Interpreting oscillations of muscle sympathetic nerve activity and heart rate variability

Computer analysis of spontaneous cardiovascular beat-by-beat variability ha s gained wide credibility as a means of inferring disturbances of autonomic cardiovascular regulation in a variety of cardiovascular conditions, inclu ding hypertension, myocardial infarction and heart failure, Recent applicat ions of spectral analysis to muscle sympathetic nerve activity (MSNA) offer a new approach to a better understanding of the relationship between cardi ovascular oscillations and autonomic regulation. However, areas of uncertai nty and unresolved debates remain, mostly concerning different methodologie s and interpretative models that we will consider in this article. Perusal of all available literature suggests that average sympathetic nerve activit y and its oscillatory components, although correlated to some extent, are l ikely to provide different types of information. In addition, the specific experimental context is of paramount importance, as the rules that seem to govern the relationship between average and oscillatory properties of MSNA appear to be different in usual conditions and in conditions of extremes of activation or disease. In general, dynamic experiments, such as with grade d tilt or with vasoactive drugs, are more suited to investigations of the c omplexity of autonomic regulation than are static comparisons. In addition, because the information is spread across variables and is affected by a po tentially large error, it appears that several different techniques should be perceived as complementary rather than as mutually exclusive. Available evidence suggests that low-frequency and high-frequency oscillations in per ipheral signals of variability might have a predominantly central, rather t han a peripheral, origin and that this applies in particular to low-frequen cy oscillations. A crucial point in the assessment of the meaning of spectr al components relates to consideration of the varying level of very-low-fre quency noise, and the mathematical manipulation of derived indices, particu larly using a normalization procedure. This appears easier to obtain with a uto-regressive than with fast Fourier techniques. With this approach, discr epant interpretations seem to be resolved, provided adequate care is taken in separating direct physiological data from derived meaning, which relates to hidden information and neural codes; in the case of sympathetic dischar ge, the latter display greater complexity than simple average spike activit y per unit time, Accordingly we believe, in conclusion, that the judicious use of spectral methodology, in addition to other techniques, might provide unprecedented, useful insights into autonomic cardiovascular regulation, i n both physiopathological and clinical circumstances. (C) 2000 Lippincott W illiams & Wilkins.