The Relationship Between Airflow and Lung
Sound Amplitude in Normal Subjects*
M.D.,
S. S. Kraman,
FCC.?
have examined
the relationship
between
and lung sound amplitude;
the available
data are
contradictory.
I measured
airflow
at the mouth and campared
the peak flow (Vmax)
to mean and peak lung sound
amplitude
(mean
AMP and peak AMP)
at four sites on the
chestwall(right
andleft
anterior
apices and posterior
bases)
in four
healthy
young
adults.
At each
site, the sounds
pmduced
by 20 breaths at Vmax ranging
between
L5 and
4 L(s (Vvar)were
measured
by an automated
technique.
Ten
breaths during nearly constant
Vmax breathing
(Vcon) also
were measured
at each site. The lung sound amplitudes
at
the four sites in each subjectwere
grouped
and compared
to
Vmax by linear regression
analysis.
The same sounds were
also submitted
to an automated
V-correction
procedure
to
evaluate
its adequacy
in automatically
adjusting
for the efFew
investigators
airflow
ung sounds
Since
define
were
that
with
the
which
basic
questions
first described
time,
many
by La#{235}nnecin 1819.1
studies
have
helped
character
of these
sounds
and
they are associated.
However,
regarding
lung
sounds
to
the diseases
some very
have
yet
to be
satisfactorily
feet
of variations
in Vmax
on lung sound amplitude.
The
showed
that lung sound amplitude
(mean
or peak) was
linearly
related
to V in all subjects
(r for mean AMP
vs
Vmax:O.77,
0.85,
0.69,
0.89; r for peak AMP vs Vmax:O.80,
0.83, 0.79, 0.88), p<l x 10 in all cases. The average
mean
AMP vs Vmax regression
line slope was 0.42,
and the
average peak AMP vs Vmax regression
line slope was 0.45.
V-correction
decreased
the coefficient
of variation
of the
Vvar sounds
by 61 percent
and flattened
the average
regression
line slopes
to 0.1.28.
For the Vcon series,
V-correction
diminished
the coefficient
of variation
from
12.2 to 10.0 percent.
The relationship
between
lung sound
amplitude
and airflow appears
tobe substantiallylinear
and
this
relationship
can
be used to adjust
effectively
for
data
variations
in airflow.
be adequately
techniques.
studied
if these
the techniques
validity
using
objective,
automated
results
are to be accepted
and
used by other investigators,
then their
be established.
Besides
the question
of
to automated
lung sound
measurement
must
applicability
addressed.
One of these
is the relation
between airflow
(V) and vesicular
lung sound
amplitude
(AMP).
It is self-evident
that a positive
rela-
procedures,
the relationship
sound amplitude
is germane
lung sound
generation
and
tionship
exists
sial subjects
of this
relationship
between
and contradictory.
linear.
Wooten
V and AMP,
is unclear
LeBlanc
et al stated
and
but the
the
character
evidence
et al2 maintained
that
that it is curvilinear,
Banaszak
et al that the logarithm
ofroot
mean
(RMS) lung sound amplitude
is linearly
related
(at
Assuming
that the relationship
the mouth)
is a simple
linear
methods
ofautomatically
measure
relative
wall
amplitudes
on the
but
ofthe
AMP
initial assumption;
that airflow at the mouth
are linearly
related.
The question
of this
is an
information
involved
important
about
in their
does
lung
production
not confirm
one.
the
It appears
sounds
and the
and transmission
From the Medical
Service,
Veterans
Administration
Center,
and Department
of Medicine,
University
Medical
Center, Lexington.
Supported
by NHLBI Grant No. 5R01 HL-26334.
Manuscript
received
November
21; revision
accepted
Reprint
requests:
D Kraman,
VA Medical
Center
ington,
Kentucky
40511
chest
of
and
relanew
physiology
can only
Medical
of Kentucky
February
(111-H),
study
was
performed
relationship
between
airflow
tude in normal
subjects
and
the correction
to examine
the
and lung sound
amplito test the assumptions
procedures
used
in previous
studies.
correctness
that
themselves.
present
underlying
to
The latest
reproduci-
bility,
tionship
alone
and
square
to V.
for V to attempt
ofairfiow
variations.
has provided
acceptable
these techniques78
this
sound
it is
between
AMP and V
one, I have developed
correcting
lung
independently
The
scant
between
airflow and lung
to the basic mechanism
of
transmission-controver-
7.
Lex-
METHODS
I studiedfourhealthy
38
years
(mean
malesubjects,
33
years).
Three
ailbetween
the ages of27 and
of the four were lifetime non-
smokers,
and one was an occasional
smoker (six pack-years).
All had
normal results of spirometric
studies.9
‘flwo identical
electret
microphone
elements,
flat in free-field
freq uency
response between
50 and 1,000 Hz, that were installed
on
top ofplastic
ECC electrodes
(cavity dimensions,
9 mm diameter,
2
mm depth) were used to record the lung sounds. The microphones
were fastened to the chest wall with double-sided
ECC tape rings.
One ofthese microphones was placed in the second right intercostal
space
and midclavicular
line and the other approximately
6 cm from
the spine and immediatelybelowtheloweredgeofthe
right scapula.
All subjects
were studied while standing
upright.
After
taldng one vital capacity
inspiration
to standardize
lung
volume
history,
the subject breathed repetitively,
20 times, through
the circuit shown in Figure
L The subjects
were instructed
to
breathe
at resting
throughout
the study. This circuit
simply allows
filled with 4 L ofroom
alt The one-way
accumulation
ofCO2 in the dead space.
spirometer
diminish
lung
volume
CHEST
Downloaded From: http://journal.publications.chestnet.org/pdfaccess.ashx?url=/data/journals/chest/20381/ on 06/17/2017
(functional
residual
capacity)
rebreathing
valve
While
was
into a
to
used
breathing,
I 86 I 2 I AUGUST, 1984
225
Tektronix
5103N
To Tape Recorder
V.
V.
Ohio
1. Equipment
FIGURE
the subject sat facing an oscilloscope,
volume (7) signals
in a flow-volume
asked
to achieve peak flows varying
spiring
only
deeply
enough
which
to measure
flow
the
cautioned
rates.
peak
flow
against
While
the
simultaneously
(i’) and
loop format. Each subject
between
1#{189}
to 4 Us while
to achieve
on each breath. He was further
inspirations
at constant
airflow
used
displayed
(Vmax)
was
in-
desired
maintained
or long
subject
performed
maneuver,
the audio signals
from the two microphones
together with the V and V signals from the spirometer
were recorded
on an FM tape recorder (HP model 3964A)
at 3#{190}in/s. The lung
sounds were high pass filtered (200 Hz cutoff, 12 db/octave
rollofi)
to
this breathing
attenuate
extraneous
muscle
and
cardiac
artifacts
that
have
been
shown
to be important at lower freq’#{176}
The recording of these
constituted
the variable flow(Vvar)part
ofthe study. After
completion
of this series, the spirometer
was flushed out with air,
and the subject was asked to take 20 more breaths at nearly identical V and V using the oscilloscope
screen display as feedback
to
ensure consistency.
These were the series ofconstant
breaths (Vcon).
The subjects were allowed to breathe at whichever
rate and depth
was comfortable
and, generally,
a Vmax of2 to 2#{189}
Ifs was chosen
by
the subjects.
After flushing out the spirometer
once again, the same
procedures
(Vvar and Vcon) were repeated
with the microphones
placed at analogous sites on the left hemithorax.
Thus,
80 Vvar and
80 Vcon breaths were recorded
in each subject.
20breaths
lung
sounds,
airflow,
volume.
similarly
processed,
butthis
time
usingaflow-correction
measurement
Table
is identical
to the
1-Coefficients
Amplitude
ofVariation
Measurements
20 Variable
Mean
CV
±
SD
was mea-
determining
the
(UC)
SD(C)
breaths
MeanCV
±
10 Steady
RUL
RLL
LUL
LLL
Mean CV ± SD (UC)
Mean CV ± SD (C)
=
4
----
UC
C
UC
C
24
27
19
19
14
18
14
12
26.5±4.44
20
25
30
30
16
22
20
22
=
16.2±3.54
=
8
12
12
8
=
=
of variation
sounds
right upper
lobe, LLL
3
----
UC
---C
UC
C
24
30
26
29
16
18
11
11
16
18
8
8
11
8
7
8
breaths
RUL
RLL
LUL
LLL
producedlung
sound
Sound
ofLung
Location*
Each
Subjects
*Ccefficients
The average intensity
of each inspiratory
lung
sured bydigitizingthe
audio signals at 5,000 Hzand
(CV)
at
2
,-‘---
Measurements
technique.7
first except that the amplitude
of each 25-msec segment
is adjusted by dividing it by the simultaneously
determined
airflow.
These
adjusted
values are then averaged. The purpose of this technique
is to correct for variations
in
airflow,
based
on the assumption
that the relation between
airflow
and lung sound
amplitude
is linea
or nearly so. The same
procedures
were used to measure the amplitude
ofthe lung sounds
that occurred during the Vcon series, although instead of 20 breaths,
the 10 breaths that had the most similar flow-volume
loops
were
This
1
After completion
ofthese
tests, ‘, V, and audio calibration
signals
were recorded.
The audio calibration was carried out by placing each
microphone
on top ofan acoustic calibrator
(Bruel and Kjaer Model
4230),
providing
1 Pa of sound pressure
at 1 kHz. The calibration
signal was used to ensure
equal gain of the two amplifier
systems as
well as to provide a reference source for estimation
ofsound pressure
recorded
on the subjects chests. The calibrator was modified by
affixingathin
foam rubber
ringon
the top toestablish
an acoustic
seal
when the microphone
was placed on it. This calibrator
is designed
to
provide reliable acoustic pressure
with varying degrees
of microphone placement.
The adequacy ofthe modification
was checked by
comparing
the output of a Bruel and Kjaer #{189}-in
condenser
microphone when placed within the calibrator cavity (as recommended
by
the manufacturer)
or on top of it using a plastic
chest
piece
(the
modified placement).
The difference
in output voltage was no more
than 5 percent between
the two placements.
226
and
mean absolute
amplitude
of the sound pressure
of each 25 msec
segment
while the V signal from the spirometer
exceeded
1.4 Us.
(Tb e lower limit of 1.4 Us was chosen
because,
at lower airfiows,
vesicular lung sounds are often near the noise level.)The
mean value
of these amplitudes
were then calculated
and taken as the mean
amplitude
(mean AMP) of the lung sound produced
by that particular
inspiration.
The signals
from the upper
and lower lobe
microphones
on each hemithorax
were measured
simultaneously
in
this manner.
The tape was run a second time and the sounds were
Calibration
Amplitude
840
=
Airflow
lobe,
14
11
14
9
12
17
10
21
12.2±4.09
31
13
33
16
26
17
30
19
61% decrease
p<1.5x107
8
7
11
14
12
9
11
15
12
12
7
6
18% decrease
10.06±2.79
p<O.05
in the amplitude
of the ‘(‘var and i’con
without(UC)
and with(C)V-correction.
RUL
RLL = right lower lobe, LUL = left upper
left lower lobe.
and Lung Sound
Amplitude
Downloaded From: http://journal.publications.chestnet.org/pdfaccess.ashx?url=/data/journals/chest/20381/ on 06/17/2017
in Normal
Subjects
(S. S. Kr&nan)
used.
This
adversely
was
done
to
determine
whether
v-correction
To determine
the effect that volume
dent variable on lung sound amplitude,
was carried
out for variable V. fixed V.
The mean absolute amplitude
of the
were then determined
by comparison
would
affect
the measurement
consistency
when airfiows were
uniform.
To determine
the peak amplitude
ofthe Vvar sounds, the
tape
was played
back, and the lung sounds were displayed
on a
storage
oscilloscope
screen.
The peak-to-peak
amplitudes
were
manually
measured
using
calipers as previously
described.
U These
were called the peak AMPofeach
breath. I do not mean to imply any
special relationship
between
the mean and peak amplitudes,
but
only to evaluate
two different
methods
of measuring
amplitude.
All amplitudes
(mean and peak) ofthe 20 variable breaths at each
site in each subject were normalized
by adjusting
them by a factor
calculated
to bring the mean (at any one location)
to unity. The
variation
from
the mean of the lung sound amplitudes
was then
related to this value regardless
ofwhat the absolute amplitude
was
and this permitted
the sounds recorded
at the four locations
on the
chest of each subject
to be plotted
on the same axes. For each
subject,
these
normalized
amplitudes
were correlated
against the
peak airflow and volume (manually
measured
on the oscilloscope
screen during playback)that
occurred
during each breath. The mean
AMP values were also correlated
by linear regression
analysis against
each other; upper lobe against lower lobe sounds. The lung sounds
analyzed
by the V-correction
technique
were similarly
plotted and
analyzed
to determine
the effect ofthis process.
may have had as an indepenmultiple regression
analysis
20 Vvar breaths
with
signals.
RESULTS
The
‘Tvar
series
airfiows
above
submit
the
involved
as wide
1.4 L/s as could
V-correction
conditions
respiration
(le,
determine
whether
a range
be achieved
procedure
to cover
as possible).
to
V-correction
the
coefficients
uncorrected
and
corrected
would
ofconsistency
tudes
at all sites,
breathing.
Figure
2 shows
“worst
to
case”
as many different
patterns
The Vcon series
were used
desired
deterioration
conditions.
Table
1 shows
(UC)
of peak
in order
case”
of variation
diagrams
of the
sound
during
Vvar
ofmean
of
to
an Un-
“best
(C) lung
in all subjects,
scatter
cause
under
ampliand
‘con
AMP,
peak
C
I
1.5
at each site
the calibration
11.0
/<“
z
a
0
o.
(/)
I
y-
O.090+O.390x
r-
0.77
0L.1
2
y--O . 050+0
. 450x
r0.80
p<IE-7
p<IE-7
3
0
4
yr-
O.580+O.180x
0.56
p<IE-7
-‘Id’
2
(L/S)
3
4
2
0
‘(L/S)
3
C
B
1.5 i
4
(L/S)
(‘-1
U
9,
0
I4
1.0
a
0.5
2
y0.69
r-
p<1E-
yr-
I
0.5
I
O.’9
Lq,.
.1,
2
3
O”2
4
3
(L/S)
O.870+O.050x
O.4
‘oS
3
-
r-
p<E-7
2
(L/S)
V(L/S)
C
1.5
1.0
I4
C,)
2
U
a
4)
0.5
0
y-o
0
290+0
0.89
r“2
. 460x
p<1E-7
-
,*!<“
r-
r-
0’
3
y--O.360+O.510x
2 0.88
3pclE-7
4
1
O.550+O.170x
0.57
p’ZIE-7
“d’-
4
V(L/S)
V(L/S)
V(L/S)
C
1.5
U
4)
2.
peak,
mean,
amplitudes
airflow
that
Scatter
plots
and V-corrected
plotted
occurred
against
during
of
normalized
lung
sound
peak
each breath.
the
0
U,
a
i
,j”
I
Ficuar
4
4
ip
o.s
I
I
y--O
y--O.310+O.510x
r-
0.85
oL//
r-
p<IE-7
4
V(L/S)
0’”
.
280+0
0.63
2
.
y.
3
0.36
4
9(1/5)
O.710+O.lIOx
r’-
490x
p<IE-7
0
“
p<IE-3
a
2
V(L/S)
CHEST
Downloaded From: http://journal.publications.chestnet.org/pdfaccess.ashx?url=/data/journals/chest/20381/ on 06/17/2017
I 86 I 2 I AUGUST, 1984
227
Subj
1.5
1
Subj 2
1.5
I.
1.0
I
1.0
I
0.5
0.5
y= 0.119+0.881x
r=
0.83
y= O.006+O.994x
0.91
r=
0.5
1.0
05
1.5
Subj 3
1.5
1:5
Subj 4
1.5
1.0
I
1.0
UPPER LOBE AlP
LJ1PER LOBE AIF
4
1.0
I
0.5
0.5
y=-0.046+1
.046x
y=-0
r= 0.93
.
052+1
.
052x
Ficuax
3. Normalized
upper
lobe mean lung
amplitude
plotted against simultaneously
measured,
normalized
lower lobe mean
lung
sound amplitude
in each ofthe four subjects.
0.90
r=
sound
0.5
1.5
AMP,
and corrected
In these
plots,
AMP
analysis
ofthem
difference
(maximum
Figure
F statistic
3 shows
against
confirmed
between
the
UC
UC
upper
all regions
should
that
them
= 1.348
the
at each
amplitude
of the
effect
of lung
was
minor
sounds
in these
two
compared
to that
of V.
In each case, the mean
lines are similar,
and
in
there
was
any
lobe
mean
no
subject).
ofregional
ventilate
nearly
et al2 studied
LeBlanc
subject
AMP
subjects,
this
between
lung
DISCUSSION
sound
at 1 and 156 df p<O.O5).
lower
lobe
mean
AMP
sured
simultaneously
in each
intensity
at each site is a function
FRC,
in the four subjects.
breaths
together.
regression
by F test
significant
plotted
vs “max
the 20 normalized
four sites were plotted
AMP and peak AMP
1.5
1.0
UPPER L08E AlP
UPPER LOBE AlP
amplitude,
rectified,
integrated
measured
(mea-
Lung
sound
airflow. At
a linear
plitude
a wide
airflow
of a
simultaneously
They demon-
between
well-established
lung
sound
am-
and V at the mouth within
They also demonstrated
concept
that
when
inspiring
that ‘ at the mouth
would
behave
as an independent
variable.
This appears
to have been
the case.
from
Table 2 shows
the mean absolute
lung sound
hide at each of the four sites in each subject.
presumably
due to delayed
opening
of the dependent
airways.
When
breathing
from FRC,
the lung sounds
Table
3 shows
analysis
assessing
In two subjects
effict
on lung
the
results
volume
( 1 and
sound
ately
ampli-
of the multiple
regression
as an independent
variable.
2)variations
amplitude.
at the
of V as an independent
variable
was signfficant
although
much
smaller
than
that of V (3:r
for
V=O.72,
for V=O.14;
4:r2 for V0.48,
r2 for
V
0.17). While inspired
volume
may have affected
the
studied,
voltage
=
Subject
RUL
RLL
LUL
LLL
Mean
Mean
228
lung sound
Lung
1
and
base
the
Sound
Subject
mouth.
Amplitude
This
was
2
0.189
0.108
0.166
V in each
subject
during
together.
breathing.
@2.67
3
to
Hz
at the
were
at all frequen-
#{149}Subject
Us
Us
0.22
Us
Abbreviations
Mrflow
to be the case
0. 10 @2.67
@2.73
0. 115 @2.73
Pa
Vvar
techniques
75 to 500
to peak air velocity
two or three
airfiows
said
Subject
0.028
and mean
ended
by Sited’
0.049 @2.75
Us
0.057 @2.75 Us
0. 15 @3.08 Us
0. 175 @3.08 Us
amphtudes
and
but the authors
concluded
that the log ofRMS
ofthe
lung sound varied linearly
with airflow
at
0.135 Pa @2.45
Us
0. 142 Pa @2.45 Us
0.32 Pa @2.22
Us
0. 16 Pa @2.22
Us
Pa
began
and related
the amplitudes
mouth.
Apparently
only
the effect
2-Mean
apex
volume,
lung
sounds
appear
immediapices
but are delayed
at the bases,
Banaszak
et al used frequency
analysis
analyze
lung sounds
at frequencies
from
in V had a negligible
In subjects
3 and 4,
Table
residual
at the
deflection
and the
and volume.
relationship
at the apex and base,
range oflung
volumes.
the now
in phase’s
as the peak
tracing,
inspired
strated
the relationship
measured
0.045
and Lung Sound
@2.54 Us
0. 10 @2.64
Us
0.105 @2.64 Us
Us
0.08
Pa
1.
Amplitude
Downloaded From: http://journal.publications.chestnet.org/pdfaccess.ashx?url=/data/journals/chest/20381/ on 06/17/2017
Us
0.071
Pa
as in Table
@2.54
4
in Normal
Subjects
(S. S. Krnan)
Table
3-Vvar
Mean
Amplitude
(AMP)
jVolume
as a Iisndion
Regression
Subject
Formula
1
y
2
y
3
y
4
y
*Regssion
=
r
mouth
0.69
0.49
0.408
0.136
<0.0001
0.72
0.381
<0.0025
coefficients
ofmean
-0.069
lung sound
the effect ofV changes
on AMP
inspired
the relationship
or peak,
and
substantially
volume
(breaths
greatest
linear
also
with
the
volume),
(Fig
varied
between
airflow
2), and
positively
greatest
Vmax
this was not
lung
at the
these nonlinear
the presented
with
Vmax
formulae
in no case
linear
models.
third
scatter
diagrams
exceeded
in each
those
part
line
would
Although
erable
flattening
nie4
of the
by a mean
had a slope
slope
in all subjects,
conditions:
the
subjects
striving
able interbreath
airflow.
During
the mean coefficient
of variation
improved
V-correction
comparable
the technique
ment.
variation
this
regional
to 10 percent
to previous
experience78
does not adversely
ventilation
sound
The
to
1);
the remaining
differences
in
variations
in tar-
of these
relationships
to loudness
in lung sounds
perceived
by the human
ear is
Generally,
a 1O-dB increase
in the power of a
frequency,
ofthe
Despite
(l’able
and showing
that
affect the measure-
random
pure
tone
is perceived
as
However,
loudness
perception
absolute
vari-
generation.
significance
changes
complex.
or
of variation
were
“worse
to achieve
presence
sounds,
a doubling
in
is also affected
of other
sounds,
the hearing
ofthe
loudness
ofthe
sound
prior
these
confounding
factors,
sound amplitude
investigators
to
ta’
loudness.
by sound
absolute
observer,
loudand the
to the change.”o’
subjective
lung
estimation
has been shown by several
be valuable
in clinical
pulmonary
CONCLUSION
The
relationship
between
lung
sound
amplitude
and
during
1.4 and
oflung
innovative
computer
regression
breathing
4 Us
sound
ACKNOWLEDGMENTS:
from
appears
was
FRC
at
to be substan-
technique
effectively
variations
to permit
the
amplitudes
The
analysis
among
without
author
different
requiring
thanks
L.O.
strict
Ong
for his
programming.
REFERENCES
De
1819
2 Leblanc
P, Macklin
tion of pulmonary
l’auscultation
mediate,
1st ed.
Paris:
Brosson
vi: Ross WED.
ventilation.
Breath sounds and distribuAm Rev Respir
Dis 1970;
102:10-16
3 Wooten
V1
Waring
WW,
Wegmann
MJ, Anderson
WE,
JD. Method for respiratory
sound analysis.
Med Instrum
12:254-7
4 Banaszak
EF, Kory RC, Snider CC. Phonopneumography.
Rev Respir Dis 1973; 107:449-50
5 O’Donnell
accompa-
the Vcon breathing,
was 12. 2 percent
and
It is difficult
to say whether
is due to small breath-to-breath
bulent
ness
(no regreswas consid-
mouth,
between
Multiple
and Chaude,
2
there
volume.
for volume.
1 La#{235}nnec RTH.
If
re-
ofzero
decrease
in the coefficient
As already
stated,
these
of 61 percent.
case”
have
this was not the case,
vs inspired
sites,
breaths,
and subjects
control
of airflow
and volume.
of
of Figure
NS
correlation
at the
comparison
sound amplitude
calculated
from
show the result oft-correction
on the Vvar breaths.
V-correction
had been
completely
effective,
the
sion).
(AMP)
the partial
NS
tially linear, and the V-correction
diminishes
the effect ofairfiow
although
tended
to have the
an important
factor
to the data and to the log ofthe
lung
vs airflow.
The correlation
coefficients
gression
=
airfiows
compared
with airflow
(Table
3). The possibility
of a
nonlinear
relationship
was further
assessed
by applying a third-degree
polynomial
curve fitting
algorithm
The
amplitudes
R (V)
airflow
was
P Value
0.57
studied.
study,
mean
Regression
0.42±0.755x
and correlation
In the present
sound
intensity,
V-Multiple
R (V) (Fixed
0.185+1.043x
0.076+0.715x
0.009+1.168x
out with fixed V to evaluate
carried
cies
equations
=
=
=
.
vs AMP)
(V
(V)#{176}
D,
Kraman
S. Vesicular
lung
sound
mapping
Conley
1978;
Am
by flow
J AppI Physiol
1982; 53(3):603-9
6 Kraman SS. Lung sounds: relative sites oforigin and comparative
amplitudes
in normal subjects.
Lung 1983; 161:57-64
7 Dosani
BA, Kraman
SS. Variability of lung sound
intensity
in
normal men studied
by contour
phonopneumography.
Chest
1983;
83:628-31
8 Kraman
SS, Austrheim
OA. Comparison
of lung sound and
transmitted
sound amplitude
in normal men. Am Rev Respir Rev
gated
phonopneumography.
1983;
128:451-4
9 Morris JA, Koski
A, Johnson
LC.
Spirometric
standards
for
healthy non smoking adults. Am Rev Respir Dis 1971; 107:57-67
10 Kraman
55. Does the vesicular lung sound come only from the
lungs? Am Rev Respir Dis 1983; 128:622-6
11 Kraman
SS. Determination
ofthe
site ofproduction
of respiratory sounds by subtraction
phonopneumography.
Am Rev Bespir
Dis 1980; 122:303-9
12 Ploysongsang
Y, Dosman
J, Macklem
PT. Demonstration
of
regional phase diflrences
in ventilation
bybreath
sounds.
JAppI
Physiol 1979; 46:361-8
13 Bendat JS, Piersol A.
Engineering
applications
of correlation
and spectral
analysis.
New York: John Wiley
and Sons,
1980:43-92
14 Cescheider
GA. Psychophysics:
method
and theory.
Hillsdale,
NJ: Lawrence
Erlbaum
Assoc,
1976:1-10
15 Berg RE and Stork DC. The physics ofsound.
Englewood
Cliffs,
NJ: Prentiss
Hall mc, 1982:137-53
16 Pardee NE, Martin
CL Morgan
EH. A test ofthe practical
value
ofestimating
breath
sound
intensity.
Chest 1976; 70:341-4
17 Bohadana
AB, Peslin R, Uflholtz H. Breath sounds in the clinical
assessment
ofairfiow
obstruction.
Thorax 1978; 33:345-351
18 Nairn JR. ilirner-Warwick
M. Breath
soundsin
emphysema.
BrJ
Dis Chest
1969; 63:29-37
CHEST
Downloaded From: http://journal.publications.chestnet.org/pdfaccess.ashx?url=/data/journals/chest/20381/ on 06/17/2017
I 86 I
2 I AUGUST.
1984
229