Computer-based training of language production: An exploratory study

NUVA: A Naming Utterance Verifier for Aphasia Treatment

•

NUVA automatically assesses online word naming attempts in aphasia therapy.

•

Significantly more accurate and faster than leading commercial speech recognition.

•

Accuracies between 83.6% and 93.6% validate use in clinical research.

Anomia (word-finding difficulties) is the hallmark of aphasia, an acquired language disorder most commonly caused by stroke. Assessment of speech performance using picture naming tasks is a key method for both diagnosis and monitoring of responses to treatment interventions by people with aphasia (PWA). Currently, this assessment is conducted manually by speech and language therapists (SLT). Surprisingly, despite advancements in automatic speech recognition (ASR) and artificial intelligence with technologies like deep learning, research on developing automated systems for this task has been scarce. Here we present NUVA, an utterance verification system incorporating a deep learning element that classifies 'correct' versus' incorrect' naming attempts from aphasic stroke patients. When tested on eight native British-English speaking PWA the system's performance accuracy ranged between 83.6% to 93.6%, with a 10-fold cross-validation mean of 89.5%. This performance was not only significantly better than a baseline created for this study using one of the leading commercially available ASRs (Google speech-to-text service) but also comparable in some instances with two independent SLT ratings for the same dataset.

Propositional speech in unselected stroke: The effect of genre and external support

Distinguished from nominal language, propositional language generation refers to the spontaneous and voluntary aspect of language that introduces novel concepts to a specific context. Propositional language can be impaired in a range of neurological disorders, including stroke, despite well-preserved nominal language. Although external support can increase speech rate in patients with reduced propositional speech, no specific investigation of propositional speech has been carried out in unselected stroke patients. The current study investigated propositional language in an unselected post-acute stroke group (N = 18) with mild cognitive impairment and prominent executive dysfunction, but without significant aphasia. Specifically, we investigated whether genre or external support affected the number of words, sentences, and novel ideas produced, compared to healthy controls (N = 27). Results showed that discourse genre was not associated with differential performances. By contrast, speech quantity increased without external support although, for stroke patients, speech novelty decreased. Overall, the novelty deficit in unselected stroke patients highlights the importance of assessing cognition and propositional speech. Our findings suggest that for stroke patients with mild cognitive deficits, including executive dysfunction, introducing external support improved speech quality but not quantity. Implications for both assessment and rehabilitation of social communication are discussed.

The Best of Both Worlds: Combining Synchronous and Asynchronous Telepractice in the Treatment of Aphasia

Telepractice is an appropriate model of service delivery for a person with aphasia (PWA). We define telepractice and its two modes of delivery (i.e., synchronous and asynchronous). We detail a technology, web-Oral Reading for Language in Aphasia (web-ORLA™), developed to provide aphasia treatment intensively from a distance and subsequently evaluated during a clinical trial. In this article, we describe our experiences using web-ORLA™, highlighting the role of the speech-language pathologist (SLP) and discussing the advantages and disadvantages of this unique combination of synchronous and asynchronous telepractice. Web-ORLA™ was delivered to PWAs in their homes by a digital agent, or “virtual therapist,” who served as a model and provided instructions similarly to a real therapist. An SLP at a distant geographical location monitored the sessions remotely, either synchronously or asynchronously, provided feedback, made appropriate adjustments to the difficulty level of the stimuli, and conducted weekly probe assessments of the participants' performance. Advantages of web-ORLA™ include increased practice, SLP oversight, guidance by the agent, program simplicity, and a level of autonomy and flexibility afforded to the PWA. Given the rapid advances in technology, current technological problems that were encountered are likely to be mitigated.

Implementation of computer-based language therapy in aphasia

A first step in evaluating the use of computers in language therapy for individuals with aphasia is to establish the treatment as active in small groups prior to large-scale clinical trials. The present study evaluated a comprehensive computer-based language therapy program in a group of eight individuals with chronic in aphasia varying broadly in age, time post onset and aphasia type. Results revealed an overall therapeutic benefit in auditory comprehension, as well as positive trends in functional communication. Findings suggest that comprehensive therapy programs may be beneficial for many individuals with aphasia, and computer-based therapy may be one feasible avenue of providing this intervention.

Sentactics®: Computer-Automated Treatment of Underlying Forms

BACKGROUND: Treatment of Underlying Forms (TUF) is a linguistically-based treatment for improving agrammatic sentence deficits, which enjoys a substantial database attesting to its efficacy for improving both sentence comprehension and production in agrammatic aphasia. However, TUF requires considerable linguistic background to administer and administration time can exceed the number of treatment sessions allotted in toto for reimbursement by third party payors in the United States. Thus, Sentactics®, an interactive computer system that enables delivery of TUF by a virtual clinician was developed. AIMS: This study tested the effects of Sentactics® on the acquisition and generalized production and comprehension of complex sentences. Additionally, a direct comparison of the results of computer-delivered Sentactics® and clinician-delivered TUF was undertaken. METHODS #ENTITYSTARTX00026; PROCEDURES: Twelve agrammatic aphasic speakers participated in the study, with six receiving Sentactics® and six serving as experimental controls, who received no treatment. All participants were administered pre- and post-treatment sentence comprehension and production tests and other measures to evaluate the effects of Sentactics®. Performance of the Sentactics® group also was compared to eight agrammatic patients who previously received clinician-delivered TUF treatment, identical to that delivered via Sentactics®, but with a human clinician. OUTCOMES #ENTITYSTARTX00026; RESULTS: Sentactics® significantly improved all six aphasic speakers' ability to comprehend and produce both trained and untrained, linguistically related, complex sentences as compared to six agrammatic control participants who did not receive Sentactics®. In addition, comparing the results of the Sentactics® to clinician-delivered TUF revealed no significant differences between approaches with regard to acquisition or generalization patterns. CONCLUSIONS: These data provide further support for the efficacy of TUF and demonstrate the viability of computer-delivered therapies in the field of aphasia treatment.

Aphasic speech with and without SentenceShaper: Two methods for assessing informativeness

BACKGROUND: SentenceShaper((R)) (SSR) is a computer program that is for speech what a word-processing program is for written text; it allows the user to record words and phrases, play them back, and manipulate them on-screen to build sentences and narratives. A recent study demonstrated that when listeners rated the informativeness of functional narratives produced by chronic aphasic speakers with and without the program they gave higher informativeness ratings to the language produced with the aid of the program (Bartlett, Fink, Schwartz, & Linebarger, 2007). Bartlett et al. (2007) also compared unaided (spontaneous) narratives produced before and after the aided version of the narrative was obtained. In a subset of comparisons, the sample created after was judged to be more informative; they called this "topic-specific carryover". AIMS: (1) To determine whether differences in informativeness that Bartlett et al.'s listeners perceived are also revealed by Correct Information Unit (CIU) analysis (Nicholas & Brookshire, 1993)-a well studied, objective method for measuring informativeness-and (2) to demonstrate the usefulness of CIU analysis for samples of this type. METHODS #ENTITYSTARTX00026; PROCEDURES: A modified version of the CIU analysis was applied to the speech samples obtained by Bartlett et al. (2007). They had asked five individuals with chronic aphasia to create functional narratives on two topics, under three conditions: Unaided ("U"), Aided ("SSR"), & Post-SSR Unaided ("Post-U"). Here, these samples were analysed for differences in % CIUs across conditions. Linear associations between listener judgements and CIU measures were evaluated with bivariate correlations and multiple regression analysis. OUTCOMES #ENTITYSTARTX00026; RESULTS: (1) The aided effect was confirmed: samples produced with SentenceShaper had higher % CIUs, in most cases exceeding 90%. (2) There was little CONCLUSIONS: That the percentage of CIUs was higher in SSR-aided samples than in unaided samples confirms the central finding in Bartlett et al. (2007), based on subjective judgements, and thus extends the evidence that aided effects from SentenceShaper are demonstrable across a range of measures, stimuli and participants (cf. Linebarger, Schwartz, Romania, Kohn, & Stephens, 2000). The data also attest to the effectiveness of the CIU analysis for quantifying differences in the informativeness of aphasic speech with and without SentenceShaper; and they support prior studies that have shown that CIU measures correlate with the informativeness ratings of unfamiliar listeners.

A comparison between computer and tabletop delivery of phonology therapy

This paper reports on the development and evaluation of a software program aimed at assisting children with phonological impairment. An experimental approach was used whereby children's speech output was assessed pre- and post-therapy. Children were randomly assigned to a computer, a tabletop or a no therapy group. Those children receiving the computer therapy were exposed to an experimental software program that mirrored the tabletop activities using interactive computer games. The results showed no significant difference between any of the three groups with regard to change in speech output. These results may relate to the amount and frequency of therapy given and also to the heterogeneous nature of children included in the study. There was considerable variation in individual performance across all three groups and the data were therefore analysed to look for patterns that might predict performance. Stimulability and gender were identified as possible predictors. Female children and those who were able to produce a greater number of consonant speech sounds in isolation were more likely to make progress in their speech output. Future research might use a similar methodology to compare the therapy conditions but with a more homogenous group in terms of stimulability and using a greater intensity of intervention.

Widening the temporal window: processing support in the treatment of aphasic language production

Investigations of language processing in aphasia have increasingly implicated performance factors such as slowed activation and/or rapid decay of linguistic information. This approach is supported by studies utilizing a communication system (SentenceShaper) which functions as a "processing prosthesis." The system may reduce the impact of processing limitations by allowing repeated refreshing of working memory and by increasing the opportunity for aphasic subjects to monitor their own speech. Some aphasic subjects are able to produce markedly more structured speech on the system than they are able to produce spontaneously, and periods of largely independent home use of SentenceShaper have been linked to treatment effects, that is, to gains in speech produced without the use of the system. The purpose of the current study was to follow up on these studies with a new group of subjects. A second goal was to determine whether repeated, unassisted elicitations of the same narratives at baseline would give rise to practice effects, which could undermine claims for the efficacy of the system.

Informativeness ratings of messages created on an AAC processing prosthesis

BACKGROUND: SentenceShaper() (SSR) is a computer program that supports spoken language production in aphasia by recording and storing the fragments that the user speaks into the microphone, making them available for playback and allowing them to be combined and integrated into larger structures (i.e., sentences and narratives). A prior study that measured utterance length and grammatical complexity in story-plot narratives produced with and without the aid of SentenceShaper demonstrated an "aided effect" in some speakers with aphasia, meaning an advantage for the narratives that were produced with the support of this communication aid (Linebarger, Schwartz, Romania, Kohn, & Stephens, 2000). The present study deviated from Linebarger et al.'s methods in key respects and again showed aided effects of SentenceShaper in persons with aphasia. AIMS: Aims were (1) to demonstrate aided effects in "functional narratives" conveying hypothetical real-life situations from a first person perspective; (2) for the first time, to submit aided and spontaneous speech samples to listener judgements of informativeness; and (3) to produce preliminary evidence on topic-specific carryover from SentenceShaper, i.e., carryover from an aided production to a subsequent unaided production on the same topic. METHODS #ENTITYSTARTX00026; PROCEDURES: Five individuals with chronic aphasia created narratives on two topics, under three conditions: Unaided (U), Aided (SSR), and Post-SSR Unaided (Post-U). The 30 samples (5 participants, 2 topics, 3 conditions) were randomised and judged for informativeness by graduate students in speech-language pathology. The method for rating was Direct Magnitude Estimation (DME). OUTCOMES #ENTITYSTARTX00026; RESULTS: Repeated measures ANOVAs were performed on DME ratings for each participant on each topic. A main effect of Condition was present for four of the five participants, on one or both topics. Planned contrasts revealed that the aided effect (SSR >U) was significant in each of these cases. For two participants, there was also topic-specific carryover (Post-U >U). CONCLUSIONS: Listeners judged functional narratives generated on SentenceShaper to be more informative than comparable narratives spoken spontaneously. This extends the evidence for aided effects of SentenceShaper. There was also evidence, albeit weaker, for topic-specific carryover, suggesting that the program might be used effectively to practise for upcoming face-to-face interactions.