Alien voices and inner dialogue: towards a developmental account of auditory verbal hallucinations

The phenomenon of auditory verbal hallucinations (AVHs) is one of the most intriguing features of the psychiatric literature. Two alternative models of the development of AVHs in both normal and psychotic populations are proposed. In the disruption to internalisation (DI) model, AVHs result from a disruption to the normal processes of internalisation of inner speech. In the re-expansion (RE) model, AVHs result when normal inner speech is re-expanded into inner dialogue under conditions of stress and cognitive challenge. Both models draw on Vygotsky's (The Collected Works Of L.S. Vygotsky, New York, Plenum Press, 1987) ideas about the development of inner speech. On this view, normal inner speech is considerably abbreviated relative to external speech, and also undergoes some important semantic transformations. In both the DI and RE models, AVHs arise when the subject's inner speech involves inappropriately expanded inner dialogue, leading the subject to experience the voices in the dialogue as alien. The two models may prove useful in explaining some of the social-developmental evidence surrounding the phenomenon, and also make a number of testable predictions which are suggested as priorities for future research.     

Refining the Inferential Model of Scientific Understanding

In this article, I use a mental models computational account of representation to illustrate some details of my previously presented inferential model of scientific understanding. The hope is to shed some light on possible mechanisms behind the notion of scientific understanding. I argue that if mental models are a plausible approach to modelling cognition, then understanding can best be seen as the coupling of specific rules. I present our beliefs as ‘ordinary’ conditional rules, and the coupling process as one where the consequent of one ordinary rule (OR) matches and activates the antecedent of the rule to which it is coupled in virtue of the activation of an intermediate ‘inference’ rule. I argue that on this approach knowledge of an explanation is the activation of ORs in a cognitive hierarchy, while understanding is achieved when those activated ORs are also coupled via correct inference rules. I do not directly address issues regarding the plausibility of mental models themselves. This article should therefore be seen as an exercise in refining the inferential model within an already presupposed computational setting, not one of arguing for the psychological adequacy of computational approaches.     

Exemplar models as a mechanism for performing Bayesian inference

Probabilistic models have recently received much attention as accounts of human cognition. However, most research in which probabilistic models have been used has been focused on formulating the abstract problems behind cognitive tasks and their optimal solutions, rather than on mechanisms that could implement these solutions. Exemplar models are a successful class of psychological process models in which an inventory of stored examples is used to solve problems such as identification, categorization, and function learning. We show that exemplar models can be used to perform a sophisticated form of Monte Carlo approximation known as importance sampling and thus provide a way to perform approximate Bayesian inference. Simulations of Bayesian inference in speech perception, generalization along a single dimension, making predictions about everyday events, concept learning, and reconstruction from memory show that exemplar models can often account for human performance with only a few exemplars, for both simple and relatively complex prior distributions. These results suggest that exemplar models provide a possible mechanism for implementing at least some forms of Bayesian inference.     

The link between speech perception and production is phonological and abstract: evidence from the shadowing task

This study reports a shadowing experiment, in which one has to repeat a speech stimulus as fast as possible. We tested claims about a direct link between perception and production based on speech gestures, and obtained two types of counterevidence. First, shadowing is not slowed down by a gestural mismatch between stimulus and response. Second, phonetic detail is more likely to be imitated in a shadowing task if it is phonologically relevant. This is consistent with the idea that speech perception and speech production are only loosely coupled, on an abstract phonological level.     

EARSHOT: A Minimal Neural Network Model of Incremental Human Speech Recognition

Despite the lack of invariance problem (the many-to-many mapping between acoustics and percepts), human listeners experience phonetic constancy and typically perceive what a speaker intends. Most models of human speech recognition (HSR) have side-stepped this problem, working with abstract, idealized inputs and deferring the challenge of working with real speech. In contrast, carefully engineered deep learning networks allow robust, real-world automatic speech recognition (ASR). However, the complexities of deep learning architectures and training regimens make it difficult to use them to provide direct insights into mechanisms that may support HSR. In this brief article, we report preliminary results from a two-layer network that borrows one element from ASR, long short-term memory nodes, which provide dynamic memory for a range of temporal spans. This allows the model to learn to map real speech from multiple talkers to semantic targets with high accuracy, with human-like timecourse of lexical access and phonological competition. Internal representations emerge that resemble phonetically organized responses in human superior temporal gyrus, suggesting that the model develops a distributed phonological code despite no explicit training on phonetic or phonemic targets. The ability to work with real speech is a major advance for cognitive models of HSR.     

Goal orientation dispositions and performance trajectories: The roles of supplementary and complementary situational inducements

Integrating goal orientation theory with interactionist approaches, this experimental study (N = 104) tested the unique and interactive effects of individual differences in goal orientations and situational goal orientation inducements on performance trajectories during skill acquisition. Results indicated that learning goal orientation predicted performance trajectories more positively when coupled with one situational inducement that captures a complementary feature (a performance, as opposed to a learning, goal frame), and when jointed with a situational inducement that captures a supplementary feature (self-referenced vs. normative-based performance feedback). There was also a complementary-like interaction between the two situational inducements, such that a learning goal frame led to more positive performance trajectories when coupled with normative, as opposed to self-referent, feedback. Implications for the motivation and skill acquisition literatures are discussed.     

Shame as a predictor of post-event rumination in social anxiety

Evidence shows that people with high social anxiety levels ruminate about distressing social events, which contributes to the maintenance of social anxiety symptoms. The present study aimed to explore the role of shame in maintaining post-event rumination (PER) following a negative social event (an impromptu speech with negative feedback) in a student sample (N?=?104). Participants reported negative rumination related to the event one day and one week after the speech. PER measured one day after the speech was not associated with social anxiety symptoms and state anxiety. One week later, participants with clinically relevant social anxiety symptoms experienced greater PER. State shame was the only significant predictor of PER in a regression equation that also included social anxiety symptoms, state anxiety and self-evaluation of performance. Possible explanations and implications are discussed in light of cognitive models of social anxiety.     

Decision-making,cognitive functions,impulsivity, and media multitasking expectancies in high versus low media multitaskers

In several studies, individuals who reported to frequently multitask with different media displayed reduced cognitive performance, for example in fluid intelligence and executive functioning. These cognitive functions are relevant for making advantageous decisions under both objective risk (requiring reflection and strategical planning) and ambiguous risk (requiring learning from feedback). Thus, compared to low media multitaskers (LMMs), high media multitaskers (HMMs) may perform worse in both types of decision situations. The current study investigated HMMs and LMMs in a laboratory setting with the Game of Dice Task (GDT; objective risk), the Iowa Gambling Task (IGT; ambiguous risk), various tests quantifying cognitive functions (logical reasoning, working memory, information processing, general executive functions), and self-report measures of impulsivity, media multitasking expectancies, and problematic Internet use. From 182 participants, 25 HMMs and 19 LMMs were identified using the Media Multitasking Index. Results show that HMMs compared to LMMs performed weaker on the IGT but not on the GDT. Furthermore, HMMs had slightly decreased performance in tests of logical reasoning and working memory capacity. HMMs tended to increased information processing speed but this difference was not significant. Furthermore, HMMs have more positive expectancies regarding media multitasking and reported higher tendencies toward problematic Internet use. HMMs and LMMs did not differ significantly with respect to impulsivity and executive functions. The results give a first hint that HMMs may have difficulties in decision-making under ambiguous but not under objective risk. HMMs may be more prone to errors in tasks that require feedback processing. However, HMMs appear not to be impaired in aspects of long-term strategic decision-making.

     

Bootstrapping the lexicon: a computational model of infant speech segmentation

Prelinguistic infants must find a way to isolate meaningful chunks from the continuous streams of speech that they hear. BootLex, a new model which uses distributional cues to build a lexicon, demonstrates how much can be accomplished using this single source of information. This conceptually simple probabilistic algorithm achieves significant segmentation results on various kinds of language corpora - English, Japanese, and Spanish; child- and adult-directed speech, and written texts; and several variations in coding structure - and reveals which statistical characteristics of the input have an influence on segmentation performance. BootLex is then compared, quantitatively and qualitatively, with three other groups of computational models of the same infant segmentation process, paying particular attention to functional characteristics of the models and their similarity to human cognition. Commonalities and contrasts among the models are discussed, as well as their implications both for theories of the cognitive problem of segmentation itself, and for the general enterprise of computational cognitive modeling.     

Distractor filtering in media multitaskers

A growing amount of modern media is consumed simultaneously, a phenomenon known as 'media multitasking'. Individuals who regularly engage in this activity, heavy media multitaskers (HMMs), are more affected by irrelevant information that can intrude into a primary task than are light media multitaskers (LMMs--Ophir et al, 2009 Proceedings of the National Academy of Sciences of the USA 106 15583). However, the locus of this deficit is unknown, as previous research is consistent with both memory and attentional explanations. Here, we isolated attentional processes by employing a singleton distractor task with low working-memory demands. In this task, LMMs used top-down information to improve their performance, yet HMMs did not. This difference in performance in an established attentional capture task argues for the presence of attentional differences in HMMs and is consistent with the idea that HMMs maintain a wider attentional scope than LMMs, even when instructed otherwise.     

Systematic studies of modified vocalization: speech production changes during a variation of metronomic speech in persons who do and do not stutter

The most common way to induce fluency using rhythm requires persons who stutter to speak one syllable or one word to each beat of a metronome, but stuttering can also be eliminated when the stimulus is of a particular duration (e.g., 1 second [s]). The present study examined stuttering frequency, speech production changes, and speech naturalness during rhythmic speech that alternated 1 s of reading with 1 s of silence. A repeated-measures design was used to compare data obtained during a control reading condition and during rhythmic reading in 10 persons who stutter (PWS) and 10 normally fluent controls. Ratings for speech naturalness were also gathered from naïve listeners. Results showed that mean vowel duration increased significantly, and the percentage of short phonated intervals decreased significantly, for both groups from the control to the experimental condition. Mean phonated interval length increased significantly for the fluent controls. Mean speech naturalness ratings during the experimental condition were approximately “7” on a 1–9 scale (1 = highly natural; 9 = highly unnatural), and these ratings were significantly correlated with vowel duration and phonated intervals for PWS. The findings indicate that PWS may be altering vocal fold vibration duration to obtain fluency during this rhythmic speech style, and that vocal fold vibration duration may have an impact on speech naturalness during rhythmic speech. Future investigations should examine speech production changes and speech naturalness during variations of this rhythmic condition.Educational objectives: The reader will be able to: (1) describe changes (from a control reading condition) in speech production variables when alternating between 1 s of reading and 1 s of silence, (2) describe which rhythmic conditions have been found to sound and feel the most natural, (3) describe methodological issues for studies about alterations in speech production variables during fluency-inducing conditions, and (4) describe which fluency-inducing conditions have been shown to involve a reduction in short phonated intervals.     

Testing between the TRACE model and the fuzzy logical model of speech perception

The TRACE model of speech perception (McClelland & Elman, 1986) is contrasted with a fuzzy logical model of perception (FLMP) (Oden & Massaro, 1978). The central question is how the models account for the influence of multiple sources of information on perceptual judgment. Although the two models can make somewhat similar predictions, the assumptions underlying the models are fundamentally different. The TRACE model is built around the concept of interactive activation, whereas the FLMP is structured in terms of the integration of independent sources of information. The models are tested against test results of an experiment involving the independent manipulation of bottom-up and top-down sources of information. Using a signal detection framework, sensitivity and bias measures of performance can be computed. The TRACE model predicts that top-down influences from the word level influence sensitivity at the phoneme level, whereas the FLMP does not. The empirical results of a study involving the influence of phonological context and segmental information on the perceptual recognition of a speech segment are best described without any assumed changes in sensitivity. To date, not only is a mechanism of interactive activation not necessary to describe speech perception, it is shown to be wrong when instantiated in the TRACE model.     

Bayesian Diagnostics of Hidden Markov Structural Equation Models with Missing Data

Cocaine is a type of drug that functions to increase the availability of the neurotransmitter dopamine in the brain. However, cocaine dependence or abuse is highly related to an increased risk of psychiatric disorders and deficits in cognitive performance, attention, and decision-making abilities. Given the chronic and persistent features of drug addiction, the progression of abstaining from cocaine often evolves across several states, such as addiction to, moderate dependence on, and swearing off cocaine. Hidden Markov models (HMMs) are well suited to the characterization of longitudinal data in terms of a set of unobservable states, and have increasingly been used to uncover the dynamic heterogeneity in progressive diseases or activities. However, the existence of outliers or influential points may misidentify the hidden states and distort the associated inference. In this study, we develop a Bayesian local influence procedure for HMMs with latent variables in the presence of missing data. The proposed model enables us to investigate the dynamic heterogeneity of multivariate longitudinal data, reveal how the interrelationships among latent variables change from one state to another, and simultaneously conduct statistical diagnosis for the given data, model assumptions, and prior inputs. We apply the proposed procedure to analyze a dataset collected by the UCLA center for advancing longitudinal drug abuse research. Several outliers or influential points that seriously influence estimation results are identified and removed. The proposed procedure also discovers the effects of treatment and individuals’ psychological problems on cocaine use behavior and delineates their dynamic changes across the cocaine-addiction states.     

Toward a Unified Model of Lexical Acquisition and Lexical Access

Much effort has gone into constructing models of how children segment speech and thereby discover the words of their language. Much effort has also gone into constructing models of how adults access their mental lexicons and thereby segment speech into words. In this paper, I explore the possibility of a model that could account for both word discovery by children and on-line segmentation by adults. In particular, I discuss extensions to the distributional regularity (DR) model of Brent and Cartwright (1996) that could yield an account of on-line segmentation as well as word discovery.     

The CHIC Model: A Global Model for Coupled Binary Data

Often problems result in the collection of coupled data, which consist of different N-way N-mode data blocks that have one or more modes in common. To reveal the structure underlying such data, an integrated modeling strategy, with a single set of parameters for the common mode(s), that is estimated based on the information in all data blocks, may be most appropriate. Such a strategy implies a global model, consisting of different N-way N-mode submodels, and a global loss function that is a (weighted) sum of the partial loss functions associated with the different submodels. In this paper, such a global model for an integrated analysis of a three-way three-mode binary data array and a two-way two-mode binary data matrix that have one mode in common is presented. A simulated annealing algorithm to estimate the model parameters is described and evaluated in a simulation study. An application of the model to real psychological data is discussed. T. Wilderjans is a Research Assistant of the Fund for Scientific Research—Flanders (Belgium). The research reported in this paper was partially supported by the Research Council of K.U. Leuven (GOA/2005/04). We are grateful to Kristof Vansteelandt for providing us with an interesting data set. We also thank three anonymous reviewers for their useful comments.     

The Effect of Team Mental Models Divergence on Creative Performance During Situational Changes

This study developed a model of team mental models’ influence on team adaptive performance at different stages. It proposed that mental models have different relationships with adaptive performance as a function of the performance stage when teams experience changes: early on (i.e., during situation assessment) more divergent mental models are needed for performance; later on (i.e., during plan execution) more convergent mental models are needed. The model was tested in an experiment with 33 teams (N= 99) faced with unforeseen change at task midway. Partially supporting the predictions of a divergence-convergence model, at situation assessment, teams that developed more dissimilar mental models attained a higher performance originality, but not higher efficiency or usefulness. However, at plan execution, team mental models’ convergence did not lead to higher performance. Implications for research on team mental models are discussed.     

A model of serial order problems in fluent, stuttered and agrammatic speech

Many models of speech production have attempted to explain dysfluent speech. Most models assume that the disruptions that occur when speech is dysfluent arise because the speakers make errors while planning an utterance. In this contribution, a model of the serial order of speech is described that does not make this assumption. It involves the coordination or 'interlocking' of linguistic planning and execution stages at the language-speech interface. The model is examined to determine whether it can distinguish two forms of dysfluent speech (stuttered and agrammatic speech) that are characterized by iteration and omission of whole words and parts of words.     

Strategy switch costs in arithmetic problem solving

Three experiments tested whether switching between strategies involves a cost. In three experiments, participants had to give approximate products to two-digit multiplication problems (e.g., 47×76). They were told which strategy to use (Experiments 1 and 2) or could choose among strategies (Experiment 3). The participants showed poorer performance when they used different strategies on two consecutive trials than when they used the same strategy. They also used the same strategy over two consecutive problems more often than they used different strategies. These effects, termed strategy switch costs, were found when the participants executed the easiest strategy and when they solved easy problems. We discuss possible processes underlying these strategy switch costs and the implications of these strategy switch costs for models of strategy choices.     

Audio Feedback with Reduced Self-focus as an Intervention for Social Anxiety: An Experimental Study

Background and Objectives: Socially anxious individuals tend to underestimate their performance largely due to attentional bias. Video and audio feedback (AF) with cognitive preparation (CP) have shown to improve socially anxious individuals′ evaluation of their performance in previous studies. In the present study, it was hypothesized that one of the three steps in CP, reduced self-focus (RS), is sufficient to cause an improved voice evaluation after AF. This was tested in a single-session randomized controlled experiment. Method: Forty-one socially anxious participants were asked to give a speech, then to listen to and evaluate a taped recording of their performance. Half of the sample were instructed to reduce their self-focus prior to AF, the rest received AF only. RS involved asking participants to listen to the audio recording as though they were listening to a stranger. Generalization effects were assessed by a second speech. Results: AF with RS led to more improved voice evaluations than AF-only after the first speech, and the effects remained in the evaluation of the second speech. More positive speech evaluations were associated with corresponding reductions of performance anxiety.Limitations: small sample, analogue study. Conclusion: One component of cognitive preparation—(RS)—appears to be sufficient to cause significant effects on voice evaluation in socially anxious individuals. If the results are replicated in clinical samples, AF with RS may be a promising intervention in the treatment of social anxiety.     

Bayesian Model Averaging for Propensity Score Analysis

This article considers Bayesian model averaging as a means of addressing uncertainty in the selection of variables in the propensity score equation. We investigate an approximate Bayesian model averaging approach based on the model-averaged propensity score estimates produced by the R package BMA but that ignores uncertainty in the propensity score. We also provide a fully Bayesian model averaging approach via Markov chain Monte Carlo sampling (MCMC) to account for uncertainty in both parameters and models. A detailed study of our approach examines the differences in the causal estimate when incorporating noninformative versus informative priors in the model averaging stage. We examine these approaches under common methods of propensity score implementation. In addition, we evaluate the impact of changing the size of Occam’s window used to narrow down the range of possible models. We also assess the predictive performance of both Bayesian model averaging propensity score approaches and compare it with the case without Bayesian model averaging. Overall, results show that both Bayesian model averaging propensity score approaches recover the treatment effect estimates well and generally provide larger uncertainty estimates, as expected. Both Bayesian model averaging approaches offer slightly better prediction of the propensity score compared with the Bayesian approach with a single propensity score equation. Covariate balance checks for the case study show that both Bayesian model averaging approaches offer good balance. The fully Bayesian model averaging approach also provides posterior probability intervals of the balance indices.