Transformer‐based Deep Learning Architecture Improves Detection of Associations between Spontaneous Speech Language Markers and Cognition
Pooyan Mobtahej, Sam Gouron, Rojan Javaheri, Annelisse El‐Khoury, Anne‐Marie C Leiby, David L Sultzer, S. Ahmad Sajjadi · 2025
Read Transformer‐based Deep Learning Architecture Improves Detection of Associations between Spontaneous Speech Language Markers and Cognition by Pooyan Mobtahej, Sam Gouron, Rojan Javaheri, Annelisse El‐Khoury, Anne‐Marie C Leiby, David L Sultzer, S. Ahmad Sajjadi (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/pmc-11716126.txt.
Description
Pooyan Mobtahej and co-authors present a 2025 research paper on using everyday speech to measure thinking skills. Speech is cheap and easy to record, but scoring it by hand takes a long time. The team built a computer model based on the transformer design to turn speech recordings into text and pull out language features. Their data came from a memory clinic group at the University of California, Irvine.
The study covered 73 healthy adults and 12 people with mild cognitive impairment or dementia. The model read the transcripts and measured five language features, then compared them with scores on a standard test of overall cognition. Richer vocabulary, longer average words, and more coherent speech each tracked with better test scores. The authors argue that automatic speech analysis could become a low-cost way to spot early decline.
Catalog record
| CREATOR | Pooyan Mobtahej, Sam Gouron, Rojan Javaheri, Annelisse El‐Khoury, Anne‐Marie C Leiby, David L Sultzer, S. Ahmad Sajjadi |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 3 KB |
| VERSION | 1.0 |
| LICENSE | CC BY 4.0 |
| RIGHTS | CC BY |
| SOURCE | Europe PMC |