Physiological Responses to Commanded Terrain Conditions in a Real-World Video-Based Indoor Cycling System

This is the authors' abstract. We don't add key points for this paper.

Background and Purpose: Indoor cycling systems can combine real-world riding footage with resistance commands derived from terrain data. This study characterized cardiovascular, heart rate variability, and metabolic responses to three commanded terrain conditions. Methods: Analyses included three 30 min sessions per participant in 30 healthy men (22 ± 2 years) under low (predominantly downhill), moderate (mixed climbing and descending), and high (predominantly uphill) terrain conditions. All participants completed the three conditions across three study visits in a fixed order: moderate, low, and high. Participants were instructed to maintain 50–60 rpm. Heart rate (HR) and heart rate variability (HRV) were analyzed in 30 participants, and respiratory gas exchange in 29. Condition and measurement period effects were assessed using two-way repeated-measures analysis of variance; all four HRV indices were natural-log-transformed. In exploratory analyses of the moderate condition, participant-level HR, oxygen uptake (VO2), and carbon dioxide output (VCO2) slopes were compared between climbing and descending segments defined from an approximately time-aligned reference terrain profile. Results: HR, respiratory frequency, VO2, VCO2, and metabolic equivalents increased across conditions, whereas the time-domain HRV indices SDNN and RMSSD decreased (all condition effects, p < 0.01). Exploratory frequency-domain analyses also showed decreases in LF and HF across conditions; interpretation remains limited by exercise nonstationarity, respiration, and unassessed sensitivity to recording quality. In the moderate condition, median HR slopes were +6.34 bpm/min during classified climbing segments and −2.32 bpm/min during classified descending segments (p = 0.005; rank-biserial rrb = 0.58); the climbing slope exceeded the descending slope in 70% of participants. Corresponding differences were observed for VO2 (+126 vs. −102 mL·min−2; p = 0.016) and VCO2 (+132 vs. −121 mL·min−2; p = 0.004). Conclusions: Physiological responses differed across the three commanded terrain conditions under the fixed-order protocol, with group-level differences between classified climbing and descending segments in the moderate condition. Delivered mechanical workload was not validated, participant-specific terrain alignment was approximate, and condition effects could not be separated from visit-order effects. These findings characterize responses under the tested protocol and do not establish mechanical load accuracy or precise terrain tracking.