Sympathovagal balance: a critical appraisal.

To the Editor: With interest, I read Dr Eckberg’s review,1 but some of his views need clarification. There are 2 principal methods for spectral analysis of heart rate variability (HRV) that differ in the determination of spectral components. Either integrals of power spectrum density over specific bands or automatic determination of components by autoregressive algorithms is used. We do not know whether 1 of these methods is superior, let alone whether 1 is the “correct” way. Although the sum of high- and low-frequency components accounts for the majority of spectral power in short-term recordings with both approaches, both methods may provide different values for individual components. Standardization efforts were organized to propose the optimum methods for spectral and other analyses of HRV and interpretation of the results.2 3 Dr Eckberg was involved in some of these discussions and knows that no consensus on the “correct” spectral analysis was reached because no study exists comparing the 2 methods in a comprehensive variety of conditions. The assumption that spectral frequency components as integrals of specific bands are equal to their automatic autoregressive detection …