Monitoring cognitive health through speech


Like a lighthouse providing early warning in the dark of night, changes in language can signal cognitive decline decades before symptoms of dementia become apparent. Currently, over 47.5 million individuals in the world have dementia, and that number is expected to triple by 2050.

At WinterLight Labs, we build technology that can quickly and accurately detect signs of cognitive impairment from a sample of speech. A short 1-5 minute snippet of speech can paint a picture about the speaker's cognitive state, including lexical diversity, syntactic complexity, semantic content, and acoustics.

Our software is based on years of experience and academic, peer-reviewed research. Our team continues to actively publish in computer science and neuroscience journals and conferences. Learn more about the research behind our speech and language analysis.

For more information about some exciting use cases of our technology, please take a look at our brochure.


Frank Rudzicz is an international expert on speech technology for individuals with speech disorders. He is the President of the international joint ACL/ISCA special interest group on Speech and Language Processing for Assistive Technology, a Young Investigator of the Alzheimer’s Society a Scientist at Toronto Rehab, and an Assistant Professor at the University of Toronto.

Liam Kaufman has published peer reviewed articles in cognitive neurology, human computer interaction and neuroscience. He is an experienced software developer, and has successfully launched Understoodit, a startup acquired by EventMobi, and featured in The Toronto Star, CTV News, and TechCrunch.

Katie Fraser has published a number of peer-reviewed papers in both computer science conferences and neuroscience journals about her research on the automatic detection of dementia, and post-stroke aphasia. She was awarded a Google Canada Anita Borg scholarship and an NSERC scholarship and was named a 2015 MIT Rising Star in Computer Science.

Maria Yancheva conducts research on the automatic detection of dementia longitudinally, for which she was awarded an NSERC scholarship. She has experience developing software for Scotiabank’s trade floor, and has co-founded GroceryGo, a startup featured on BlogTO.

WinterLight News

Selected publications

  • Fraser, K.C., Meltzer, J.A., Rudzicz, F. (2016). Linguistic features differentiate Alzheimer's from controls in narrative speech. Journal of Alzheimer's Disease.
  • Yancheva, M., Fraser, K., Rudzicz, F. (2015) Using linguistic features longitudinally to predict clinical scores for Alzheimer's disease and related dementias. 6th Workshop on Speech and Language Processing for Assistive Technologies, 11 September 2015, Dresden Germany.
  • Fraser, K.C., Meltzer, J.A., Graham, N.L., Leonard, C., Hirst, G., Black, S.E., and Rochon, E. (2014). Automated classification of primary progressive aphasia subtypes from narrative speech transcripts. Cortex, pp. 43-60.
  • Fraser, K., Rudzicz, F., Rochon, E. (2013) Using text and acoustic features to diagnose progressive aphasia and its subtypes. Proceedings of Interspeech 2013, pp. 2177-2181, 25-29 August, Lyon France.