otometrics Madsen AccuScreen PRO, Руководство пользователя

Страница: 16 / 98

16 GN Otometrics A/S
Test Methods AccuScreen PRO User Manual
the criterion for this outcome is met in any of the quality buffers. A
PASS is indicated by the presence of 8 data points at which a time-
locked signal has been detected.
The advantage of this system over the usual averaging techniques is
that the tests are more robust with regard to noise and the test dura-
tion can be reduced. Also, the statistical criterion reduces the proba-
bility for a false negative result down to less than 0.1%.
Artifact is defined in the AccuScreen PRO as a sweep which, based
on its relatively large amplitude, is likely to be influenced mostly by
noise. Thus, including it in the analysis would lower the signal-to-
noise ratio and prolong test time. The automatic identification and
exclusion of artifact results in a dramatic shortening of the test dura-
tion, especially in the case of restless infants. An artifact rate, which
exceeds 20%, indicates that conditions are too noisy to carry out re-
liable tests.
The Stimulus Stability indicates the proportion of recorded sweeps
in which the tested stimulus level deviates significantly from the cal-
ibrated level. Low stimulus stability indicates that the probe has
moved during testing.
2.1.5 How the AccuScreen PRO D-module determines a PASS
Applies to AccuScreen PRO D, DA, TD, TDA.
Distortion Products are the result of cochlear activity, which origi-
nates from processing a two-tone signal. The two tones are called
"primaries", and their frequencies are designated as f
1
and f
2
, and the
levels are termed L
1
and L
2
.
The challenge of DPOAE evaluation is to discriminate environment
sounds - the "noise" - from the cochlear sound response, the Distor-
tion Product (DP). This cannot be achieved with absolute certainty.
However, by applying an appropriate statistical algorithm, the
chances of wrongly classifying noise as a cochlear signal can be re-
duced to a known probability - in the AccuScreen PRO the probabil-
ity of erroneously identifying noise as a DP is 0.3% for each
frequency applied.
Conventional criteria used commonly in DP screening devices are
based on algorithms which attempt to estimate the "noise floor" by
averaging the amplitude of the response at frequencies adjacent to
the DP frequency. This value is then compared to the amplitude of
the response at the DP frequency, and the difference is taken to be
the Signal-to-Noise Ration (SNR). Such methods show a false
PASS
rate that is considerably higher than the AccuScreen method. In fact,
the false
PASS rate can be up to 10% for the commonly used criteria
of 6 dB SNR.