Association Between Obesity Indices and Heart Rate Variability: A Cross‐Sectional Study in Asian Young College Students

Figures

• 

  FIGURE 1. Sex‐stratified heatmaps of correlation matrices between indicators (N = 3180). The figure displays only indicator pairs with statistically significant correlations (FDR p < 0.05). Bubble size represents the absolute value of the correlation coefficient, and color indicates the direction of correlation (teal: negative correlation; pink: positive correlation). Solid borders denote male participants, while dashed borders denote female participants. (A) Correlation matrix among HRV parameters (lower‐left triangle: males; upper‐right triangle: females). (B) Correlation matrix among obesity indicators (lower‐left triangle: males; upper‐right triangle: females). (C) Correlation matrix between obesity indicators and HRV parameters in males. (D) Correlation matrix between obesity indicators and HRV parameters in females. All nonsignificant correlation pairs and same‐indicator pairs (diagonal cells in A and B) are left blank. Complete correlation coefficients and statistical test results are provided in supporting Table S1.

• 

  FIGURE 2. Association between standardized obesity indicators and HRV parameters across three autonomic function (N = 3180) domains. Tornado plots display standardized β coefficients (points) and 95% CIs (horizontal lines) from sex‐stratified linear regression models adjusted for age, HR, physical activity (PACE), and measurement time. Data points represent results from different analytical groupings: circles = total sample, squares = male subgroup, triangles = female subgroup. Obesity indicators are sorted in descending order of absolute β coefficient magnitude from the total sample analysis (strongest association at the top). Significance levels: ∗FDR p < 0.05, ∗∗FDR p < 0.01, ∗∗∗FDR p < 0.001 after false discovery rate correction.

• 

  FIGURE 3. Four‐track parallel analysis evaluating obesity indicators for predicting HRV (N = 3180). (A) Best subset regression showing R 2 values for different predictor combinations. (B) Relative importance analysis quantifying the percentage contribution of each indicator. (C) Elastic net variable selection with cross‐validation. (D) Incremental predictive value of adding each indicator to a baseline model. Highlighted elements indicate optimal or selected components. All analyses are based on total sample and adjusted for covariates. ∗FDR p < 0.05, ∗∗FDR p < 0.01, ∗∗∗FDR p < 0.001. All models were adjusted for age, HR, physical activity (PACE), and measurement time.