多変量自己回帰モデルによるヒト神経性循環調節のシステム同定に関する研究

采用多变量自回归模型进行闭环分析，对受试者RR间隔（RRI）、平均血压（BP）、血管舒张性交感神经活动（肌肉交感神经活动（MSNA））和呼吸的时间序列数据进行分析，以综合确定仰卧休息、平头倾斜和老年受试者体循环的神经控制。自功率谱密度曲线显示马约尔节律和呼吸节律的存在，其中相干性表明这两个变量之间存在线性联系。噪声贡献分析表明，BP中的Mayer波是BP和MSNA共同作用的结果。在仰卧休息时，假设血压升高1mmhg，脉冲反应显示MSNA被短暂抑制（BP-to-MSNA反应），RRI在接下来的一秒内延长了4ms （BP-to-RRI反应）。另一方面，msnna -to-BP反应显示，MSNA放电引起长时间（长达30秒）的血压升高，而RRI-to-BP反应引起短暂的血压下降。与仰卧位相比，平卧时msnna -to- bp反应增强并延长，BP-to-RRI反应被抑制。RRI-to-BP和BP-to-MSNA不变。在老年受试者中，BP-to-RRI反应减弱，而BP-to-MSNA反应不变。MSNA-to-BP反应明显降低，而RRI-to-BP反应不变。这些结果表明，采用多变量自回归模型的闭环分析可以定性和定量地描述人体体循环的神经控制。然而，需要进一步的研究来证实该方法的有效性。

Closed loop analysis using a multivariate auto-regressive model was applied to time series data for the RR interval (RRI), mean blood pressure (BP), vasomotor sympathetic nerve activity recorded as muscle sympathetic nerve activity (MSNA) and respiration with subjects, to determine integratively the neural control of human systemic circulation at supine rest, during head-up tilt and in elderly subjects. The auto-power spectral density curves showed the existence of Mayor's and respiratory rhythms, in which the coherence indicated a linear linkage among these variables. A noise contribution analysis revealed that Mayer's waves in BP are derived from the combination of both BP and MSNA. At supine rest, impulse responses, when assuming 1 mmHg elevation in BP, showed that MSNA is transiently suppressed (BP-to-MSNA response) and RRI is elongated by 4 ms in the following second (BP-to-RRI response). On the other hand, the MSNA-to-BP response revealed that MSNA discharge induces a prolonged (up to 30 s) elevation in BP, while RRI-to-BP response elicited a transient decrease in BP. During head-up tilt, impulse response analysis showed that MSNA-to-BP response was accentuated and prolonged, and that BP-to-RRI response was suppressed, compared with those in a supine position. RRI-to-BP and BP-to-MSNA were unchanged. In elderly subjects, BP-to-RRI response was diminished, whereas BP-to-MSNA response was unchanged. MSNA-to-BP response was markedly reduced, while RRI-to-BP response was unchanged. These results suggest that closed loop analysis using a multivariate auto-regressive model can describe qualitatively and quantitatively the neural control of human systemic circulation. However, further investigation is necessary to confirm the validity of the method.