An experimental setting and software were developed to evaluate cardiac autonomic function in unrestrained rats. Subcutaneously implanted ECG electrodes and an indwelling venous catheter were tunneled to a tail cuff in five rats. The ECG was A/D converted at 1000 Hz. After peak detection, a time series of RR intervals was obtained. Programs for the analysis of heart rate variability (HRV) were implemented in LabVIEW. Statistical properties were determined in the time domain. After cubic spline function curve fitting, resampling at 0.1 s and test for stationarity, power spectral analysis was performed on sampled records of 30 min duration after applying a sliding Hanning window (Welch method: 256 points (duration 25.6 s), 50% overlap and 0.039 Hz resolution). Algorithms were tested with simulated signals consisting of isolated frequency components, which were retrieved at their exact locations. Physiological validation of the system was performed by, beta-adrenergic and cholinergic blockade and by forced breathing at a fixed rate. Measurements were performed on five unrestrained rats under basal conditions. Mean RR was 174.2 +/- 3.6 ms; S.D., 13.3 +/- 4.6 ms; rMSSD, 5.2 + /- 1.2 ms; pNN10, 3.5 +/- 1.9% and pNN5, 18.7 +/- 6.4%. Low (0.19-0.74 Hz) and high frequency (0.78-2.5 Hz) power were determined (and also percent of low to total and high to total): 18.42 +/- 10.74 ms2 (22.9 +/- 6.5%) and 15.66 +/- 5.56 ms2 (19.9 +/- 2.7%), and the ratio low/high: 1.16 +/- 0.39. In conclusion, HRV analysis programs were developed and thoroughly tested through simulations and in vivo, under basal conditions and after pharmacological blockades. Using this software, HRV data from unrestrained rats were obtained.