In this new blog series, we give the floor to scientists engaged in technology related research and valorization activities within the sports technology and innovation field at Ghent University. In this way we want to provide insight into what is brewing behind the scenes. In short, we put the researcher under the microscope.
This blog is based on a chat with Kevin Caen.
Kevin is a postdoctoral researcher at the Department of Movement and Sport Sciences, currently working on a project situated on the border of general exercise physiology, training and sports analytics: PACE.
PACE: modelling real-time energy balance during exercise
PACE is a computational model designed to estimate an athlete’s real-time energy balance based on power output.
“During cycling, a power meter quantifies the mechanical power an athlete produces,” Kevin explains. “PACE uses that value as an input parameter to predict how quickly an athlete depletes energy above a certain intensity, and how efficiently they recover below it.”
The model captures the dynamic interplay between exertion and recovery using a series of physiological determinants linked to exercise thresholds. By quantifying this balance, PACE enables personalized pacing strategies based on an athlete’s moment-to-moment energetic state.
The foundation of PACE lies in Kevin’s doctoral work on exercise thresholds. “At the time, one model existed in the literature attempting to predict similar dynamics,” he recalls. “However, it was quite simplistic. That motivated us to test its validity and to explore how a more accurate and physiologically grounded model could be developed.”
Over several years, extensive laboratory testing was conducted on different groups of athletes, each study targeting a specific physiological component. These datasets were later combined to create a single, refined model containing multiple parameters and improved predictive accuracy.
Research coordination and ongoing developments
Since completing his PhD, Kevin has continued to coordinate a broad research programme around PACE. This includes the development, validation, and application of the model across different contexts, supported by several funded projects.
One recent study involved recreational cyclists performing a Zwift time trial with and without a PACE-based pacing plan. “The PACE-based strategy resulted in faster performances,” Kevin notes, demonstrating practical value beyond the laboratory setting.
In parallel, the team is developing a prototype mobile application that integrates power meter data and visualizes an athlete’s real-time energy balance outdoors. This tool is currently being prepared for pilot testing among recreational cyclists.
While the initial applications of PACE focused on elite sport, particularly track cycling, current research is extending into the healthcare domain.
In track cycling, Kevin and his colleagues support Belgian Cycling in optimizing pacing strategies for the team pursuit. “Since 2023, our team has provided data analysis and simulation support at major competitions in collaboration with Sport Vlaanderen,” he explains.
However, a potentially larger societal impact lies in cardiac rehabilitation.“Patients in cardio-rehabilitation already undergo exercise testing, which allows us to determine the threshold values required for the model,” Kevin says. “Current rehabilitation protocols can be conservative, often resulting in under-stimulation. Using PACE, training sessions could be tailored based on an individual’s energy balance, improving safety and personalization.”
This dual applicability, elite sports performance and clinical rehabilitation, positions PACE as a versatile tool with broad relevance.
Stakeholders and valorization: exploring a spin-off trajectory
The development of PACE involves collaborations with numerous partners:
- Belgian Cycling for elite track cycling applications
- UZ Gent – Cardiology for rehabilitation research
- Lotto Cycling Team for outdoor validation in road cycling
- Recreational athletes involved in app testing and user evaluation
These partnerships help evaluate the model’s performance in varied real-world environments, ranging from controlled indoor settings to competitive contexts.
Two IOF-funded projects (StarTT and Advanced) have enabled the team to explore pathways toward commercialization. A key requirement of the current project phase is the development of a go-to-market strategy and business plan.
“Recently we presented PACE during a Ghent University–Vlerick event, after which a group of students contacted us to use PACE as a real-life business case,” Kevin explains. “Their analysis will support our decision on whether a spin-off is the appropriate next step.”
The team expects to make strategic decisions on the future valorization trajectory in the coming months.
Scientific communication and publication considerations
The research underpinning the model’s physiological determinants has been published in scientific journals. However, the full PACE formula itself has not yet been made public.
“We’re still evaluating how to balance scientific openness and potential commercial value,” Kevin notes. “A formula is difficult to patent, which complicates the decision. For now, the model remains internal.”
Nevertheless, the team frequently presents the research and its applications at conferences, workshops, and sports science events to ensure broader knowledge dissemination.
Technical and scientific challenges
Developing PACE required integrating heterogeneous datasets collected across multiple studies. “Combining five to six years of data into a single coherent model was one of the main challenges,” Kevin reflects. “Each dataset targeted a different aspect of energy dynamics, so bringing them together required extensive analysis and iterative refinement.”
Applying the model in practice, particularly in track cycling, introduced additional complications. “In laboratory settings, power output is stable. On the track, power meters show artefacts, riders draft behind each other, and aerodynamics strongly influence performance. These factors made the translation from power output to predicted finishing time far more complex than expected.”
To address these issues, the team added specialized data-processing expertise to the project.
Role models in valorization
When asked about successful examples of valorization, Kevin points to OnTracx. “Seeing how they transformed scientific research into a functioning startup is impressive,” he says. “Their trajectory is a strong example of what academic research can evolve into.”
PACE in the media
Read more about PACE in these previously published articles.
Blog on Kennismakers.be: ‘Vijf weetjes uit het sportlabo’ (in Dutch)