Effect of the data sampling rate on accuracy of indices for heart rate and blood pressure variability and baroreflex function in resting rats and mice

Abstract

The aim of this study was to determine the minimal sampling rate (SR) required for blood pressure (BP) waveform recordings to accurately determine BP and heart rate (HR) variability indices and baroreceptor reflex sensitivity in rats and mice. We also determined if an 8-bit (versus 12-bit) analog-to-digital converter (ADC) resolution is sufficient to accurately determine these hemodynamic parameters and if spline interpolation to 1000 Hz of BP waveforms sampled at lower SRs can improve accuracy. BP and ECG recordings (1000 Hz SR, 12-bit ADC resolution) from two strains of rats and BP recordings (1000 Hz SR, 12-bit ADC resolution) from two strains of mice were mathematically converted to lower SRs and/or 8-bit ADC resolution. Time-domain HR variability and frequency-domain HR and BP variability indices and baroreflex sensitivity (using the sequence technique) were determined and the results obtained from the original files were compared to the results obtained from the mathematically altered files. Our results demonstrate that an ADC resolution of 8 bit is not sufficient to determine HR and BP variability in rats and mice and baroreceptor reflex sensitivity in mice. Average values for systolic, mean and diastolic BP and HR can be accurately derived from BP waveforms recorded at a minimal SR of 200 Hz in rats and mice. Spline interpolation of BP waveforms to 1000 Hz prior to extracting derived parameters reduces this minimal SR to 50 Hz in rats but still requires 200 Hz in mice. Frequency-domain BP variability (very low and low frequency spectral powers) can be estimated accurately at a minimum SR of 100 Hz in rats and mice and spline interpolation of BP waveforms to 1000 Hz reduces this minimal SR to 50 Hz in rats but does not reduce the minimal SR in mice. Time- and frequency-domain HR variability parameters require at least a SR of 1000 Hz in rats and mice. Spline interpolation of BP waveforms to 1000 Hz reduces this minimal SR to 100 Hz in rats and to 200 Hz in mice. Estimation of baroreflex sensitivity using the sequence technique requires a SR of at least 1000 Hz in rats and mice. Spline interpolation of BP waveforms to 1000 Hz reduces this minimal SR to 100 Hz in rats but does not reduce the minimum SR in mice. Finally, our results indicate that HR time series derived from BP waveforms are not totally consistent with HR time series derived from the ECG in rats. In conclusion, accurate assessment of HR variability and baroreflex sensitivity from BP waveform recordings requires a SR of at least 1000 Hz in rats and mice. If lower SRs are used for BP waveform recordings, a cubic spline interpolation to 1000 Hz (or an even higher SR) prior to extracting derived parameters significantly improves accuracy.