EVO+ 1.0

EVO+ 1.0 is a personalised, online 12-week behaviour change program designed to help adults become more physically active, especially those at risk of developing non-infectious chronic diseases.

The program adapts to each person’s needs, preferences, motivation, and confidence levels. Over 12 weeks, participants are guided step-by-step through setting goals, tracking their activity, as well as planning, and staying on track over time. The program is easy to use and is based on theory, methods, and prior evidence in behavioural science to make healthy habits easier to build and maintain.

Developed using behaviour change science, motivational communication, and user-centered design, the program tailors content based on users’ behavioral profiles, context, interests, type of behaviour to change and readiness to change.

Scientific Foundations of the Program

EVO 1.0 guides participants through a structured 12-week program integrating a range of evidence-based behaviour change techniques (BCTs) from the BCT taxonomy. Core modules include goal setting, action planning, self-monitoring, problem solving, relapse and lapse prevention, and long-term maintenance. The intervention architecture is grounded in the ORBIT model, the COM-B framework, and Self-Determination Theory, with content and progression dynamically adapted to user engagement, interests, and progress.

Current Phase 

Refining the content and structure of the EVO 1.0 intervention, including the personalized physical activity and behavior change modules, with preparations also underway for expert consensus meetings to validate the theoretical framework before moving to the proof-of-concept phase. Concurrently, collaborating with developers and UX/UI designers to integrate the modules into the digital platform and ensure a user-centered experience.

Call for Experts

Experts interested in contributing to this process are welcome to contact the Research Coordinator: laura.ortiz.cnmtl@ssss.gouv.qc.ca

EVO 1.0 Team 

• Simon Bacon , PhD – Co-supervisor, Professor at Concordia University
• Kim Lavoie , PhD – Co-supervisor, Professor at Université du Québec à Montréal
• Florence Coulombe Raymond, B.Sc. – Project Lead
• Fanie Saint-Jean Miron, M.Sc. Kin. – Project Co-lead
• Laura Lucia Ortiz, M.Ed, CAPM – Research Coordinator
• Hafedh Mili, PhD – Co- Supervisor, Professor at Université du Québec à Montréal
• Edilton Lima dos Santos, PhD – Application Development
• Julien Champagne, M. CompSci – Application Development
• Djawad Mohamed Abi Ayad,  M. CompSci  – Application Development

EVO+ Colombia

Constructs of behaviour change are universal, but context is not.

The EVO+ Colombia project is a collaborative effort between EAFIT University in Medellin, Colombia, and the EVO+ 1.0 team in Montreal, Canada.

Project objectives 

1. To adapt EVO1.0, specifically for the Colombian setting. This involves not only creating new, context-specific content but also testing its implementation rigorously across the province of Antioquia.
2. To develop a comprehensive scale-up strategy for the intervention’s implementation across all of Colombia. This strategy is designed with broader applicability in mind, serving as a blueprint that could be effectively utilised in other similar Low and Middle-Income Countries (LMIC) seeking to implement comprehensive digital behaviour change interventions.

By working closely with various healthcare and government officials in Colombia, we aim to develop and refine implementation strategies that resonate with the local context, ensuring the intervention’s relevance, accessibility, and ultimately, its success in fostering positive health behaviours.

Current ongoing activities

• Establishing an executive committee with potential users, specialists, and health organization leaders, and holding regular meetings every semester.
• Conducting focus groups with potential users and health workers in different subregions of Antioquia to assess how context influences health behaviors.
• Adapting existing nutrition workbooks and developing new ones in collaboration with potential users and nutritionists, ensuring alignment with local guidelines to make them relevant and appealing to the population.
• Interviewing potential stakeholders to design the implementation plan.
• Disseminating health content on social networks focused on behavioral change, healthy nutrition, and physical activity.

More online about the EVO+ Colombia project

• Follow the EVO+ Colombia project on Instagram 
• The EVO+ Colombia project as a Life Courses featured on the Global Alliance for Chronic Diseases website

Team Members (EAFIT University)  

• Mariantonia Lemos, PhD – Principal Investigator
• Elizabeth Rendon Velez, PhD – Co-investigator
• Ana Maria Peralta Velez, MSc – Research Assistant
• Gloria Carolina Usuga Arcila, MSc – Research Assistant
• Isabela Cano Aguilar, BA – Research Assistant
• Maria del Pilar Guevara Gonzalez, BA – Research Assistant
• Johanna Marcela Blanco, BA – Research Assistant
• Guillermo Alejandro Peña Gómez, BA – Communications
• Mariana Escudero Arbelaez, BSc – Nutritionist
• Margarita María Duque Trujillo, MA – UX/UI Designer

Automatic Expression Recognition (AER)

The AER project aims to develop a coding system for annotating multimodal cues that indicate ambivalence and hesitancy in videos of individuals discussing behaviours.

Preventing chronic diseases, and stopping them from getting worse, is possible when engaging in healthy behaviours (eating a balanced diet, doing regular physical activity, etc.), but changing a behaviour is not easy.

For most people, things get in the way of their process, such as wanting to exercise but not finding the time, or being unsure how exactly to go about having a healthier diet which can be signs of someone feeling ambivalent or hesitant.

When face-to-face with their healthcare providers, patients experiencing ambivalence have a chance to talk about it (and work on finding solutions), but digital health interventions, such as health apps, can’t tell when people are experiencing ambivalence or hesitancy.

Thanks to new technologies, computer systems can be created and trained to identify people’s emotional states and expressions. This hasn’t been done yet for ambivalence and hesitancy. The AER project is working to develop and apply codes that can help computer systems identify ambivalence and hesitancy. It is using videos of people talking about their behaviours to watch they facial, body, and verbal expressions. These tags will be used by the project’s computer science collaborators to help computer systems identify ambivalence and hesitancy. This work could ultimately be used to personalize digital health interventions, that is, to adapt it to someone’s specific needs at the right time and help them make positive behaviour changes.

Scientific Foundations of the Project

The AER project is a complex, multi-phase, intersectoral collaboration between the Montreal Behavioural Medicine Centre (MBMC), Concordia University, and École de technologie supérieure (ÉTS). Its goal is to develop a coding system for annotating multimodal cues that indicate ambivalence and hesitancy from videos of individuals discussing health behaviour change. The project also aims to create robust deep learning algorithms for detecting these expressions by providing a foundational dataset. These technologies may be integrated into existing e-health platforms to enable more personalized behaviour change pathways, enhancing user engagement and health outcomes.

Relevant publications:

• BAH Dataset for the Ambivalence/Hesitancy Recognition Challenge in the 8th ABAW
  • Published abstract about an academic challenge at the ABAW 8 conference where teams working in affecting computing accessed an annotated dataset containing ambivalence and hesitancy to create better (more efficient and confident) deep learning algorithms to identify A/H in audio visual data
• Advancements in Affective and Behavior Analysis: The 8th ABAW Workshop and Competition
  • Overview of all academic challenges and workshops for the ABAW8 conference, including the one from the AER team (BAH, Behavioural Ambivalence and Hesitancy). 
• Textualized and Feature-based Models for Compound Multimodal Emotion Recognition in the Wild
  • Produced by the AER team for a compound emotion recognition challenge at the 2024 ECCV conference. Teams were asked to find better ways to identify compound emotions. The AER team used a model that focused on language cues, facial expression and context cues, and used an expert annotated dataset to identify where compound emotions were present. 

Current Project Activities

• Online data collection – Creation of a diverse representative pan-Canadian sample (600 people across Canada)
• Data annotation – Codebook refinement
• Annotator training
• Deep learning algorithm development
• Behavioural Ambivalence/Hesitancy (BAH) dataset
• Academic challenges (ABAW-8 workshop)
• Ambivalence/Hesitancy Scoping reviews

Call for Volunteers

The AER team is currently recruiting volunteers to help annotate video data. Volunteers interested in contributing to this activity are welcome to contact the Research Coordinator: laura.ortiz.cnmtl@ssss.gouv.qc.ca

Team members 

• Simon Bacon , PhD – Professor at Concordia University
• Kim Lavoie , PhD – Professor at Université du Québec à Montréal
• Manuela González-González, MSc – Project lead 
• Jessica Almeida, BSc – Research Assistant and Expert annotator
• Laura Lucia Ortiz, MEd, CAPM – Research Coordinator and Expert Annotator
• Eric Granger, PhD – Professor at ÉTS and Director of LIVIA Lab
• Marco Pederesoli, PhD – Professor at ÉTS
• Alessandro Lameiras Koerich, PhD – Professor at ÉTS

News Coverage

• Concordia professor lands unique co-chair in artificial intelligence and health
• Changing Habits With Artificial Intelligence