As the population of older adults is increasing in Belgium, Europe and around the world, the number of persons with dementia is also increasing significantly. It is estimated that currently around 132,000 persons in Flanders have dementia, a number that is estimated to increase to 188,000 by 2035. Dementia is characterized by a loss of cognitive functioning (memory, language, judgement, etc.). However, behavioral and psychological symptoms of dementia, also known as neuropsychiatric symptoms of dementia, are clinically as relevant as cognitive symptoms. One of the most important behavioral problems is agitation, which leads to significantly decreased quality of life, the (over)use of medication and is one of the main reasons for early institutionalization due to the burden put on the caregiver. However there today no other way than observer-based ratings to measure agitation. Thus, there is a clear need to measure agitation more directly, objectively and continuously and in this way also try to gain more insight into the mechanisms of agitation. This in turn will lead to better prevention of agitation and aggression and give way to better monitoring of the effect of (both pharmacological and non-pharmacological) interventions.
In this PhD research the PhD candidate will develop a methodology to measure, and even predict levels of agitation. Unfortunately, there is no direct observable metric or biomarker that can be linked to agitation. The approach to be investigated is to combine multiple physiological measurements with environmental and contextual measurements. These can come from a variety of devices but should be suitable for the target population. Within imec, there is a strong background in various physiological measurements using wearable and non-contact technology. These inputs (contextualized physiological data) will be used to develop a suitable model to measure and ultimately predict levels of agitation. To this end machine-learning algorithms will be developed. The measurement of agitation in essence can be seen as a classification problem based on multi-sensor input. In order to do so, the candidate will be involved in various data collection trials to generate a sufficiently large database. The candidate will then develop low-level signal analysis algorithms to extract relevant features. Finally, a personalized machine-learning model will be developed to classify levels of agitation.
- Algorithm development
- Digital signal processing
- AI & machine-learning
- Basic knowledge of human physiology
- Affinity for electronic devices
Required background: Computer Science, Electrical engineering, Data science, Strong in DSP, AI & machine-learning.
Type of work: 30% data collection with patients, 30% signal processing, 40% algorithm development/simulation/validation)
Supervisor: Chris Van Hoof
Daily advisor: Nick Van Helleputte
The reference code for this position is 2020-102. Mention this reference code on your application form.
Chinese nationals who wish to apply for the CSC scholarship, should use the following code when applying for this topic: CSC2020-51.