Return on Main Page

CUNSTCAMERA (CABINET OF CURIOSITIES) OF CLINICAL SCIENCE

Exhibit ¹2
Results of independent metrological testing of programs for calculation of the heart rate variability parameters

7.5. Comparison of dispersion of RR-intervals and general power of spectrum of HRV

In the theory of spectral transformation there is a postulate in accordance with which the sum of powers of all of accordions of spectral transformation is equivalent the undisplaced dispersion (or dispersions for a general aggregate) of the studied row of sequences. In accordance with this condition in the high-quality algorithm of spectral transformations general power of spectrum (TP) will be always a few the less undisplaced dispersion (after minus of powers of accordions which are disposed outside standard ranges).
At the calculation of correlation of general power of spectrum and undisplaced dispersion in tests with the set powers 2 mc2 in most programs on the known reasons (see is a next section) found out rough errors.
Therefore tests with extremely low clinically unmeaningful powers are excluded from this analysis, and other results (for 16 tests for every program) are generalized as the small box graphs with pointing of median and kvartiley of the calculated relations on each of sixteen tests.

On the resulted picture a median determines the relative generalized error of the program in times (1 - means absence of errors), and quartile characterizes the size of variation of its separate indexes. A zero height of quartile testifies to absence of variation.

Among three programs of the independent testing the program CardioLab+ shows the best results. Programs UM300 and CardioSpectum give the understated and overpriced results of calculations accordingly, that is confirmed the first group of tests of estimation of correlation of the set and calculated powers.


Return to contens Next part
Hosted by uCoz