| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Department of Biomedical Engineering, Boston University, 02215 Boston, Massachusetts; and Department of Pediatrics, University of Bern, Inselspital, 3010 Berne, Switzerland
ABSTRACT
INTRODUCTION
METHODS
RESULTS
DISCUSSION
ACKNOWLEDGEMENTS
FOOTNOTES
REFERENCES
ABSTRACT 
Frey, Urs, Bela Suki, Richard Kraemer, and Andrew C. Jackson. Human respiratory input impedance between 32 and 800 Hz,
measured by interrupter technique and forced oscillations. J. Appl. Physiol. 82(3):
1018-1023, 1997.
Respiratory input impedance (Zin) over a wide
range of frequencies (f) has been
shown to be useful in determining airway resistance (Raw) and tissue
resistance in dogs or airway wall properties in human adults. Zin
measurements are noninvasive and, therefore, potentially useful in
investigation of airway mechanics in infants. However, accurate
measurements of Zin at these f values
with the use of forced oscillatory techniques (FOT) in infants are
difficult because of their relatively high Raw and large compliance of
the face mask. If pseudorandom noise pressure oscillations generated by
a loudspeaker are applied at the airway opening (FOT), the power of the
resulting flow decreases inversely with
f because of capacitive shunting into
the volume of the gas in the speaker chamber and in the face mask. We
studied whether high-frequency respiratory Zin can be measured by using rapid flow interruption [high-speed interrupter technique
(HIT)], in which we expect the flow amplitude in the respiratory
system to be higher than in the FOT. We compared Zin measured by HIT with Zin measured by FOT in a dried dog lung and in five healthy adult
subjects. The impedance was calculated from two pressure signals
measured between the mouth and the HIT valve. The impedance could be
assessed from 32 to 800 Hz. Its real part at low
f as well as the
f and amplitude of the first and
second acoustic resonance, measured by FOT and by HIT, were not
significantly different. The power spectrum of oscillatory flow when
the HIT was used showed amplitudes that were at least 100 times greater
than those when FOT was used, increasing at
f > 400 Hz. In conclusion,
the HIT enables the measurement of high-frequency Zin data ranging from 32 to 800 Hz with particularly high flow amplitudes and, therefore, possibly better signal-to-noise ratio. This is particularly important in systems with high Raw, e.g., in infants, when measurements have to
be performed through a face mask.
respiratory mechanics; tube technique; wave-tube technique; forced
oscillation technique
INTRODUCTION INPUT IMPEDANCE (Zin) of the respiratory system is
routinely measured at frequencies between 2 and 32 Hz. However, Zin
extending to higher frequencies has been shown to contain additional
information about the mechanical properties of the respiratory system.
Zin at frequencies between 2 and ~120 Hz can be used to separate
airway (Raw) and tissue resistance (Rti) as well as to estimate
thoracic gas volume (Vtg) in dogs (8) but not in adult
humans (7). In adult humans, Zin up to even higher frequencies (256 Hz)
has been shown to contain information about airway wall properties (5).
It is not yet known whether high-frequency Zin measurements in human
infants can provide information about Raw, Rti, and Vtg or about airway
wall properties. Accurate measurements of Zin at these frequencies with
the use of forced oscillatory techniques (FOT) in infants are difficult
because of low signal-to-noise ratios due to several causes (11).
First, the energy content of the forcing function in the FOT decreases
with increasing frequency (unpublished observations) because of
capacitive shunting into the gas volume in the loudspeaker. Second,
measurements in infants must be made through a face mask that acts as a
shunt compliance, which further degrades the energy content of the flow
into the infant's respiratory system as frequency increases. An
alternative broad spectrum signal that could be used to excite the
respiratory system is rapid airway interruption [high-speed
interrupter technique (HIT)]. Rapidly interrupting the flow at
the airway opening causes a step change in flow, the frequency content
of which, like the FOT, decreases with increasing frequency. However,
the magnitude of the flow-forcing function is expected to be much
larger in the HIT compared with the FOT. In this methodological paper,
we studied whether high-frequency Zin can be measured during airflow interruption. We performed these HIT measurements in a dried dog lung
model and in healthy adults and compared them with Zin measurements when using the FOT, where the FOT has been proven to measure
high-frequency Zin reliably (7).
METHODS Measurement system (Fig.
1). If the flow
interruption occurs instantaneously, the input flow signal applied to
the lungs is a step function, the frequency content of which varies as
1/frequency. Thus the more rapid the interruption (i.e., as close as
possible to a square wave), the higher the frequency content of the
flow. With a noninstantaneous interruption, the content at higher
frequencies is reduced. To preserve as much frequency content as
possible, a high-speed interrupter system was developed. This
interrupter is driven by a stepper motor (start speed = 1,200 Hz,
maximum speed 7,500 Hz, ramp time 13 ms). The shutter consists of a
rotating blade that closes (opens) the airway within 1 ms and remains
closed (open) for 14.5 ms. The complete interruption-closure cycle
occurs once every 31 ms. The motor is controlled by a digital-to-analog converter (model AT-MIO-16, National Instruments). The position of the
shutter valve (open or closed) is measured by a photo-optic resistor to
ensure that the shutter is reopened after an interruption so that the
subject is able to breathe between measurements. Each time the
interrupter is triggered, the valve mechanism rotates four times
producing four separate interruptions. The shutter mechanism is
connected to the mouthpiece through a tube of 0.6 cm in radius and 30 cm in length.
Zin was measured by using the wave-tube technique described in detail
elsewhere (2, 12). Briefly, in this technique, pressures are measured
at two locations (12 cm apart) along the tube between the shutter and
the mouthpiece. The proximal pressure transducer
(P1) was placed 5 cm and the
distal pressure transducer (P0)
17 cm from the airway opening (mouthpiece). Zin was computed as the
load impedance of the tube by
Fig. 1.
High-speed interrupter setup. P1, proximal pressure
transducer; P0, distal pressure transducer.
[View Larger Version of this Image (12K GIF file)]
where
L is the distance between the two
pressure transducers, Zc is the characteristic impedance of the tube,
and
(1)
is the propagation coefficient of the tube. The
pressures P1 and
P0 were measured with Microswitch
(model 164) transducers that were matched within ±2% in magnitude
and ±2° of phase up to 1,024 Hz. The electrical output of these
transducers was band-pass filtered (8-2,000 Hz) and
analog-to-digital converted at 8,258 Hz (model AT-MIO-16, National
Instruments). Data were stored during four complete cycles of the
interrupter (4 × 31 ms = 124 ms, with a sampling rate of 8,258 Hz = 1,024 points). The ratio of
P1/P0
was estimated from the cross power spectra of
P0P1
and the auto power spectrum of P1.
The method for measuring Zin by the FOT and its validation up to 2,000 Hz has been described elsewhere (12). Briefly, Zin(
) was measured by
using loudspeaker-generated, low-amplitude pressure oscillations
applied to the airway opening. A pseudorandom noise signal containing
frequencies between 32 and 800 Hz in 8-Hz increments was generated by
computer and output via a digital-to-analog converter. Zin was measured
by using the same wave-tube technique described above.
The quality of the measurements for both HIT and FOT was assessed by computing the coherence function according to Michaelson et al. (14).
Experimental protocol. Zin of a dried dog lung was measured by HIT and FOT and compared at each frequency. The measurements of Zin by using the HIT in a dried dog lung were possible because the alveoli on the pleural surface of a dried dog lung are permeable, presumably because the surface had dried and cracked. As a consequence, a positive pressure applied to the trachea results in airflow through the dog lung. The airflow interruption was performed at a flow rate of 0.1 l/s. Zin was also measured by using FOT and HIT in five healthy nonsmoking adult volunteers whose biometric data are given in Table 1. Informed consent was obtained, and the study was approved by the Institutional Review Board at Boston University. The measurements in the human subjects were made as follows: during tidal breathing, the airflow interruption was triggered at an inspiratory flow rate of 0.1 l/s during 10 respiratory cycles. The subjects were seated, wearing noseclips, and they supported their cheeks by their hands. The mean and SD as well as the coherence of the 10 measurements were calculated.
|
|||||||||||||||||||||||||||||||||||||||||||||
RESULTS
Figure 2 shows the Zin and coherence of the
dried dog lung measured by the HIT and by the FOT. A paired
t-test
(P < 0.01) was used to determine
those frequencies where there was a significant difference between
impedances measured by the two different techniques. For nearly all
frequencies, the differences were insignificant, except at those
frequencies surrounding the antiresonances and between 160 and 290 Hz
in the imaginary part. Coherence was >0.95 for all frequencies.
Figures 3 and 4 show the means ± SD and coherence of the Zin measured by the FOT
and by the HIT in two representative healthy subjects. Apart from
single frequency points close to the antiresonances and at very high
frequencies, the coherence of the measurements was >0.95 between 32 and 800 Hz in both techniques. In all five subjects, there were two
dominant antiresonances when using both the HIT [first
antiresonance (far,1) at 266 ± 39 Hz, second antiresonance (far,2)
at 653 ± 104 Hz] and FOT (255 ± 36 and 650 ± 114 Hz,
respectively). Antiresonances occurred where the imaginary
part was zero and the real part showed a relative maximum. The
frequencies and amplitudes of the antiresonances were not significantly
different in the group of five subjects when the two techniques were
compared (Table 2). As a means of comparing
the variability of the two techniques, we computed the average SD for
all frequencies in all five subjects. The average SD in the real part
measured by HIT was 3.59 ± 1.61 cmH2O · l
1 · s1
and measured by FOT was 1.37 ± 1.01 cmH2O · l1 · s1,
which was not significantly different (paired
t-test,
P > 0.05). In the imaginary part,
the averaged SD was significantly different between the HIT (4.03 ± 1.13 cmH2O · l1 · s1)
and the FOT (0.99 ± 0.62 cmH2O · l1 · s1),
P < 0.05.
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
DISCUSSION
Zin data over a wide range of frequencies (2 < f < 100 Hz) have been shown to be useful in providing separate estimates of Raw, Rti, and Vtg in dogs (13). However, to obtain reliable estimates of these parameters one must obtain data up to an antiresonance (at ~80 Hz in dogs) associated with gas compression compliance (Cg) and tissue inertance (Iti). In dogs, there is a second antiresonance at ~180 Hz, which is an acoustic antiresonance that is primarily dependent on airway path length and airway wall properties (12). Unlike dogs, adult humans do not have a Cg-Iti-related antiresonance, but their first antiresonance (at ~180 Hz) is acoustic related (7). As a consequence, Zin data in adult humans cannot be used to estimate Raw, Rti, and Vtg.
Zin measurements are particularly attractive for investigating lung function in infants, since they are noninvasive, do not require patients' cooperation, can be performed during tidal breathing, and have the potential of providing estimates of clinically useful parameters (Raw, Vtg). However, it is not known whether infants have a Cg-Iti-related antiresonance and thus whether infant Zin data could provide estimates of Raw or Vtg. Zin measurements have been made up to 256 Hz in infants by using the FOT, and an antiresonance was found at ~113 Hz (11). However, this antiresonance has shown to be influenced by the parallel combination of face-mask gas-compression compliance and total respiratory system inertance. It is possible that a Cg-Iti antiresonance occurs at frequencies >256 Hz, but measurements with the use of the FOT at these high frequencies are problematic because of shunting of face-mask dead space (11). Infants have relatively high Raw in comparison with adults (18). High-frequency Zin measurements in infants are difficult. If pseudorandom noise pressure oscillations generated by a loudspeaker are applied at the airway opening (FOT), the power of the resulting flow decreases inversely with frequency because of capacitive shunting into the volume of the gas in the speaker chamber and in the face mask. Because the energy content of the forcing function decreases with increasing frequency in the FOT, we expect low flows at higher frequencies, particularly in the presence of a parallel shunt impedance (gas in the face mask). But to measure high-frequency respiratory impedance accurately we need high oscillatory flow amplitudes at higher frequencies.
To make measurements to higher frequencies, we considered
using an alternative broad spectral forcing function, a pseudostep function generated by the HIT, that provides increased energy of the
forcing function at these higher frequencies. The HIT was derived from
the standard interrupter technique introduced by von Neergaard and Wirz
(15). As proposed by von Neergaard and Wirz and as implemented by
several others since (1, 9, 10, 17), estimates of resistance were
provided from analysis of the airway opening pressure (Pao) as a
function of time following interruption and the airflow at airway
opening (
ao) signal just before
interruption, assuming that there is a complete equilibration between
alveolar and mouth pressure. In a more recent study by Romero et al.
(16) in dogs, the Pao signal was analyzed in frequency domain. Here it
was shown that an antiresonance in the power spectrum of Pao occurred
at ~80 Hz and a gas density-dependent antiresonance occurred at 180 Hz. They speculated that these antiresonances corresponded to the
Cg-Iti tissue resonance (80 Hz) and the first acoustic resonance (180 Hz) described by Jackson and Lutchen (8), respectively. Frey et al. (3,
4) found in adult humans only a single antiresonance and speculated
that it corresponds to the acoustic antiresonance in the Zin spectrum
found by Jackson et al. (7).
On the basis of these previous studies in dogs and humans, it appears
that similar information can be measured by the FOT and by the HIT. In
fact, if ao and Pao are measured before and during
interruption, Zin could be computed. Although the HIT and FOT
techniques are similar in that they both provide estimates of Zin, they
are different in that they use different forcing functions. The FOT
uses a pseudorandom noise that has, in most cases, equal energy at each
frequency of interest, whereas the HIT uses a pseudostep function the
energy of which is distributed as 1/frequency. Another difference
between the two techniques is the magnitude of the forcing function;
the flow and pressure signals are much larger when the HIT is used than
they are in the FOT (Fig. 5). In both
techniques, the power of the pressure and flow signals decreases with
frequency, but the HIT provides flow signals with two orders of
magnitude higher energy at higher frequencies than the FOT.
Furthermore, the decrease of the flow power with frequency is less in
the HIT than in the FOT. This effect can be seen even more distinctly
in the power spectrum of pressure and flow signals measured in a
12-mo-old infant through a face mask (Fig.
6). Whereas the power of the FOT flow
signal drops considerably at frequencies >400 Hz, the power of the
HIT flow signal is much higher at these frequencies.
, Power spectrum of pressure (HIT); 
, power spectrum of resulting flow (HIT); 
, power spectrum of pressure (FOT); 
,
power spectrum of resulting flow (FOT).
, Power spectrum of pressure (HIT); , power spectrum of
resulting flow (HIT); , power spectrum of pressure (FOT); , power
spectrum of resulting flow (FOT). Coherence: solid line, HIT; dotted
line, FOT.
These results show that the HIT enables the excitement of higher flow amplitudes at high frequencies than does the FOT. The amplitude of the FOT flow signal could theoretically have been increased if larger loudspeakers were used. However, even with much larger speakers, it is unlikely that the flow amplitudes could be increased to the level possible with the HIT. The question arises whether the large flow amplitudes produced by the HIT cause nonlinear behavior of the respiratory system. Because the pseudostep function produced by the HIT is a continuous signal and thus has a continuous spectrum in frequency domain, nonlinear behavior will cause a decrease in the coherence function. However, this was not the case (Figs. 4 and 6). We further explored the possibility of introducing nonlinear behavior by performing flow interruption at higher flows. We found that at flows higher than 100 ml/s the coherence decreased, which suggests that the respiratory system may begin to behave nonlinearly at flows >100 ml/s. If the forcing function is a pseudorandom noise signal with a discrete spectrum (FOT), the coherence function would not necessarily decrease when the respiratory system behaves nonlinearly. In this respect, an advantage of the HIT over the FOT is that its coherence is a function of nonlinear behavior. Thus decreased coherence could be used as an indication of nonlinear behavior.
By using the HIT it was possible to measure Zin from 32 to 800 Hz with a coherence of 0.95 in a dried dog lung and in humans. The comparison of impedance measurements in a dried dog lung done by FOT and HIT showed a congruence of the spectra over a large range. At frequencies <32 Hz, the HIT shows a large SD, and, therefore, the data <32 Hz are not reliable. This is due to the fact that each interruption lasted for only 31 ms, and, as a consequence, there is no information below 32 Hz when one is using the HIT. However, the interrupter time could be increased to investigate lower frequency ranges.
In humans, the correlation between the measurements obtained by using
the FOT and the HIT was not as good as in the dog model. In humans,
measurements have to be made through the upper airways, which might
contribute to the variability of the measurements and might explain
differences between the techniques. However, we tried to minimize these
effects by stabilizing the cheeks (6) and standardizing the head
position. Because the coherence function was 0.95 in these measurements
apart from single frequencies, nonlinear effects cannot be a major
reason for the differences in the Zin measured by these two techniques.
As discussed above, a major difference between the two techniques is
the amount of energy contained in the forcing function, which could
result in differences in the variability of the Zin estimated by the
different techniques. However, since the HIT forcing function contains
more energy, one would expect the SD of Zin measured by HIT to be
smaller than the SD of Zin measured by FOT. This was not
the case in adult subjects (e.g., Figs. 3 and 4). We cannot completely
exclude the possibility that entrance effects of the oscillations into
the respiratory system [i.e., due to abrupt changes in airway
geometry in the upper airways (glottis) as well as bifurcations in the central airways where Reynold's numbers are high] are different in both techniques. This might contribute to differences between Zin
measurements in both techniques and remains to be shown. Differences in
the variability of the Zin could also be due to differences in the lung
volume at which the measurements were made. As shown by Frey et al.
(4), the oscillatory pressure transients after airflow interruption are
sensitive to the change in lung volume. Although changes in
postinterruption Pao do not necessarily indicate that Zin is volume
dependent, this may be the case. The HIT measurements were made at the
beginning of a breath where ao was <100 ml/s, whereas the FOT measurements were made throughout the entire breathing cycle. If Zin is sensitive to lung volume, then we would have expected
the SD of the HIT-measured Zin to be less than the SD of the Zin
measured by FOT, which was not the case.
In summary, the HIT enables the measurement of high-frequency Zin data from 32 to 800 Hz in healthy adults. Impedance data in this frequency range potentially enable the noninvasive measurements of resistive and elastic properties of the airways. The forcing function derived from the HIT excites higher flow amplitudes at high frequencies. Measurements in systems with high impedance, such as, for example, the infant lung where measurements have to be performed through a face mask, might have a better signal-to-noise ratio and, therefore, better coherence than the FOT (see Fig. 6). However, this effect cannot clearly be seen in the impedance spectrum of healthy adults. Further advantage for use in infants consists of the easy use and cleaning of the device, the small dead space, the short measurement period, and the fact that the forcing function in the HIT provides a continuous spectrum, which causes a drop in coherence function if the system behaves nonlinearly. Because the HIT device is a flow-through system, whereas the FOT is not, it can be more easily implemented into ventilatory circuits.
ACKNOWLEDGEMENTS
We thank A. Fritschi, who built the high-speed interrupter valve following our suggestions.
FOOTNOTES
Address for reprint requests: U. Frey, Dept. of Child Health, Univ. of Leicester School of Medicine, Robert Kilpatrick Clinical Sciences Bldg., Leicester Royal Infirmary, PO Box 65, Leicester LE2 7LX, UK.
Received 5 November 1995; accepted in final form 28 November 1996.
REFERENCES
| 1. |
Bates, J. H. T.,
M. S. Ludwig,
P. D. Sly,
K. Brown,
J. G. Martin,
and
J. J. Fredberg.
Interrupter resistance elucidated by alveolar pressure measurements in open-chest normal dogs.
J. Appl. Physiol.
65:
408-414,
1988.
|
| 2. | Franken, H., J. Clement, M. Cauberghs, and K. P. Van de Woestijne. Oscillating flow of a viscous compressible fluid through a rigid tube: a theorethical model. IEEE BME 28 (5): 417-420, 1981. |
| 3. | Frey, U., and R. Kraemer. Analysis of oscillatory pressure transients after flow interruption in healthy and asthmatic children. Pediatr. Pulmonol. 19: 379-388, 1995. [Medline] |
| 4. | Frey, U., A. Schibler, and R. Kraemer. Pressure oscillations after flow interruption in relation to lung mechanics. Respir. Physiol. 102: 225-237, 1995. [Medline] |
| 5. |
Habib, R. H.,
R. B. Chalker,
B. Suki,
and
A. C. Jackson.
Airway geometry and wall mechanical properties estimated from subglottal input impedance in humans.
J. Appl. Physiol.
77:
441-451,
1994.
|
| 6. |
Habib, R. H.,
and
A. C. Jackson.
Total respiratory input impedance with the upper airway wall shunt minimized.
J. Appl. Physiol.
74:
1045-1055,
1993.
|
| 7. |
Jackson, A. C.,
C. A. Giurdanella,
and
H. L. Dorkin.
Densitiy dependance of respiratory system impedances between 5 and 320 Hz in humans.
J. Appl. Physiol.
67:
2323-2330,
1989.
|
| 8. |
Jackson, A. C.,
and
K. R. Lutchen.
Modeling of respiratory system impedances in dogs.
J. Appl. Physiol.
62:
414-420,
1987.
|
| 9. | Jackson, A. C., and H. T. Milhorn. Digital computer simulation of respiratory mechanics. Comput. Biomed. Res. 6: 27-56, 1973. [Medline] |
| 10. |
Jackson, A. C.,
H. T. Milhorn,
and
J. R. Norman.
A reevaluation of interrupter technique for airway resistance measurements.
J. Appl. Physiol.
36:
264-268,
1974.
|
| 11. | Jackson, A. C., K. M. Neff, K. R. Lutchen, and H. R. Dorkin. Interpretation of respiratory system impedances (4-256 Hz) in healthy infants. Eur. Respir. Rev. 4 (19): 165-166, 1994. |
| 12. |
Jackson, A. C.,
B. Suki,
M. Ucar,
and
R. Habib.
Branching airway network models for analyzing high-frequency lung input impedance.
J. Appl. Physiol.
75:
217-227,
1993.
|
| 13. |
Lutchen, K. R.,
and
A. C. Jackson.
Reliability of parameter estimates from models applied to respiratory impedance data.
J. Appl. Physiol.
62:
403-413,
1987.
|
| 14. | Michaelson, E. D., E. D. Grassman, and W. R. Peters. Pulmonary mechanics by spectral analysis of forced random noise. J. Clin. Invest. 56: 1210-1230, 1975. |
| 15. | Neergaard, K. von, and K. Wirz. Die Messung der Strömungswiderstände in den Atemwegen des Menschen, insbesondere bei Asthma und Emphysem. Z. Clin. Med. 105: 51-82, 1927. |
| 16. |
Romero, P. V.,
J. Sato,
F. Shardonowski,
and
J. H. Bates.
High-frequency charactersistics of respiratory mechanics determined by flow interruption.
J. Appl. Physiol.
69:
1682-1688,
1990.
|
| 17. | Shepard, R. J. Mechanical characteristics of the human airway in relation to use the interrupter valve. Clin. Sci. Lond. 25: 263-279, 1962. |
| 18. |
Stocks, J.,
and
S. Godfrey.
Specific airway conductance in relation to postconceptional age during infancy.
J. Appl. Physiol.
43:
144-154,
1977.
|
This article has been cited by other articles:
|
|
 |
|
  M. Henschen, J. Stocks, I. Brookes, and U. Frey New aspects of airway mechanics in pre-term infants Eur. Respir. J., May 1, 2006; 27(5): 913 - 920. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
E. Oostveen, D. MacLeod, H. Lorino, R. Farre, Z. Hantos, K. Desager, and F. Marchal The forced oscillation technique in clinical practice: methodology, recommendations and future developments Eur. Respir. J., December 1, 2003; 22(6): 1026 - 1041. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 U. FREY, K. MAKKONEN, T. WELLMAN, C. BEARDSMORE, and M. SILVERMAN Alterations in Airway Wall Properties in Infants with a History of Wheezing Disorders Am. J. Respir. Crit. Care Med., June 1, 2000; 161(6): 1825 - 1829. [Abstract] [Full Text]
|
|
|
|
 |
|
 V. Kessler, G. Mols, H. Bernhard, C. Haberthur, and J. Guttmann Interrupter airway and tissue resistance: errors caused by valve properties and respiratory system compliance J Appl Physiol, October 1, 1999; 87(4): 1546 - 1554. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Visit Other APS Journals Online |
