Influence of low sampling rate on heart rate variability analysis based on non-linear dynamics

The aim of this study was to examine how sampling rates below 1000 Hz influence the precision of heart rate variability analysis. Simulated tachograms as well as tachograms from healthy persons were investigated. With parameters from time, frequency domain and non-linear dynamics these tachograms were classified. The results were compared with theoretical assumptions and special error measures were derived. New developed measures from non-linear dynamics showed an insensitive behavior to a reduced sampling frequency down to 128 Hz. The major results of this study lead to the conclusion to use sampling rates only above 120 Hz which implies a total error in the range of approximately 1.8/spl plusmn/1.5%. However only the use of sampling rates above 200 Hz provides nearly error free results.