The present study investigated the effects of self-criticism, dependency, and attachment variables in depression among couples. We utilized a multisource design that involved self-reports and spouse reports of personality and depression. This approach enabled us to explore the patterns of relations between self-reported and the spouse's report of the partner's view of self-criticism, dependency, and attachment dimensions, as well as the contribution of the latter to the moderation of distress. Participants were 120 couples in their first marriages. It was found that: (1) Self- and spouses' reported self-criticism are both associated with depression; (2) negative assessments of personality factors and attachment models by the self and spouse contribute uniquely in predicting depressive symptomatology; and (3) beyond the covariation between target's depression and marital maladjustment, attachment models of self and of other as reported by both the self and spouse moderate the effects of self-reported personality vulnerability on depressive symptomatology. Our results indicate that self-ratings and ratings by others must both be considered in the context of depression in close interpersonal relationships. Beyond the methodological implications of multisource data, our findings support the view of depression as an interpersonal process.     

Dual commitment to organization and supervisor: A person-centered approach

A recent trend in commitment research has been to use person-centered analytic strategies to identify homogeneous subgroups with varying configurations of commitment mindsets (affective, normative, continuance) or targets (e.g., organization, supervisor, team). A person-centered approach takes a more holistic perspective than the traditional variable-centered approach and can reflect potentially complex interactions among commitment mindsets and/or targets. We extend application of the person-centered approach to investigate profiles of commitment to two interrelated targets, the organization and supervisor, in two studies (Ns = 481 and 264) involving Belgian university graduates. Using latent profile analyses, we found that a similar 5-profile model fit best in both studies. The mindset pattern for the two targets was similar for some profile groups, but differed for others. The groups differed on perceived organizational and supervisory support and voluntary turnover largely as expected from commitment and support theory. Implications for future research and management practice are discussed.     

A new approach to the formulation and testing of learning models

It is argued that current attempts to model human learning behavior commonly fail on one of two counts: either the model assumptions are artificially restricted so as to permit the application of mathematical techniques in deriving their consequences, or else the required complex assumptions are imbedded in computer programs whose technical details obscure the theoretical content of the model. The first failing is characteristic of so-called mathematical models of learning, while the second is characteristic of computer simulation models. An approach to model building which avoids both these failings is presented under the title of a black-box theory of learning. This method permits the statement of assumptions of any desired complexity in a language which clearly exhibits their theoretical content.Section II of the paper is devoted to the problem of testing and comparing alternative learning theories. The policy advocated is to abandon attempts at hypothesis testing. It is argued that, in general, we not only lack sufficient data and sufficiently powerful techniques to test hypotheses, but that the truth of a model is not really the issue of basic interest. A given model may be true in the sense that on the basis of available evidence we cannot statistically reject it, but not interesting in the sense that it provides little information about the processes underlying behavior. Rather, we should accept or reject models on the basis of how much information they provide about the way in which subjects respond to environmental structure. This attitude toward model testing is made precise by introducing a formal measure of the information content of a model. Finally, it is argued that the statistical concept of degrees-of-freedom is misleading when used in the context of model testing and should be replaced by a measure of the information absorbed from the data in estimating parameters.I would like to express my indebtedness to Ernest W. Adams and Julian Feldman for their guidance and encouragement of the research reported here.Much of the research reported here was conducted at the University of California, Berkeley, and incorporated in a Ph.D. dissertation in Logic and the Methodology of Science [15]. The research was supported in part by the Computer Institute for Social Science Research, Michigan State University.     

Talking nets: a multiagent connectionist approach to communication and trust between individuals

A multiagent connectionist model is proposed that consists of a collection of individual recurrent networks that communicate with each other and, as such, is a network of networks. The individual recurrent networks simulate the process of information uptake, integration, and memorization within individual agents, and the communication of beliefs and opinions between agents is propagated along connections between the individual networks. A crucial aspect in belief updating based on information from other agents is the trust in the information provided. In the model, trust is determined by the consistency with the receiving agents' existing beliefs and results in changes of the connections between individual networks, called trust weights. These weights lead to a selective propagation and thus to the filtering out of less reliable information, and they implement H. P. Grice's (1975) maxims of quality and quantity in communication. The unique contribution of communicative mechanisms beyond intrapersonal processing of individual networks was explored in simulations of key phenomena involving persuasive communication and polarization, lexical acquisition, spreading of stereotypes and rumors, and a lack of sharing unique information in group decisions.