Psychological Review

A neurocomputational account of taxonomic responding and fast mapping in early word learning.
We present a neurocomputational model with self-organizing maps that accounts for the emergence of taxonomic responding and fast mapping in early word learning, as well as a rapid increase in the rate of acquisition of words observed in late infancy. The quality and efficiency of generalization of word–object associations is directly related to the quality of prelexical, categorical representations in the model. We show how synaptogenesis supports coherent generalization of word–object associations and show that later synaptic pruning minimizes metabolic costs without being detrimental to word learning. The role played by joint-attentional activities is identified in the model, both at the level of selecting efficient cross-modal synapses and at the behavioral level, by accelerating and refining overall vocabulary acquisition. The model can account for the qualitative shift in the way infants use words, from an associative to a referential-like use, for the pattern of overextension errors in production and comprehension observed during early childhood and typicality effects observed in lexical development. Interesting by-products of the model include a potential explanation of the shift from prototype to exemplar-based effects reported for adult category formation, an account of mispronunciation effects in early lexical development, and extendability to include accounts of individual differences in lexical development and specific disorders such as Williams syndrome. The model demonstrates how an established constraint on lexical learning, which has often been regarded as domain-specific, can emerge from domain-general learning principles that are simultaneously biologically, psychologically, and socially plausible. (PsycINFO Database Record (c) 2009 APA, all rights reserved)
A motivational theory of life-span development.
This article had four goals. First, the authors identified a set of general challenges and questions that a life-span theory of development should address. Second, they presented a comprehensive account of their Motivational Theory of Life-Span Development. They integrated the model of optimization in primary and secondary control and the action-phase model of developmental regulation with their original life-span theory of control to present a comprehensive theory of development. Third, they reviewed the relevant empirical literature testing key propositions of the Motivational Theory of Life-Span Development. Finally, because the conceptual reach of their theory goes far beyond the current empirical base, they pointed out areas that deserve further and more focused empirical inquiry. (PsycINFO Database Record (c) 2009 APA, all rights reserved)
A neural network model of the structure and dynamics of human personality.
We present a neural network model that aims to bridge the historical gap between dynamic and structural approaches to personality. The model integrates work on the structure of the trait lexicon, the neurobiology of personality, temperament, goal-based models of personality, and an evolutionary analysis of motives. It is organized in terms of two overarching motivational systems, an approach and an avoidance system, as well as a general disinhibition and constraint system. Each overarching motivational system influences more specific motives. Traits are modeled in terms of differences in the sensitivities of the motivational systems, the baseline activation of specific motives, and inhibitory strength. The result is a motive-based neural network model of personality based on research about the structure and neurobiology of human personality. The model provides an account of personality dynamics and person–situation interactions and suggests how dynamic processing approaches and dispositional, structural approaches can be integrated in a common framework. (PsycINFO Database Record (c) 2009 APA, all rights reserved)
Serial position curves in free recall.
The scenario for free recall set out in Laming (2009) is developed to provide models for the serial position curves from 5 selected sets of data, for final free recall, and for multitrial free recall. The 5 sets of data reflect the effects of rate of presentation, length of list, delay of recall, and suppression of rehearsal. Each model accommodates the serial position curve for first recalls (where those data are available) as well as that for total recalls. Both curves are fit with the same parameter values, as also (with 1 exception) are all of the conditions compared within each experiment. The distributions of numbers of recalls are also examined and shown to have variances increased above what would be expected if successive recalls were independent. This is taken to signify that, in those experiments in which rehearsals were not recorded, the retrieval of words for possible recall follows the same pattern that is observed following overt rehearsal, namely, that retrieval consists of runs of consecutive elements from memory. Finally, 2 sets of data are examined that the present approach cannot accommodate. It is argued that the problem with these data derives from an interaction between the patterns of (covert) rehearsal and the parameters of list presentation. (PsycINFO Database Record (c) 2009 APA, all rights reserved)
Allostasis and the human brain: Integrating models of stress from the social and life sciences.
We draw on the theory of allostasis to develop an integrative model of the current stress process that highlights the brain as a dynamically adapting interface between the changing environment and the biological self. We review evidence that the core emotional regions of the brain constitute the primary mediator of the well-established association between stress and health, as well as the neural focus of wear and tear due to ongoing adaptation. This mediation, in turn, allows us to model the interplay over time between context, current stressor exposure, internal regulation of bodily processes, and health outcomes. We illustrate how this approach facilitates the integration of current findings in human neuroscience and genetics with key constructs from stress models from the social and life sciences, with implications for future research and the design of interventions targeting individuals at risk. (PsycINFO Database Record (c) 2009 APA, all rights reserved)
The central role of recognition in auditory perception: A neurobiological model.
The model presents neurobiologically plausible accounts of sound recognition (including absolute pitch), neural plasticity involved in pitch, loudness and location information integration, and streaming and auditory recall. It is proposed that a cortical mechanism for sound identification modulates the spectrotemporal response fields of inferior colliculus neurons and regulates the encoding of the echoic trace in the thalamus. Identification involves correlation of sequential spectral slices of the stimulus-driven neural activity with stored representations in association with multimodal memories, verbal lexicons, and contextual information. Identities are then consolidated in auditory short-term memory and bound with attribute information (usually pitch, loudness, and direction) that has been integrated according to the identities’ spectral properties. Attention to, or recall of, a particular identity will excite a particular sequence in the identification hierarchies and so lead to modulation of thalamus and inferior colliculus neural spectrotemporal response fields. This operates as an adaptive filter for identities, or their attributes, and explains many puzzling human auditory behaviors, such as the cocktail party effect, selective attention, and continuity illusions. (PsycINFO Database Record (c) 2009 APA, all rights reserved)
Context, learning, and extinction.
A. Redish et al. (2007) proposed a reinforcement learning model of context-dependent learning and extinction in conditioning experiments, using the idea of “state classification” to categorize new observations into states. In the current article, the authors propose an interpretation of this idea in terms of normative statistical inference. They focus on renewal and latent inhibition, 2 conditioning paradigms in which contextual manipulations have been studied extensively, and show that online Bayesian inference within a model that assumes an unbounded number of latent causes can characterize a diverse set of behavioral results from such manipulations, some of which pose problems for the model of Redish et al. Moreover, in both paradigms, context dependence is absent in younger animals, or if hippocampal lesions are made prior to training. The authors suggest an explanation in terms of a restricted capacity to infer new causes. (PsycINFO Database Record (c) 2009 APA, all rights reserved)
Intrusive images in psychological disorders: Characteristics, neural mechanisms, and treatment implications.
Involuntary images and visual memories are prominent in many types of psychopathology. Patients with posttraumatic stress disorder, other anxiety disorders, depression, eating disorders, and psychosis frequently report repeated visual intrusions corresponding to a small number of real or imaginary events, usually extremely vivid, detailed, and with highly distressing content. Both memory and imagery appear to rely on common networks involving medial prefrontal regions, posterior regions in the medial and lateral parietal cortices, the lateral temporal cortex, and the medial temporal lobe. Evidence from cognitive psychology and neuroscience implies distinct neural bases to abstract, flexible, contextualized representations (C-reps) and to inflexible, sensory-bound representations (S-reps). We revise our previous dual representation theory of posttraumatic stress disorder to place it within a neural systems model of healthy memory and imagery. The revised model is used to explain how the different types of distressing visual intrusions associated with clinical disorders arise, in terms of the need for correct interaction between the neural systems supporting S-reps and C-reps via visuospatial working memory. Finally, we discuss the treatment implications of the new model and relate it to existing forms of psychological therapy. (PsycINFO Database Record (c) 2009 APA, all rights reserved)
Toward an integrated gender-linked model of aggression subtypes in early and middle childhood.
An integrative model is proposed for understanding the development of physical and relational aggression in early and middle childhood. The central goal was to posit a new theoretical framework that expands on existing social-cognitive and gender schema models (i.e., Social Information-Processing Model of Children’s Adjustment [N. R. Crick & K. A. Dodge, 1994] and the Schematic-Processing Model of Sex Role Stereotyping [C. L. Martin & C. F. Halverson, 1981]). The proposed model suggests several individual- and group-level effects and the available evidence for each of these hypotheses is discussed. The ways in which the proposed model may guide future research in the field are presented. (PsycINFO Database Record (c) 2009 APA, all rights reserved)
A dual system model of preferences under risk.
This article presents a dual system model (DSM) of decision making under risk and uncertainty according to which the value of a gamble is a combination of the values assigned to it independently by the affective and deliberative systems. On the basis of research on dual process theories and empirical research in Hsee and Rottenstreich (2004) and Rottenstreich and Hsee (2001) among others, the DSM incorporates (a) individual differences in disposition to rational versus emotional decision making, (b) the affective nature of outcomes, and (c) different task construals within its framework. The model has good descriptive validity and accounts for (a) violation of nontransparent stochastic dominance, (b) fourfold pattern of risk attitudes, (c) ambiguity aversion, (d) common consequence effect, (e) common ratio effect, (f) isolation effect, and (g) coalescing and event-splitting effects. The DSM is also used to make several novel predictions of conditions under which specific behavior patterns may or may not occur. (PsycINFO Database Record (c) 2009 APA, all rights reserved)
Computational modeling of reading in semantic dementia: Comment on Woollams, Lambon Ralph, Plaut, and Patterson (2007).
Woollams, Lambon Ralph, Plaut, and Patterson (see record 2007-05396-004) reported detailed data on reading in 51 cases of semantic dementia. They simulated some aspects of these data using a connectionist parallel distributed processing (PDP) triangle model of reading. We argue here that a different model of reading, the dual route cascaded (DRC) model of Coltheart, Rastle, Perry, Langdon, and Ziegler (2001), not only provides a more accurate simulation of these aspects of reading in semantic dementia than does the PDP model but also provides highly accurate simulations of other aspects of reading in this disorder that the PDP approach has not simulated. We conclude that our findings add to evidence both from simulations of normal skilled reading and from simulations of other kinds of acquired dyslexia that the nonconnectionist DRC model of reading offers a better account of normal and disordered reading than the connectionist PDP models of reading. (PsycINFO Database Record (c) 2009 APA, all rights reserved)
Postscript: Reading in semantic dementia—A response to Woollams, Lambon Ralph, Plaut, and Patterson (2010).
The current authors reply to a response by Woollams, Lambon Ralph, Plaut, and Patterson (see record 2009-25263-010) on a comment by the current authors (see record 2009-25263-008) on the original article (see record 2007-05396-004). The current authors list their agreements and disagreements with Woollams, Lambon Ralph, Plaut, and Patterson's response on the topics of the human reading system, cognitive architecture, experimental psychology of reading, and computational models. (PsycINFO Database Record (c) 2009 APA, all rights reserved)
SD-squared revisited: Reply to Coltheart, Tree, and Saunders (2010).
The connectionist triangle model of reading aloud proposes that semantic activation of phonology is particularly important for correct pronunciation of low-frequency exception words. Our consideration of this issue (Woollams, Lambon Ralph, Plaut, & Patterson, 2007) (see record 2007-05396-004) reported computational simulations demonstrating that reduction and disruption of this semantic activation resulted in the marked deficit in low-frequency exception word reading that is characteristic of surface dyslexia. We then presented 100 observations of reading aloud from 51 patients with semantic dementia (SD) demonstrating a universal decline into surface dyslexia, a phenomenon we termed “SD-squared.” Coltheart, Tree, and Saunders (see record 2009-25263-008) have more recently provided a simulation of the SD-squared data within the dual route cascaded (DRC) model, achieved by varying the amount of damage to components of the lexical and nonlexical pathways. Although they suggested that these simulations provide a closer fit to the SD patients’ reading data than our own, we demonstrate here that this is not the case. Moreover, we argue that the connectionist triangle model account has substantially greater explanatory and predictive power than the DRC account. (PsycINFO Database Record (c) 2009 APA, all rights reserved)
Postscript: SD-squared revisited again.
The current authors reply to a postscript by Coltheart, Tree, and Saunders (see record 2009-25263-009) which was in response to the current authors response (see record 2009-25263-010) on a comment by the current authors (see record 2009-25263-008) on the original article (see record 2007-05396-004). The current authors begin by responding to the final challenge posed by Coltheart, Tree, and Saunders (2010). They believe that both experimental and computational methodologies should be aimed at explaining human behavior by uncovering the computational principles that govern the underlying cognitive and neural processes. While it is important to ensure that the explanation and principles remain self-consistent, there is no need for a single uber-simulation, just as there is no need for a single behavioral experiment that manipulates all relevant variables simultaneously. Detailed comparison between model and human performance is critical for evaluating and improving current understanding, but data fitting per se is not the goal of the enterprise. (PsycINFO Database Record (c) 2009 APA, all rights reserved)
Locating object knowledge in the brain: Comment on Bowers’s (2009) attempt to revive the grandmother cell hypothesis.
According to Bowers (see record 2009-00258-008), the finding that there are neurons with highly selective responses to familiar stimuli supports theories positing localist representations over approaches positing the type of distributed representations typically found in parallel distributed processing (PDP) models. However, his conclusions derive from an overly narrow view of the range of possible distributed representations and of the role that PDP models can play in exploring their properties. Although it is true that current distributed theories face challenges in accounting for both neural and behavioral data, the proposed localist account—to the extent that it is articulated at all—runs into more fundamental difficulties. Central to these difficulties is the problem of specifying the set of entities a localist unit represents. (PsycINFO Database Record (c) 2009 APA, all rights reserved)
Postscript: Parallel distributed processing in localist models without thresholds.
The current authors reply to a response by Bowers (see record 2009-25263-016) on a comment by the current authors (see record 2009-25263-012) on the original article (see record 2009-00258-008). Bowers (2010) mischaracterizes the goals of parallel distributed processing (PDP research)—explaining performance on cognitive tasks is the primary motivation. More important, his claim that localist models, such as the interactive activation model, “recognize” their inputs when a threshold is reached runs directly counter to an essential feature of these models. This undermines his attempt to distinguish between what a neuron responds to and what it codes for and, indeed, undermines the whole localist argument he has proposed. Bowers’s (2010) also continues to face difficulty in specifying what localist units correspond to, and all of the possible choices face problems. In the paragraphs below we substantiate these points. (PsycINFO Database Record (c) 2009 APA, all rights reserved)
Measuring sparseness in the brain: Comment on Bowers (2009).
Bowers (see record 2009-00258-008) challenged the common view in favor of distributed representations in psychological modeling and the main arguments given against localist and grandmother cell coding schemes. He revisited the results of several single-cell studies, arguing that they do not support distributed representations. We praise the contribution of Bowers (2009) for joining evidence from psychological modeling and neurophysiological recordings, but we disagree with several of his claims. In this comment, we argue that distinctions between distributed, localist, and grandmother cell coding can be troublesome with real data. Moreover, these distinctions seem to be lying within the same continuum, and we argue that it may be sensible to characterize coding schemes with a sparseness measure. We further argue that there may not be a unique coding scheme implemented in all brain areas and for all possible functions. In particular, current evidence suggests that the brain may use distributed codes in primary sensory areas and sparser and invariant representations in higher areas. (PsycINFO Database Record (c) 2009 APA, all rights reserved)
Postscript: About grandmother cells and Jennifer Aniston neurons.
The current authors reply to a response by Bowers (see record 2009-25263-016) on a comment by the current authors (see record 2009-25263-014) on the original article (see record 2009-00258-008). A typical problem in any discussion about grandmother cells is that there is not a general consensus about what should be called as such. Here, we discuss possible interpretations in turn and contrast them with what we find in our own data (arguably the closest experimental evidence of grandmother cells so far). A first and naïve interpretation of the term grandmother cell is that one and only one neuron encodes for one and only one concept (a face, an object, an animal, etc.). We agree with Bowers (2010) that this is a straw-man version of this idea—although some people still take this view when (incorrectly) arguing that if we would have grandmother cells then the concept of grandma would disappear if her dedicated cell dies—which clearly does not apply to our data. (PsycINFO Database Record (c) 2009 APA, all rights reserved)
More on grandmother cells and the biological implausibility of PDP models of cognition: A reply to Plaut and McClelland (2010) and Quian Quiroga and Kreiman (2010).
Replies to the comments by Plaut and McClelland (see record 2009-25263-012) and Quian Quiroga and Kreiman (see record 2009-25263-014) on the authors original article (see record 2009-00258-008) both challenged my characterization of localist and distributed representations. They also challenged the biological plausibility of grandmother cells on conceptual and empirical grounds. This reply addresses these issues in turn. The premise of my argument is that grandmother cells in neuroscience are the equivalent of localist representations in psychology. When defined in this way, grandmother cells are biologically plausible, given the neuroscience to date. By contrast, the neurophysiology is shown to be inconsistent with the distributed representations often learned in existing parallel distributed processing (PDP) models, and it poses a challenge to PDP theories more generally. (PsycINFO Database Record (c) 2009 APA, all rights reserved)
Postscript: Some final thoughts on grandmother cells, distributed representations, and PDP models of cognition.
The author briefly responds to a number of terminological, theoretical, and empirical issues raised in some postscripts. The goal is not to respond to each outstanding point but rather to address some comments that in my view confuse rather than clarify matters. He responds to Plaut and McClelland (see record 2009-25263-012) and Quian Quiroga and Kreiman (see record 2009-25263-014) in turn. (PsycINFO Database Record (c) 2009 APA, all rights reserved)

Psychological Review - Vol 117, Iss 1