AI and Aphasia in the Digital Age: A Critical Review.

Aphasiology has a long and rich tradition of contributing to understanding how culture, language, and social environment contribute to brain development and function. Recent breakthroughs in AI can transform the role of aphasiology in the digital age by leveraging speech data in all languages to model how damage to specific brain regions impacts linguistic universals such as grammar. These tools, including generative AI (ChatGPT) and natural language processing (NLP) models, could also inform practitioners working with clinical populations in the assessment and treatment of aphasia using AI-based interventions such as personalized therapy and adaptive platforms. Although these possibilities have generated enthusiasm in aphasiology, a rigorous interrogation of their limitations is necessary before AI is integrated into practice. We explain the history and first principles of reciprocity between AI and aphasiology, highlighting how lesioning neural networks opened the black box of cognitive neurolinguistic processing. We then argue that when more data from aphasia across languages become digitized and available online, deep learning will reveal hitherto unreported patterns of language processing of theoretical interest for aphasiologists. We also anticipate some problems using AI, including language biases, cultural, ethical, and scientific limitations, a misrepresentation of marginalized languages, and a lack of rigorous validation of tools. However, as these challenges are met with better governance, AI could have an equitable impact.

Spoken discourse in episodic autobiographical and verbal short-term memory in Chinese people with dementia: the roles of global coherence and informativeness.

Introduction: Memory and discourse production are closely related in healthy populations. A few studies in people with amnestic mild cognitive impairment and people with dementia (PWD) suggested similar links, although empirical evidence is insufficient to inform emerging intervention design and natural language processing research. Fine-grained discourse assessment is needed to understand their complex relationship in PWD. Methods: Spoken samples from 104 PWD were elicited using personal narrative and sequential picture description and assessed using Main Concept Analysis and other content-based analytic methods. Discourse and memory performance data were analyzed in bivariate correlation and linear multiple regression models to determine the relationship between discourse production and episodic autobiographical memory and verbal short-term memory (vSTM). Results: Global coherence was a significant predictor of episodic autobiographical memory, explaining over half of the variance. Both episodic autobiographical memory and vSTM were positively correlated with global coherence and informativeness, and negatively with empty speech indices. Discussion: Coherence in personal narrative may be supported by episodic autobiographical memory and vice versa, suggesting potential mechanism of interventions targeting personhood through conversation. Indices of global coherence, informativeness, and empty speech can be used as markers of memory functions in PWD.

Measuring pragmatic competence of discourse output among Chinese-speaking individuals with traumatic brain injury.

OBJECTIVE: Discourse analysis is one of the clinical methods commonly used to assess the language ability of individuals with traumatic brain injury (TBI). However, the majority of published analytic frameworks are not geared for highlighting the pragmatic aspect of discourse deficits in acquired language disorders, except for those designed for quantifying conversational samples. This study aimed to examine how pragmatic competence is impaired and reflected in spoken monologues in Chinese speakers with TBI. METHODS: Discourse samples of five tasks (personal narrative, storytelling, procedural, single- and sequential picture description) were elicited from ten TBI survivors and their controls. Each discourse sample was measured using 16 indices (e.g., number of informative words, percentage of local/global coherence errors, repeated words or phrases) that corresponded to the four Gricean maxims. Twenty-five naïve Chinese speakers were also recruited to perform perceptual rating of the quality of all 50 TBI audio files (five discourse samples per TBI participant), in terms of erroneous/inaccurate information, adequacy of amount of information given, as well as degree of organization and clarity. RESULTS: The maxim of quantity best predicted TBI's pragmatic impairments. Naïve listeners' perception of pragmatics deficits correlated to measures on total and informative words, as well as number and length of terminable units. Clinically, personal narrative and storytelling tasks could better elicit violations in pragmatics. CONCLUSION: Applying Gricean maxims in monologic oral narratives could capture the hallmark underlying pragmatic problems in TBI. This may help provide an additional approach of clinically assessing social communications in and subsequent management of TBI.

Sentence types and complexity of spontaneous discourse productions by Cantonese-speakers with traumatic brain injury- a preliminary report.

Previous investigations on sentence production in English-speaking individuals with traumatic brain injury (TBI) have yielded mixed conclusions based on their findings. While some studies found comparable sentence complexity between speakers with TBI and control speakers, others reported more syntactic and lexical errors, reduced sentence complexity, and erroneous word order transpositions in the sentence production of speakers with TBI. These contradictory findings could possibly be due to the use of language measures that were less sensitive to subtle syntactic impairments among speakers with TBI. In this preliminary report, the language samples obtained from 11 Cantonese-speaking participants with mild-moderate TBI in Guangzhou, with a mean age of 37.6 and mean years of education of 10 years, and nine control speakers with a similar age range and education background were analyzed using in-depth linguistic-oriented frameworks adopted from pervious works in Cantonese. The results indicated that the TBI group produced more errors, different varieties of sentence types, and lower syntactic complexity in their sentence production compared with the control group. The findings suggested that the more refined and linguistic-oriented measures used in the present study were more sensitive in identifying the subtle syntactic impairments produced by the participants with TBI.

Adaptation and validation of the main concept analysis of spoken discourse by native Japanese adults.

Many people with aphasia demonstrate problems of oral production at the discourse level. The Main Concept Analysis (MCA) for oral discourse production is a published evidence-based battery for quantifying the degree of presence, accuracy, completeness, and efficiency of targeted main concepts in oral discourse. In Japan, such a standardized tool specialized for assessing spoken discourse is currently lacking. The purpose of this study was to adapt the Japanese version of MCA for oral discourse production (the Japanese-MCA) and examine its validity and reliability. Stage 1 of the study involved the establishment of linguistically-specific main concepts (MCs) of the Japanese-MCA. Ten speech-language-hearing therapists and 60 healthy participants who were native monolingual Japanese speakers were recruited to determine MCs. Stage 2 examined the criterion validity and reliability of the Japanese-MCA. Language samples of 20 participants with aphasia, as verified by Standard Language Test of Aphasia (SLTA), and 20 healthy older participants were used. Results of Stage 1 of the study yielded normative data with a set of target MCs that were geographically and linguistically specific for use in Japan. The results also revealed the comparability of the Japanese-MCA and previously reported versions of other languages. Stage 2 findings indicated not only a high correlation of criterion validity, but also good reliability of the test. With established norms and specific scoring criteria of the Japanese-MCA, it is believed that this new tool will become a useful addition to clinical management and research of aphasia in Japan.

COVID-19 and Aphasia.

PURPOSE OF REVIEW: Aphasia is an acquired neurological language disorder after brain damages. Persons with aphasia (PWA) are more susceptible to behavioral and emotional implications due to inherent communication and/or cognitive difficulties. Currently, little is known regarding the impact of COVID-19 on PWA. RECENT FINDINGS: There are now growing reports with evidence of neurological and dysexecutive syndromes subsequent to interference of brain functions in acute patients with COVID-19, leading to variable aphasia-like symptoms. COVID-19 affected chronic PWA more in terms of disrupted communication and daily routines, worsened psychosocial well-being, and difficulties getting aphasia services that adequately addressed their needs. Acute versus chronic PWA were disproportionately affected by COVID-19. Recognizing, examining, and managing COVID-19-related neurological and behavioral problems in PWA is not straightforward. As we passed the 1-year mark and approaching the 2-year mark of the onset of COVID-19, more research is necessary to prioritize strategies for improving current evidence-based care and rehabilitation of aphasia.

The Dutch Main Concept Analysis: Translation and Establishment of Normative Data.

Purpose The main concept analysis (MCA; Kong, 2009) quantifies the effectiveness and efficiency of information transfer during verbal discourse by means of four sets of sequential pictorial stimuli. This test was originally developed for a Cantonese-speaking population. The main goals of this study were (a) to translate and adapt the MCA to Dutch; (b) to establish normative data for healthy native Dutch-speaking adults; (c) to assess the effect of age, education level, and gender on MCA outcome; and (d) to establish inter- and intrarater reliability of the Dutch MCA. Method Language samples were collected from 60 healthy native Dutch speakers, equally recruited in different age (20-39 years, 40-59 years, 60-79 years) and education (middle and highly educated) categories through administration of the MCA. Utterances produced by at least 75% of the participants were included as a main concept in the Dutch MCA. Subsequently, age-specific normative data were established for each of the MCA parameters. Finally, an ICC was calculated in order to verify inter- and intrarater reliability of the Dutch MCA. Results The translated MCA consisted of 19 main concepts. Age-specific normative data were obtained. Both age and education level had a significant effect on MCA outcome. Information transfer in elderly was both less effectively and efficiently compared to young- and middle-aged adults. In addition, highly educated participants transferred information less efficiently compared to middle educated participants. Based on inter- and intrarater reliability measures, the Dutch MCA proved to be a reliable measuring instrument. Conclusions The MCA was translated to Dutch, and age-specific normative data were established for a healthy, Dutch-speaking population. The Dutch MCA is a reliable tool for eliciting and quantifying discourse production. Validation of the test for people with aphasia is necessary in order for the test to be useful in a clinical practice.

Systematic Review of Training Communication Partners of Chinese-speaking Persons With Aphasia.

OBJECTIVE: To determine the aims, participants, and outcomes of training communication partners of Chinese-speaking persons with aphasia (PWA). DATA SOURCES: Sixty search terms related to communication partner training (CPT) in Chinese characters were searched in 8 electronic databases (published 1991-2020). STUDY SELECTION: Journal articles written in Chinese that primarily target the Chinese audience and university theses were selected for review. Studies involving CPT and training to enable communication partner to deliver language tasks were included, but reports without PWA or direct training of communication partners were excluded. A final corpus of 37 articles, representing publications of group studies, case studies, qualitative studies, and opinion articles, were selected for full review. DATA EXTRACTION: For all articles, 2 reviewers independently reviewed abstracts, excluding those without PWA or those that did not involve training of communication partners. One reviewer extracted descriptive data of participants with aphasia, communication partners, intervention details of the intervention (purpose, amount, setting, description), outcome measures, results, and clinical guidelines. A second reviewer performed accuracy verifications. DATA SYNTHESIS: Quality of reviewed articles were classified using the American Academy of Neurology levels of evidence. The current review suggested an evidence base of low to medium quality supporting 2 intervention groups: (1) training partners to deliver therapy tasks and (2) training to improve communication between PWA and their communication partners. There was a higher proportion of persons with acute and subacute aphasia involved in these investigations, suggesting evidence on treatment efficacy of CPT in the acute stage. CONCLUSIONS: Additional high-quality research with a better methodological quality, for example, randomized controlled trials or experimental design, are required to strengthen the current evidence of CPT. This systematic review suggests that the inclusion of studies published in languages other than English may influence the findings of mainstream reviews relating to aphasia.

The Impact of COVID-19 on Speakers With Aphasia: What Is Currently Known and Missing?

Purpose Enhancing social participation and reducing emotional distress in persons with aphasia (PWA) are a critical rehabilitation goal. Social relationships and meaningful activities performed by PWA are also crucial to promote positive psychosocial well-being. As a precautionary measure specific to the COVID-19 pandemic, most PWA worldwide have generally followed the guidelines of going out less, restricting when and where to gather with friends and peers, reducing social activities, and maintaining appropriate social distance; these acts are contrary to the traditional principles of managing aphasia. This article aims to (a) highlight and add to our understanding of issues related to the impact of the currently evolving COVID-19 pandemic on PWA, (b) direct readers to relevant reports in the literature of telerehabilitation for aphasia to look for useful information regarding remote assessment and therapy to be considered during the pandemic, (c) summarize support initiatives developed and resources compiled thus far as well as provide links for caregivers and PWA to find more information about COVID-19 in their communities, and (d) offer recommendations to potentially move the field of aphasia research and clinical PWA services forward in a positive way to endure the pandemic and in the forthcoming post-COVID world. Conclusions At present, there are significant knowledge gaps regarding the short and long-term impacts of COVID-19 on PWA and their caregivers. It is crucial that different stakeholders be sensitive and flexible when addressing the psychosocial and rehabilitation needs of PWA to mitigate the negative effects during and after the COVID-19 era.

The Effect of Metalinguistic Sentence Combining on Eighth-Grade Students' Understanding and Written Expression of Comparison and Contrast in Science.

Purpose The purpose of this study was to examine whether sentence combining with an explicit metalinguistic approach in comparison to typical science instruction was effective in improving written expression and understanding of comparison/contrast in science for eighth-grade students who struggle with literacy. Method Eighty-four eighth-grade students who struggle with literacy participated in this study. The experimental group (n = 36) received the writing intervention of metalinguistic sentence combining (MSC) during their science class for a total of 400 min (20 intervention sessions, 20 min each), while the comparison group (n = 48) participated in their typical science instruction. Total science instruction time was held constant for both groups. All students completed pretests and posttests to determine an increase in (a) syntactic factors of academic science text such as longer sentence length and use of syntactic forms of connectives, targeted connectives, left embeddedness, and agentless passive voice when responding to a science compare and contrast writing prompt; and (b) listing similarities and differences between two science concepts on a graphic organizer. Results Treatment was effective in improving the experimental group's score in listing similarities and differences between two science concepts on a graphic organizer. There were no significant differences between the two groups in their use of syntactic factors typical of academic text when responding to a science compare and contrast writing prompt. Conclusions MSC was effective in improving the experimental student's ability to demonstrate understanding of comparison and contrast in science. Modifications to the MSC intervention may yield better results in the experimental group's posttreatment writing in future studies. Supplemental Material https://doi.org/10.23641/asha.12735950.

Automatic Assessment of Speech Impairment in Cantonese-speaking People with Aphasia.

Aphasia is a common type of acquired language impairment resulting from dysfunction in specific brain regions. Analysis of narrative spontaneous speech, e.g., story-telling, is an essential component of standardized clinical assessment on people with aphasia (PWA). Subjective assessment by trained speech-language pathologists (SLP) have many limitations in efficiency, effectiveness and practicality. This paper describes a fully automated system for speech assessment of Cantonese-speaking PWA. A deep neural network (DNN) based automatic speech recognition (ASR) system is developed for aphasic speech by multi-task training with both in-domain and out-of-domain speech data. Story-level embedding and Siamese network are applied to derive robust text features, which can be used to quantify the difference between aphasic speech and unimpaired one. The proposed text features are combined with conventional acoustic features to cover different aspects of speech and language impairment in PWA. Experimental results show a high correlation between predicted scores and subject assessment scores. The best correlation value achieved with ASR-generated transcription is .827, as compared with .844 achieved with manual transcription. The Siamese network significantly outperforms story-level embedding in generating text features for automatic assessment.

RELEASE: a protocol for a systematic review based, individual participant data, meta- and network meta-analysis, of complex speech-language therapy interventions for stroke-related aphasia.

Background: Speech and language therapy (SLT) benefits people with aphasia following stroke. Group level summary statistics from randomised controlled trials hinder exploration of highly complex SLT interventions and a clinically relevant heterogeneous population. Creating a database of individual participant data (IPD) for people with aphasia aims to allow exploration of individual and therapy-related predictors of recovery and prognosis. Aim: To explore the contribution that individual participant characteristics (including stroke and aphasia profiles) and SLT intervention components make to language recovery following stroke. Methods and procedures: We will identify eligible IPD datasets (including randomised controlled trials, non-randomised comparison studies, observational studies and registries) and invite their contribution to the database. Where possible, we will use meta- and network meta-analysis to explore language performance after stroke and predictors of recovery as it relates to participants who had no SLT, historical SLT or SLT in the primary research study. We will also examine the components of effective SLT interventions. Outcomes and results: Outcomes include changes in measures of functional communication, overall severity of language impairment, auditory comprehension, spoken language (including naming), reading and writing from baseline. Data captured on assessment tools will be collated and transformed to a standardised measure for each of the outcome domains. Conclusion: Our planned systematic-review-based IPD meta- and network meta-analysis is a large scale, international, multidisciplinary and methodologically complex endeavour. It will enable hypotheses to be generated and tested to optimise and inform development of interventions for people with aphasia after stroke. Systematic review registration: The protocol has been registered at the International Prospective Register of Systematic Reviews (PROSPERO; registration number: CRD42018110947).

Analysing coherence of oral discourse among Cantonese speakers in Mainland China with traumatic brain injury and cerebrovascular accident.

Purpose: Coherence can reflect subtle language deficits in individuals with traumatic brain injury (TBI) and cerebrovascular accident (CVA). This study aimed at investigating whether global and local coherence in Cantonese-speaking adults with CVA and TBI differ from non-brain-injured (NBI) speakers. Factors contributing to the coherence ratings and impacts of elicitation tasks on coherence were examined.Method: Two clinical groups with fluent aphasia (7 CVA and 11 TBI) and 18 controls matched in age and education, who were Cantonese speakers living in China participated. Language samples of single and sequential picture description and storytelling were elicited, and subsequently analysed on global and local coherence, content sequence, and informativeness.Result: TBI speakers had impaired global and local coherence, while CVA speakers had poor global coherence. Sequence of main events produced by the three groups correlated significantly with global coherence. Attention and visuospatial skills were also significantly related to global coherence in both clinical groups. Finally, impaired language integrity was associated with problems of local coherence.Conclusion: The results were consistent with previous studies. Linguistic deficits of coherence in discourse in the two clinical groups and possible impacts of elicitation tasks on the cognitive demands and coherence ratings were discussed.

Use of co-verbal gestures during word-finding difficulty among Cantonese speakers with fluent aphasia and unimpaired controls.

BACKGROUND: Co-verbal gestures refer to hand or arm movements made during speaking. Spoken language and gestures have been shown to be tightly integrated in human communication. AIMS: The present study investigated whether co-verbal gesture use was associated with lexical retrieval in connected speech in unimpaired speakers and persons with aphasia (PWA). METHODS & PROCEDURES: Narrative samples of 58 fluent PWA and 58 control speakers were extracted from Cantonese AphasiaBank. Based on the indicators of word-finding difficulty (WFD) in connected speech adapted from previous research, and a gesture annotation system with independent coding of gesture forms and functions, all WFD instances were identified. The presence and type of gestures accompanying each incident of WFD were then annotated. Finally, whether the use of gesture was accompanied by resolution of WFD, i.e., the corresponding target word could be retrieved, was examined. OUTCOMES & RESULTS: Employment of co-verbal gesture did not seem to be related to the success of word retrieval. PWA's naming ability at single-word level and their overall language ability (as reflected by the aphasia quotient of the Cantonese version of the Western Aphasia Battery) were found to be the two strongest predictors of success rate of resolving WFD. CONCLUSIONS: The Lexical Retrieval Hypothesis highlighting the facilitative functions of iconic and metaphoric gestures in lexical retrieval was not supported. Challenges in conducting research related to WFD, and the clinical implications in gesture-based language intervention for PWA were discussed.

Cantonese AphasiaBank: An annotated database of spoken discourse and co-verbal gestures by healthy and language-impaired native Cantonese speakers.

This article reports the construction of a multimodal annotated database of spoken discourse and co-verbal gestures by native healthy speakers of Cantonese and individuals with language impairment: the Cantonese AphasiaBank. This corpus was established as a foundation for aphasiologists and clinicians to use in designing and conducting research investigations into theoretical and clinical issues related to acquired language disorders in Chinese. Details in terms of the purpose, structure, and levels of annotation of the database (containing part-of-speech-annotated orthographic transcripts with Romanization and the corresponding videos) are described. The discussion presents the challenges of building a spoken database of a language that is not linguistically well-researched and that does not have a standardized written form for many of its lexical items, as well as presenting how these issues were addressed. Most importantly, the article highlights the potential of Cantonese AphasiaBank as a powerful research tool for linguists and psycholinguists.

An Integrative Analysis of Spontaneous Storytelling Discourse in Aphasia: Relationship With Listeners' Rating and Prediction of Severity and Fluency Status of Aphasia.

Purpose: This study investigated which of the three analytic approaches of oral discourse, including linguistically based measures, proposition-based measures, and story grammar, best correlated with aphasia severity and with naïve listeners' ratings on aphasic productions. The predictive power of these analytic approaches to aphasia severity and fluency status of people with aphasia (PWA) was examined. Finally, which approach best discriminated fluent versus nonfluent PWA was determined. Method: Audio files and orthographic transcriptions of the storytelling task "The Boy Who Cried Wolf" from 68 PWA and 68 controls were extracted from the Cantonese AphasiaBank. Each transcript was analyzed using these 3 systems. Results: The linguistic approach of discourse analysis best correlated with aphasia severity and naïve listeners' subjective ratings. Although both linguistically based and proposition-based measures significantly predicted aphasia severity, a subset of linguistic measures focusing on the quantity and efficiency of production were particularly useful for clinical estimation of the fluency status of aphasia. Conclusions: The linguistically based measures appeared to be the most clinically effective and powerful in reflecting PWA's performance of spoken discourse.

Verbal short-term memory and language impairments in Cantonese speakers after stroke.

PURPOSE: The study examined the relationship between verbal short-term memory (STM) and language impairment in Cantonese speakers after stroke. It is hypothesised that Cantonese speakers with left-hemisphere (LH) stroke would perform worse than those with right hemisphere (RH) stroke and normal controls. Specific linguistic factors of Cantonese might affect results in the tasks. METHOD: Fifteen participants with LH stroke, 10 with RH stroke and 25 healthy controls were tested with auditory-verbal immediate serial recall (ISR) tasks and auditory linguistic tasks. All stroke participants were assessed with the Cantonese version of Western Aphasia Battery (CAB). RESULT: The LH group performed significantly worse than the RH and healthy control groups in the auditory verbal ISR and auditory linguistic tasks. There were significant lexicality, frequency and imageability effects in most tasks. Auditory discrimination and word comprehension tasks, but not the auditory word recognition task had correlations with ISR tasks. CONCLUSION: Verbal STM and language performance of Cantonese-speakers with history of LH stroke were inferior to RH stroke and healthy controls. The effects of lexicality, word frequency and imageability on verbal STM memory performance were found. Cantonese tones have effects on performance in auditory word recognition task, similar to onset, nucleus and rime.

Measuring discourse coherence in anomic aphasia using Rhetorical Structure Theory.

PURPOSE: The existing body of work regarding discourse coherence in aphasia has provided mixed results, leaving the question of coherence being impaired or intact as a result of brain injury unanswered. In this study, discourse coherence in non-brain-damaged (NBD) speakers and speakers with anomic aphasia was investigated quantitatively and qualitatively. METHOD: Fifteen native speakers of Cantonese with anomic aphasia and 15 NBD participants produced 60 language samples. Elicitation tasks included story-telling induced by a picture series and a procedural description. The samples were annotated for discourse structure in the framework of Rhetorical Structure Theory (RST) in order to analyse a number of structural parameters. After that 20 naïve listeners rated coherence of each sample. RESULT: Disordered discourse was rated as significantly less coherent. The NBD group demonstrated a higher production fluency than the participants with aphasia and used a richer set of semantic relations to create discourse, particularly in the description of settings, expression of causality, and extent of elaboration. People with aphasia also tended to omit essential information content. CONCLUSION: Reduced essential information content, lower degree of elaboration, and a larger amount of structural disruptions may have contributed to the reduced overall discourse coherence in speakers with anomic aphasia.