Studies using spectral analysis of cardiovascular variability as a noninvasive means for assessing autonomic nervous system activity have provided controversial results in athletes. One reason is that a slow breathing rate-a common feature in athletes-affects spectral estimation because it causes the low-frequency (LF) and high-frequency (HF) components to overlap. Low-frequency power increases during sympathetic activation; high-frequency corresponds to respiratory sinus arrhythmia. In this study, to assess how controlled respiration influences autonomic nervous system activity, we determined the effect of controlled and uncontrolled breathing conditions on cardiovascular variability. Our aim was to identify a standard respiratory rate for spectral estimation of cardiovascular neural control in athletes. During electrocardiographic recordings, subjects lay supine and breathed at their spontaneous frequency and at rates of 15, 12, and 10 to 14(random) breaths·min-1. Uncontrolled and random breathing rates significantly altered spectral sympathetic indices; conversely, 15 and 12 breaths·min-1 redistributed respiratory related power through the HF, thus yielding correct LF power estimation. None of the breathing conditions significantly changed mean heart rate, arterial blood pressure, or spectral total power of cardiovascular variability. In conclusion, when power spectral analysis is used for assessing autonomic activity in athletes, respiration should be standardized at 15 breaths·min-1. Controlled respiration at this rate leaves autonomic nervous system activity unchanged.