Download Fiber-Optic Stethoscope: A Cardiac Monitoring and Gating System

Magnetic Resonance in Medicine 47:314 –321 (2002)
Fiber-Optic Stethoscope: A Cardiac Monitoring and
Gating System for Magnetic Resonance Microscopy
Anja C.S. Brau, Charles T. Wheeler, Laurence W. Hedlund, and G. Allan Johnson*
A fundamental problem associated with using the conventional
electrocardiograph (ECG) to monitor a subject’s cardiac activity
during magnetic resonance imaging (MRI) is the distortion of
the ECG due to electromagnetic interference. This problem is
particularly pronounced in MR microscopy (MRI of small animals at microscopic resolutions (< 0.03 mm3)) because the
strong, rapidly-switching magnetic field gradients induce artifacts in the animal’s ECG that often mimic electrophysiologic
activity, impairing the use of the ECG for cardiac monitoring
and gating purposes. The fiber-optic stethoscope system offers
a novel approach to measuring cardiac activity that, unlike the
ECG, is immune to electromagnetic effects. The fiber-optic
stethoscope is perorally inserted into the esophagus of small
animals to optically detect pulsatile compression of the esophageal wall. The optical system is shown to provide a robust
cardiac monitoring and gating signal in rats and mice during
routine cardiac MR microscopy. Magn Reson Med 47:
314 –321, 2002. © 2002 Wiley-Liss, Inc.
Key words: MRI; MR microscopy; ECG; EKG; cardiac monitoring; cardiac gating
The development of transgenic small-animal models of
cardiac disease has focused attention in recent years on
MR microscopy as a tool to characterize cardiac structure
and function (1–5). A crucial component of in vivo MR
microscopy is the ability to continuously assess the animal’s cardiac activity, for purposes of both cardiac monitoring and cardiac gating (the careful synchronization of
imaging to the heart cycle). However, the method conventionally used to measure cardiac function, the electrocardiograph (ECG), is corrupted during MR microscopy by the
imaging process itself. According to Faraday’s Law, the
time-varying magnetic field gradients applied during imaging induce an electromotive force (EMF) in the conductive loop formed by the animal’s body and the ECG circuit.
Since the ECG circuit is designed to measure the differential voltage between electrodes on the subject’s body, it
detects not only true electrophysiologic potentials, but
also any superimposed voltage artifacts. While ECG distortion can occur with all MRI techniques, it is exacerbated
in cardiac MR microscopy, in which pulse repetition times
(TRs) are short and gradient slew rates are high— up to two
orders of magnitude larger than those in clinical MR systems. The resulting voltage artifacts can span the entire
cardiac cycle, often having similar amplitude (⬇1.0 mV)
Center for In Vivo Microscopy, Duke University Medical Center, Durham,
North Carolina.
Grant sponsor: NIH/NCRR; Grant number: P41 RR05959; Grant sponsor:
NIH/NHLBI; Grant number: R01 HL 55348.
*Correspondence to: G. Allan Johnson, Center for In Vivo Microscopy, Duke
University Medical Center, Box 3302, Durham, NC 27710. E-mail:
[email protected]
Received 12 July 2001; revised 21 September 2001; accepted 22 September
2001.
© 2002 Wiley-Liss, Inc.
DOI 10.1002/mrm.10049
and bandwidth (⬇200 Hz) to the ECG of the small animal
being imaged, rendering thresholding and filtering techniques ineffective at discerning the true ECG.
The consequences of a corrupted ECG are twofold. First,
the ECG is undermined as a tool to assess the animal’s
health. The ability to monitor heart rate and activity is
essential in animal studies, particularly when anesthesia
or disease may alter cardiac function. Yet because ECG
voltage artifacts tend to mimic electrophysiologic potentials, cardiac activity is obscured. Second, a corrupted ECG
impairs cardiac gating. Voltage artifacts may mistakenly be
interpreted as R-waves and trigger erroneous cardiac gating, resulting in noisy, temporally misregistered images.
Several strategies can be employed to improve the quality of the ECG during MRI, including the use of complex
signal processing techniques (6,7) and high-resistance or
even fiber-optic leads to electrically isolate the ECG circuit
(8). Despite these techniques, the ECG, being an inherently
electrical measurement, will always suffer from electromagnetic interference. The subject’s body alone acts as a
conductive medium though which magnetic flux lines
pass and induce an EMF. Alternative measures of cardiac
activity could be used, including peripheral blood pressure and blood perfusion techniques. In general, however,
such superficial measurements suffer from low sensitivity,
particularly in small animals (9). Moreover, the arterial pulse
from peripheral tissue can vary over the course of an experiment due to temperature- or drug-induced vasoconstriction/
dilation. Legendre et al. (10) successfully detected cardiac
activity in rats using a fiber-optic reflectometer; however,
their technique was invasive, requiring arterial catheterization. Thus, none of the existing techniques have been shown
to provide reliable, noninvasive cardiac monitoring ability,
much less gating ability, in small animals.
The purpose of this study was to design and implement
a cardiac monitoring and gating system that was immune
to electromagnetic interference during MR microscopy.
The system was required to be: 1) inherently nonelectrical,
2) MR-compatible, 3) noninvasive (requiring no surgery),
and 4) small enough for use on rats and mice.
METHODS
Fiber-Optic Stethoscope System Design
As shown in the MR image of the mouse thorax (Fig. 1), the
mammalian esophagus shares close anatomical proximity
with the heart, lying dorsal to the aortic arch (AA) and left
atrium (LA), then adjoining the descending aorta until
emptying into the stomach. As a result of this proximity,
the heart imposes significant pulsatile compression of the
esophagus, as has been observed in humans both directly
via endoscopy (11) and indirectly via esophageal pressure
signals (12). The premise for this work was that this me-
314
Fiber-Optic Stethoscope for MR Microscopy
315
FIG. 1. Mid-sagittal slice of the mouse
thorax revealing the proximity of (a) the
esophagus (arrows) to (b) the heart. This
image was extracted from an isotropic 3D
MR volume of a fixed mouse, part of a
larger project (20) to provide a web-based
atlas for morphologic phenotyping. The
animal is prone, with its head towards the
left. The spine is labeled for clarification
(d). The esophagus is not clearly visible in
the cervical section because the esophageal lumen is collapsed there. At the level
of the esophageal hiatus, the esophagus
curves leftwards, out of the plane shown,
and then empties into (c) the stomach.
The esophagus is visible in cross section
above the stomach.
chanical esophageal compression could be detected optically as an indication of cardiac activity. Optical displacement sensors have previously been employed for respiratory gating in MRI (13,14), and techniques such as
transesophageal ECG monitoring and echocardiography
have exploited the heart– esophagus proximity for closer
access to the heart; however, to the authors’ knowledge,
esophageal wall motion has never before been detected as
a means to gauge cardiac activity.
A fiber-optic “stethoscope” was created that could be
inserted into the esophagus of a small animal to optically
detect esophageal wall pulsations related to cardiac activity. A schematic of the fiber-optic stethoscope system is
shown in Fig. 2. Two step-index multimode optical fibers
(OFs) (Thorlabs, Newton, NJ) were used: one for transmission and the other for detection of light. These silica fibers
had numerical apertures (NA) of 0.37, and were 5 m in
FIG. 2. Schematic of the fiber-optic stethoscope system. The transmit fiber carries 650-nm light from the laser diode to the esophagus
of the animal in the magnet. Reflected and backscattered light is
detected by the receive fiber and conveyed to an amplified photodetector outside the magnet. A signal-processing circuit then generates two signals from the detected optical signal: 1) an amplified
monitoring signal that is sent to a physiologic monitor for display,
and 2) a 5-ms, 5-V gating pulse that is sent to the scanner to trigger
cardiac-gated imaging.
length. The last 8 cm of each fiber were stripped of buffer
to a diameter of 125 ␮m, and then the two bare (core ⫹
cladding) fibers were bundled together with epoxy for a
total diameter of 250 ␮m. The fiber tips were cleaved and
polished to maximize light detection. In some experiments, polyethylene tubing (inner diameter ⫽ 1.2 mm)
filled with a 1:500 concentration of Magnevist (gadopentetate dimeglumine; Berlex Laboratories, Wayne, NJ)
was secured along the length of the OFs with the tubing tip
at the same level as the OF tips. While the OFs themselves
do not yield signal in MR images, the Magnevist-filled
tubing generates a strong signal and therefore provides a
distinct marker of stethoscope position.
Collimated continuous-wave light from a 40-mW,
650-nm laser diode (Thorlabs) was focused into the transmit fiber using an optical lens. As light from the transmit
fiber impinged upon a moving surface, such as the esophageal wall, the amount of diffusely reflected and backscattered light detected by the receive fiber increased as the
surface moved closer to the probe tip and decreased as it
moved away from the tip. The detected light was conveyed
by the receive fiber to an amplified p-i-n photodiode
(gain ⱕ 125 dB) (Thorlabs) outside the imaging magnet.
The resulting electrical signal was passed to a custom-built
signal-processing circuit for further amplification, and
then to a physiologic monitor for recording and display.
Unlike the ECG, the more slowly-varying optical signal
did not contain a high-frequency spike from which to
trigger cardiac-gated imaging. Instead, a second signal—a
5-ms, 5-V square-wave pulse—was generated by the circuit
whenever the slope of the optical signal exceeded a certain
threshold. This signal, similar to the ECG in appearance but
fundamentally different in origin, was used to trigger cardiacgated imaging and to measure the animal’s heart rate.
Animal Preparation
All animal studies were conducted at the Duke Center for
In Vivo Microscopy (CIVM) in accordance with institu-
316
Brau et al.
FIG. 3. Anesthetized rat lying prone in the cradle. Anesthesia is maintained by the delivery of isoflurane by the MR-compatible valve on the
left. Pediatric electrodes are taped to the animal’s footpads to acquire a reference ECG. The fiber-optic stethoscope is shown in front. The
two fibers, transmit and receive, are bundled together for a length of 8 cm. The plastic tubing in the animal’s mouth permits easy insertion
of the fiber-optic stethoscope into the animal’s esophagus.
tional animal care regulations. Twenty-eight rats (Charles
River, Wilmington, MA) weighing 150 –250 g, and one
C57BL/6J mouse (Jackson Lab, Bar Harbor, ME) weighing
40 g were anesthetized with methohexital (45 mg/kg,
Brevital; Eli Lilly, Indianapolis, IN) and placed prone in a
plexiglas cradle. Following intubation, anesthesia was
maintained with 2–3% isoflurane (IsoFlo; Abbott Laboratories, North Chicago, IL) delivered by a computer-controlled, MR-compatible pressure ventilator set at 40 – 80
breaths/min (15). A solid-state pressure transducer on the
breathing valve measured airway pressure. Pediatric electrodes were taped to the animal’s footpads to acquire a
reference ECG signal. Body temperature was measured
using a rectal thermistor. In some experiments, arterial
pressure was measured by inserting a pressure transducer
into the AA via the right carotid artery. All physiologic
signals were processed (Coulbourn Instruments, Allentown, PA) and displayed on a computer using LabVIEW
software (National Instruments, Austin, TX) for continuous monitoring throughout the experiments.
An 8-cm length of plastic tubing, slightly larger in diameter than the bundled OFs, was lubricated with Surgilube (Altana Inc., Melville, NY) and perorally inserted into
the animal’s esophagus to the mid-thorax. The purpose of
the tubing was to protect the walls of the esophagus from
perforation as the OFs were inserted. The protective tubing
was then withdrawn to expose the fibers to the esophageal
wall. The laser diode was turned on and the ambient room
lights were turned off to reveal a red glow emanating from
inside the animal’s body, permitting easy positioning of
the fiber tip near the heart. Photothermal effects were
negligible because tissue absorption of 650-nm light is
minimal (16). The probe was then secured to the cradle to
remain stationary throughout the experiment. Figure 3
shows a rat anesthetized by ventilator-delivered isoflurane, with the fiber-optic stethoscope shown in front, prior
to insertion in the animal’s esophagus.
Imaging Setup
For this study, cardiac-gated imaging using the fiber-optic
stethoscope was performed on rats only. The animal in the
cradle was placed inside a 7-cm-diameter birdcage RF coil
and positioned in a 30-cm bore 2.0 T magnet (Oxford
Instruments, Oxford, UK) with shielded gradients (up to
0.18 T/m) controlled by a Signa console (General Electric
Medical Systems, Milwaukee, WI). Body temperature was
maintained by a feedback system that controlled heated
airflow through the magnet (17). The combination of
closely regulated body temperature and anesthesia delivery kept the animal’s heart rate to within 5% of its baseline
rate. Baseline heart rates averaged 300 beats/min. To minimize image motion artifacts from breathing, data acquisition was enabled by a 600-ms window during end-expiration of the respiratory cycle. Cardiac-gated, radial-acquisition (RA) cine imaging (18) was then triggered by the
fiber-optic gating pulse rather than the conventional ECG.
RESULTS AND DISCUSSION
Initial benchtop experiments revealed that the optical signal detected by the fiber-optic stethoscope exhibited periodic amplitude oscillations matching the frequency of the
cardiac cycle as measured by the ECG. The shape and
Fiber-Optic Stethoscope for MR Microscopy
317
FIG. 4. Physiologic signals (left column) and corresponding axial MR images (right column) from a 250-g rat with the tips of the OFs and
the Magnevist marker tube inserted to esophageal depths of 7 cm (top row) and 6 cm (bottom row), respectively. In the MR images, the
OFs are visible as dark circular regions directly below the bright MT. The ECG, detected optical signal, and aortic pressure waveforms are
arbitrarily amplified and offset for better viewing (timing and gain settings are the same in a and c). b: The OF tips are at the same axial level
as the LV and RV. d: The OF tips neighbor the AA just below the tracheal bifurcation. The data in a and b indicate that, at a depth of 7 cm,
ventricular contraction causes the variations in optical signal, while in c and d, at a depth of 6 cm, aortic pulsation appears to modulate the
signal.
timing of these variations were observed to vary with the
position of the stethoscope tip in the thorax, indicating
that the nature of esophageal compression depended
greatly on the surrounding tissue. To further examine this
point, the OFs were inserted to several different esophageal depths and the resulting optical signal was recorded
and compared to the ECG and aortic pressure waveforms.
The relative timing of the optical signal with respect to
these other known metrics of cardiac activity helped to
elucidate the anatomical relationship between the heart
and the stethoscope tip in the esophagus. As a secondary
verification of fiber-tip position, a series of MR images was
acquired at incremental axial levels through the thorax
from head to tail, and the path of the Magnevist tubing
(MT) was followed to its end to pinpoint the exact position
of the stethoscope tip and therefore the site of optical
signal detection.
Figure 4a and b shows sample results from an experiment in which the fiber-optic stethoscope was inserted to
a depth of 7 cm in the esophagus of a 250-g rat. The ECG,
detected optical signal, and aortic pressure waveforms are
plotted in Fig. 4a. The optical signal amplitude decreases
suddenly after the R-wave of the ECG, reaching a minimum just before the peak in aortic pressure. Because it
occurs between ventricular depolarization and peak aortic
pressure, the optical signal trough is likely due to contraction of the ventricles. Systolic ventricular contraction involves a decrease in chamber volume (19), which then
causes the neighboring tissue, including the esophagus, to
experience a corresponding expansion. An expanded
esophageal diameter lengthens the distance between the
esophageal wall and stethoscope tip and decreases the
amount of reflected light detected by the receive fiber. As
the ventricles fill again in diastole, the heart expands and
compresses the esophagus, causing the amount of detected
light to steadily increase.
The corresponding axial MR image (Fig. 4b) shows the
position of the OF tips relative to the cardiac ventricles.
The OF tips are visible as two dark circular regions directly below the bright MT. The proximity of the esophagus to the left ventricle (LV) and right ventricle (RV) is
apparent in the image, and supports the theory that ventricular contraction is responsible for optical signal variations at this depth.
318
Brau et al.
FIG. 5. (a) Physiologic signals and (b) corresponding axial MR image from a 250-g rat with the fiber-optic stethoscope inserted to an
intermediate esophageal depth of 6.5 cm. The ECG, optical signal, optically-derived gating pulse, and airway pressure waveforms are
shown in a. Signals are arbitrarily amplified and offset for better viewing. The optical signal exhibits distinct peaks in end-diastole,
suggestive of atrial activity. b: The corresponding axial MR image confirms the proximity of the LA to the OFs below the bright MT. The
optically-derived gating pulses consistently coincide with the R-wave of the ECG, as represented by the dotted line.
The fiber-optic stethoscope was then withdrawn 1 cm to
an esophageal depth of 6 cm, and a very different optical
signal was detected, as shown in Fig. 4c. Whereas in Fig.
4a the optical signal troughs precede peaks in aortic pressure, in Fig. 4c the optical signal troughs coincide with
aortic pressure peaks (the two signals essentially mirror
each other), suggesting that aortic rather than ventricular
motion was being detected in this case. In addition, the
signal-to-noise ratio (SNR) of the optical signal is greater in
c than in a, suggesting that the extent of cardiac-induced
esophageal compression has increased because of closer
proximity. Indeed, the corresponding axial MR image in
Fig. 4d confirms that the OF tips border the AA at this
level. It is known that just below the tracheal bifurcation,
the AA partially encircles the esophagus and, together
with the left main bronchus, causes left-sided compression
of the esophagus known as the aortic narrowing. Significant pulsatile motion has been observed at the site of aortic
narrowing due to dorsal shifting of the aorta (11). Thus,
systolic aortic shifting may momentarily increase esophageal diameter and decrease the amount of light detected by
the receive fiber at this depth.
The optimal OF position for cardiac gating purposes in
the rat was found to be at an intermediate esophageal
depth of 6.5 cm. At this level, the detected optical signal
had the greatest SNR and provided a cardiac trigger pulse
that consistently coincided with the R-wave of the ECG.
The recorded waveforms are shown in Fig. 5a. The amplitude of the optical signal increases in end-diastole, reaching a maximum coincident with the R-wave, and then
decreases again just after the R-wave. This behavior suggests that atrial contraction, which occurs immediately
prior to ventricular contraction to supply 20 –30% additional filling of the ventricles, is responsible for decreasing
esophageal diameter and causing the optical signal peaks
in end-diastole. The corresponding axial MR image in Fig.
5b reveals the proximity of the OFs to the LA, supporting
the idea that atrial activity is detectable at this depth.
Because the optical signal began to drop in amplitude
just after the R-wave of the ECG, the falling edge was used
to generate a cardiac gating pulse that would closely track
the R-wave. As illustrated by the dotted line in Fig. 5a, the
optically-derived gating pulses generated in this manner
do, in fact, consistently coincide with the R-wave. The
standard deviation between the two signals was measured
to be less than 1% of the cardiac cycle, and over 95% of the
falling edges of the optical signal were detected. While the
coincidence of the optically-derived gating pulse and the
R-wave makes the gating pulse easily interpretable, it is
not necessary for purposes of cardiac gating. Cardiac gating can be achieved no matter where the trigger pulse
occurs in the cardiac cycle—as long as the position of the
pulse remains consistent throughout the scan.
An interesting feature of the system is its ability to detect
breathing motion in addition to cardiac motion. As seen in
Fig. 5a, the optical signal amplitude increases during inspiration, indicating that increased lung volume further
compresses the esophagus and increases the amount of
detected light. If desired, the slower-frequency breathing
signal could be filtered out to measure the animal’s respiratory rate, further extending the capability of the system.
Screen saves of the physiologic monitor taken during
cardiac-gated MR microscopy of a rat demonstrate the
value of the fiber-optic stethoscope over the ECG (Fig. 6).
In Fig. 6a, imaging gradients are off, the ECG appears
normal, and the optically-derived gating pulses are coincident with the R-wave. In Fig. 6b, however, imaging gradients are turned on and the ECG is visibly distorted by
induced voltage artifacts, whereas the optically-derived
gating pulse is unaffected and continues to provide a robust signal for cardiac gating.
The fiber-optic stethoscope was used in a series of cardiac MR microscopy experiments to test its ability to provide a reliable cardiac gating pulse for time-resolved functional imaging. Each optically-derived gating pulse triggered the cine acquisition of data at six 30-ms intervals
across the cardiac cycle, for a total imaging time of 10 min.
Figure 7 shows a temporal series of the same 2-mm coronal
slice of the rat thorax at six phases of the cardiac cycle.
Without a consistent cardiac gating pulse, the data would
Fiber-Optic Stethoscope for MR Microscopy
319
FIG. 6. Screen saves of the physiologic monitor taken during ventilatory- and cardiac-gated MR microscopy of a rat. The physiologic
signals, graphed from top to bottom, are the ECG, optically-derived gating pulse, and ventilatory pressure waveform (unlabeled). a: Imaging
gradients are off and optically-derived gating pulses are coincident with the R-wave of the ECG. b: Imaging gradients are on and the ECG
is visibly distorted by induced voltage artifacts, whereas the optically-derived gating pulses continue to provide a reliable trigger for cardiac
gating.
be misregistered in time. But, as seen in Fig. 7, the distinct
cardiac phases are discernible as the heart cycles through
diastole (a), systole (b and c), and then returns to diastole
(d and f), confirming the ability of the fiber-optic stethoscope to provide an accurate temporal reference frame for
cardiac imaging.
Once the operation of the system was validated on rats,
it was then tested on a 40-g mouse to verify its feasibility
on an animal with one-fifth the mass and up to twice the
heart rate. The results were very similar to those in the rat,
with apparent detection of atrial contraction and coincidence of the optically-derived gating pulse to the R-wave
of the ECG. This finding demonstrates that cardiac-induced esophageal compression is not species-dependent,
and that the fiber-optic stethoscope can be used with rats,
mice, and (potentially) other small animals.
CONCLUSIONS
It is important to note that the ECG and the fiber-optic
stethoscope provide fundamentally different kinds of information about cardiac activity. Whereas the ECG assesses electrical aspects of impulse generation and conduction, the fiber-optic stethoscope provides a measure of
mechanical cardiac contractility; both types of information
are useful in monitoring cardiac activity. Therefore, the
fiber-optic stethoscope is not meant to replace the ECG.
Though virtually useless during some imaging sequences,
the ECG does provide important cardiac information when
imaging gradients are off. However, when imaging gradients are on, the fiber-optic stethoscope’s ability to provide
a robust signal makes it a valuable tool for cardiac monitoring and gating purposes. For this reason, the system has
320
Brau et al.
FIG. 7. Temporal series of the same 2-mm coronal slice of the rat thorax acquired at six equally spaced intervals in the cardiac cycle using
the RA-CINE sequence, triggered by optically-derived gating pulses over a period of 10 min. Dynamic changes in ventricular volume and
myocardial wall thickness are visible as the heart cycles through (a) diastole and (b-d) systole, and then (e-f) returns to diastole. Images are
separated in time by 30 ms and have an in-plane spatial resolution of ⬍200 ␮m.
been incorporated into routine microscopy studies beyond
those mentioned here. Furthermore, the system’s ability to
detect esophageal compression from multiple sources, including the aorta, atria, ventricles, and lungs, makes it a
versatile monitoring device with other possible applications not explored in this work.
ACKNOWLEDGMENTS
The authors thank Dr. Brett Hooper for useful technical
discussions.
REFERENCES
1. Franco F, Dubois SK, Peshock RM, Shohet RV. Magnetic resonance
imaging accurately estimates LV mass in a transgenic mouse model of
cardiac hypertrophy. Am J Physiol 1998;274:H679 –H683.
2. Ruff J, Wiesmann F, Hiller K-H, Voll S, von Kienlin M, Bauer WR,
Rommel E, Neubauer S, Haase A. Magnetic resonance microimaging for
noninvasive quantification of myocardial function and mass in the
mouse. Magn Reson Med 1998;40:43– 48.
3. Siri FM, Jelicks LA, Leinwand LA, Gardin JM. Gated magnetic resonance imaging of normal and hypertrophied murine hearts. Am J
Physiol 1997;272:H2394 –H2402.
4. Slawson SE, Roman BB, Williams DS, Koretsky AP. Cardiac MRI of the
normal and hypertrophied mouse heart. Magn Reson Med 1998;39:
980 –987.
5. Brau ACS, Cofer GP, Hedlund LW, Tajik JK, Johnson GA. Functional
cardiac microscopy using projection encoding. In: Proceedings of the
8th Annual Meeting of ISMRM, Denver, 2000. p 659.
6. Chia JM, Fischer SE, Wickline SA, Lorenz CH. Performance of QRS
detection for cardiac magnetic resonance imaging with a novel vectorcardiographic triggering method. J Magn Reson Imaging 2000;12:678 –
688.
7. Laudon MK, Webster JG, Frayne R, Grist TM. Minimizing interference
from magnetic resonance imagers during electrocardiography. IEEE
Trans Biomed Eng 1998;45:160 –164.
8. Morris Sr GR, Morris Jr GR, Valentine JW. Electrocardiograph sensor
and sensor control system for use with magnetic resonance imaging
machines. 2000. U.S. patent 6,052,614.
9. Oberg AP, Lindberg L-G. Method and apparatus for analyzing heart and
respiratory frequencies photoplethysmographically. 1995. U.S. patent
5,396,893.
10. Legendre JP, Misner R, Forester GV, Geoffrion Y. A simple fiber optic
monitor of cardiac and respiratory activity for biomedical magnetic
resonance applications. Magn Reson Med 1986;3:953–957.
11. Savary M, Miller G. The esophagus. Handbook and atlas of endoscopy.
Bern, Switzerland: Verlag Gassmann; 1978.
12. Cheng YP, Wu HD, Jan GJ, Wang CY. Removal of cardiac beat artifact in
esophageal pressure measurement via a modified adaptive noise cancellation scheme. Ann Biomed Eng 2001;29:236 –243.
Fiber-Optic Stethoscope for MR Microscopy
13. Lemieux SK, Glover GH. An infrared device for monitoring the respiration of small rodents during magnetic resonance imaging. J Magn
Reson Imaging 1996;6:561–564.
14. Wilson S, Brereton I, Hockings P, Roffmann W, Doddrell D. Respiratory
triggered imaging with an optical displacement sensor. Magn Reson
Imaging 1993;11:1027–1032.
15. Hedlund LW, Cofer GP, Owen SJ, Johnson GA. MR-compatible ventilator for small animals: computer-controlled ventilator for proton and
noble gas imaging. Magn Reson Imaging 2000;18:753–759.
16. Hale GM, Querry MR. Optical constants of water in the 200 nm to
200-micron wavelength region. Appl Opt 1973;12:555–563.
321
17. Qiu H, Cofer GP, Hedlund LW, Johnson GA. Automated feedback
control of body temperature for small animal studies with MR microscopy. IEEE Trans Biomed Eng 1997;44:1107–1113.
18. Viallon M, Cofer GP, Suddarth SA, Möller H, Chen XJ, Chawla MS,
Hedlund LW, Crémillieux Y, Johnson GA. Functional MR microscopy of
the lung with hyperpolarized 3He. Magn Reson Med 1999;41:787–792.
19. Smith JJ, Kampine JP. Circulatory physiology—the essentials.
Baltimore: Williams & Wilkins; 1990.
20. Johnson GA, Cofer GP, Gewalt SL, Hedlund LW. Morphologic phenotyping with magnetic resonance microscopy: the visible mouse. Radiology, in press.