Reported results are from -100 ms to +120 ms, with the center of the flash at 0 ms. The extended
pre-stimulus baseline is used so that baseline stability can be better assessed.
Quantitative analysis is performed as follows. The a-wave and b-wave cursors are placed on the
reported waveform at their respective peaks. PhNR cursors are placed at 72 ms and at the
minimum point between 30 ms and 100 ms. Two PhNR ratios are computed. The P-ratio was
introduced in Preiser et al. (2013) and was found therein to be reduced in subjects with glaucoma.
The W-ratio is the inverse of “PTR” as defined in Mortlock et al. (2010) and was found therein to
have the lowest level of inter-individual, inter-session, and inter-ocular variability of the 5 ERG
measurement techniques tested. Neither publication tested both ratios. The ratios are defined
as follows:
P-ratio = p72 / b
W-ratio = (b – pmin) / (b – a)
where a, b, p72, and pmin are the voltages relative to baseline defined as a: a-wave peak, b: b-
wave peak, p72: voltage at 72 ms, pmin: minimum voltage between 30 ms and 100 ms. Note the
b-wave voltage typically reported (including in the RETeval device) is equal to (b-a).
To generate the displayed waveform, novel and proprietary processing methods are used that
are based on the study of 92,000 sweeps from 219 subjects across 2 sites, who were randomly
split into a training set (80%) and a validation set (20%). Each subject was measured at least
twice, and the processing methods were tuned to minimize the test-retest variability of the PhNR.
Results from training and validation sets were the same, indicating the new (firmware 2.12.0)
processing methods reflect an improved understanding of both the signal and noise profiles of
this test. Compared to earlier firmware versions, the PhNR minimum, PhNR @ 72 ms, P-ratio,
and W-ratio variabilities have all been reduced. For the validation set, the mean test-retest
variability improved by 4× from 61% CV to 14% CV (standard deviation divided by the mean) for
the PhNR minimum voltage, improved by 2× from 1.3 µV to 0.64 µV standard deviation for the
PhNR @ 72 ms, improved by 3× from 0.08 to 0.03 standard deviation for the P-ratio and
improved by 8× from 0.15 to 0.02 standard deviation for the W-ratio. These changes affected
the expected value for the a-wave amplitude and W-ratio, so care has to be taken when
comparing these values across firmware versions. The reference data has been recomputed to
account for the new processing method.
S-cone protocols
Two S-cone protocols are provided, which may be useful in the detection of enhanced s-cone
syndrome (Yamamoto, Hayashi, and Takeuchi 1999). These protocols use a background of
To Next Page
Loading...
