# Assessing semantic encoding and decoding models in stroke-induced aphasia

> **NIH NIH F32** · UNIVERSITY OF TEXAS AT AUSTIN · 2024 · $73,408

## Abstract

Project Summary
People with aphasia struggle to translate their thoughts into language. One potential way to help people with
aphasia is using brain-computer interfaces (BCIs) that decode intended speech from brain recordings. Recent
studies have shown that continuous language can be decoded from semantic representations that encode the
meaning of intended speech. However, semantic decoding has only been demonstrated in neurologically
healthy participants, and current approaches do not accommodate the language comprehension impairments
that often accompany language production impairments in aphasia. The long-term goal of this proposal is to
develop BCIs that can improve communication in people with aphasia.
This study has three goals: 1) to adapt existing semantic decoding approaches for people with aphasia, 2) to
develop semantic decoding approaches that do not require any language training data from the person being
decoded, and 3) to involve people with aphasia in the design of BCIs. To accomplish these goals, ten
participants with aphasia and ten neurologically healthy participants will be recruited. Semantic decoders will
be trained on functional MRI (fMRI) responses while participants listen to stories and watch movies. Semantic
decoders will be tested on fMRI responses to perceived speech, perceived movies, and imagined speech.
Collaborative design workshops will be held to assess when and how participants with aphasia envision using
BCIs. These findings will evaluate the potential for using semantic decoding to improve communication in
people with aphasia.
This fellowship will provide the applicant with a unique interdisciplinary training experience, which will include
the development of the necessary skills for a) administering language assessments, b) conducting
neuroimaging experiments in participants with aphasia, and c) collecting and analyzing qualitative feedback
from participants with aphasia. The applicant's sponsors and collaborators will provide mentorship in the areas
of participant recruitment, language assessment, experimental design, functional neuroimaging, and thematic
analysis. Together, these experiences will prepare the applicant for a successful independent research career.

## Key facts

- **NIH application ID:** 11070609
- **Project number:** 1F32DC022178-01A1
- **Recipient organization:** UNIVERSITY OF TEXAS AT AUSTIN
- **Principal Investigator:** Jerry Tang
- **Activity code:** F32 (R01, R21, SBIR, etc.)
- **Funding institute:** NIH
- **Fiscal year:** 2024
- **Award amount:** $73,408
- **Award type:** 1
- **Project period:** 2024-09-01 → 2027-08-31

## Primary source

NIH RePORTER: https://reporter.nih.gov/project-details/11070609

## Citation

> US National Institutes of Health, RePORTER application 11070609, Assessing semantic encoding and decoding models in stroke-induced aphasia (1F32DC022178-01A1). Retrieved via AI Analytics 2026-05-24 from https://api.ai-analytics.org/grant/nih/11070609. Licensed CC0.

---

*[NIH grants dataset](/datasets/nih-grants) · CC0 1.0*