Automatic analysis of speech and language for detecting signs of dementia

Main Investigators

Bahman Mirheidari

Heidi Christensen

Dan Blackburn

Markus Reuber

Annalena Venneri

About the project

Early diagnosis of dementia is difficult. Currently, when a person with memory-worries sees their GP, they will be asked about their memory and given pen-and-paper memory tests.

These tests are not accurate, and the GP is likely to lack expertise to tell who is developing dementia. As a result, GPs refer too many patients, who are not developing dementia, for specialist assessment, or they fail to identify patients who are developing dementia. This is costly, causes unnecessary worry and delayed diagnosis.

This project is developing tools to help GPs make better referral decisions. It is based on the analysis of the speech and language of patients with memory problems. The production of spoken language and interaction in conversation are complex cognitive achievements, and an extensive body of literature describes the progressive lexical, grammatical and phonological changes associated with neurodegenerative disorders. We have shown that we can accurately predict whether a patient has got Functional Memory Disorder (FMD) or Neurodegenerative Dementia (ND) by extracting information from patient-neurologist conversations reminiscent of the typical history-taking part of a consultation.