Not available / text

Not available / text

Not available / text

The current investigation examined perturbation and adaptation during language comprehension in young normal subjects. Using a dynamic system framework, induced instability was studied by increasing perceptual demand (compressed sentences), syntactic demand, or both. Two experiments were conducted, one behavioral and one using fMRI technology, to explore the relations between brain responses and behavior. This study examines if changes in rate of speech, syntax, or both induce an instability, or perturbation, with subsequent adaptation in which subjects regain a previous stable state. Dependent measures in the behavioral study were accuracy and reaction time based indices of perturbation, adaptation, and stability. Results of the behavioral study demonstrated that language comprehension can be perturbed by changes in syntactic complexity or syntactic + perceptual complexity. Further, it was found that subjects adapted to being perturbed. The more complex the stimulus, the longer it took for subjects to adapt. The second experiment used fMRI to measure brain activation associated with perturbation and adaptation of language. Several brain regions showed increases in activation with increasing complexity (i.e., perturbation). Some regions (e.g., the superior parietal lobule) appeared more active when the perturbation was perceptual and others (e.g., the left inferior frontal gyrus) more active when the perturbation was syntactic in nature. These regions either remained active during the adaptation process, or reduced in activation during adaptation suggesting a role specific role in perturbation. These results suggest that subjects develop and maintain a representation of either the syntactic frame (i.e., via priming), a conscious strategy for accommodating syntactic complexity, or rate normalization schema. Thus, the brain regions that remain active during adaptation may be used to maintain the linguistic or perceptual frame. Within a dynamic system framework, the development of these representations, which occurs over a few items, serves as an attractor to which subjects are drawn each time they are perturbed. Like other complex systems, once instability occurs, there needs to be a strong attractor state to pull subjects into stability that permits appropriate performance to continue.

The purpose of this study was to investigate the relationship between aphasia severity and cerebral perfusion and lesion size in stroke. Nine subjects with acute ischemic stroke were examined within 24 hours of symptom onset and six were reexamined at one-month post-stroke. Examination included aphasia testing, testing of face discrimination ability, administration of the National Institutes of Health Stroke Scale, and perfusion MRI (PI), diffusion MRI (DWI), and T2-weighted MRI (T2-MRI). Subjects with a variety of aphasia types and a large range of aphasia severity participated in the study. MR images were visually inspected to verify perfusion and diffusion abnormalities. Perfusion abnormality was quantified by calculating a perfusion signal ratio of the affected hemisphere over the whole image (left/whole = ratio). Lesion volume was calculated from the DWI and T2-MRI. A perfusion abnormality larger than a DWI lesion was observed in 8 of 9 subjects. Minimal lesions were observed on DWI in three of the subjects while their PI revealed significant perfusion abnormality. Correlation coefficients (Spearman) between aphasia severity and hypoperfusion were significant in the acute stage and again at one-month post-stroke. Five of six subjects that were reexamined at one-month post-stroke experienced significant aphasia recovery. Visual inspection of their PI scans suggests that aphasia recovery was accompanied by increase in cerebral perfusion. The correlation between aphasia severity and lesion size was not statistically significant in the acute stage or at one-month post-stroke. Consequently, it is probable that cerebral hypoperfusion is a better predictor of aphasia severity and recovery in early stroke than lesion volume.

<p> This study uses ultrasound to image onset velar stop consonant articulation in words. By examining tongue body placement, the extent of velar closure variation across vowel contexts provides for the measurement of anticipatory coarticulation while productions within the same vowel context provide measurement of extent of token-to-token variation. Articulate Assistant Advanced 2.0 software was used to semi-automatically generate midsagittal tongue contours at the initial point of maximum velar closure and was used to fit each contour to a curved spline. Patterns of lingual coarticulation and measures of speech motor stability, based on curve-to-curve distance (Zharkova, Hewlett, &amp; Hardcastle, 2011), are investigated to compare the speech of typically fluent speakers to the speech of people who stutter. Anticipatory coarticulation can be interpreted as a quantitative measure indicating the maturity of the speech motor system and its planning abilities. Token-to-token variability is examined from multiple velar vowel productions within the same vowel context, describing the accuracy of control, or stability, of velar closure gestures. Measures for both speaking groups are examined across the lifespan at stages during speech development, maturation, and aging. Results indicate an overall age effect, interpreted as refinement, with increased speech stability and progressively more segmental (less coarticulated) productions across the lifespan. A tendency toward decreased stability and more coarticulated speech was found for younger people who stutter, but this difference was small and absent among older adults. Outcomes of this study suggest the articulatory maturation trajectories of people who stutter may be delayed, but overall maturation of the speech mechanism is evident by older adulthood for typically fluent speakers and those who stutter. Applications to intervention are discussed in closing. </p>

<p> <i>Purpose:</i> The current study sought to investigate the relationship between two metrics of sentence intelligibility in adults with Parkinson&rsquo;s Disease (PD), Multiple Sclerosis (MS), and healthy controls. An objective measure of intelligibility, orthographic transcription, and a subjective measure of intelligibility, Visual Analog Scaling (VAS), were the two metrics of intelligibility examined. Areas of interest included 1) comparisons of the pattern of intelligibility change in transcription and VAS, 2) strength of the relationship between these two types of intelligibility measures, and 3) differences in intralistener and interlistener reliability between the two metrics. </p><p> <i>Methods:</i> 78 speakers and the speech samples reported in Tjaden, Sussman, and Wilding (2014) and Kuo, Tjaden, and Sussman (2014) were used in the current study. The pool of 78 speakers consisted of 32 healthy control speakers, 16 speakers with PD, and 30 speakers with MS. Speakers read Harvard Psychoacoustic Sentences in habitual, clear, fast, loud, and slow conditions. In Tjaden et al. (2014) and Kuo et al. (2014), 50 naive listeners used a VAS on a computer to estimate how much of the speaker&rsquo;s message was understood (e.g., from &lsquo;didn&rsquo;t understand anything&rsquo; to &lsquo;understand everything&rsquo;). In the current study, 50 naive listeners heard the same stimuli, but were instructed to type exactly what they heard. Responses were scored to obtain a percentage of key words transcribed correctly for each stimulus. Results from the current study were compared to results from the VAS task studies (Tjaden et al., 2014; Kuo et al., 2014) using descriptive statistics (e.g., mean, standard deviation, etc.), parametric statistics (e.g., multivariate linear model fit to the data in this repeated measured design), correlation analyses (e.g., between the two metrics), and metrics of reliability. </p><p> <i>Results and Discussion:</i> Results revealed that the pattern of transcription intelligibility scores was very similar to scaled intelligibility derived from VAS. However, transcription scores were higher in magnitude than the VAS scores. In addition, correlation analyses showed the two intelligibility measures were highly correlated. Last, both interlistener and intralistener reliability were marginally higher for the VAS reported in Tjaden et al. (2014) and Kuo et al. (2014) than for the transcription data in the current study. These results suggest that a less time-consuming task, such as the VAS task, may be a viable substitute for a more time-consuming transcription task when documenting intelligibility in a clinical population to obtain an overall metric of severity for tracking disease progression and/or treatment progress. </p>

Over the last decade, the internet industry has rapidly grown with regard to infrastructure and bandwidth. Widespread internet networks with large bandwidth connect people-to-people, people-to-machines, and machine-to-machine. Like other multimedia services, large bandwidth enables voice services to be provided over IP networks where network connectivity is not consistent. In this context, research on service quality monitoring is necessary to satisfy customers by providing consistent service quality. The major contribution of this dissertation is the development of three novel techniques to improve or measure voice quality over IP networks. This dissertation first addresses an adaptive playout buffer scheduling algorithm that enables systems to lossen delay jitter due to the legacy of packet-switched networks. The scheduling algorithm is operated by a desired quality of service, minimizing the end-to-end delay by adjusting playout delay times. Secondly, this dissertation also explores a parameter-based nonintrusive speech quality measure to monitor the quality of VoIP. During the lifetime of sound, the network parameters are estimated and used to predict the quality of speech. As a cognitive model, a machine-learning technique is exploited to map features in the feature space into the perceived speech quality scale space. Finally, this dissertation introduces a signal-based nonintrusive speech quality measure. Features for the proposed measurement are extracted from observations of the characteristics of natural speech sounds and artificial noises. The calculated features are mapped into the perceived speech quality scale. The proposed parameter-based measure achieves a high prediction accuracy while the signal-based measure reaches to a comparable performance to the official International Telecommunication Union (ITU) standard, P.563. The contributions described in this dissertation provides smart methodologies for monitoring or enhancement of VoIP service qualities. / text

published_or_final_version / Philosophy / Master / Master of Philosophy

This study investigated the fast-mapping ability of young children with normal language (NL) and specific language impairment (SLI). It compared their ability to fast-map semantic and lexical information in different conditions. Children had to fast map visual information only, visual plus non-linguistic auditory information, and visual plus linguistic auditory information. Children with SLI performed worse than children with NL overall. They showed specific deficits when the task did not meet their expectations and when they were asked to map phonologically infrequent linguistic information. A nonword repetition task was correlated with both semantic and lexical fast-mapping. The findings are discussed in light of their support for a limited capacity model of processing, and for the need to evaluate children with SLI for semantic deficits.