The effect of paced breathing at the resonance frequency on nonlinear features of heart rate variability

Abstract

IntroductionHeart rate variability (HRV) biofeedback with paced breathing at the resonance frequency is a method for increasing vagal heart rate (HR) regulation and psychological functioning. However, the nature and mechanism underlying this effect are not well understood yet. The investigation of nonlinear dynamics of HR is a useful research and diagnostic instrument for understanding intrinsic mechanisms of resonances in the cardiovascular system, caused by paced breathing. The aim of this study was to test how the paced respiration at the resonance frequency affects the nonlinearities in HRV.MethodsThe study group consisted of 70 healthy students, who underwent paced breath training. HR was recorded in volunteers, breathing at different respiratory rates (spontaneous breathing, 4.5, 5, 5.5, 6, 6.5 breaths/min). The resonant frequency was determined using HR spectral analysis. Short‐term 5‐min HRV indices from the linear time domain (SDNN) and from nonlinear methods ((ApEn, SampEn, MSE), Correlation dimension D2, Detrended Fluctuation Analysis (DFA), Recurrence Plot (REC, DET, Lmean, Lmax, ShanEn), Poincaré plot (SD1, SD2), time asymmetry (GI, P)) were computed. Statistical analysis was performed using the Sign test.Results and discussionSDNN was significantly lower during free breathing (42.08 ± 2.25 ms) than during resonance (79.64 ± 3.4; p < 0.001). Correlation dimension analysis of HRV showed that the D2 does not differ in control compared with resonance breathing (p > 0.05). SD1 and SD2 Poincaré parameters during spontaneous breathing were significantly smaller (respectively, 28.91 ± 1.67 ms and 51.51 ± 2.83 ms) than those during resonance breathing (respectively, 35.61 ± 1.68 ms and 106.79 ± 4.55 ms; p < 0.001). SampEn and ApEn were significantly higher during free breathing (respectively, 1.65 ± 0.03 and 1.22 ± 0.02) than in resonance breathing (1.02 ± 0.03; p < 0.001). DFA alpha‐1 was smaller during free breathing (1.03 ± 0.03) than that during paced breathing (1.55 ± 0.02; p < 0.001). In relation to long‐term exponents (DFA alpha‐2), there was a significant difference between free breathing (0.31 ± 0.01) and resonance breathing (0.14 ± 0.01; p < 0.01). Resonance breathing induces significant increases in time irreversibility indices (GI: from 50.88 ± 0.61 to 57.38 ± 0.67, p < 0.001; P: from 49.01 ± 0.44 to 54.92 ± 0.54, p < 0.001). Recurrence plot analysis reveals a decrease in complexity during resonance breathing: REC increases from 24.77 ± 0.82 to 30.38 ± 0.27 (p < 0.01), DET increases from 96.79 ± 0.18 to 98.81 ± 0.06 (p < 0.01), ShanEn increases from 2.97 ± 0.03 to 3.35 ± 0.02 (p < 0.01). The higher level of Lmean during resonance breathing (17.03 ± 0.51 vs 9.33 ± 0.29; p < 0.01) indicates the transition from chaotic to periodic RR sequences. The complexity of RR signals, as measured by the MSE method, was significantly lower during resonance breathing than during free respiration.ConclusionsResonance breathing is characterized by simultaneous increasing of HRV and time asymmetry and decreasing complexity. Cardiovascular resonance is accompanied by a transition of the HR regulation system from chaos to order. These results are inconsistent with previous studies, showing that decreased nonlinear complexity is a dynamical signature of disrupted physiologic control systems.Support or Funding InformationThis work was supported by the Ministry of Education (№19.9737.2017/BCh).This abstract is from the Experimental Biology 2018 Meeting. There is no full text article associated with this abstract published in The FASEB Journal.