The International Society for Clinical Electrophysiology of Vision (ISCEV) standard for full-field electroretinography (ERG) describes a minimum set of tests [1], but encourages the use of additional protocols for clinical ERG testing. This extended protocol describes extracting ERG components related to rod phototransduction from the a-wave response to a strong flash. The protocol was prepared by the authors in accordance with ISCEV procedures (

Scope and applications

The normal human dark-adapted ERG to a strong brief flash is a mixed response with both rod- and cone-mediated contributions. The cone-driven contribution is not negligible; in diseases in which rod responses are selectively attenuated, the cone system is likely to contribute proportionately more to the response, making accurate assessment of rod photoreceptor signals more challenging. Rod-isolated ERG components may be obtained by subtracting cone-mediated responses as detailed below. Information about the dynamics of rod phototransduction can be derived from analysis of the initial portion of the rod-isolated ERG a-wave [2,3,4,5,6,7,8,9,10,11,12,13,14,15]. This extended protocol describes the recording of ERGs to the same strong flash under both dark-adapted and light-adapted (LA) or photopic rod-saturating conditions. The cone-driven LA ERG is then subtracted from the mixed response obtained in the dark.

The protocol is applicable to patients in whom diseases affecting the rod photoreceptors are suspected. Specifically, the protocol enables a more focused evaluation of the light responses of rods, not obtainable from the standard ISCEV protocol, and thus allows for more detailed phenotyping and understanding of disease mechanisms. This includes characterization of diseases which affect rod quantal catch (e.g. shortening of the outer segment), diseases that affect the dynamics of any component of the activation phase of rod phototransduction, or diseases that reduce the total number of rod photoreceptors. This protocol has thus far only been used as a research tool and is not currently used in a clinical setting for the diagnosis of retinal disease.

Patient population

Patients of all ages able to tolerate Ganzfeld stimulation referred for investigation of possible retinal or rod photoreceptor dysfunction, including those with reduced or delayed DA3 and DA10 ERG waves, consistent with abnormal rod function.

Technical issues

This protocol uses many of the recording parameters specified in the ISCEV protocol for the full-field ERG [1]. Additional considerations include the following:

  1. (a)

    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.

  1. 1.

    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.

  2. 2.

    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 [7]. 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.

  3. 3.

    A flash delivered immediately (300 ms) after extinction of a rod-saturating background [18]. Such stimulus delivery may not be readily achievable with standard equipment and also increases testing time, especially if averaging is needed.

  4. 4.

    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.

  1. (b)

    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 [18]. Strong flashes presented immediately after onset of the photopic background can also be difficult to tolerate.

  1. (c)

    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]).

  1. (d)

    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.

  1. (e)

    Flash duration.

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).

  1. (f)

    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].

  1. (g)

    Photopic background.

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).

  1. (h)

    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).

  1. (i)

    Photopic inter-stimulus interval.

Cones recover within 2 s even in response to strong flashes [20]. 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.

  1. (j)

    Signal averaging.

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].

  1. (k)

    Response analysis.

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 [15]. 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 [15] to such analysis (discussed in “Appendix 2”), this protocol does not endorse the use of a particular mathematical model at this time.


The stimulus parameters are specified in scotopic trolands with conversion to photopic units for a white spectral light source with spectral composition similar to a xenon flash. If a source with different spectral composition is used, scotopic and photopic stimulus strength should be measured directly using appropriate filters. All other calibration issues are identical to those specified in the ISCEV ERG standard [26].

Protocol specifications

Patient preparation follows that for the current ISCEV standard ERG [1]. This extended protocol has the following additional specifications.

  1. (a)

    Sequence and timing of scotopic and photopic ERGs.

    Suitable scotopic strong flash responses for a-wave analysis are recorded after 20 min DA or following the standard DA ERG protocol. Photopic responses are recorded after 10 min light adaptation after the standard LA ERG protocol.

  2. (b)

    Scotopic stimulus flash strengths.

    This protocol specifies a flash strength equivalent to 75 scotopic cd s m−2 or 30 photopic cd s m−2. Assuming a pupil diameter of 8 mm and a broad-spectrum white (7000 k) light source, this equates to a flash of 3.6 log sc Td s. To be consistent with ISCEV stimulus notation, this stimulus should be referred to as DA30. If the pupils are smaller, a higher flash strength may be needed. If a range of flash strengths is used, the DA30 ERG should be included.

  3. (c)

    Flash duration.

    This should be a brief flash not longer than 4 ms.

  4. (d)

    Scotopic inter-stimulus interval.

    For the DA30 ERG, a minimum inter-stimulus interval of 30 s is specified.

  5. (e)

    Photopic background.

    The photopic background luminance is 30 cd m−2, as for the ISCEV standard ERG protocol.

  6. (f)

    Photopic flash strength.

    The same flash strength should correspond to that used to elicit the DA strong flash ERG. The stimulus in this context is referred to as LA30, assuming it is delivered on the same background and after at least 10 min LA or after the ISCEV standard LA ERG protocol.

  7. (g)

    Photopic inter-stimulus interval.

    The inter-stimulus interval is a minimum of 3 s. This allows cone recovery between flashes and is comfortable for most patients, and an interval of 3 s allows efficient averaging if needed.

  8. (h)

    Averaging may be used to improve the signal/noise ratio, and typically 3–10 responses have been averaged in previous studies [5,6,7,8,9,10,11,12,13,14,15,16,17,18].


The dark-adapted (DA30) mixed response, the light-adapted (cone-mediated) response (LA30) and the rod-isolated ERG should be shown. The time axis should be chosen so that the ascending portion of the a-wave can be clearly visualized as well as the pre-stimulus baseline (Fig. 1). Trough amplitude and peak time of the isolated rod-driven a-wave together with normative ranges (if available) should be reported. Stimulus and recording parameters should be specified as for the ISCEV full-field ERG. Pupil diameter should be reported. Mathematical modelling of the a-wave is not included in this extended protocol (see “Appendix 2” for discussion of previous studies).

Fig. 1
figure 1

Recordings from a healthy volunteer using a Burian–Allen electrode in response to identical flash strengths (30 cd s m2) under dark-adapted (DA30; black trace) and light-adapted conditions (LA30; red trace). The blue trace is the result of subtracting the LA30 response from the DA30 response and represents the rod contribution to the mixed response