Best Paper Award - Pervasive Health 2018

Publish on 22/01/2020
Alexandra König (Stars Team) received a "Best paper award" for her paper "Telephone-based Dementia Screening I: Automated Semantic Verbal Fluency Assessment" à la conférence "Pervasive Health" in New York (United States).

Could you tell us a few words about the conference?

The overall goal of the Pervasive Health Conference is to take a multidisciplinary approach to Pervasive Healthcare Technology research and development. This conference aims to gather technology experts, practitioners, industry and international authorities contributing towards the assessment, development and deployment of pervasive medical based technologies, standards and procedures. This includes domains such as Sensing/Actuating Technologies and Pervasive Computing, Medicine, Nursing, and Allied Health Professions, Human-Computer Interaction (HCI) and Computer Supported Cooperative Work (CSCW) and Hardware and Software Infrastructures. The acceptance rate was this year at around 20%.

What research project is this paper part of? 

It is coming from and EIT (European Institute of Innovation and Technology) Digital Innovation Activity entitled ‘ELEMENT’. The Digital Wellbeing Action Line leverages digital technologies to stay healthy (prevention and early detection) or cope with an existing chronic condition. ELEMENT stands for Early detection of cognitive disorders on the basis of speech analysis  for which the project is developing an app that enables neuropsychological assessment for cognitive decline, such as dementia. As it is based on speech analysis in non-clinical settings, it results in faster and earlier diagnosis and intervention. For this paper, we examined the technologic feasibility of automatically assessing a neuropsychological test via a telephone-based solution. We investigate its suitability for inclusion into an automated dementia frontline screening and global risk assessment, based on concise telephone-sampled speech, ASR and machine learning classification with encouraging results. 


What are the current or future collaborations within the framework of this project? 

This project represents a strong collaboration between INRIA Sophia Antipolis, the German Research Center on Artificial Intelligence in Saarbruecken and the Cobtek Lab and Memory Clinic in Nice allowing currently a continuous stream of speech data collection and analysis for the further improvement of our detection results. 

If the project has an end, could there be follow-up? if so, in what form?

We hope to continue our ongoing work on speech analysis and apply it also to other populations and symptoms such as stress, pain, and anxiety for which we received already several requests from potential future partners such as pharmaceutical companies and clinical research . We are at the moment submitting proposals for future research projects together with these partners and at the same time since we founded a start up during this project, interested partners can purchase basically the services and analysis which will be another avenue for us to pursue. 

What about you? What is your background?  


I studied first psychology at the University of Montréal in Canada, and then moved back to Europe to undergo a Master programme in Neuropsychology at Maastricht University in the Netherlands followed there by a PhD for which I started working in parallel at the Memory Clinic and Alzheimer Research Center (CMRR) and the Cobtek (Cognition-Behaviour-Technology) Lab in Nice, France within the framework of the European FP7 project Dem@care. My thesis was on the use of ICT for the assessment of Alzheimer’s disease. After this, I went for a postdoctoral fellowship back to Canada, where I worked mainly in Toronto, at the Intelligent Assistive Technology and Systems Lab, focusing more on how to improve the adoption of assistive technologies in elderly people and namely dementia patients. 

How did you get to Inria? 

Since 2017, I am back in Nice again working now more focused on the use of speech analysis for detection and monitoring of cognitive decline in close collaboration with a computer linguistic team at the German Research center for artificial intelligence within a EIT (European Institute of Innovation and Technology) funded project called ‘ELEMENT’ and for which I was hired by INRIA. The Cobtek Lab consists of a research group of clinical (different hospitals in Nice) and technical researchers from INRIA, and for a few years now we are working closely hand in hand on project in the health technology field.  

What are your future projects?  

We founded a start-up called ki elements in order to commercialize the software for automatic speech analysis and detection of cognitive impairment. The telephone based screening scenario described in our paper attracted a few pharmaceutical companies already which are interested in testing out the solution for potential early screening and remote monitoring of people for clinical trials. Furthermore, we are applying for further projects with clinical partners all over Europe to be able to develop our application in several languages (already French, English, German available) and extent our data collection to improve detection accuracies.