Calibration with fixed
volume bags
The calibration analysis allows ANSLAB to translate a subject’s
respiration band
movements into calibrated respiration variables such as tidal volume or
inspiratory flow rate. During an
experiment, all subjects need to perform a calibration procedure
conisting of repeated (e.g., 8) inhalations and exhalations of a
standardized volume
(e.g., 750 ml). ANSLAB matches the
change
in respiration bands with the known volume and uses this information to
calculate lung volume throughout the experiment. Your job during
the calibration analysis is to locate these breath cycles and select
those which
most consistently approximate the bag volume. After telling
ANSLAB
which breaths to use, you may proceed to editing respiration data.
Select "Find breath maxima and minima automatically"....
----- Choose method for finding extremes: -----
1) Find breath maxima and minima automatically
2) Manual setting of maxima and minima
Select "Find maxima with default thresholds" ....
-----
Choose threshold method: -----
1)
Find maxima with default thresholds
2)
Change default thresholds
3) Quit
A window will open, displaying thoracic (yellow) and abdominal
(magenta) breathing cycles. Your goal now is to exclude all
non-calibration cycles, and all “bad” calibration cycles. In the
optimal case, you keep 6-8 breathing cycles. But it is better to leave
out breaths that are not
done correctly than to have the full set. Therefore: When
in doubt, exclude cycles. Minimally, you should have 3 good breaths
altogether. “Good” are several regular breaths that are consistent and
will probably add up to the same amplitude.
This window shows one good calibration set:
You can zoom into the signal by drawing a box with the mouse, and
unzoom with a right mouseclick.
Now, you can step-by-step exclude all the “bad” cycles by typing in the
corresponding numbers:
Take out cycle no. [e.g., 3, or: 1-12; <enter>=end] ==> 1
Take out cycle no. [e.g., 3, or: 1-12; <enter>=end] ==> 5
When you are done, you just hit
<enter>:
Then [select "Quit"]
Next, you are asked to decide how many channels were “OK”....
-----
Specify qualitiy of thorax and abdomen: -----
1)
Both channels ok
2)
Only thorax ok
3)
Only abdomen ok
Depending on
your decision select "Both channels o.k.” “only thorax o.k.” or “only
abdomen o.k.” Most of the time you will be able to say “both channels
o.k.,” but sometimes you might only be able to find 3 regular breaths
for thorax, and have to say “only thorax o.k.” This might be the case
if the other signal is bad due to technical difficulties, e.g. frequent
resets or a zero signal.
Then, two new windows will open that will show the voltage change of
the breathing bands and their “beta weights” predicting lung volume
(don’t worry too much about their meaning). This second window,
“beta-weights”, can additionally help you exclude outliers, since an
outlier will be seen as a peak in this display, which is outside the
broken line. Normally, though, if you have thoroughly excluded breaths
in the first step, you can just move on by hitting <enter>
Note: In this figure, all breaths fit within an acceptable deviation
from 750 ml (or cubic centimeters, ccm). Keep all.
In some cases, though, there might still be an outlier. Keep in mind
that the scale on your left can show a rather large range (e.g. 0-2500
ccm), in which case a peak will be relevant, or a rather small range
(e.g. 400-1100 ccm), in which case a peak will be less relevant. When
you exclude an outlier by typing in the corresponding number, the
beta weight (displayed above the window) will change. Only if they
change a lot, the outlier was relevant, but a rule of thumb is: “when
in doubt, exclude.”
When you are done with excluding outliers, hit <enter>
Then select 'Last betas'...
Special Cases
What Do I Do When Thorax and Abdomen
Move in Opposite Directions? This is called ‘paradoxical
breathing’ and it makes calibration difficult. If you can find
some cycles or breaths within cycles that are synchronous, use them; if
not, select ‘only thorax o.k.’ option.
What If There is No Usable
Respiration Calibration File? You can still salvage respiration
data. The easiest ‘fix’ is to manually enter estimated
calibration coefficients into the [study]c.m file using Excel.
Estimate these coefficients by looking at the coefficients of subjects
of same age and sex. As a constraint in the regression model
estimating the beta weights, thoracic coefficients are set to twice the
value of abdominal coefficients (this is supported by a publication
that showed that this provides the most consistently accurate
calibrations). Now check the baseline respiration
data in ANSLAB. Have most breaths been correctly detected?
You can modify the coefficients you entered if anything about the
breaths or minute ventilation seems unreasonable. For example,
decreasing the coefficients by a factor of 3 will decrease baseline
minute ventilation by about a factor of 3. Typical baseline minute
volumes are in the order of 4-8 liters per minute. Using estimated
coefficients will certainly affect the
accuracy of respiratory volume variables (tidal volume, minute
ventilation, inspiratory flow rate), but all timing related variables
will be correct. The volume inaccuracy will be
systematic within the subject. Since the error is multiplicative rather
than additive, change scores for within subject analyses should be
percent change scores.