Abstract
The purpose of this study was to investigate the limitation for applying a linear model to the cardiorespiratory control system. Four subjects performed the two types of exercise bouts, constant (CONST) and pseudorandom (PRBS) exercise, on an electrically braked cycle ergometer at three different work rates. The target work rate of CONST were set to 80. 100, 120 % of the individual anaerobic threshold (AT). In PRBS, the work rates were varied between i 10% of the individual AT around the respective target work rates of CONST. Although the spectral density of beat-to-beat heart rate fluctuations showed the conventional patterns for most cases, there was no obvious difference between CONST and PRBS. These results indicated that the variation of ±10% of AT did not affect the heart rate variation as the output response, suggesting a dilemma inevitable to apply a linear model based on the transfer function.
