Summary
Objectives
: This paper examines the operational characteristics of the multivariate autoregressive analysis applied to the simultaneous recordings of the instantaneous heart rate (IHR) and the change in systolic blood pressure (SBP).
Methods
: The multivariate autoregressive model has been utilized to reveal the feedback characteristics between IHR and SBP. The model assumes the presence of independent set of driving forces to activate the system. However, it is likely that the driving forces may have correlation due to the presence of a common fluctuation source. This paper examines the effect of the presence of correlated components in the driving forces to the estimation accuracy of impulse responses characterizing the feedback properties. The twodimensional autoregressive model driven bytwo correlated 1/f noises was chosen for the analysis of operational characteristics. The driving force was generated by a moving average system which simulates non-integer order integration.
Results
: Computer simulation revealed that the mean square estimation errors of impulse responses sharply increase as relative power of common driving force exceeds 50%. However, the estimation accuracy and bias are found to be in permissible range in practice.
Conclusions
: These findings ensure the practical validity of utilizing multivariate autoregressive models for the feedback analysis between IHR and SBP where both signals have the common driving force.
Keywords
Autonomic nervous system - heart rate variability - blood pressure - feedback analysis - multivariate autoregressive model - operational characteristics
