This protocol uses many of the recording parameters specified in the ISCEV protocol for the full-field ERG . Additional considerations include the following:
Methods of estimating the dark-adapted cone system contribution to the scotopic ERG. The rod-isolated a-wave can be obtained by removing the cone-mediated contribution to the scotopic ERG a-wave using one of the four methods ([4, 6, 7, 16,17,18,19,20]; see “Appendix 1”), outlined below.
An identical flash may be delivered shortly after the DA very strong flash, at a time interval sufficient for the cones to recover from the first flash, but prior to the onset of rod recovery. A practical consideration is that the use of such paired bright flashes in quick succession can be difficult for some subjects to tolerate.
A photopically matched red flash may be used to elicit an ERG mediated largely by L- and M-cones, to be subtracted from the mixed DA rod and cone system response. However, for strong flashes, it is likely the red flash would also stimulate the rods substantially, and so an estimation of the cone-derived component would be compromised . It is possible to estimate the rod-derived component in the red flash response by delivering blue flashes that are scotopically matched to the red flashes [7, 9], but this increases the range of stimuli required and consequent testing time.
A flash delivered immediately (300 ms) after extinction of a rod-saturating background . Such stimulus delivery may not be readily achievable with standard equipment and also increases testing time, especially if averaging is needed.
An identical flash to that used under DA conditions, delivered in the presence of a steady rod-saturating background. Such stimuli may be conveniently integrated into the ISCEV standard ERG method, and this technique forms the basis of the extended protocol specified below.
Sequence and timing of scotopic and photopic ERGs.
Suitable scotopic strong flash responses for a-wave analysis may be recorded after 20 min DA or after the ISCEV standard DA ERG protocol. The ISCEV standard LA ERGs are recorded after 10 min light adaptation, to allow the cone-mediated responses to reach a steady-state amplitude. During the 10 min of light adaptation, the cone-mediated ERG grows in amplitude [21,22,23], and so a larger a-wave may be recorded. Thus, it could be argued that for estimation of the cone component in the dark, the strong flash should be delivered immediately after onset of the photopic background rather than following adaptation to this background as the latter may lead to over-estimation. However, there is evidence that the ERG following a period of light adaptation may be closer to the dark-adapted cone-driven ERG than that obtained immediately following onset of the rod-saturating background . Strong flashes presented immediately after onset of the photopic background can also be difficult to tolerate.
Spectral characteristics of scotopic flash.
Shorter wavelength (blue) flashes preferentially stimulate the rods while minimizing stimulation of L- and M-cones and have been used in a number of studies (e.g. [2,3,4,5,6,7,8,9]). However, white flashes are consistent with ISCEV standard ERG stimuli and have also been widely used to derive the rod-isolated ERG a-wave (e.g. [10, 12,13,14]).
Scotopic stimulus flash strengths.
Studies isolating and fitting models to the rod-mediated ERG a-wave have used a series of strong flashes [2,3,4,5,6,7,8,9,10,11,12,13,14,15]. However, shortened protocols based on one or two flashes have been proposed for clinical use [7, 10]. The stimulus should be strong enough to saturate the rods and suitable to probe changes in rod sensitivity. The flashes should be well-tolerated by most patients and chosen to avoid the flash artefacts that may be associated with the strongest flash stimuli.
The trough of the ERG a-wave to a 3.6 log sc Td-s flash occurs at approximately 10 ms. The duration of the flash must be considerably less than the photoreceptor integration time. Xenon flashes typically have flash durations of less than 1 ms and are ideal. Longer duration flashes produced by LED sources should be not more than 4 ms (i.e. at least 2 times shorter than a-wave peak time).
Scotopic inter-stimulus interval.
The interval between flashes should be a minimum of 30 s which allows for complete recovery of the rod ERG a-wave to a 3.6 log sc Td s flash [17, 19].
The photopic background is used to saturate rods, enabling recording of cone-mediated ERGs. A white background of 3.3 log sc Td (15 ph cd m2) eliminates the rod a-wave and has a small effect on the cone a-wave [4, 19]. To minimize cone desensitization, a blue background which has a high scotopic, but low photopic luminance, has also been used [18, 20]. The ISCEV standard for photopic ERGs stipulates a 30 ph cd m−2 background which will produce a reasonable estimate of the cone-isolated ERG over the rising phase of the ERG a-wave (although there may be some desensitization of the cone photoreceptors).
Photopic flash strength.
Photopic ERGs in this protocol are intended to estimate the cone-mediated contribution to the mixed DA strong flash ERG. To achieve this, stimulus strength should be unchanged from that used under scotopic conditions (see protocol specification below).
Photopic inter-stimulus interval.
Cones recover within 2 s even in response to strong flashes . Thus, an interval of ≥ 3 s between flashes is sufficient when averaging photopic responses. Flashes of this strength delivered at shorter intervals are less likely to be comfortable for subjects.
Averaging may be used to improve the signal/noise ratio, and typically 3–10 responses have been averaged [5,6,7,8,9,10,11,12,13,14,15,16,17,18].
Once the LA (cone-driven) response has been subtracted from the DA (rod- and cone-driven) response to obtain the isolated rod-driven a-wave, the conventional measurements of a-wave amplitude and peak time can be applied to the derived response. In addition, the time taken by the response to proceed from 10% to 90% of the peak amplitude has been proposed as an indicator of rod photoreceptor sensitivity . When LED flashes are used, the timing of the delay to the a-wave peak should be determined from the mid-time of the stimulus.
A large number of studies have used mathematical models to fit the leading edge of the rod-isolated a-wave to extract parameters relating to rod phototransduction [2,3,4,5,6,7,8,9,10,11,12,13,14]. However, as there are some important limitations  to such analysis (discussed in “Appendix 2”), this protocol does not endorse the use of a particular mathematical model at this time.