Capturing sex differences in frequent and spontaneous fluctuations of heart rate

Background Autonomic control of the heart is a key indicator of self-regulation and overall mental and physical health. The vagus nerve plays a central role in this regulation and helps maintain psychophysiological balance. Resting-state heart rate variability (HRV), reflecting fluctuations in inter-beat intervals (IBIs), is the primary noninvasive marker of vagal activity. Because males and females differ in aspects of self-regulation, HRV may help clarify underlying neurobiological differences. Yet sex differences in short-term HRV metrics such as natural log transformed root mean square of successive RR interval differences (lnRMSSD) and high-frequency HRV (lnHF-HRV) appear smaller in young adults than in other age groups. These short-term metrics capture vagally mediated activity as linear, averaged metrics and may therefore overlook rapid spontaneous IBI fluctuations. We tested in a sample of young adults whether a similarity graph theory algorithm could better capture sex differences in nonlinear, rapid IBI variability within 2 to 5 second time windows. Methods Electrocardiogram (ECG) recordings of 269 young, healthy adults between 18 and 30 years old (M = 21.5, SD = 3.0) were pooled from three different studies. Males accounted for 52.4% of participants, indicating a comparable distribution between sexes. Similarity graph–theory metrics were computed to quantify nonlinear, rapid interbeat interval (IBI) variability using sliding windows of 2–5 seconds and 10 seconds. In addition, conventional linear and nonlinear heart rate variability metrics, including lnRMSSD and lnHF-HRV, were calculated. Logistic regression models were used to assess the predictive value of graph-theory and HRV metrics for sex, both separately and in combined models for comparison. All models were adjusted for age, body mass index, mean heart rate, and respiratory rate. Results Males showed higher values of the graph metric, indicating lower IBI variability, with a substantial effect size (odds ratio 4.53; 95% Confidence Interval (CI) 1.83–11.20). Neither lnRMSSD nor lnHF-HRV distinguished males from females alone, although lnRMSSD became predictive when combined with the graph metric (odds ratio 1.73; 95% CI 1.06–2.81). Conclusions These findings suggest that nonlinear methods sensitive to rapid spontaneous IBI changes can complement traditional short-term HRV metrics when assessing the physiological mechanisms underlying sex differences in self-regulation.