A novel technique for achieving plethysmography measurements utilizing noncontact laser displacement sensors is described. This method may have utility in measuring respiratory and pulmonary function similar to that of respiratory inductive plethysmography. The authors describe the apparatus and method and provide results of a validation study comparing respiratory excursion data obtained by (1) the laser sensor technique, (2) standard respiratory inductive plethysmography (RIP), and (3) lung volume measurements determined by pressure variations in a control volume. Six healthy volunteers (five female, one male, ages ranging from 19 to 23 years) were measured for tidal breathing excursions simultaneously via all three measurement techniques. Results: Excellent correlation between the techniques was shown. Pairwise comparisons among all three measurement techniques across all subjects showed intraclass correlation coefficients of 0.995 in each case. These results indicate the laser plethysmograph (LP) system provides results that are, at a minimum, equivalent to those of the RIP at the two sites commonly measured by RIP. Use of the LP system has the potential to provide much more extensive and precise measurements of chest wall function and the respiratory musculature.
Plethysmography in general refers to the measure of volume and volume displacement.1 The technique is widely used in the fields of vascular and pulmonary medicine to provide diagnosis and evaluation of conditions associated with measurable changes in body volume. In the latter case, respiratory plethysmography is used to provide accurate means of measuring parameters such as tidal volume, respiratory frequency, and various ventilation times. Because of the ability of this technique to provide remarkably accurate measurements under a variety of conditions, the respiratory inductive plethysmograph (RIP) is the most widely used method of monitoring ventilation in a noninvasive and quantitative manner.2
Respiratory inductive plethysmograph utilizes two or more insulated coils of wire sewn into elastic bands that are then positioned around the chest and abdomen. The coils of wire in the bands are electrically connected to oscillator and amplifier circuits. As the subject breathes, excursions cause the bands to alternately stretch and contract, causing a change in the inductance of the oscillator circuits. Additional circuitry transforms the inductance change into a voltage signal that is proportional to the change in the cross-sectional area of the rib cage and abdomen where the bands have been placed.3
From the voltage signal, a number of volume and time measures can be determined. For example, determination of tidal volume is achieved by mathematically summing the signals from both the rib cage band and the abdominal band. The validity of this technique relies on the assumption that lung volume changes can be accurately described with two degrees of freedom: changes in thoracic volume at the rib cage and the abdomen.4 This assumption is widely accepted and has been reasonably well-validated for subjects under a wide set of circumstances.5,6 However, there are a number of conditions present that negatively impact this assumption. Of particular interest in the field of pulmonary medicine, it has long been recognized that patients with chronic obstructive pulmonary disease (COPD) possess additional degrees of freedom.7 Even with the presence of such additional degrees of freedom, several investigators have shown that the principle of two degrees of freedom does provide reasonable accuracy in determining volumetric data.8–11
However, there are a number of circumstances present in the clinic where the use of RIP bands may not provide accurate measures or may not be feasible for use at all. These include cases involving respiratory measurements during sleep12,13 and in COPD cases where large changes in the distribution of rib cage and abdominal contributions to tidal volume occur.14 Furthermore, difficulties in using the RIP system have been described during measurements when thoracoabdominal asynchrony presents15 and when being used as a means of measuring respiratory function during some forms of exercise.16
In clinical application, determination of volumes measured with RIP systems requires careful attention to detail and the ability to calibrate both the electronics and the obtained measurements from the subject to a known volume.2 This can prove to be both time consuming and challenging in a clinical setting. In addition to the potential difficulties calibrating the RIP system and utilizing it on all clinical presentations, it should be noted that a number of technology limitations will be encountered in clinical use. Specifically, these include short life of the elastic bands, necessitating costly replacement; sensitivity to metals worn by the subject (if not detected by the clinician and removed); poor or incorrect cable connections; saturation of amplifier gain; and the possibility of cross-talk between cables and other equipment in near proximity.
As an alternative to RIP systems, a method has been developed utilizing noncontact laser displacement sensors to obtain measurements of chest and abdominal movement during tidal breathing in normal subjects. Unlike RIP systems, the laser plethysmograph (LP) sensor method obtains single-point displacements. Therefore, a broad assumption is made that single-point measurement of the expansion and contraction of the chest and abdomen correlates to the overall peripheral displacements measured by the RIP elastic bands. If such is the case, then accurate measures of respiratory activity similar to those made by RIP systems could be obtained without the use of bands or any other form of contact with the patient. Furthermore, the laser sensors offer the ability to gather absolute measurements, potentially eliminating the need for calibration, and have a much higher frequency response, making measurement of short duration events (e.g., the chest wall motion of a cough or nasal inspiratory effort) much more feasible.
Another potentially attractive feature of the proposed LP method is the ability to assess measures at a specific, local level that may facilitate relevant assessments not feasible with the more global measures achieved by RIP bands. For example, symptoms of respiratory muscle dysfunction are subtle in the early stages of illness, causing a number of conditions to go undetected until late in their evolution.17 However, a number of clinical findings provide evidence of increased patient effort in compensatory muscle groups. An increase in the activity of the sternocleidomastoid and scalene muscles and recession of the suprasternal, supraclavicular, and intercostal spaces may all be indicative of neuromuscular disease affecting the respiratory system.
Accordingly, this paper reports the apparatus and method of the LP system and provides results of a validation study comparing tidal volume measures in normal subjects obtained by the LP technique and a standard RIP system. Furthermore, both displacement methods are compared to actual lung volume measurements determined by pressure variations in a control volume.
Apparatus and Experimental Method
The LP system was developed utilizing two noncontact laser displacement sensors (AR200-50, Acuity Research, Portland, OR) with a differential sensing range of 50 mm. The sensors were configured to provide a zero- to 10-volt analog signal proportional to sensed distance. Because of this high voltage level and range, no additional signal conditioning circuitry was necessary. The output signal from each laser sensor was interfaced to a computer-based data acquisition system (PMD-1208LS, Measurement Computing, Norton, MA). A readily portable, multiple degree-of-freedom mechanical holding fixture was provided for simple positioning of each laser sensor at the desired point of measurement on each subject.
To accomplish validation, a standard RIP system was utilized (RespiTrace 200, Non-Invasive Monitoring Systems, North Bay Village, FL). Analog voltage output from the RespiTrace ranges from ± 2.048 volts, with 100% tidal volume nominally ranging up to 400 millivolts. The analog outputs of the RespiTrace for both the rib cage band and the abdominal band were individually interfaced to the same data acquisition system utilized with the LP system.
Figure 1 shows the LP apparatus and its articulation with a subject. The subject is seated, and the lasers are positioned consistent with the testing method described. Although not shown, RIP bands were placed around the chest and abdomen at the same level as the laser sensors.
As the final element of the overall experimental apparatus, a cubical pressure chamber was constructed (24 in on each side) from acrylic sheets (Plexiglas) measuring 0.5 in thick. The chamber was completely sealed with dual layers of acrylic cement and silicone sealant. A manifold was constructed by cementing an additional 0.5 x 2 x 6 in piece of acrylic to the outer surface of the cube near the bottom. This was drilled and tapped for 0.75-in national pipe threads (NPT). A breathing tube with a 0.75-in threaded fitting on one end was mounted in a manifold hole and sealed. Remaining holes were sealed with pipe plugs. Another manifold block, measuring 0.5 x 1 x 1 in, was cemented to the opposite side of the chamber near the top. This manifold was drilled and tapped for 0.125-in NPT. A fitting and tube were connected to this hole and further affixed to a signal conditioned pressure sensor (SenSym ASCX15DN, Honeywell, Golden Valley, MN) providing analog voltage output of 0.3 volts/pounds per square inch (PSI). The pressure sensor's output voltage was amplified with a Burr-Brown INA114P instrumentation amplifier with gain of 101. The voltage from the amplifier circuit was then interfaced to the same data acquisition system as the RIP system and the laser sensors, providing a total of five channels of data acquisition for the overall apparatus.
Prior to use, the chamber was pressurized and the pressure sensor used to ensure that no leaks occurred over a significant period of time. Further, the 0.5-in acrylic thickness was chosen to permit sufficient rigidity to minimize flexing associated with pressure variations occurring as a subject breathes through the tube. The pressure test ensured that this was the case and that the volume of the chamber would effectively be constant. Thus, with the further assumption that the temperature was effectively constant, Boyle's Law could be used as the basis for determining the volume of air breathed in and out of the lungs as a function of measured pressure variations. Finally, the overall size of the chamber was chosen so that breathing in and out of such volume would not be significantly impeded due to pressure loading (verified by lead author).
For the data acquisition system, custom software was written by the lead author to acquire all five signals at a sample rate of 20 samples per second with 16-bit (0.015 millivolt) resolution. The samples were displayed on a strip-chart graphic user interface to permit visual inspection of all five traces simultaneously and ensure signal integrity. Acquired data were then written to comma separated variable (CSV) format for later import into analysis software.
In the experimental method to validate the LP system, six university student volunteers (five females, one male), ranging in age from 19 to 23 years, underwent a series of tidal breathing tests. Each subject gave written consent to participate and was screened as having no history of respiratory illness and no history of injury to any region that may affect the muscles of respiration or associated body mechanics, was a nonsmoker, and was in generally good health (see Table 1).
During testing, the subject was seated on a hard chair and instructed to sit upright so that no contact was made with the back of the chair. This was done to ensure that the RIP bands were not interfered with in any way. The bands were placed in areas common to tidal breathing measurements (i.e., directly superior to the umbilicus and across the chest just above the breasts). The subjects were further instructed to rest their arms on the chair arms to further ensure that no interference with the RIP bands occurred. In all cases, the bands were tightened by the same investigator to ensure consistency in over-all tension and continuity of elasticity throughout the breathing cycle.
Once the RIP bands were in place, the LP system was positioned so that one laser was pointed at the rib cage and another at the abdomen. Positioning of the lasers was further done to put the measurement “dot” (i.e., the laser light) directly on the RIP bands at the anterior mid-line. Thus, measurement of the rib cage and abdomen took place at exactly the same level by both techniques. In both cases, the lasers were oriented so that the beam was normal to the surface being measured. Nominal distance from the laser sensor to the skin surface was approximately 30 mm.
Once the RIP system and the LP system were in place, the subject was provided a breathing tube and instructed to breathe only through the mouth and to make sure no air escaped around the interface between the tube and lips. A research assistant held the tube up to the subject and ensured that seal integrity was maintained and that no perceptible breathing was taking place through the nose. At that point the subject was instructed to perform tidal breathing at the pace they were comfortable with. The acquisition system was started, and measurements of tidal breaths were taken for a period of several minutes. This process was repeated for each subject.
Collected data was imported into Microsoft Excel 2003 for graphic analysis and some statistical analysis. Other statistical analyses were done using MiniTab (Release 14) and SAS software version 8.2. All analyses were done with an α level of .05. Figure 2 shows a sample of traces taken from a subject. The rib cage and abdominal components of both the RIP system and the LP system were mathematically summed consistent with standard methods of determining tidal volume. Note also that the data for both the LP system and the volume has been scaled down to more closely match the smaller signal of the RIP system and provide relatively equal trace representations in this graphic. Data were not scaled for statistical analysis.
Data from the beginning to the end of each breathing excursion were analyzed for each subject for a minimum of seven breathing cycles. Preliminary correlation analyses were done to determine the initial pairwise relationships (LP-RIP, RIP-Vol, LP-Vol) between methods. Results of these correlations for each subject are summarized in Table 2.
To determine overall relationships, data for every breathing excursion of all six subjects were combined (57 total excursions) into a single set and intraclass correlation coefficients (ICC) were calculated (Shrout and Fleiss Case 3, since these instruments are the only measurement methods of interest).18 These findings indicate an excellent correlation among data from all methods. The ICCs for the pairwise combinations were each .995 (Table 2). Confidence intervals (CI) (95%) were calculated, but with the ICC being so close to 1.0 and the large number of observations involved, the CI was extremely small. The ICC calculated across all three measurement methods also was very high (ICC = .994).
The results of this study reinforce the long-accepted belief that respiratory inductive plethysmography is a reasonably accurate means of relating thoracoabdominal motion caused by breathing excursions to actual lung volumes in tidal breathing of normal subjects. A high correlation value (ICC = .995) between RIP excursion data and volume data determined by pressure variations in a control volume demonstrates this.
This study also tested a laser displacement sensor system utilized to acquire point measures of chest and abdominal excursions on the body's anterior midline at the same levels where RIP bands were placed. Analysis also showed high correlation value (ICC = .995) between the LP system's excursion measurements and volume data determined by pressure variations in a control volume.
Further, comparison between the data from the RIP system and the LP system also showed a very high correlation value (ICC = .995), suggesting that both methods produce remarkably similar results. This supports the notion that single-point measures obtained by an LP system are valid for representing tidal breathing in a normal population. Further testing to determine the robustness of the technique in populations where respiration is compromised is warranted.
These results indicate the LP system provides results that are, at a minimum, equivalent to those of the respiratory-inductive plethysmograph at the two anatomical sites commonly measured by the inductive plethysmograph. The potential for extending the range of measurements (e.g., focusing on specific respiratory muscles) using the laser plethysmograph raises the possibility that this approach will be able to provide much more extensive and precise measurements of chest wall function and respiratory musculature.
Jeff Hargrove is an associate professor of mechanical engineering at Kettering University; an adjunct assistant professor in the Department of Internal Medicine, College of Osteopathic Medicine, Michigan State University; and an adjunct assistant professor in the Department of Medicine, College of Human Medicine, Michigan State University. He holds a BS in electrical engineering and an MS and PhD in mechanical engineering.
Eric D. Zemper is a research assistant professor in the Department of Physical Medicine and Rehabilitation at the University of Michigan. He holds a BS and MS in microbiology and public health and a PhD in educational psychology.
Mary L. Jannausch is a senior statistician with the Department of Epidemiology, School of Public Health, at the University of Michigan. She holds a BS in statistics and an MS in biostatistics.