There have been limited direct tests of the hypothesis that self-injurious behavior (SIB) regulates arousal. In this study, two autonomic biomarkers for physiological arousal (heart rate [HR] and the high-frequency [HF] component of heart rate variability [HRV]) were investigated in relation to SIB for 3 participants with intellectual disabilities. Second-by-second correlations were examined using time series statistical models. The probabilities of HR changes preceding or following SIB were derived using sequential analyses and compared using resampling procedures. Significant correlations and sequential dependencies were found between SIB and arousal parameters. Combining within-subject statistical methods with single-subject experimental designs may provide a replicable methodology for use across larger samples to examine relationships between SIB and arousal in real-world settings.