Karen Chenausky

Automatic babble recognition for early detection of speech related disorders

We have developed a program, the Early Vocalization Analyzer (EVA), that analyses digitized recordings of infant vocalizations. The purpose of such a system is to automatically and reliably screen infants who may be at risk for later communication problems. EVA applies the landmark detection theory of Stevens et al., for the recognition of acoustic features in adult speech, to detect syllables in vocalizations produced by typically developing six to thirteen month old infants. We discuss the differences between adult-specific code and code written to analyse infant vocalizations. In a validity test, EVA achieved 90% agreement in marking 128 landmarks commonly identified by two human judges, was often closer to one or both judges than the humans were to each other. In a second test EVA and a human judge had 86% agreement in identifying 150 landmarks.

Utilization of microprocessors in voice quality improvement: the electrolarynx

Electrolarynges are a common form of post-laryngectomy speech prosthesis. However, the quality of the resulting voice is not as acceptable or intelligible as normal speech, and technology has changed very little since it was introduced in the 1950s. In the past five years, speech science and microtechnology have produced some major advances. The chief of these involve digital signal processing techniques that allow noise reduction and clarity improvement despite the substantial differences between normal and electrolaryngeal speech. There are several avenues for enhancement of electrolaryngeal voice and speech, although some will require far more research before they can contribute to improving speech quality for electrolarynx users.

Auditory-motor mapping training: comparing the effects of a novel speech treatment to a control treatment for minimally verbal children with autism

This study compared Auditory-Motor Mapping Training (AMMT), an intonation-based treatment for facilitating spoken language in minimally verbal children with autism spectrum disorder (ASD), to a matched control treatment, Speech Repetition Therapy (SRT). 23 minimally verbal children with ASD (20 male, mean age 6;5) received at least 25 sessions of AMMT. Seven (all male) were matched on age and verbal ability to seven participants (five male) who received SRT. Outcome measures were Percent Syllables Approximated, Percent Consonants Correct (of 86), and Percent Vowels Correct (of 61) produced on two sets of 15 bisyllabic stimuli. All subjects were assessed on these measures several times at baseline and after 10, 15, 20, and 25 sessions. The post-25 session assessment timepoint, common to all participants, was compared to Best Baseline performance. Overall, after 25 sessions, AMMT participants increased by 19.4% Syllables Approximated, 13.8% Consonants Correct, and19.1% Vowels Correct, compared to Best Baseline. In the matched AMMT-SRT group, after 25 sessions, AMMT participants produced 29.0% more Syllables Approximated (SRT 3.6%);17.9% more Consonants Correct (SRT 0.5); and 17.6% more Vowels Correct (SRT 0.8%). Chi-square tests showed that significantly more AMMT than SRT participants in both the overall and matched groups improved significantly in number of Syllables Approximated per stimulus and number of Consonants Correct per stimulus. Pre-treatment ability to imitate phonemes, but not chronological age or baseline performance on outcome measures, was significantly correlated with amount of improvement after 25 sessions. Intonation-based therapy may offer a promising new interventional approach for teaching spoken language to minimally verbal children with ASD.

White Matter Integrity and Treatment-Based Change in Speech Performance in Minimally Verbal Children with Autism Spectrum Disorder

We investigated the relationship between imaging variables for two language/speech-motor tracts and speech fluency variables in 10 minimally verbal (MV) children with autism. Specifically, we tested whether measures of white matter integrity—fractional anisotropy (FA) of the arcuate fasciculus (AF) and frontal aslant tract (FAT)—were related to change in percent syllable-initial consonants correct, percent items responded to, and percent syllable insertion errors (from best baseline to post 25 treatment sessions). Twenty-three MV children with autism spectrum disorder (ASD) received Auditory-Motor Mapping Training (AMMT), an intonation-based treatment to improve fluency in spoken output, and we report on seven who received a matched control treatment. Ten of the AMMT participants were able to undergo a magnetic resonance imaging study at baseline; their performance on baseline speech production measures is compared to that of the other two groups. No baseline differences were found between groups. A canonical correlation analysis (CCA) relating FA values for left- and right-hemisphere AF and FAT to speech production measures showed that FA of the left AF and right FAT were the largest contributors to the synthetic independent imaging-related variable. Change in percent syllable-initial consonants correct and percent syllable-insertion errors were the largest contributors to the synthetic dependent fluency-related variable. Regression analyses showed that FA values in left AF significantly predicted change in percent syllable-initial consonants correct, no FA variables significantly predicted change in percent items responded to, and FA of right FAT significantly predicted change in percent syllable-insertion errors. Results are consistent with previously identified roles for the AF in mediating bidirectional mapping between articulation and acoustics, and the FAT in its relationship to speech initiation and fluency. They further suggest a division of labor between the hemispheres, implicating the left hemisphere in accuracy of speech production and the right hemisphere in fluency in this population. Changes in response rate are interpreted as stemming from factors other than the integrity of these two fiber tracts. This study is the first to document the existence of a subgroup of MV children who experience increases in syllable- insertion errors as their speech develops in response to therapy.

Auditory-Motor Mapping Training in a More Verbal Child with Autism

We tested the effect of Auditory-Motor Mapping Training (AMMT), a novel, intonation-based treatment for spoken language originally developed for minimally verbal (MV) children with autism, on a more-verbal child with autism. We compared this child’s performance after 25 therapy sessions with that of: (1) a child matched on age, autism severity, and expressive language level who received 25 sessions of a non-intonation-based control treatment Speech Repetition Therapy (SRT); and (2) a matched pair of MV children (one of whom received AMMT; the other, SRT). We found a significant Time × Treatment effect in favor of AMMT for number of Syllables Correct and Consonants Correct per stimulus for both pairs of children, as well as a significant Time × Treatment effect in favor of AMMT for number of Vowels Correct per stimulus for the more-verbal pair. Magnitudes of the difference in post-treatment performance between AMMT and SRT, adjusted for Baseline differences, were: (a) larger for the more-verbal pair than for the MV pair; and (b) associated with very large effect sizes (Cohen’s d > 1.3) in the more-verbal pair. Results hold promise for the efficacy of AMMT for improving spoken language production in more-verbal children with autism as well as their MV peers and suggest hypotheses about brain function that are testable in both correlational and causal behavioral-imaging studies.

Automatic methods to monitor the speech of Parkinson’s patients with deep brain stimulators.

Parkinson’s disease (PD) is a neurodegenerative disease causing hypokinetic dysarthria, associated with “blurred” or underarticulated speech, imprecise consonants, and, sometimes, irregular syllable trains. Within the past decade, deep brain stimulation (DBS) of the subthalamic nucleus (STN) has provided substantial benefit to PD patients. DBS treatment has largely been directed toward the motoric features of PD: bradykinesia, rigidity, and tremor, but its effects on speech vary. The speech of PD patients receiving DBS treatment, with or without accompanying medical therapy, was analyzed for rate (syllables per second), regularity (relative deviation of syllable length), stop consonant spirantization (a measure of stop consonant precision), vowel ratio (length of vowel to length of syllable), and other features using automatic routines written specifically for the purpose. Patients’ speech is more variable on DBS stimulation than on medication or no treatment. It is possible to find a combination of DBS s...

Preliminary comparison of infants speech with and without hearing loss

The speech of ten children with hearing loss and ten children without hearing loss aged 12 months is examined. All the children with hearing loss were identified before six months of age, and all have parents who wish them to become oral communicators. The data are from twenty minute sessions with the caregiver and child, with their normal prostheses in place, in semi‐structured settings. These data are part of a larger test battery applied to both caregiver and child that is part of a project comparing the development of children with hearing loss to those without hearing loss, known as the Early Development of Children with Hearing Loss. The speech comparisons are in terms of number of utterances, syllable shapes, and segment type. A subset of the data was given a detailed acoustic analysis, including formant frequencies and voice quality measures. [Work supported by NIDCD R01 006237 to Susan Nittrouer.]

Behavioral predictors of improved speech output in minimally verbal children with autism: Chenausky et al./Predictors of Speech Output in ASD

We investigated the relationship between eight theoretically motivated behavioral variables and a spoken‐language‐related outcome measure, after 25 sessions of treatment for speech production in 38 minimally verbal children with autism. After removing potential predictors that were uncorrelated with the outcome variable, two remained. We used both complete‐case and multiple‐imputation analyses to address missing predictor data and performed linear regressions to identify significant predictors of change in percent syllables approximately correct after treatment. Baseline phonetic inventory (the number of English phonemes repeated correctly) was the most robust predictor of improvement. In the group of 17 participants with complete data, ADOS score also significantly predicted the outcome. In contrast to some earlier studies, nonverbal IQ, baseline levels of expressive language, and younger age did not significantly predict improvement. The present results are not only consistent with previous studies showing that verbal imitation and autism severity significantly predict spoken language outcomes in preschool‐aged minimally verbal children with autism, but also extend these findings to older minimally verbal children with autism. Autism Res 2018, 11: 1356–1365.

From intuition to intervention: developing an intonation‐based treatment for autism

Autism affects ∼1.5% of children under age 8; its core symptoms include impairment in social‐communicative functioning and repetitive behaviors/restricted interests. Music‐based interventions have been considered one modality through which to treat autism. This report discusses considerations to take into account when developing a music‐based intervention for a core symptom of autism. Treatment modality must be matched to symptom both clinically and theoretically, the behavior to be treated must be carefully defined and assessed, and outcome measures must be capable of showing improvement in that behavior over the course of the study. Fidelity assessment and rater blinding reduce experimenter bias. High inter‐rater reliability for perceptually determined outcome measures helps obtain accurate estimates of treatment response. Later stages of testing compare the experimental intervention to matched control treatments or other validated therapies, isolating the interventions “active ingredients.” Such systematic investigation of a new music‐based intervention can provide information of different types, ranging from an assessment of whether the intervention has any effect at all to an assessment of its outcomes and risks in uncontrolled community settings. Findings ultimately compose the evidence base that clinicians and families can use to decide the most effective way of addressing symptoms of autism for particular children.

Effect of auditory-motor mapping training and speech repetition training on consonant and vowel accuracy in minimally verbal children with autism spectrum disorder

Various therapies exist for teaching first words to minimally verbal (MV) children with autism spectrum disorder (ASD). Previous outcome measures have focused on number of words imitated or produced spontaneously, or on communication rate. No studies thus far have examined phonetic accuracy in MV ASD as a result of therapy, yet understanding whether therapy improves articulation in MV ASD is important for understanding how speech is affected in ASD and how best to treat these children. In this preliminary study of 30 children with MV ASD, we report on perceptual analysis of their speech after 25 sessions of one of two therapies, employing bisyllabic stimuli with a variety of consonant types. Twenty-three children received Auditory Motor Mapping Training (AMMT), where stimuli are intoned at approximately one syllable per second. Seven children, matched on age and cognition to 7 AMMT participants, received Speech Repetition Training (SRT), where non-intoned stimuli are presented at a normal speech rate. ANO...