Towards modern, theory-driven approaches to grammar in aphasia

Organizers: Danielle Fahey1, Jeremy Yeaton2; 1University of Montana, 2University of California, Irvine
Presenters: Nicoletta Biondo, Jeremy Yeaton, Laura Giglio, Sladjana Lukic, Charalampos Themistokleous, Alex Krauska

Traditional approaches to grammatical deficits in aphasia harken back to categorical definitions conceived in the 19th century. Though the methodologies have improved, much research has continued to integrate these definitions. We seek to move beyond nebulous characterizations of comprehension and production deficits in aphasia—receptive and expressive agrammatism and paragrammatism—towards theory-driven, modern empirical approaches to understanding what underlies syntactic deficits in aphasia. The investigations presented ask the basic questions of whether there are distinct syndromes of agrammatism and paragrammatism, where in the brain and/or linguistic process deviations arise, whether the deviations are strictly syntactic, and what methodologies will best serve to improve our understanding of these conditions, and how the brain does syntax. Therefore, we draw together researchers examining the issue from diverse theoretical perspectives, disciplines and training backgrounds, and empirical evidence from patients with various diagnoses, etiologies and language experiences.


Bridging the gap: From linguistic theory to psycholinguistics in the assessment of post-stroke comprehension deficits

Nicoletta Biondo1,2; 1Basque Center for Cognition, Brain and Language, 2University of California, Berkeley

Why are we still debating grammatical deficits in aphasia? In this talk, I argue that this debate stems from a tension between the "applied" needs for swift language tools to assess deficits and guide treatment plans, and the "theoretical" needs for investigating grammar deficits in depth to provide a more reliable formalization of the neurobiology of language. I demonstrate the effects of this tension by first reviewing previous studies that have examined grammar deficits in post-stroke aphasia through lesion symptom mapping analyses. I then propose a sequential approach, where linguistic theory offers finer-grained analyses, and psycholinguistics provides mechanistic insights into how grammar is affected in aphasia, bridging linguistic concepts with brain mechanisms. Finally, I outline future potential implementations, aiming to contribute both to more reliable formalizations of grammatical processing and improved tools for the assessment of grammatical abilities.

Characterization of expressive grammar deficits in aphasia using SUBA coding

Jeremy Yeaton1, Danielle Fahey2; 1University of California, Irvine, 2University of Montana

There are two main types of expressive syntactic deficit in aphasia: agrammatism and paragrammatism. Agrammatism is often characterized by reduction or simplification of morphosyntax, while paragrammatism is characterized by over and misuse of morphosyntax, rather than reductions. Paragrammatism has received less attention than agrammatism with no large-scale structural description available in the literature. We developed novel error coding, the Syntactic Utterance-Based Analysis (SUBA), based on mechanistic accounts of sentence production and focused on capturing breakdowns in morphosyntax and hierarchical structure, analyzing comparable discourse sample transcripts of people with aphasia and healthy controls. Productive errors were used to cluster participants, resulting in several distinct phenotypes. These phenotypes localize to distinct regions of the inferior frontal and posterior temporal cortex. Results highlight the importance of integrating linguistic theory into structural descriptions of disordered output, as well as how that can inform our understanding of the neurobiology of syntax.

Characterizing verb argument structure production in discourse in aphasia

Laura Giglio1, Julius Fridriksson1, Dirk B. Den Ouden1; 1University of South Carolina

Verb production and verb argument structure production are often affected in people with aphasia that present with grammatical deficits. Measuring verb argument structure production in spontaneous speech can help characterize grammatical deficits in aphasia, but it usually requires intensive manual coding. I will discuss the use of dependency parsers to automatically extract verb use and argument structure production in controls and people with aphasia. Dependency parsers allow for a fast and use-based quantification of verb choice and use by identifying verbs in sentences and their dependent relations. These automatically-derived measures identify differences across aphasia types classified by the Western Aphasia Battery-Revised diagnosis relative to baseline measures derived from a large control group from AphasiaBank. This approach also captures variability within aphasia types, thus potentially providing a tool to better characterize grammatical behavior from discourse in people with aphasia.

Neural correlates of morphosyntax: evidence from acquired and progressive aphasias

Sladjana Lukic1; 1Florida State University

A longstanding debate in neurolinguistics revolves around understanding how and where the brain integrates morphosyntactic elements to encode and decode language. Morphosyntactic deficits are the defining features of nonfluent subtypes of aphasia, in contrast to fluent subtypes which primarily exhibit lexical retrieval deficits regardless of the etiology of aphasia. Various perspectives challenge this traditional distinction, with some arguing that syntax is inseparable from lexical processing, while others advocate for a "syntax-centric" view. Over the last decade, we observed the flourishing of quantitative linguistic approaches aided by AI tools enabling analysis of morphosyntactic features in speech samples. This talk delves into lesion-mapping and neural disorders, addressing key questions in neurolinguistics: 1)Are there clear functional-anatomical dissociations? 2)What roles do other linguistic and non-linguistic capacities play in morphosyntactic processing? 3)How do task differences and subtypes impact morphosyntax performance? Answers to these inquiries offer a comprehensive framework for understanding the anatomical-functional roles of morphosyntax.

Automated rscores for dementia assessment using Open Brain AI

Charalampos Themistokleous1; 1University of Oslo

Mild Cognitive Impairment (MCI) can significantly impact language comprehension, creating challenges for patients. Developing tailored materials to assess language comprehension is crucial for effective care. This study assesses readability measures in selecting tailored assessment texts for evaluating language comprehension in patients with MCI. We employed Open Brain AI (OBAI)’s readability scoring, semantic analysis, and text simplification to select assessment items. These items undergo rigorous testing and validation with dementia patients to ensure their sensitivity and appropriateness for capturing language comprehension deficits. Although automated readability scores have been typically standardized based on students' educational levels, our results show that they can be employed to assess patients with dementia to support assessment and therapy. Standardized assessment materials based on their comprehension level enhance assessment accuracy and streamline the process, benefiting both clinicians and patients by facilitating timely diagnosis and tailored interventions.

Moving Away from Lexicalism in Aphasiology

Alex Krauska1; 1University of Maryland

Recent work in non-lexicalist linguistic theory asks for a deep re-thinking of the format of linguistic representations and the kinds of processes which operate over them. This shift has significant implications for how we think about language disorders, given that the aphasia literature is deeply steeped in a tradition where the grammar and the lexicon are fundamentally distinct. Reinterpreting clinical data in a non-lexicalist lens - where elements above and below the “word” level are retrieved and combined via the same mechanisms, where meaning, syntax, and form are fully independent data structures, and where linear order is not fully determined by the syntax - suggests that the surface form of speech errors and language impairments may not transparently reflect the underlying cause. I discuss the issues with lexicalism as they pertain to language production in aphasia, and how a non-lexicalist approach can be useful in the analysis of language disorders.

SNL Account Login

Forgot Password?
Create an Account


Abstract Submissions extended through June 10

Meeting Registration is Open

Make Your Hotel Reservations Now

2024 Membership is Open

Please see Dates & Deadlines for important upcoming dates.