BACKGROUND: In-task affective responses to moderate-intensity continuous exercise training (MICT) have been shown to predict future physical activity behavior. However, limited research has investigated whether this affect-behavior relationship is similar for high-intensity interval training (HIIT) and whether it holds true over the longer term. This study aims to determine (1) if in-task affect during 2 weeks of supervised MICT and HIIT predicted changes to unsupervised moderate-to-vigorous physical activity (MVPA) behavior 12 months post-intervention and (2) if this predictive relationship was moderated by exercise type (MICT vs. HIIT). METHOD: Ninety-nine adults (69.7% female
50.9 ± 9.4 years) who were low active and overweight were randomized to 2 weeks of exercise training of MICT (n = 52) or HIIT (n = 47), followed by 12 months of accelerometry-assessed free-living MVPA. RESULTS: The pooled moderation model was not significant, F(3, 94) = 2.54, p = .07 (R CONCLUSION: The current findings raise important questions about whether the affect-behavior relationship may vary depending on exercise type. For HIIT-based exercise in particular, additional psychological constructs beyond in-task affect should be considered when attempting to predict future physical activity behavior.