Time‐Frequency Analysis of Hemodynamics Oscillations during Presyncopal Lower Body Negative Pressure (LBNP)

Abstract

Introduction Greater low frequency (LF) oscillations (~0.1 Hz) in mean arterial pressure (MAP) and cerebral blood flow (indexed by middle cerebral artery velocity; MCAv) have been associated with higher tolerance to central hypovolemia. As the fast Fourier transform approach was used in these prior studies, measuring exact timing of oscillatory amplitude was not possible. In the present study, the continuous wavelet transform was used for time localization and measurement of the magnitude of LF oscillations during application of lower body negative pressure (LBNP) to presyncope to reduce cerebral perfusion. We hypothesized that the magnitude of LF oscillations in MAP and MCAv would be higher in subjects with greater tolerance to this stress. We also hypothesized that the time of instantaneous maximum magnitude of LF oscillations would be further from baseline for high tolerant subjects. Methods 31 healthy human subjects (14 F: 17 M; 25.1 ± 3.0 y) underwent a stepwise LBNP protocol to presyncope. Subjects were classified as high tolerant (HT) if they completed the -60 mmHg stage of LBNP, and low tolerant (LT) if they did not. Continuous beat-to-beat MAP and mean MCAv were recorded. The continuous wavelet transform was used to extract magnitude of oscillations within the LF range (0.07-0.15 Hz) over time for both MAP and mean MCAv. The instantaneous maximum magnitude of LF oscillations was then extracted. Pearson's correlation was calculated for instantaneous LF magnitude of MAP and mean MCAv versus presyncopal time. A t-test was then used to compare magnitude and relative time of LF oscillations between HT and LT subjects. Results The maximum magnitude of MAP LF oscillations was positively correlated with presyncopal time (R = 0.39, P = 0.03). The maximum magnitude of MCAv LF oscillations, however, was not correlated with presyncopal time (R = -0.03, P = 0.86). When assessing HT and LT subjects, no differences were observed for maximum magnitude of LF oscillations in MAP (HT 4.3 ± 1.0, LT 3.7 ± 1.2, P=0.15) or MCAv (HT 3.1 ± 0.9, LT 3.1 ± 0.7, P=0.91). Time to maximum magnitude of LF oscillations was greater in high tolerant subjects for both MAP (HT 1384.5 ± 329.6 s, LT 798.6 ± 416.8 s, P<0.01) and MCAv (HT 1064.3 ± 431.6 s, LT 662.4 ± 380.1 s, P=0.01). When expressing the time of maximum amplitude as a percent of presyncopal time, the average time in all subjects for MAP was 73.6 ± 22.0 % and 59.6 ± 26.1 % for MCAv, and there were no differences between tolerance groups in either MAP (HT 79.5 ± 14.0 %, LT 67.4 ± 27.3 %, P=0.14) or MCAv (HT 61.8 ± 24.8 %, LT 57.3 ± 28.1 %, P=0.64). Conclusions Using wavelet analysis, only the maximum magnitude of MAP LF oscillations was modestly correlated to presyncopal time. Contrary to previous measures using fast Fourier transform, there was no difference in instantaneous maximum magnitude of oscillations in MAP and mean MCAv between high and low tolerant subjects indicating that both groups could mount similar responses. Interestingly, the relative timing of maximum magnitude of oscillations in MAP and MCAv was no different between tolerance groups.