Estimation of respiratory frequency from autoregressive spectral analysis of heart period

Variability in heart period is modulated by respiration. Using an autoregressive spectral analytic approach on the heart period time series produces a components analysis. The central frequency of the heart period spectral component in the frequency of respiration (HFFreq) has been suggested as an index of respiratory frequency. The present experiment measured heart period variability (HPV) using an autoregressive algorithm and respiratory frequency using a mercury strain gauge (SGResp) to assess the relationship between the respiratory frequency indexed by the two methods. Twenty human participants were studied during resting baseline, relaxation, and mental effort conditions. Within- and between-person Pearson correlations and t-tests were used to assess the concordance between HFFreq and SGResp. Both within-person and between-person correlations indicated good concordance among the two methods. Moreover, t-tests, corrected for Type I error, indicated no significant differences between the respiratory frequency estimates derived from the two methods. The resolution of the estimates was approximately 1 breath per minute. These results suggest that the HFFreq may be a useful index of respiratory frequency.