University Carlos III de Madrid
Advisor: Antonio Artés
Thanks to the advent of eHealth and mHealth technologies, abundant environmental information can be collected from patients in a non-invasive and passive manner. Using healthcare observations from electronic medical records and the digital footprint of the patients machine learning methods can be developed and applied to health data for the conscious assessment of the disease evolution. Data access to patients’ information will be done under all legal guarantees and respecting the privacy policies for health care data protection.
Mood related disorders, such as mood disorder, depression and elation, severely affect the everyday activities of individuals. Common symptoms are decreased energy, loss of interest or pleasure in hobbies and activities, moving or talking more slowly, difficulty sleeping, early-morning awakening or oversleeping. These symptoms vary over time as a function of the stage of the disease and the individual. This project aims to develop both supervised and unsupervised machine learning models to understand mental disease evolution based on behavioural data, collected through the a mobile app running in the background of the patients phone, and clinician assessments at the time of the follow-up visits of the patient, which provide ground truths. The methods to be developed have to cope with irregularly sampled, high dimensional and heterogeneous time series data.
- Developing network models of health trajectories
Host: University of Liege (virtually)
Date: Postponed to Spring 2021 due to COVID
- Inferring causal relationships from personalized health trajectories
Host: Siemens Healthcare GmbH
Planned date: December 2021 – February 2022
- Emese Sükei, Agnes Norbury, M Mercedes Perez-Rodriguez, Pable M Olmos, and Antonio Artés. (2021). Predicting Emotional States Using Behavioral Markers Derived From Passively Sensed Data: Data-Driven Machine Learning Approach. JMIR Mhealth Uhealth 2021; 9(3): e24465. https://mhealth.jmir.org/2021/3/e24465