R. L. Pimmel

Estimating Respiratory Mechanical Parameters in Parallel Compartment Models

Four iterative parameter estimation algorithms were used to obtain estimates in three parallel compartment models of the respiratory system. The stability of the parameter estimates and the agreement between the forced random noise impedance data and the models response were evaluated for each algorithm-model combination. The combination of a two-stage simplex algorithm with a five element model provided the most stable parameter estimates and the second best fit to the data.

Instrumentation for Measuring Respiratory Impedance by Forced Oscillations

Recent reports have suggested that the frequency dependence of the respiratory impedance may provide a sensitive method for characterizing early changes in pulmonary mechanics. A modification of the forced-oscillation technique provides an experimental method for obtaining the necessary data. A loudspeaker was used to provide the pressure oscillations, and the magnitudes and phase angle of the transduced pressure and flow signals were measured with a special electronics unit. A test comparing predicted values of a standard impedance (a 5-g bottle) to experimental data indicates that the measured amplitude is within 10 percent and the measured phase within over the frequency range of 1-16 Hz. Dog studies showed that measurements at all frequencies up to 16 Hz were reproducible within a few percent of the mean value on a given animal. Data obtained following bronchoconstriction and its reversal in six dogs indicate that the measurements are sensitive to alterations in pulmonary mechanics. Data from two dog models of clinical disease suggest that the technique may provide meaningful diagnostic information.

Computation of Respiratory Resistance, Compliance, and Inertance from Forced Oscillatory Impedance Data

Relationships for computing total respiratory resistance, compliance, and inertance from the frequency dependence of forced oscillatory impedance are presented. These equations result by minimizing the sum squared difference between the experimental impedance data and the response of the model.

Automatic Classification of Spirometric Data

Pattern recognition principles have been applied to 200 sets of spirometric data obtained from pulmonary function laboratory patients. Each patient was classified by a pulmonary specialist as normal, restricted, or mildly, moderately, severely, or very severely obstructed. Each patient was represented by a five-element pattern vector consisting of forced vital capacity (FVC), forced expiratory volume in one second (FEV1), midmaximum flow rate (MMFR), and flow rates with 50 and 25 percent of the vital capacity remaining (V¿50 and V¿25) normalized by predicted values. By Karhunen-Loeve expansion techniques, this vector was reduced to a two-feature pattern vector with only a 6 percent residual mean square representation error. The more important feature essentially represented the average of the three flow rates, while the second feature depended on FVC and FEV1. Data were divided into training and testing sets, and using the former, a parametric Bayes classifier and one-and two-layer pair-wise Fisher linear classifiers, were designed to assign patterns described by the two derived features to one of the six categories. With the testing set, overall recognition rates were 81 to 82 percent, with most errors representing misclassifications within the four obstructive categories. If the four obstructive classes were considered as a single class, the recognition rate increased to about 94 percent.

An evaluation of recovery of ventilation-perfusion ratios from inert gas data.

Abstract The recoverability of the distribution of ventilation-perfusion ratios from calculated retention (or excretion) for six inert gases was studied. Least square error minimization was investigated using noise-free artificial data and data with simulated experimental error. The accuracy of the recovered distribution was quantified by a distance function. Both unimodal and bimodal distributions were recovered from noise-free data and from data with a simulated ±1% error. Using data with a ±3% error, the original distributions were not recovered, in fact, the distance function increased while the error function decreased with repeated iterations. Since ±3% represents current experimental error in retention (or excretion) ratio measurements, great care must be taken in applying this technique to real data.

Enforced Smoothing Techniques for Recovering V A /Q Distributions from Inert Gas Data

Enforced smoothing techniques for computing V_A/Q distributions from inert gas retentions were evaluated by recovering distributions from artificial data. The value of the smoothness parameter z was varied with smoothness constraints that minimized the area either under the square of the distribution or under the square of the second derivative of the distribution. The accuracy of the recovered distributions were quantified by the RMS difference between the original and the recovered distributions. Values of z outside of a certain range (typically, 0.01 <z <32) produced underflow or overflow interrupts with the IBM 370/165. Since the optimum value of z varied with the smoothness constraint, the form of the distribution, and the level of random error, it is difficult to select a best value of z for all situations. RMS differences were similar to those achieved with the gradient descent method, and some compartments had smiall negative flows Using linear programming, upper bounds were found for the solution space that contains al possible solutions. The non-negativity constraint extremely limits this space, application of a smoothness constraint further reduces this space, and the addition of 3% random error to the retention values expands it.

Changes in Bronchial Reactivity of Asthmatics and Normals Following Exposures to 0.1 Ppm No2

ABSTRACT This study examined the effects of a one-hour exposure to 0.1 ppm NO 2 on respiratory mechanics and airways reactivity of normal subjects and of atopic asthmatics. Fifteen normal and 15 asthmatic subjects were exposed to air and to NO 2 on consecutive days in a randomized double-blind, cross-over design. Airways responsiveness to methacholine was assessed by plethysmography measurement of specific airway resistance (SRaw) and by measurement of total respiratory resistance by a forced oscillatory technique immediately after each methacholine dose. Following exposure to NO 2 , there was a slight, but not significant increase in mean baseline airway resistance (Raw) and SRaw in both normals and asthmatics. The overall resistive properties of the respiratory system derived from the forced random noise impedance spectrum (5–30 Hz) were not affected by NO 2 inhalation either. Similarly, despite good dose-response curve reproducibility, no change in response to methacholine related to NO 2 exposure was observed in either group. These findings indicate that 0.1 ppm NO 2 exposure for one hour without exercise had no significant airways effects, either in young atopic asthmatics or in young normal subjects.

Forced Oscillatory Resistance and Compliance Parameters following Pneumonectomy in Beagle Dogs

Forced oscillatory resistance (RFO) and compliance (CFO) were measured in 3 groups of adult beagle dogs. Group I (n = 12) had undergone a left pneumonectomy at 6-10 weeks of age; group II (n = 11) underwent a similar procedure at 1 year of age; and group III (n = 8) served as unoperated controls. The mean value of RFO for group II was significantly higher than corresponding values for groups I and III. Similarly, the mean value of CFO was significantly lower for group II. Mean values for RFO and CFO between groups I and III were not statistically different. When RFO and CFO were normalized by functional residual capacity, mean values for all three groups were similar. However, when RFO and CFO were normalized by body weight, mean values for groups I and II were significantly different from those of group III. These findings cannot be explained directly by postoperative hyperinflation of the remaining lung. They suggest that there is an adaptive change in the mechanical characteristics of the remaining lung following pneumonectomy and that the degree of adaptation varies with the age of the animal undergoing resection.

Forced Oscillatory Parameters of the Canine Respiratory System with Altered Vagal Tone

Total respiratory impedance was obtained by forced oscillations ranging from 0.9 to 16 Hz in six anesthetized intubated dogs during a control period, after bilateral vagotomy, and during bilateral vagal stimulation. Values for forced oscillatory resistance, compliance, and inertance were calculated using regression analysis with a linear model. Mean values ±SD for the control period were 2.49 ±0.18 cmH2O ·L¿1 ·s, 0.0254 ±0.0039 L ·cmH2O¿1, and 0.0849 ±0.0055 cmH2O ·L¿1·s2, respectively. These were similar to previously reported values. Vagotomy produced only small changes in these parameters. Vagal stimulation produced a 33 percent increase in resistance, a 20 percent decrease in compliance, and an 8 percent increase in inertance. Changes in resistance and compliance were consistent with reported effects in the literature. Thus, the transient mechanical changes induced by vagal stimulation can be characterized by this technique.

Difference between occluded airway pressure and generated muscle pressure

During the measurement of occluded airway pressure (Pao), the gas in the lung expands. This volume change causes Pao to underestimate the generated inspiratory muscle pressure (Pmus), which has been recommended as an index of respiratory drive. Analysis shows that the error is proportional to the thoracic gas volume at occlusion and to the effective respiratory system elastance (Ers′) and is inversely proportional to the dry atmospheric pressure. Examples of potential error in various species are presented. In normal human subjects at sea level measurement of Pao would underestimate Pmus on the order of 5 percent depending on Ers′. In subjects with large functional residual capacity due to obstructive lung disease, the error can be as large as 20 percent. It is recommended that, in all studies of occlusion pressure, an estimation of this error be made.