J. P. Saul

Assessment of autonomic response by broad-band respiration

The authors present a technique for introducing broadband respiratory perturbations so that the response characteristics of the autonomic nervous system can be determined noninvasively over a wide range of physiologically relevant frequencies. A subjects respiratory bandwidth was broadened by breathing on cue to a sequence of audible tones spaced by Poisson intervals. The transfer function between the respiratory input and the resulting instantaneous heart rate was then computed using spectral analysis techniques. Results using this method are comparable to those found using traditional techniques but are obtained with an economy of data collection.<<ETX>>

A time domain approach for the fluctuation analysis of heart rate related to instantaneous lung volume

A time domain technique for estimating transfer characteristics from fluctuations of instantaneous lung volume (ILV) to heart rate (HR) is presented. An effective procedure for estimating the impulse response of HR to ILV is proposed. Pre- and post-processing procedures, including prefiltering of the HR signal, preenhancement of the high frequency content of the ILV signal, and post-filtering of the estimated impulse response, together with a random breathing technique, are shown to effectively reduce spurious transfer gain so as to get a stable estimate of the impulse response. Analysis of the data collected from fourteen healthy male subjects in various conditions revealed that there are three components in the impulse response: fast positive, delayed slow negative, and oscillatory. The effects of the autonomic blocking agents propranolol and atropine on these transfer characteristics are also described.<<ETX>>

Altered Cardiac Repolarization in Some Victims of Sudden Infant Death Syndrome

Abnormal prolongation of cardiac repolarization, as reflected by a long QT interval with respect to the RR interval on the electrocardiogram, is known to be associated with ventricular tachyarrhythmias. To test the hypothesis that prolonged cardiac repolarization may characterize some babies who die of sudden infant death syndrome (SIDS), we studied the dependence of the QT interval on the preceding RR interval in 10 babies with SIDS and 29 healthy control babies. We analyzed approximately 5000 pairs of QT and RR intervals in each subject over a wide range of RR intervals. We found that the QT intervals demonstrated less dependence on the preceding RR intervals in 5 of 10 babies who subsequently died of SIDS than in normal controls. No ventricular arrhythmias were observed, however, during the six-hour recording period. Our data suggest that in some babies with SIDS the ability to shorten the QT interval as the heart rate increases is impaired. These observations are consistent with the hypothesis that relatively prolonged cardiac repolarization may predispose such babies to ventricular arrhythmias.

Closed-loop identification of cardiovascular regulatory mechanisms

An approach to assessing closed-loop physiologic regulation by the analysis of multiple fluctuating physiologic signals, is presented. In particular, the couplings among heart rate, instantaneous lung volume, and arterial blood pressure have been studied. The procedure used consists of several steps. First, a block diagram model is constructed to display the assumed interrelationships among the measured data. Secondly, linear, constant-coefficient, autoregressive moving-average equations are written to describe each block in the model. Thirdly, the physiologic signals are recorded under conditions which ensure that their frequency content is broadband (e.g. during random interval breathing). Finally, the set of coefficients that result in an optimum least squares fit to the data is found. These parameters specify the transfer functions that correspond to the blocks in the original model. Preliminary results suggest that the time constants and sensitivities inferred from the estimated transfer functions agree with more invasive direct measurements reported in the literature,.<<ETX>>

Fluctuation Analysis Of The Heart Rate Related To Instantaneous Lung Volume: A Time Domain Approach

This paper presents a time domain technique to estimate transfer characteristics from fluctuations of instantaneous lung volume (ILV) to heart rate (HR). Pure moving average (MA) and autoregressive moving average (ARMA) system identification techniques are introduced to characterize the system. Pre-processing (pre-filtering of the HR fluctuations and pre-enhancement of the high frequency contents of ILV fluctuations) are shown to be effective in obtaining stable estimates of MA coefficients. Model order selection of the ARMA system based on a priori knowledge of the system characteristics is proposed. Data analysis using the time domain techniques revealed some temporal transfer characteristics which are poorly resolved using frequency domain analysis.

Transfer function analysis of cardiovascular regulation in an open-loop animal model

Cardiovascular regulation involves feedback and feedforward limbs of a control loop. In order to understand the behavior of this control system, it is desirable, although difficult to disentangle the dynamical properties of the feedback and feedforward limbs. The authors have developed a conscious animal model in which the control loop is functionally opened by interruption of atrioventricular conduction, thereby dissociating fluctuations in cardiac output and arterial pressure from reflexogenic changes in atrial rate. Cross-spectral techniques are used to comput the frequency response of the mechanical system that couples ventricular activity to arterial pressure and of the baroreflex system that governs sinoatrial rate. The infusion of various pharmacologic agents is used to study the behavior of the system under conditions of selective autonomic blockade. Results from this work will enable validation of analysis techniques that attempt to derive open-loop system properties from data collected in closed-loop preparations.<<ETX>>

An Application Of A Neural Network For The Analysis Of Heart Rate Fluctuations Related To Instantaneous Lung Volume

This paper presents an application of a neural network for the analysis of the heart rate fluctuations. A three layer neural network was trained by the back propagation algorithm to relate fluctuations in heart rate to instantaneous lung volume. Considerable reduction of the residual variance was attained compaired to the optimum linear system as increasing the number of hidden layer units up to 10. The result suggests the involvement of a nonlinear factor in the transfer characteristics from the instantaneous lung volume to heart rate.

Assessment of the HR baroreflex in human subjects using parametric system identification

Quantitative assessment of the coupling between arterial blood pressure (ABP) and heart rate (HR) is discussed. Traditionally, the HR baroreflex has been studied directly by controlling sensed ABP independent of HR changes. A method is presented that, by combining broadband stimulation and parametric system identification, makes it possible to assess baroreceptor function without disturbing the normal closed-loop function of the cardiovascular system. Preliminary results suggest that the methodology can detect changes in baroreceptor sensitivity in response to the administration of atropine and propanolol.<<ETX>>

A simple analytical model mimics complex physiological behavior

The authors have previously described the use of nonparametric transfer function analysis in combination with broadband respiration to compute the magnitude and phase relations between respiration, heart rate (HR), and arterial blood pressure (ABP). These experimental results have led to the formulation of a model of HR and ABP control which includes the mechanical feedforward from HR to ABP, the baroreflex feedback from ABP to HR, and respiration as a noise source which directly affects HR through central modulation of autonomic activity while affecting ABP through mechanical thoracic coupling. Despite the fact that causality was not built into the ABP-HR relation, analytic transfer functions from the model closely simulate many of the observed experimental phenomena. The findings suggest that the complex transfer relations observed with a relatively simple analysis of respiration, HR, and ABP can be understood in terms of feedback loop between ABP and HR, with respiration as an external noise source.<<ETX>>