A novel metric linking stellate ganglion neuronal population dynamics to cardiopulmonary physiology

Abstract

Cardiopulmonary sympathetic control is exerted via stellate ganglia (SG); however, little is known about how neuronal firing patterns in the stellate ganglion relate to dynamic physiological function in the heart and lungs. We performed continuous extracellular recordings from SG neurons using multielectrode arrays in chloralose-anesthetized pigs (n = 6) for 8-9 h. Respiratory and left ventricular pressures (RP and LVP, respectively) and the electrocardiogram (ECG) were recorded concomitantly. Linkages between sampled spikes and LVP or RP were determined using a novel metric to evaluate specificity in neural activity for phases of the cardiac and pulmonary cycles during resting conditions and under various cardiopulmonary stressors. Firing frequency (mean 4.6 ± 1.2 Hz) varied spatially across the stellate ganglion, suggesting regional processing. The firing pattern of most neurons was synchronized with both cardiac (LVP) and pulmonary (RP) activity indicative of cardiopulmonary integration. Using the novel metric to determine cardiac phase specificity of neuronal activity, we found that spike density was highest during diastole and near-peak systole. This specificity was independent of the actual LVP or population firing frequency as revealed by perturbations to the LVP. The observed specificity was weaker for RP. Stellate ganglion neuronal populations exhibit cardiopulmonary integration and profound specificity toward the near-peak systolic phase of the cardiac cycle. This novel approach provides practically deployable tools to probe stellate ganglion function and its relationship to cardiopulmonary pathophysiology.NEW & NOTEWORTHY Activity of stellate ganglion neurons is often linking indirectly to cardiac function. Using novel approaches coupled with extended period of recordings in large animals, we link neuronal population dynamics to mechanical events occurring at near-peak systole. This metric can be deployed to probe stellate ganglion neuronal control of cardiopulmonary function in normal and disease states.

Keywords: cardiopulmonary; hemodynamics; neural recordings; spike activity; stellate ganglion.

Figure 1.

Experimental workflow. A: representative image of the recording linear microelectrode array in the pig left stellate ganglion. B: example of recordings of neural activity, respiratory pressure, left ventricular pressure, and electrocardiogram. C: the approach to spike sorting. D: examples of identified spikes. ECG, electrocardiogram; LVP, left ventricular pressure.

Figure 2.

Spatiotemporal dynamics of stellate ganglion neuronal populations. A: schematic of the 16-electrode array inserted into the ganglion (upper) and actual size of recording electrode relative to neurons (lower). Spike dynamics over 8 h in eight channels spanning the stellate ganglion (B); plot of firing frequencies across all 16 electrodes (E1–E16) in six male study animals upon insertion of the probe, during steady state, and during cardiopulmonary stressors (C); mean firing frequencies across the ganglion in study animals (D); mean firing frequencies across the ganglion in the study animals (E); distribution of responses to several cardiovascular stressors including aorta occlusion, rapid ventricular pacing, ischemia, apnea plotted as differences in mean firing between steady and stressed states (F).

Figure 3.

Cardiopulmonary integration is reflected in stellate ganglion neural activity. A and B: representative recordings from several electrodes (blue and red tracings) using the 16-channel linear array, along with respiration (Resp), left ventricular pressure (LVP), and the electrocardiogram (ECG) in purple tracings. Yellow bars highlight respiration (A) and the cardiac cycle (B). Activity in the blue and red channels are locked to cardiac and/or pulmonary. C: stellate ganglion (SG) neuron activity (red dots are individual spikes) shows increased firing rate at peak and near-peak LV pressures; however, firing is inhibited during respiration, as reflected by increased spike period (heavy blue line) mirroring inspiration and expiration. D: representative response of SG neural activity to apnea (60 s). Black trace is scaled up respiratory activity, and black arrows identify baseline peak and trough of the spike period (green trace). E: spike period oscillation (peak-trough values) while breathing at baseline, and over the same period in apnea. n = 6 animals; *P = 0.016, two-tailed Wilcoxon rank-sum test.

Figure 4.

Periodicity in spiking activity reflects specificity for a narrow range of LVP and respiratory rate. A: left ventricular pressure (LVP) is plotted in blue tracings along with individual spikes as red dots. B: respiratory pressure is plotted in black tracings with individual spikes superimposed as blue dots. C and D: autocorrelation of spiking activity calculated for equally spaced windows for different lags over the course of an experiment. The autocorrelation value (between −1 and 1) is represented by the color scheme shown in C and D. Maximum autocorrelation (one) of the spiking activity in a window for a particular lag represents periodicity in the spiking activity for that lag. The plot of the sliding autocorrelation reveals the heart rate and the respiration rate with the spiking activity being highly correlated for the respiratory period (5 s) and the heart period (0.5 s). The respiration rate (D) is revealed more clearly than the heart rate (C). E and F: the neural bias toward LVP and respiratory pressure for the duration of the experiment (near-zero bias is teal, positive bias is yellow, and negative bias is blue). E: the degree of bias in neural activity toward specific LVP over the course of the experiment. F: the same with respect to the respiratory pressure.

Figure 5.

Regionality of cardiac cycle specificity in neural activity. Bias in neural activity plotted for the duration of the experiment across five channels in one animal subject, with the spike rate superimposed as a red line. The five channels plotted are arranged according to the spatial arrangement of the electrodes across the ganglion. This plot indicates that within each channel the bias in neural activity is relatively consistent over time while spike is highly variable. On the other hand, spike rate across channels shows some degree of consistency, whereas the bias in neural activity is relatively more variable.

Figure 6.

Dynamic changes in neuronal spiking density does not impact cardiac cycle phase specificity. The number of spikes (red line) and associated variation (blue line) are computed from the mean and standard deviation of spiking frequency calculated from sliding windows of 10-min width. These are superimposed over the remaining phase selectivity image. Changes in the number of spikes and their variation are not uniquely associated with the phase selectivity or changes in systolic pressure. Changes in the same computation across different animals show little consistency within, or among, animals of the spiking activity and its variation with respect to LVP. The degree of phase selectivity to LVP just below the systolic pressure is consistent across animals despite wide variations in systolic pressure. LVP, left ventricular pressure.

Figure 7.

Near-peak systolic phase selectivity is independent of blood pressure. A: spike periodicity (green tracings) plotted along with the left ventricular pressure (LVP) in blue tracings and respiration in black tracings for two stressors—inferior vena cava (IVC) and aortic occlusion use to lower and raise LVP, respectively. Stellate ganglion neuron activity (individual spikes plotted as red dots) is shown in response to IVC and aortic occlusion. Near-peak systolic phase selectivity is indicated by the arrows. B: spike fraction is plotted as a histogram and a distribution for normalized LVP before (blue tracings), during (green tracing) and after (red tracings) for the interventions plotted in A. C: spike probability (area under the curve of spike fraction distribution) is plotted for systole (left) and diastole (right) for before (blue), during (green) and after (red) the two interventions. n = 5 animals/condition, two-tailed analysis of variance for all comparisons were not statistically significant.