Extracting autonomic information from oscillations in MSNA

Over the last few decades, thanks to the rapid development of computer applications, several laboratories provided evidence suggesting that spectral analysis of time series of RR interval and arterial pressure variabilities could be used as a tool capable of providing information on autonomic dysfunction of clinical and prognostic relevance (Malliani et al. 1991)). Initially, the investigative focus was especially on RR variability, as a means to estimate the sympathovagal balance regulating systolic arterial (SA) node beat-by-beat variability. However, still now there are discrepant positions regarding techniques (fast Fourier transfer (FFT) vs. autoregression (AR)) and mathematical handling of data (e.g. absolute vs. normalized units). Concerning this latter, we reported that focusing on the information provided by normalized units and using non-parametric statistics, it was possible to recognize, individual by individual, well-defined physiological conditions related to the different autonomic profiles associated with posture (Malliani et al. 1997)). We also stated that low-frequency (LF) oscillation evaluated from systolic arterial pressure (SAP) variability in various physiopathological conditions appears to be a convenient marker of the sympathetic modulation of vasomotor activity. This statement was corroborated by studies in which low-frequency (LF) variability of SAP was contrasted with a simultaneous measure of sympathetic efferent activity to muscles (MSNA) (Pagani et al. 1997; Furlan et al. 2000)) showing that, across small perturbations, oscillation characteristics of MSNA (in particular after normalization) were closely mirrored by oscillations of SAP.

Overall our results indicate a compelling relationship between changes in autonomic drive and cardiovascular oscillations.

Interpretation of MSNA is, however, not immediate. The multiunit signal depends heavily on the specific electrode, the (chance) tip position within the nerve bundle, the destination of (efferent) sympathetic nerve fibres, progressive recruitment of units during activation, possible damage and electronic settings (amplification, filtering and modality of integration). In addition, measurements require standards to be used across different conditions, subjects and laboratories, which are still not available for MSNA. It may thus be surmised that with MSNA analysis the experimental variance is still very high. Single-unit recordings might solve some of these problems, as classically done in animals (Pagani et al. 1974)) and recently tested in humans. Lower experimental error with multiunit recordings might also be obtained using high-frequency acquisition of the raw MSNA signal without filtering. The key issue with MSNA, however, is in regard to the code of interpretation (Pagani & Malliani, 2000)) in order to attach meaning to measures: average sympathetic nerve activity and its oscillatory components, although correlated to some extent, are likely to provide different, yet complementary, types of information, related to different modalities considering not only the simple spike activity per unit time (intensity code), but also more complex codes, based on pattern, correlation, rhythm, oscillation and synchronization, which are rarely considered. What can be taken for certain is the non-linearity of the system, inclusive of the quantal release of synaptic transmitter (a local event) and the influence of the humoral–endothelial milieu in the generation of arterial oscillations (a global phenomenon).

Ryan et al. (2011), reporting in a recent issue of The Journal of Physiology, should accordingly be praised for their attempt to specifically map the relationship between MSNA and LF oscillations of arterial pressure, considering data not only as a group but rather individually. These authors provide further evidence that across a wide range of sympathetic drive, as produced by progressively intense lower body negative pressure, there is overall a significant link between average MSNA and LF component of arterial pressure, and a very strong (r²= 0.97) relationship between LF oscillations of blood pressure and similar LF oscillations of MSNA.

Regarding individual relationships, using parametric statistics, however, these authors, observe that only in a portion of the population is this relationship highly significant. On this basis, the authors conclude that data do not support the use of low-frequency oscillations in arterial pressure as a non-invasive surrogate of MSNA.

We believe that these data are of importance as they confirm the stronger role of oscillatory rather than amplitude code in autonomic nerve activity. We also would like to conceive that the data of Ryan et al., if anything, corroborate the use of low-frequency oscillations of systolic arterial pressure SAP_LF in providing information on autonomic vascular oscillatory modulation. We might argue that the suboptimal coherence between LF oscillations of SAP and MSNA in the data of Ryan et al. might depend on different techniques (FFT vs. AR) and on the strong non-linear nature of the MSNA signal, which might not allow a consistent recognition of individual (linear) correlations, although from their Fig. 4C it appears evident that in all subjects, both SAP_LF and low-frequency oscillations of muscle sympathetic nerve activity (MSNA_LF) change in the same direction. Individual non-linearities might not preclude, however, a clear detection of global correlation.

Strictly speaking, and considering the above limitations, it may even be difficult to clearly characterize the pathophysiology of autonomic regulation from MSNA (Fatouleh & Macefield, 2011)).

We would like to conclude that SAP_LF is not a measure of MSNA, but more simply a broad index of the vascular sympathetic modulation. As such it appears as one of the multiple non-invasive, non-intrusive, autonomic indices that might be useful in the clinical management of conditions characterized by a disturbed autonomic regulation, such as hypertension (Lucini et al. 2002)) or diabetes (Lucini et al. 2009)).

We hope that in throwing away the water we will avoid throwing away the baby too.