arXiv:2606.15696v1 Announce Type: cross Abstract: Correct Information Units (CIUs) are central to discourse assessment in aphasia because they quantify communicative informativeness rather than linguistic form alone. However, CIU scoring is time intensive and requires trained raters. This study examined whether instruction-tuned large language models (LLMs) can reliably perform token-level CIU classification from aphasic discourse transcripts. Sixteen picture-description transcripts elicited with the Cat Rescue stimulus were annotated for CIU status according to Nicholas and Brookshire (1993).
The proliferation of more capable instruction-tuned large language models (LLMs) makes this kind of applied research possible and increasingly relevant for real-world applications.
This research explores the growing capability of AI to perform nuanced, specialized human tasks, potentially transforming fields like clinical assessment and healthcare support by automating time-intensive processes.
LLMs might reliably take on complex diagnostic or assessment tasks in healthcare, moving beyond general text generation to more precise information identification in specialized domains.
- · Healthcare providers
- · Patients with aphasia
- · AI developers
- · Clinical researchers
- · Human raters for CIU scoring
- · Traditional assessment methods
LLMs demonstrate proficiency in a specialized clinical assessment function, specifically identifying correct information units in aphasic discourse.
The automation of such assessments could significantly reduce costs and improve the speed of diagnosis and treatment planning for neurological conditions.
This could lead to a broader integration of AI into other sensitive, human-centric clinical assessments, raising questions about ethical oversight and the evolving role of human experts.
This signal links to a primary source. Continuum Brief monitors and indexes it as part of the live intelligence stream — we do not republish source content.
Read at arXiv cs.CL