Using AI-generated Adaptive Training Recommendations to Improve Physical Performance
Using AI-generated Adaptive Training Recommendations to Improve Physical Performance
In this study two different training strategies are compared; one is adaptive where the training is adjusted up or down on a daily basis to better match the recovery status (readiness) of the subject. The other strategy is static, i.e. no changes are made depending on readiness level. Instead the subjects in the static group are encouraged to perform the prescribed training.
The training intervention will last for 8 weeks.
Thorough physiological tests will be performed pre and post the training intervention together with muscle biopsies for assessment of mitochondrial function.
40 recreational athletes, 20 male and 20 females, will be recruited to this study. The research subjects will be stratified into two groups: an Adaptive and a Static training group. The training sessions will be mainly home-based training according to daily instructions from an online application. Testing sessions will be performed on a laboratory treadmill. During the entire period (run-in + training period) the participants will wear continuous glucose monitors for minute-by-minute glucose tracking, an Oura ring for sleep measurements, a Garmin GPS watch with heart rate chest-strap for training data and answer questiions regarding subjective well-being using an online application (Readiness Advisor).
The training program will be a mixture of training sessions based on the experience levels of the research subject.
During the onboarding session, data from each persons Garmin watch, will be used to create an optimal training program that takes into account recent training history and which type of training each subject responds the best to. For the subjects in the Adaptive group the training will be adjusted up or down on a daily basis based on subjective readiness, sleep or glucose control. The subjects in the Static group will not get any adjustments in their training program regardless of readiness level.
The length of the training intervention will be 8 weeks.
In addition to adjusting the training for the adaptive group based on daily readiness, the training in that group will be adjusted based on actual training done. For example if a participant fails to reach his suggested heart rate range during a prescribed session, the training load for upcoming sessions will be recalculated so that the weekly overall training load goal is achieved. This kind of adjustments will not be done in the static training group.
Both pre and post the training period a battery of physiological tests will be done including determination of VO2max, lactate threshold and running economy. A performance test (3000m time-trial) will be done on a running track.
A muscle biopsy will be donated from the Vastus Lateralis for assessment of mitochondrial function, efficiency, capacity and density using a combination of methods including respirometry, western blot and enzymatic methods. Capillarization and number of myonuclei will be assessed using histochemical staining. In addition muscle buffering capacity will be investigated using an acid-titration assay.
In addition to the physiological and biochemical investigation, parameters such as subjective training tolerability, perceived pleasure of the training and number of dropouts are important objectives in the proposed studie.
Statistical power calculation. Based on the parallel group design, 80% power, an alpha of 0.05 and a coefficient of variation of 2.4% for VO2max, 4.2% for performance and 14% for mitochondrial function, a differences can be detected between groups of 6%, 11% and 26% for the three outcomes if 15 participants per group are included. 20 subjects will be recruited per group to allow for 5 dropouts per group.
Inclusion Criteria:
Exclusion Criteria: