The aim of this study was to test whether the heart rate (HR) control in infants at risk differs in comparison with healthy infants. Twelve former preterm infants suffering from bronchopulmonary dysplasia and 18 control infants, matched for their postconceptional age, were examined polygraphically during quiet and active sleep. HR, low-frequency (LF) power, high-frequency (HF) power, total power, and the ratio of LF to HF power (LF/HF) of the instantaneous HR spectra were calculated for linear analysis. The largest Lyapunov exponent (LLE) of the R-R interval time series was calculated to determine a nonlinear property of HR. Infants at risk had significantly lower LF power (median: 0.51 x 10(-3) vs. 1.16 x 10(-3) Hz2) and lower LF/HF (median: 1.05 vs. 1.94) during quiet sleep. LLE was positive, revealing low-dimensional chaotic behavior of HR control, and did not differ between both groups (median: quiet sleep, 0.05 bit/s vs. 0.06 bit/s; active sleep, 0.16 bit/s vs. 0.15 bit/s). Sleep state-related changes in spectral parameters and LLE were similar in both groups. In infants at risk, the lower LF/HF during quiet sleep can be interpreted in terms of changes in the rhythmic components of the sympathovagal balance of the autonomic system, which is an expression of linear properties of HR control. Conversely, the lack of differences in LLE between both groups indicates similar nonlinear properties of the control system.