Consider what a typical metabolic analysis software
package offers regarding the determination of the gas exchange threshold.
Almost invariably, you get only one method, typically the
VCO2 by VO2 plot.
You push a button and the program "magically" decides where
the supposed breakpoint in the data is, shows you an estimate of the
time and percentage of VO2max at which this supposed breakpoint occured
and, if you are lucky, plots regression lines through the before-the-breakpoint
and after-the-breakpoint data. Then, you are allowed to manually shift
the supposed breakpoint higher or lower to fit what "looks better"
Those who have worked extensively with expired gases, particularly
when these gases are collected on a breath-by-breath basis, know very well
that this manual adjustment of the "automatically" determined
breakpoint is very often necessary, as the erratic nature of ventilatory
data make it very easy for statistical methods to lead to physiologically
irrelevant and erroneous solutions. Perhaps most importantly, the typical
metabolic analysis software packages do not reveal precisely what
algorithm they use in making the "automatic" determination of the
gas exchange threshold and do not disclose the full computational details of the
iterative process that leads to this determination. An examination of the research
literature that pertains to gas exchange analyses shows that there is
a distinct lack of specificity in describing the methodology used to
determine the gas exchange threshold. It seems that, at best, authors
resort to using general and imprecise terms, such as "the V-Slope
method", which have nearly no substantive meaning unless exact
procedural details are described, and at worst, they admit that they
relied exclusively on subjective and, thus, essentially impossible to
precisely document and replicate criteria. Naturally, this situation
can only lead to tremendous potential for inconsistency and confusion in
the literature, as well as to justified scepticism toward the reliability
and validity of any findings based on such poorly documented methodology.
Against this bleak backdrop, the features of WinBreak
and the progress that they represent are astounding. Some of the features
of WinBreak 3.7 include:
First, WinBreak is the only software
program that is specifically designed to assist in the determination
of the gas exchange threshold. The alternatives are (a) the
continued reliance upon the very limited (due to the reasons explained
above) methods for the determination of the gas exchange
threshold incorporated into most of the commercial metabolic
analysis software packages, (b) trying to obtain copies
of computer code from researchers who, according to their
publications, have developed computerized methods for determining
the gas exchange threshold (unfortunately, for various reasons,
most researchers are not willing to share their code), and (c) writing one's
own computer program from scratch, an extremely time-consuming
and, therefore, inefficient solution.
Second, WinBreak offers an automated and
integrated analysis and data-presentation solution. All the
analytical and graphical methods are incorporated into the
program, so there is no need to have access to auxiliary
mathematical software packages. All the common data-processing
and graphing options related to the determination of the gas
exchange threshold are available, usually with one click of
the computer mouse.
Third, with WinBreak, there are no secret,
arbitrary, or untested methods. All the methods and algorithms
used in the program have been validated and published in the
exercise physiology literature. This not only allows researchers
and practitioners to have confidence in the methods on which
they are basing their decisions, but also allows authors of
scientific papers to document their methodologies in a simple,
precise, and defensible fashion.
Fourth, WinBreak does not rely on a single
method of determination. Idiosyncracies in data sets or excessive
noise in the data can derail any statistical method of determination,
leading to an erroneous solution. For this reason, WinBreak combines
multiple methods of determination, thus giving users more options
for viewing and analyzing their data. Specifically, WinBreak incorporates
the 3 most popular and powerful graphical methods (the V-slope method,
the method of ventilatory equivalents, and the Excess CO2
method) and can place markers on each graph, indicating the location
of the gas exchange threshold determined via the remaining two methods.
Furthermore, in the V-slope module, in particular, users can select
among five published algorithms (Jones & Molitoris, 1984; Orr et al.,
1982; Beaver et al., 1986; Cheng et al., 1992; Sue et al., 1988).
Of course, WinBreak also allows users to manually override the
results of the "automatic" methods of determination
if these results are deemed unsatisfactory. As the breakpoint
is shifted higher or lower, the graphs are re-plotted and the
regression slopes are re-calculated to enable users to observe
the impact of their changes.
Fifth, to eliminate the uncertainty about
the decision-making process that the program follows (i.e., how it
arrives at a certain solution and why it considers this solution
preferrable over all others), WinBreak provides the complete
computational details for each method requested by the user.
This output can be voluminous, particularly for breath-by-breath
data, but it is important that users have access to it and be able
to choose whether they wish to save it, print it, or delete it.
Sixth, to help users obtain a clear representation
of their data and facilitate the process of determining the gas exchange
threshold by minimizing some of the noise and clutter of metabolic
data, WinBreak offers an extensive array of data manipulation
options, including averaging, interpolation, outlier removal, and
five smoothing options (running-window average, low-pass FFT filter,
Savitzky-Golay least squares filter, cubic spline filter, polynomial
from second to tenth order). With one click of the mouse, the data can
revert to the original set.
Seventh, all graphs produced by WinBreak
are presentation-quality and fully customizable (labels,
fonts, axis scaling, symbol styles, line styles, sizes, and colors).
Users can also turn some elements on or off (e.g., data series,
grid, legend, etc.). Furthermore, WinBreak "memorizes"
all the stylistic choices you make (separately for each graph,
amounting to dozens of graph elements), so that you will not
have to spend time re-customizing each graph.
Eighth, WinBreak can save all graphs
in Windows® Metafile (WMF) format. This format entails
no loss of quality (as with bitmap formats, for example) and is
compatible with all major word processors and presentation packages.
In conjunction with WinBreak's graph customization options, this means
that authors can prepare and incorporate graphs in their manuscripts
in one easy step.
Ninth, WinBreak can produce not only printouts
of its results in text format, but also graphical printouts accompanied
by a summary of computational results. This type of printout is
very convenient for hard-copy record-keeping, as it can provide an
one-page overview of an analysis for future reference.
Tenth, WinBreak offers a flexible interface
to import data from any metabolic analysis software package that can
generate ASCII files and, given the fact that WinBreak can perform
several data analysis and manipulation tasks that are difficult to
replicate with common spreadsheet software packages, it is also important
that WinBreak can save its data sheet in a variety of ASCII formats
(e.g., comma-delimited, tab-delimited) and in the popular Microsoft®
Excel® file format.
Finally, WinBreak, despite its technical sophistication and
rich feature set, is very intuitive and easy to use, having a very short
learning curve, is great as an educational tool for exercise physiology
classes and laboratories, and its low price is a very welcome surprise!