Computational modeling of stuttering - Supplementary material

Brain and Language, 2013

Atypical white-matter integrity and elevated dopamine levels have been reported for individuals who stutter. We investigated how such abnormalities may lead to speech dysfluencies due to their effects on a syllable-sequencing circuit that consists of basal ganglia (BG), thalamus, and left ventral premotor cortex (vPMC). ''Neurally impaired'' versions of the neurocomputational speech production model GOD-IVA were utilized to test two hypotheses: (1) that white-matter abnormalities disturb the circuit via corticostriatal projections carrying copies of executed motor commands and (2) that dopaminergic abnormalities disturb the circuit via the striatum. Simulation results support both hypotheses: in both scenarios, the neural abnormalities delay readout of the next syllable's motor program, leading to dysfluency. The results also account for brain imaging findings during dysfluent speech. It is concluded that each of the two abnormality types can cause stuttering moments, probably by affecting the same BGthalamus-vPMC circuit.

2009

Speakers plan the phonological content of their utterances prior to their release as speech motor acts. Using a finite alphabet of learned phonemes and a relatively small number of syllable structures, speakers are able to rapidly plan and produce arbitrary syllable sequences that fall within the rules of their language. The class of computational models of sequence planning and performance termed competitive queuing (CQ) models have followed Lashley (1951) in assuming that inherently parallel neural representations underlie serial action, and this idea is increasingly supported by experimental evidence. In this paper we develop a neural model that extends the existing DIVA model of speech production in two complementary ways. The new model includes paired structure and content subsystems (cf. MacNeilage, 1998) that provide parallel representations of a forthcoming speech plan, as well as mechanisms for interfacing these phonological planning representations with learned sensorimotor programs to enable stepping through multi-syllabic speech plans. On the basis of previous reports, the model's components are hypothesized to be localized to specific cortical and subcortical structures, including the left inferior frontal sulcus, the medial premotor cortex, the basal ganglia and thalamus. The new model, called GODIVA (Gradient Order DIVA), thus fills a void in current speech research by providing formal mechanistic hypotheses about both phonological and phonetic processes that are grounded by neuroanatomy and physiology. This framework also generates predictions that can be tested in future neuroimaging and clinical case studies.

Speech Communication, 1987

A theoretical account of stuttering is presented in which an inadequacy of neuronal resources for sensory-motor information processing is seen as the basis of the disorder. It is proposed that stutterers are deficient in the processing resources normally responsible for determining and adaptively maintaining the internal models which subserve speech production. A general description of such computational processes is detailed in the form of circuitry for an adaptive controller which can calibrate itself to control any variable, nonlinear, dynamic, multiple input, multiple output system. Zusammenfassung. Eine theoretische Darstellung des Stotterns wird pr~isentiert, in der die Unzulfinglichkeit neuronaler Ressourcen fiir senso-motorische Informationsverarbeitung als Basis der Strrung aufgefaBt wird. Als Aussage wird formuliert, dab Stotterer defizit~ir sind im Bereich der Verarbeitungsressourcen, die normalerweise die der Sprachproduktion zuarbeitenden internen Modelle bestimmen und adaptiv erhalten. Eine allgemeine Beschreibung solcher Rechenprozesse wird erl~iutert anhand eines Schaltkreises fiir einen adaptiven Regler, der fiber eine Selbstkalibrierung verfiigt und jedes variable, nicht-lineare und dynamische System mit Mehrfachein-und ausgang steuert. Rrsumr. Nous expliquons le b6gaiement par un drficit au niveau de la quantit6 de neurones pouvant vrhiculer l'information sensorimotrice. Nous pensons que le brgue manque des ressources de traitement nrcessaires au maintien et ~t l'adaptation des modules internes qui sous-tendent la production de la parole. Nous prrcisons les mrcanismes computationnels sous la forme d'un circuit d'un contr61eur adaptatif autorrgul6 pour le contr61e d'un systrme h sorties et h entrres dynamiques multiples, non-linraires et variables.

Journal of Fluency Disorders, 2010

This paper investigates the hypothesis that stuttering may result in part from impaired readout of feedforward control of speech, which forces persons who stutter (PWS) to produce speech with a motor strategy that is weighted too much toward auditory feedback control. Over-reliance on feedback control leads to production errors which if they grow large enough, can cause the motor system to "reset" and repeat the current syllable. This hypothesis is investigated using computer simulations of a "neurally impaired" version of the DIVA model, a neural network model of speech acquisition and production. The model's outputs are compared to published acoustic data from PWS' fluent speech, and to combined acoustic and articulatory movement data collected from the dysfluent speech of one PWS. The simulations mimic the errors observed in the PWS subject's speech, as well as the repairs of these errors. Additional simulations were able to account for enhancements of fluency gained by slowed/prolonged speech and masking noise. Together these results support the hypothesis that many dysfluencies in stuttering are due to a bias away from feedforward control and toward feedback control.

Journal of Neurolinguistics, 1997

Journal of Memory and Language, 1990

Using a response-priming procedure, five experiments examined the effects of vowel similarity on the motor programming of spoken syllables. In this procedure, subjects prepared to produce a pair of spoken syllables as rapidly as possible, but sometimes had to produce the syllables in reverse order instead. The spoken responses consisted of consonant-vowel-consonant (CVC) syllables whose medial vowels were /i/, /I/, /A/, and /a/. Performance was measured as a function of the phonetic relationship between the vowels in a syllable pair. Longer response latencies occurred for syllable pairs that contained similar vowels (e.g., /i/ and /I!) than for syllable pairs that contained dissimilar vowels (e.g., /i/ and /A/). This inhibitory vowel-similarity effect occurred regardless of whether the initial consonants of the syllables in a pair were the same or different. However, it decreased substantially when the final consonants of the paired syllables were different. These results suggest that a lateral-inhibition mechanism may modulate the motor programming of vowels during speech production. They also provide evidence for the integrity of vowel-consonant (VC) subunits in syllables. 8

In a speech production model proposed by Levelt a distinction is made between two routes of phonetic implementation in speech. A syllabary route is used to retrieve the stored motor programs for the most frequent syllables of a language, and segmentby-segment assembly is used for the implementation of low-frequency syllables. One of the predictions of the model is that there should be a difference in coarticulation between motor programs retrieved from the syllabary and programs that are computed online. In this paper we present two laboratory experiments and a corpus study on German which were designed to verify this prediction. Our results support the hypothesis that articulatory programs for high-frequency syllables are implemented differently than those for rare syllables.

2008

ABSTRACT It is tempting to think of the brain as functioning very much like a computer. Like the digital computer, the brain takes in data and outputs decisions and conclusions. However, unlike the computer, the brain does not store precise memories at specific locations. Instead, the brain reaches decisions through the dynamic interaction of diverse areas operating in functional neural circuits. The role of specific local areas in these functional neural circuits appears to be highly flexible and dynamic.

2002

white matter morphology, which is otherwise difficult Montreal Neurological Institute due to the lack of clear boundaries between adjacent McGill University white matter subregions. Voxel-based morphometry has Montreal, Quebec been used by other investigators to demonstrate norma-Canada H3A 2B4 tive brain asymmetries (Watkins et al., 2001), maturation of white matter tracts (Paus et al., 1999), structural correlates of arithmetic abilities (Isaacs et al., 2001), and dif-Summary ferences in brain morphology in normal versus clinical groups (Vargha-Khadem et al., 1998; Wright et al., 1995; We examined the relationship between brain anatomy Thompson et al., 2001). and the ability to learn nonnative speech sounds, as We trained 59 healthy individuals to distinguish a phowell as rapidly changing and steady-state nonlinguisnetic contrast not present in their native language: the tic sounds, using voxel-based morphometry in 59 dental and retroflex sounds used in Hindi. Previous funchealthy adults. Faster phonetic learners appeared to tional imaging work on phonetic perception has shown have more white matter in parietal regions, especially the involvement of several temporoparietal regions of in the left hemisphere. The pattern of results was simithe left hemisphere (Dé monet et al., 1994; Zatorre et al., lar for the rapidly changing but not for the steady-state 1992, 1996; Binder et al., 1996, 1997; Petersen et al., nonlinguistic stimuli, suggesting that morphological 1988; Paulesu et al., 1993), the superior temporal gyri correlates of phonetic learning are related to the ability bilaterally (Binder et al., 1994; Mazoyer et al., 1993; Jä nto process rapid temporal variation. Greater asymmecke et al., 1998; Mummery et al., 1999), and left inferior try in the amount of white matter in faster learners frontal regions in and around Broca's area (Zatorre et may be related to greater myelination allowing more al., 1992, 1996; Paulesu et al., 1993; Fiez et al., 1995; efficient neural processing, which is critical for the Burton et al., 2000). We predicted that phonetic learning ability to process certain speech sounds. measures would be correlated with differences in brain morphology in language-related cortical areas. In addi

Neurocase, 2014