An extensive collection of mobile software was developed and used in multiple experiments exploring self-regulation during exercise leading to a number of key findings. The present thesis represents a multidisciplinary approach including elements of software engineering, exercise physiology, computational biology, and biofeedback. In all, the work presented in this thesis includes:
• Three complete mobile software applications.
• Two human experiments including a 5-week intervention.
• One comprehensive software library for the study of heart rate variability (HRV).
• One case study involving human exercise and HRV.
• One validation study focused on HRV analysis.
A novel form of continuous acoustic feedback, developed as part of an ad hoc mobile software application, outperformed the current “gold standard” as implemented in consumer heart rate monitors. Participants were exposed, in a repeated measures fashion, to one of 2 different feedback conditions and a silent control during 20 minutes of moderate intensity exercise on an indoor stationary bicycle. Feedback conditions included either a boundary-style feedback emulating current heart rate monitors consisting of a sequence of auditory “beeps” heard only when heart rate fell outside the target training zone, or a novel form of continuous feedback developed specifically as part of this study. Time in Zone (TIZ), defined as the ratio of the time spent within the target training zone divided by the overall time of exercise, rating of perceived exertion (RPE), and subjective measures of association, dissociation, and distress were compared across feedback and control conditions using one-way, repeated measures ANOVA. TIZ increased significantly in both the continuous (M ± SD: .99 ± 0.01) and boundary (.94 ± 0.05) conditions over control (.52 ± 0.40) and was significantly greater for continuous versus boundary, all ps <= .01. RPE did not change significantly under different feedback conditions. Dissociation was significantly reduced under continuous feedback compared to both control, p < .01, and boundary, p < .05. No significant differences in association or distress across conditions were present, though a moderate effect size, Cohen’s d = .55, for association was observed between continuous and boundary conditions.
The ability to exercise autonomously at a moderate to vigorous intensity can be entrained through repeated exposure to auditory feedback. Healthy adults performed 10 indoor exercise sessions on cycle ergometers over 5 weeks following a twice-weekly schedule. During these sessions, participants received auditory feedback (FB) designed to maintain heart rate within a personalized, moderate-intensity training zone between 70 and 80% of maximum heart rate. All feedback was delivered via a custom, mobile software application. Participants underwent an initial assessment (PREFB) to measure their ability to maintain exercise intensity defined by the training zone without use of feedback. After completing the feedback training, participants performed three additional assessments identical to PREFB at 1 week (POST1), 2 weeks (POST2), and 4 weeks (POST3) after their last feedback session. TIZ, RPE, instrumental attitudes, and affective attitudes were then evaluated to assess results using two-way, mixed-model ANOVA with session and gender as factors. Training with feedback significantly improved TIZ (p < .01) compared to PREFB. An absence of significant differences in TIZ between FB, POST1, POST2, and POST3 (p >= .35) indicated that these improvements were maintained after feedback was removed. No significant differences in RPE, p >= .4, or attitude measures, p >= .3, were observed.
Real-time analysis of HRV parameters during exercise was realized as both a software library (HRVKit) and a standalone mobile application (HRVLab). Comparing results from HRVKit and a “gold standard” HRV analysis application revealed overall high levels of agreement with intra-class correlations (ICCs) >= .81. A case study using HRVLab for real-time analysis during a graded exercise test was performed, and results are consistent with existing literature and suggest anaerobic threshold may be determined using this non-invasive technology.
|Qualification||Doctor of Philosophy|
|Publication status||Unpublished - Dec 2014|