Work Packages

WEALTH consists of 5 interrelated work packages (WPs). Each WP will be led by one partner country, co-led by another and all other partners will contribute to the WP.

WP1 Novel method development integrating sensor and EMA protocols for monitoring and understanding daily behaviour

WP2 Collection of data using accelerometer devices, wearables and EMA

WP3 Machine learning model development for prediction of physical and eating behaviours

WP4 Proof of concept and data analyses based on newly developed ML and EMA techniques and web-based infrastructure

WP5 Dissemination, guidelines and toolbox creation

Work Package 1

Novel method development integrating sensor and EMA protocols for monitoring and understanding daily behaviour

Lead: Tomas Vetrovsky (UHK, Czech Republic)

Co-leads: Janas Harrington UCC (Ireland), Greet Cardon (UGent, Belgium); All Partners

Objective 1.1: To develop and test a protocol for the measurement of PB and EB and the context in which they occur by combining EMA and sensor data in an independent feasibility sample in each WEALTH centre.

Work Package 2

Collection of data using accelerometer devices, wearables and EMA

Lead: Alan Donnelly (UL, Ireland)

Co-Leads: Tomas Vetrovsky (UHK, Czech Republic); All partners

Objective 2.1: To implement data collection studies for the development of ML models (WP3).

Objective 2.2: Collect data to be used in WP4 to determine the feasibility of the use of a combined ML and EMA measurement.

Work Package 3

Machine learning model development for prediction of physical and eating behaviours

Lead: Antje Hebestreit & Christoph Buck (BIPS, Germany)

Co-Leads: Alan Donnelly (UL, Ireland); All Partners

Objective 3.1: Training and validation of ML models to predict eating, and physical behaviours as well as EE derived from tri-axial accelerometry and further sensors from protocol and free-living data collected in WP2.

Objective 3.2: Development of web-based solution for collection and data processing of raw sensor data from wearable devices and opensource provision of ML algorithms and preliminary data analysis in order to classify PB and EB.

Work Package 4

Proof of concept and data analyses based on newly developed ML and EMA techniques and web-based infrastructure

Lead: Jean-Michel Oppert (UP13, Paris, France)

Co-Leads: Greet Cardon (UGent, Belgium); All Partners

Objective 4.1: To assess the feasibility and acceptability by citizens and usability by analysing entities of innovative measures developed in WEALTH.

Objective 4.2: To identify multi-behavioural patterns using devicebased PA measures predicted from ML algorithms (developed in WP3) and data provided by EMA (as collected in WP2).

Work Package 5

Dissemination, guidelines and toolbox creation

Lead: Alan Donnelly (UL, Ireland)

Co-Leads: Greet Cardon (UGent, Belgium); All Partners

Objective 5.1: To use both integrated and end-of-project knowledge translation strategies to disseminate the findings of WEALTH as

broadly as possible.

Objective 5.2: To encourage maximum use of this new knowledge and new methods, an online toolbox will be created.