By 2050, more than 150 million people will suffer from a neurodegenerative disorder, such as Alzheimer’s disease, Parkinson’s disease or frontotemporal dementia. This motivates the search for objective, affordable and scalable markers to improve early detection and monitoring of symptoms. Among these efforts, automated speech and language analysis (AAHL) stands out. In fact, multiple investigations support its potential to improve current evaluation procedures. Now, how to convert scientific results into clinical resources? How to go from paper to practical tools? From these questions emerged Toolkit to Examine Lifelike Language (TELL), an app to evaluate speech at the service of brain health.
Every neurodegenerative disorder impacts, in one way or another, various aspects of speech and language. Therefore, the AAHL reveals clinically useful patterns. To do this, it is enough to record the participants’ speech and extract acoustic and linguistic characteristics from the audios and their transcriptions, respectively. Multiple studies indicate that AAHL can identify and differentiate people with Alzheimer’s disease, Parkinson’s disease, and frontotemporal dementia in early stages; discriminate between subtypes of these disorders; and predict the severity and progression of your symptoms. However, the potential of AAHL is limited by several factors, as research protocols are highly heterogeneous, centralized systems for harmonized data management are lacking, and the underlying technologies are not available for clinical use.
TELL is a strategic effort to address such challenges. This web application offers an intuitive platform to collect, encrypt, analyze, download and visualize speech and language data for clinical and research purposes. The application includes a linguistic profile questionnaire and a general protocol with various tasks, which include routine description, video narration and text reading, among others. Each task involves different motor and cognitive demands. State-of-the-art algorithms are used to capture sensitive acoustic measures (such as speech rate, duration of silences, and tonal patterns), as well as word frequency, affective valence, and emotional magnitude. These tasks and metrics have been widely validated through interdisciplinary studies, including work on underrepresented populations such as Latinos. Audio data and automated transcripts are automatically analyzed and graphed for clinical use, and can be downloaded for offline analysis. Likewise, TELL complies with current regulations on the protection of confidential information. The app is free to use for research purposes.
The app is growing in scope and potential. It is currently used in 10 centers in six different countries, a number that is increasing. Tasks and metrics are available in English, Spanish and Portuguese, and new languages will be added based on the specific needs of each project. In January 2023, an investment of $500,000 has been secured (by SF500, an Argentine life sciences fund) to develop its version 2.0, which will include new metrics and added functions. In summary, with its current and future features, TELL is a promising initiative to advance equity in the global fight against neurodegenerative disorders.
To request access and for details, news and publications, visit the TELL website at https://tellapp.org/ .
