The electroretinogram (ERG) is a useful tool for objective, noninvasive assessment of retinal function both in the clinic and the laboratory. It is a mass electrical potential that represents the summed response of all the cells in the retina to a change in illumination. Recordings can be made in vivo under physiologic or nearly physiologic conditions using electrodes placed on the corneal surface. For standard recordings in the clinic, ERGs such as those illustrated in Fig. 24.1 are recorded from alert subjects who are asked not to blink or to move their eyes. Anesthesia, selected for having minimal effects on retinal function, may be required for recordings from some very young human subjects, and is generally used for recordings in animals. The positive and negative waves of the ERG reflect the summed activity of overlapping positive and negative component potentials that originate from different stages of retinal processing. The choice of stimulus conditions and method of analysis will help to determine which of the various retinal cells and circuits are generating the response. Information about retinal function provided by the ERG is useful for diagnosing and characterizing retinal diseases, monitoring disease progression, and evaluating the effectiveness of therapeutic interventions.

Fig. 24.1

Standard tests for full-field clinical electrophysiology. The Society for the Clinical Electrophysiology of Vision (ISCEV) has recommended five standard electroretinogram ( ERG ) tests with full-field stimulation for use worldwide in clinical electrodiagnostic facilities. This figure, from the 2015 update of ISCEV Standard for Full-Field Clinical Electroretinography shows examples, but not norms, of ERGs elicited using the recommended tests in normal humans. Light calibrations in candela seconds per meter squared (cd.s.m –2 ) for each test appear above the ERGs. Large arrowheads indicate the time at which the stimulus flash occurred. Dashed arrows show common ways to measure time-to-peak (“t,” implicit time), and a- and b-wave amplitudes. For some purposes, for example, quantitative analyses to separate various ERG components, amplitude measurements at a fixed time after stimulus onset are more appropriate (see Fig. 24.12 ).

From McCulloch DL, Marmor MF, Brigell MG, et al. ISCEV Standard for full-field clinical electroretinography (2015 update). Doc Ophthalmol . 2015;130:1–12, used with permission.

Much of the basic research on the origins, pathophysiology, and treatment of retinal diseases that occur in humans has been carried out in animal models. The ERG provides a simple objective approach for assessing retinal function in animals. It has been of particular benefit in studies of mouse, rat, other species, various models of retinal disease, and retinal drug toxicity. The ERG also has been very useful for characterizing changes in retinal function that occur as a consequence of genetic alterations in mice, in other animal models for human disease, and in humans themselves, that affect the generation, transmission, and processing of visual signals. This chapter provides information on retinal origins and interpretation of the ERG, with emphasis on advances in our understanding of the ERG that have occurred through pharmacologic dissection studies in macaque monkeys, whose retinas are similar to those of humans. The similarity of the waveforms of flash ERGs of humans and macaques can be seen in Fig. 24.2 . The chapter also will examine origins of the mouse ERG and consider similarities and differences between mouse and primate ERG. Although the focus of this chapter is on primate and rodent retinas, it is important to note that the ERG is a valuable tool for assessing retinal function of all classes of vertebrates. This includes amphibians and fish, in which classical studies of the retina were carried out (for a review, see Dowling ) and in which current studies continue to improve our understanding of retinal function under normal, genetically altered, and pathologic conditions.

Fig. 24.2

Full-field flash electroretinograms ( ERGs ) of human subjects and macaque monkeys. ( A ) Top : Dark-adapted ERG responses to brief high energy flashes of ~400 scotopic troland seconds (sc td.s) from darkness that occurred at time zero for a normal alert human subject (left) and an anesthetized macaque (right) . For macaque ERGs, and ERGs from all animals in subsequent figures, subjects were anesthetized.

Bottom : Light-adapted ERGs in response to longer duration flashes (150 or 200 ms). For both subjects, white Ganzfeld flashes of 4.0 log photopic (ph) td were presented on a steady rod-saturating background of 3.3 log sc td. ( Top modified from Robson JG, Frishman LJ. Dissecting the dark-adapted electroretinogram. Doc Ophthalmol . 1998; 95(3–4):187–215 used with permission; bottom modified from Sieving PA, Murayama K, Naarendorp F. Push-pull model of the primate photopic electroretinogram: a role for hyperpolarizing neurons in shaping the b-wave. Vis Neurosci . 1994;11(3):519–532, used with permission.)

( B ) ERG recording setup: recordings made using a traditional Ganzfeld bowl (left) or more modern LED-based full-field stimulator. Use of DTL fiber electrodes is illustrated. ERG recordings are amplified and sent to a computer for averaging, display, and analysis.

This chapter provides useful background information for interpreting the ERG but does not provide a comprehensive review of the characteristics of ERGs associated with retinal disorders encountered in the clinic. A guide to electrodiagnostic procedures and the more comprehensive texts of Heckenlively and Arden and Lam are useful resources for learning more about clinical applications.

Generation of the electroretinogram

Radial current flow

The ERG is an extracellular potential that arises from currents that flow through the retina as a result of neuronal signaling and, for some slower ERG waves, potassium (K + ) currents in glial cells. Local changes in the membrane conductance of activated cells give rise to inward or outward ion currents, and cause currents to flow in the extracellular space (ECS) around the cells, creating extracellular potentials. Although all retinal cell types can contribute to ERGs recorded at the cornea, the contribution of a particular type may be large, or hardly noticeable in the recorded waveform, depending upon several factors detailed further in the chapter.

The orientation of a cell type in the retina is a major factor in determining the extent to which its activity will contribute to the ERG. A schematic drawing of the mammalian retina with the various cell types labeled can be found in Chapter 21 ( Figs. 21.1 and 21.2 ). When activated synchronously by a change in illumination, retinal neurons that are radially oriented with respect to the cornea, i.e., the photoreceptors and bipolar cells, make larger contributions to the major waves of the ERG than the more laterally oriented cells and their processes (i.e., the horizontal and amacrine cells). The major waves at light onset are, as marked in Figs. 24.1 and 24.2 , initial negative-going a-waves (mainly from photoreceptors), which are followed by positive-going b-waves, mainly generated by ON (depolarizing) bipolar cells. For longer-duration flashes ( Fig. 24.2 , bottom row ), the light-adapted ERG includes the b-wave at light onset, and another positive-going wave, the d-wave, at light offset, with major contribution from OFF (hyperpolarizing) bipolar cells. Currents that leave retinal cells and enter the ECS at one retinal depth (the current source) will leave the ECS to re-enter the cells at another (the current sink), creating a current dipole. These retinal currents also travel through the vitreous humor to the cornea, where the ERG can be recorded noninvasively, as well as through the extraocular tissue, sclera, choroid, and high resistance of the retinal pigment epithelium (RPE), before returning to the retina. Local ERGs can be recorded near the retinal generators using intraretinal microelectrodes in animals, while simultaneously recording the global ERG elsewhere in the current path, for example, from the corneal surface, or using an electrode in the vitreous humor, with a reference electrode behind the eye. Historically, such recordings have provided useful information about the origins of the various waves of the ERG.

Glial currents

The ERG waveform is also affected by glial cell and RPE cell currents. Retinal glial cells whose currents affect the ERG include Müller cells and radial astrocytes in the optic nerve head. One crucial function of glia is to regulate extracellular K + concentration, [K + ] o , to maintain the electrochemical gradients across cell membranes that are necessary for normal neuronal function. Membrane depolarization and spiking in retinal neurons that occur in response to changes in illumination lead to leak of K + from the neurons and to K + accumulation in the ECS. Membrane hyperpolarization, in contrast, leads to lower [K + ] o as the membrane leak conductance is reduced, but the Na + K + -ATPase in the membrane continues to transport K + into the cell.

K + currents in Müller cells move excess K + from areas of high [K + ] o to areas of lower [K + ] o by a process called spatial buffering. Return currents in the retina are formed by Na + and Cl . The regional distribution and electrical properties of inward-rectifying K + (Kir) channels in Müller cells (see Fig. 24.3C ) are critical for the spatial buffering capacity of the cells. Studies by Kofuji and coworkers have shown that strongly rectifying Kir2.1 channels in Müller cells are located in “source” areas, particularly in synaptic regions where [K + ] o is elevated due to local neuronal activity. In contrast, Kir4.1 channels (weakly rectifying) are densest in Müller cell endfeet, in inner and outer limiting membranes, and in processes around blood vessels. K + enters Müller cells through Kir2.1 channels that minimize K + outflow and K + exits the Müller cells via the more bidirectional Kir4.1 channels to enter the extracellular “sink” areas with low [K + ] o : the vitreous humor, subretinal space (SRS), and blood vessels.

Fig. 24.3

Distal retinal components of the direct-current electroretinogram (dc-ERG), and distribution of inward-rectifying potassium channels, Kir2.1 and Kir4.1, in Müller cells.

( A ) Simultaneous intraretinal (local) and vitreal (global) ERG recordings from the intact eye of the cat. Top: Transretinal potentials and transepithelial potentials (TEP) were recorded in response to a 5-minute period of illumination. The a- and b-waves cannot be seen due to the compressed time scale. The intraretinal recordings show the two (sub)components of the c-wave: (1) the retinal pigment epithelium (RPE) c-wave, which is the TEP recorded between a microelectrode in the subretinal space and a retrobulbar reference, and (2) slow PIII, the cornea-negative transretinal component generated by Müller cell currents (see text), recorded between the same microelectrode and the vitreal electrode. The vitreal ERG (bottom) , recorded between the vitreal electrode and retrobulbar reference, is the sum of the RPE c-wave and slow PIII. The c-wave is followed by the fast oscillation ( FO ) trough and then the light peak ( LP ), both of which are exclusively RPE responses.

(From Steinberg RH, Linsenmeier RA, Griff ER. Retinal pigment epithelial cell contributions to the electroretinogram and electrooculogram. Prog Retin Res . 1985;4:33–66, used with permission.)

( B ) Direct-current electroretinogram (dc-ERG) recorded from a C57BL/6 mouse in response to a 7-minute period of illumination. FO, LP, and OFF-response amplitudes are marked by arrows. Upper record shows the early portion of the response on an expanded (×5) time scale. Amplitude calibration: 0.5 mV.

(From Kofuji P, Biedermann B, Siddharthan V, et al. Kir potassium channel subunit expression in retinal glial cells: implications for spatial potassium buffering. Glia . 2002;39(3):292–303, used with permission.)

( C ) Strongly rectifying Kir2.1 channels are widely distributed in the Müller cell membrane. K + in synaptic regions enters Müller cells through strongly rectifying Kir2.1 channels and exits through weakly rectifying (bidirectional) Kir4.1 channels concentrated in the ILM, inner limiting membrane OLL, outer limiting lamina and Müller cell processes around blood vessels. IPL , Inner plexiform layer; OPL , outer plexiform layer.

Modified from Kofuji P, Biedermann B, Siddharthan V, et al. Kir potassium channel subunit expression in retinal glial cells: implications for spatial potassium buffering. Glia . 2002;39(3):292–303, used with permission.

ERG waves associated with glial K + currents have a slower time course than waves related to currents around the neurons, the activity of which causes the changes in [K + ] o . As [K + ] o increases or decreases with neuronal activity, the resulting glial K + currents will be related to the integral of the K + flow rate. Other waves generated by glial K + currents in Müller cells, or other retina cells that move K + such as the RPE cells, include the c-wave and slow PIII ( Fig. 24.3A,B ), both related to the reduction in [K + ] o in the SRS that occurs when photoreceptors hyperpolarize in response to a strong flash of light. The negative scotopic threshold response (nSTR) and photopic negative response (PhNR), both originating from inner retinal activity, are also thought to be mediated by glial K + currents resulting from that activity, in the retina or optic nerve head.

Stimulus conditions

Aside from structural and functional aspects of the retina, stimulus conditions are of great importance in determining the extent to which particular retinal cells and circuits contribute to the ERG. Signals will be generated in rod pathways, in cone pathways, or in both, depending upon the stimulus energy, wavelength, and temporal characteristics, as well as upon the extent of background illumination, with rods responding to, and being desensitized by, lower light levels than cones. Fully dark-adapted ERGs driven by rods only (i.e., scotopic ERGs) are thus useful for assessing rod pathway function (dark-adapted ERGs in Figs. 24.1 and 24.2 , top ), and light-adapted ERGs driven only by cones (i.e., photopic ERGs), for assessing cone pathway function ( Figs. 24.1 and 24.2 , bottom ). Fig. 24.1 , bottom right , also shows responses to 30 Hz flicker, which isolates cone-driven responses because rod circuits do not resolve high frequencies well. Bright light flashes elicit small wavelets superimposed on the b-wave called oscillatory potentials (OPs) that are generated by circuits proximal to bipolar cells. OPs can be isolated by bandpass filtering: 75 to 300 Hz in Fig. 24.1 , top right .

The spatial extent of the stimulus is an important factor in ERG testing. For standard clinical tests, as illustrated in Fig. 24.1 , and listed in Box 24.1A , as well as for most testing of animal models, a full-field (Ganzfeld) flash of light is used ( Fig. 24.2B ). A full-field stimulus generally elicits the largest responses because more retinal cells are activated and the extracellular current is larger than for focal stimuli. Full-field stimulation also has the advantage that all regions of retina are evenly illuminated and, with respect to background illumination, evenly adapted. Pupils are generally dilated for full-field stimulation. More spatially localized (focal) stimuli are useful for analysis of function of particular retinal regions, for example, foveal versus peripheral regions in primates. Multifocal stimulation allows assessment of many small regions simultaneously.

Box 24.1

Standard and more specialized electroretinogram tests

Standard electroretinogram (ERG) tests

Described by ISCEV Standard for Full-Field Clinical Electroretinography (2015 Update) . All numbers are stimulus calibrations in cd.s.m –2

  • Dark-adapted 0.01 ERG (“rod response”)

  • Dark-adapted 3.0 ERG (“maximal or standard combined rod-cone response”)

  • Dark-adapted 3.0 oscillatory potentials (“oscillatory potentials”)

  • Light-adapted 3.0 ERG (“single-flash cone response”)

  • Light-adapted 3.0 flicker ERG (“30 Hz flicker”)

  • Recommended additional response: either dark-adapted 10.0 ERG or dark-adapted 30.0 ERG

Specialized types of ERG and recording procedures

  • Macular or focal ERG

  • Multifocal ERG (see published standard )

  • Pattern ERG (see published standard )

  • Early receptor potential (ERP)

  • Scotopic threshold response (STR), negative and positive

  • Photopic negative response (PhNR*)

  • Direct-current (dc) ERG

  • Electro-oculogram (see published standard )

  • Long-duration light-adapted ERG (ON-OFF responses a

    a Published Extended Protocols for specialized nonstandard diagnostic testing are listed on the ISCEV website: https://iscev.wildapricot.org/standards


  • Paired-flash ERG to isolate cone contributions

  • Chromatic stimulus ERG (including S-cone ERG* and dark-adapted red flash ERG*)

  • Dark and light adaptation of the ERG

  • Dark-adapted* and light-adapted luminance-response analyses*

  • Strong flash a-wave analysis*

  • Specialized procedures for young and premature infants

ERG responses for a minimum set of stimuli were selected by the International Society for the Clinical Electrophysiology of Vision (ISCEV) to efficiently acquire standard, comparable data on rod and cone pathway function from clinics and laboratories around the world (illustrated in Fig. 24.1 ). Names of the standard tests are listed in Box 24.1A . Tests using stimuli presented over a fuller range of stimulus conditions to allow more complete or specific evaluation of retinal function, are listed in Box 24.1B , and a few of these tests will be described later in this chapter.

Noninvasive recording of the electroretinogram

ERGs can be recorded from the corneal surface using various types of electrodes. A commonly used electrode, with good signal-to-noise characteristics, is a contact lens with a conductive metal electrode set into it (Burian Allen electrode). It has a lid speculum to reduce effects of blinking and eye closure. In the bipolar form of the electrode, the outer surface of the lid speculum is coated with conductive material that serves as the reference. This type of electrode is best tolerated (in alert subjects) when a topical anesthetic is used. Other types of contact lens electrodes have also been used, for example, the jet electrode, which is disposable. Some clinicians and researchers use thin mylar fibers impregnated with silver particles, called DTL electrodes, as illustrated in Fig. 24.2B , or gold foil, or wire loop electrodes (H-K loop) that hook over the lower eyelid. For rodents, metal wires in loops or other configurations, placed in contact with the corneal surface, are often used. Some labs use cotton wick electrodes and some use DTL fibers under contact lenses or another form of contact lens electrode. Corneas are kept hydrated with a lubricating conductive solution in all cases. The reference electrode can be placed under the eyelid, for example, the speculum of a contact lens electrode as described previously, or remotely, for example, on the temple, the forehead, or the cornea of the fellow eye. ERG signals ranging from microvolts to a millivolt or more, peak to peak, for responses to strong stimuli, are amplified and digitized for computer averaging and analysis. Filtering is carried out to remove signals outside the frequency range of retinal responses to stimulation (<1 and >300 Hz), and to remove line frequency noise (e.g., 50 or 60 Hz).

Classical definition of components of the electroretinogram

The origins of the various waves of the ERG have been of long-standing interest to clinicians and researchers. Our current understanding of the cellular origins of the ERG profits from extensive knowledge, as described in previous chapters, of the functional microcircuitry of the retina, and particularly of the physiology and cell biology of the retinal cell types, and the identity and action of retinal neurotransmitters, their receptors, transporters, and release mechanisms. However, a classical study using ether anesthesia provided the first pharmacologic separation of ERG components.

Granit’s classical pharmacologic dissection of the ERG (illustrated in Fig. 24.4 ) provided valuable insights on origins of ERG waves, as well as a nomenclature for waves based on their distinct retinal origins. Component “processes” were found to disappear from the ERG during the induction of ether anesthesia in the following order: process (P)I—the slow c-wave response that follows the b-wave, generated predominantly by the RPE; PII—the b-wave, generated by bipolar cells; and eventually PIII—the photoreceptor-related responses that remained the longest during ether anesthesia. PII and PIII are still commonly used terms for ERG components generated by ON bipolar cells and photoreceptors, respectively.

Fig. 24.4

Dark-adapted electroretinogram ( ERG ) of the cat and its three component processes revealed by induction of ether anesthesia. The solid red line indicates the ERG of the intact cat eye in response to a 2 s flash (14 Lamberts) from darkness. Dashed lines indicate three processes ( PI , PII , and PIII ), numbered in the order in which they disappeared from the ERG during induction of ether anesthesia.

Modified from Granit R. The components of the retinal action potential in mammals and their relation to the discharge in the optic nerve. J Physiol . 1933; 77(3):207–239, used with permission.

Slow PIII, the c-wave, and other slow components of the direct-current (dc)- electroretinogram

PIII of the ERG can be separated into a fast and a slow portion: fast PIII is the a-wave, which reflects photoreceptor current (see further), and slow PIII results from Müller cell currents induced by photoreceptor-dependent reduction in [K + ] o in the SRS ( Fig. 24.3C ). Negative-going slow PIII and the positive-going pigment epithelial response to the same reduction in subretinal [K + ] o add together to form the c-wave, which is positive-going in the dark-adapted ERG of the cat ( Figs. 24.3 and 24.4 ). This is because the positive-going RPE contribution is larger than the negative-going Müller cell contribution (illustrated by the intraretinal recordings from intact cat eye in Fig. 24.3A ). In mice, the c-wave is also positive-going. ( Fig. 24.3B ). In humans and monkeys, slow PIII and the RPE c-wave are more equal in amplitude, and the corneal c-wave is less positive. Two slower potentials that arise from the RPE, the fast oscillation potential (FO) and light peak (LP), are also present in cat and mouse dc-ERG recordings ( Fig. 24.3 ). The mouse LP is much smaller in amplitude than that in the cat. The cellular mechanisms that generate these slow waves were reviewed in more detail in previous reviews of ERG origins.

In alert human subjects it is not possible to obtain stable dc-ERG recordings necessary for recording slow events, for example, those arising from the RPE, that occur over seconds or minutes, because the eyes move too frequently. Therefore, to measure slow potentials, the electro-oculogram (EOG), an eye movement–dependent voltage, is recorded. The EOG is a corneo-fundal potential that originates largely from the RPE; its amplitude changes with illumination, being maximal at the peak of the LP. Use of EOGs to evaluate retinal/RPE function is described in an ISCEV standard publication on clinical EOG.

Full-field dark-adapted (Ganzfeld) flash electroretinogram

In Fig. 24.5 , full-field dark-adapted ERG responses to a range of stimulus strengths for a human subject (left) , a macaque monkey, whose retina and ERG are similar to that of human (middle) , and a C57BL/6 mouse (right) . The mouse ERG is similar to the primate ERG but larger in amplitude (see calibrations). For higher stimulus strengths than shown in the figure, the mouse ERG develops larger OPs than those generally seen in the primate ERGs (e.g., Fig. 24.6 ). The ERGs in Fig. 24.5 were generated almost entirely, except for responses to the strongest stimuli, by the primary rod circuit, which is the most sensitive retinal circuit. For all three species, the strongest stimuli evoked an a-wave, followed by a b-wave. For stimuli more than two log units weaker than the strongest one, b-waves were still present, but a-waves were no longer visible. B-waves can be seen in responses to weaker stimuli than a-waves partly because of the convergence of many rods (20–40) onto each rod bipolar cell, which increases their sensitivity, and partly because of the large radial extent of ON bipolar cells in the retina. The slow negative wave in the ERGs of the three subjects in response to the weakest stimuli, called the (negative) scotopic threshold response (nSTR), and the equally sensitive positive (p)STR are related to amacrine and/or ganglion cell activity. The high sensitivity of the STRs relative to b-waves (and a-waves) reflects the additional convergence of rod signals in the primary rod circuit in the inner retina proximal to the rod bipolar cells.

Fig. 24.5

Dark-adapted full-field electroretinogram (ERGs) of a human subject, a macaque monkey, and a mouse. The ERGs are of a human, macaque, and C57BL/6 mouse in response to brief (<5 ms) flashes from darkness generated by computer-controlled LEDs. Responses to the weakest stimuli were rod-driven (scotopic), whereas for the strongest stimuli they were driven by both rods and cones.

Modified from Robson JG, Frishman LJ. Dissecting the dark-adapted electroretinogram. Doc Ophthalmol . 1998; 95(3–4):187–215, used with permission.

Fig. 24.6

Negative electroretinogram (ERG) of mouse lacking a b-wave. Single-flash responses on short (left) and long (right) time scales from a control C57BL/6 mouse ( A ) and a Nob1 mouse ( B ). ERGs on the left are responses to flashes of 0.11, 0.98, 2.57, 4.37, 17.4, 27.5, and 348 sc cd.s.m –2 , and on the right to flashes of 40.0, 68.4, 102, 170, and 348 sc cd.s.m –2 .

From Kang Derwent JJ, Saszik SM, Maeda H, et al. Test of the paired-flash electroretinographic method in mice lacking b-waves. Vis Neurosci . 2007; 24(2):141–149, used with permission.

Dark-adapted a-wave

It has long been appreciated that the dark-adapted a-wave primarily reflects the rod receptor photocurrent. The a-wave generator was localized to photoreceptors in classical intraretinal recording studies in mammalian retinas, some of which included current source density (CSD), or source-sink, analyses. The most direct demonstration of the a-wave’s cellular origin was provided in such experiments by Penn and Hagins in isolated rat retina. These experiments produced evidence that light suppressed the circulating (i.e., dark) current of the photoreceptors, and the investigators proposed that this suppression is seen in the ERG as the a-wave.

Negative electroretinograms

The receptoral origin of the a-wave was also demonstrated in early studies in amphibians using compounds that blocked synaptic transmission, Mg 2+ , Co 2+ , and Na + -aspartate, and isolated photoreceptor signals in the ERG while abolishing responses of postreceptoral neurons. These manipulations also caused the b-wave to disappear, indicating its postreceptoral origin. As our understanding of synaptic pharmacology has improved, it has become more common to use glutamate agonists and antagonists to block transfer of signals from photoreceptors to specific second-order neurons. For example, blocking metabotropic transmission to depolarizing (ON) bipolar cells with l -2-amino-4-phosphonobutyric acid (APB or AP4), an mGluR6 receptor agonist, eliminates the b-wave and produces a negative ERG. Much of the remaining ERG is the rod photoreceptor response, although late negative signals also arise from OFF pathway neurons.

An essentially identical negative ERG to that after APB administration occurs in mice in which the mGluR6 receptor is genetically deleted, or when there are mutations in the mGluR6 receptor or other proteins whose function is necessary for normal signal transduction in ON bipolar cells. For example, Fig. 24.6 shows the typical negative ERG of the dark-adapted Nob1 , that is, no b-wave, mouse. The Nob1 mouse has a mutation in the Nyx gene that encodes nyctalopin, a protein found in ON bipolar cell dendrites. Mutation of this protein produces a negative ERG in human patients who are diagnosed with X-linked complete congenital stationary night blindness (CSNB-1). Negative ERGs also occur for other forms of CSNB caused by mutations in mGlur6 receptors, and in Nob3 and Nob4 mice with such mutations, as well as in Nob2 mice in which glutamatergic transmission from photoreceptors is compromised. Although the ERG of the Nob1 and other mice lacking b-waves is almost entirely negative-going, it rises from its trough at the time course of the c-wave. Photoreceptor-dependent slow responses such as c-wave and slow PIII are not affected by blockade of postreceptoral responses in neural retina.

Retinal ischemia owing to compromised inner retina circulation also isolates the a-wave and eliminates postreceptoral ERG components. This was demonstrated in early experiments in monkeys by occlusion of the central retinal artery. A “negative ERG” in which b-waves are reduced or missing is a common clinical readout of central retinal artery and vein occlusions, as well as other disorders affecting postreceptoral retina such as melanoma-associated retinopathy, X-linked retinoschisis, complete CSNB, or toxic conditions.


The utility of the a-wave in studies of normal and abnormal photoreceptor function was advanced by the development of quantitative models based on single-cell physiology that could predict the behavior both of the isolated photoreceptor cell outer segment currents and ERG a-wave in the same or similar species. Hood and Birch demonstrated that the behavior of the leading edge of the dark-adapted a-wave in the human ERG can be predicted by a model of photoreceptor function derived to describe in vitro suction electrode recordings of currents around the outer segments of single primate rod photoreceptors. Lamb and Pugh followed a simplified kinetic model of the leading edge of the photoreceptor response (in vitro current recordings initially in amphibians) that took account of the stages of the biochemical phototransduction cascade in the outer segments of vertebrate rods. This model was subsequently shown to predict the leading edge of the human dark-adapted a-wave generated by strong stimuli and has been used extensively in clinical studies of retinal disease, and in analyses of photoreceptor function in animal models. A simplified formulation presented by Hood and Birch is often used (see legend of Fig. 24.7 ) and can also be adjusted to application to cone signals. Fig. 24.7 shows fits of Hood and Birch’s model to the dark-adapted a-wave of a normal human subject and a patient with retinitis pigmentosa (RP).

Fig. 24.7

Modeling the leading edge of the dark-adapted a-wave. ( A ) and ( B ) Rod-isolated electroretinogram (ERG) of a normal human subject ( A ), and a retinitis pigmentosa (RP) patient ( B ) in response to flashes of increasing strengths near a-wave saturation. Responses are fit with the Hood and Birch adaptation of the Lamb and Pugh model, expressed in the following equation:
R ( I , t ) = ( 1 − exp [ − I ∗ S ∗ ( t − t d ) 2 ] ) ∗ R max For t > t d

where response amplitude, R , is a function of flash strength, I , and the time, t , after the occurrence of brief (impulse) flash of light. S is the sensitivity that scales the stimulus strength, R max is the maximum amplitude, and t d , a brief delay commonly between 2.4 to 4 ms. ( C ) and ( D ) A simple approach is shown for estimating A max , and t max , indicators for model parameters, R max and S for the normal human subject ( C ) and RP patient ( D ).

Heckenlively JR, Arden GB, eds. Principles and Practice of Clinical Electrophysiology of Vision . 2nd ed. Fig. 35.13 (p. 498), © 2006 Massachusetts Institute of Technology, by permission of The MIT Press.

Photoreceptor models of Hood and Birch, Lamb and Pugh, and recent ones with improved fits provide parameters to represent the maximum amplitude of the a-wave, R max in equations of Hood and Birch, and the sensitivity ( S ) of the response. R max and S vary depending upon the pathology and stimulus conditions (e.g., adaptation level), and both parameters may be affected by RP. For example, S is thought to be more affected than R max in eyes in which the photoreceptors are hypoxic and more generally for abnormalities in the transduction cascade or increases in retinal illumination, whereas R max is more affected by photoreceptor loss (as illustrated in Fig. 24.7 ).

Although models of the leading edge of the a-wave yield useful parameters for describing the health of the photoreceptor outer segments, simpler approaches using stronger flashes than those advised by the ISCEV standard ( Fig. 24.1 and Box 24.1 ) are helpful. Hood and Birch and other investigators have noted that a pair of strong flashes, or even a single flash that nearly or just saturates the rod response, can be analyzed without fitting a model to make a rough estimate of R max and to measure a peak time that is related to S . Such measurements for a single flash are illustrated in the bottom row of Fig. 24.7 . The flash strength used was 4.0 log sc td.s, which is 63 times (1.8 log units) higher than the current ISCEV standard flash for mixed rod-cone ERG (assuming an 8-mm pupil).

Mixed rod-cone a-wave

For weak to moderate stimulus flashes to the dark-adapted retina, a-waves are dominated by rod signals, but stronger flashes elicit mixed rod-cone ERGs ( top row of Fig. 24.5 ). To investigate relative contributions of rod- and cone-driven responses to the ERG, it is necessary to separate them. Fig. 24.8 shows ERG (red circles) of a dark-adapted macaque in response to two different flash strengths and individual rod- and cone-driven contributions.

Fig. 24.8

Mixed rod-cone dark-adapted electroretinogram (ERG) of macaque monkey showing separate components. ( A ) ERG responses of a macaque to a brief blue LED flash of 188 sc td.s (57 ph td.s) and ( B ) to a xenon-white flash of 59,000 sc td.s (34,000 ph td.s). In both plots, the largest response (red circles) is the mixed rod-cone a-wave; the red dashed line shows the modeled mixed rod/cone response. The second largest response is the (isolated) rod-driven response ( filled blue squares and circles ), and the dashed blue line is the modeled rod photoreceptor contribution. The second to smallest response ( filled triangles ) is the isolated cone-driven response, and the dashed green line shows the modeled response (including postreceptoral portion). The smallest response (purple line) is the modeled cone photoreceptor contribution, based on findings after intravitreal injection of PDA (5 mM intravitreal concentration, assuming a vitreal volume of 2.1 mL in macaque eye in this and subsequent figures). The stimulus for part A (for 8.5 mm pupil) was 1 ph cd.s.m –2 , that is, about three times weaker (for cones) than the ISCEV standard flash of 3 cd.s.m –2 , whereas for part B it was 200 times stronger than the standard flash.

From Robson JG, Saszik SM, Ahmed J, Frishman LJ. Rod and cone contributions to the a-wave of the electroretinogram of the macaque. J Physiol . 2003;547(Pt 2):509–530, used with permission.

Rod-driven responses ( blue circles in Fig. 24.8 ) are extracted by subtracting the isolated cone-driven response to the same stimulus from the full mixed rod-cone response. Isolated cone-driven responses (triangles) are obtained by briefly (1 s) suppressing rod-driven responses with an adapting flash and then measuring the response to the original test stimulus presented 300 ms after offset of the rod-suppressing flash. Cone-driven responses in primates recover to full amplitude within about 300 ms, whereas rod-driven responses take at least 1 second, making it possible to isolate the cone-driven responses. The cone photoreceptor–driven portion of the leading edge of the a-wave represents about 20% of the saturated response ( Fig. 24.8 ). Model lines for the rod ( Fig. 24.8 , blue) and cone (purple) photoreceptor responses are modifications of Lamb and Pugh’s model. The entire cone-driven a-wave (green line) is larger than the modeled photoreceptor contribution (purple line) because it includes additional negative-going signals from the postreceptoral OFF pathway that can be eliminated with ionotropic glutamate receptor antagonists, as described in a later section on light-adapted ERG.

Time course of the rod photoreceptor response recorded as the a-wave

The leading edge of the a-wave is the only portion of the photoreceptor response that normally is visible in the brief flash ERG. As seen in the Nob1 mouse that has no b-wave ( Fig. 24.6 ), elimination of the major postreceptoral contributions (i.e., the b-wave) to the dark-adapted ERG leaves the c-wave present at late times. For high energy stimuli, the mouse ERG a-wave recovers toward baseline, forming a “nose” in response to strong stimuli ( Fig. 24.6B ). The a-wave nose is a common feature of the mammalian ERG. It also can be seen in the macaque ( Fig. 24.9 ) when post receptor responses are eliminated by pharmacologic blockade, and in humans in whom complete CSNB has eliminated the ON bipolar cell–dependent b-wave. This indicates that the b-wave does not have a role in truncating the negative-going a-wave response, although this has been commonly assumed to be the case. The presence of a nose generated by the photoreceptors also is in conflict with the models of a-wave generation described previously in the chapter and illustrated in Fig. 24.7A . These models assume that the a-wave reflects only the outer segment photocurrent response and not currents generated by more proximal portions of the photoreceptor cells. Note that the model line in Fig. 24.7A , based only on outer segment photocurrents, does not recover toward baseline despite the presence of the a-wave nose. In a study investigating the origin of the “nose” of the a-wave Robson and Frishman provided evidence that recovery of the a-wave toward baseline is shaped by currents around more proximal portions of the rod photoreceptor cell than the outer segments. They used SPICE (Simulation Program and Integrated Circuit Emphasis) to generate a-waves that took account of the electrical properties (resistance and capacitance) and spatial dimensions of the rat rod (outer segments, axons, nucleus, spherule) from classical studies of Hagins et al. and Penn and Hagins. The simulations demonstrated that the recovery from the a-wave trough (shaping the a-wave nose) can be attributed to the rod photoreceptor itself, as is suggested by the presence of a nose in ERGs of subjects lacking b-waves ( Fig. 24.9 ) . They proposed that the initial recovery of the a-wave nose reflects the large capacitive currents observed by Hagins et al. to occur in the outer nuclear layer in rat retina. The capacitive currents are largest around the stout nuclear region of the photoreceptor cell, when strong stimulus flashes are used.

Fig. 24.9

Electroretinograms (ERGs) lacking ON bipolar cell components. Normalized ERGs from three patients with complete congenital stationary night blindness ( cCSNB ) and a normal control (upper records redrawn from Nakamura M, Sanuki R, Yasuma TR, et al. TRPM1 mutations are associated with the complete form of congenital stationary night blindness. Mol Vis . 2010:16;425–437) and the ERG from a macaque (lower records from Robson JG, Frishman LJ. The rod-driven a-wave of the dark-adapted mammalian electroretinogram. Prog Retin Eye Res . 2014;39:1–22.) before and after pharmacologic suppression of both ON and OFF bipolar cell responses with a mixture of APB and PDA.

Dark-adapted b-wave (PII)

It is well accepted that the dark-adapted b-wave arises primarily from ON bipolar cells. Results of intraretinal recording and CSD analyses were consistent with the b-wave originating primarily from bipolar cells with contributions from Müller cell currents. Also, pharmacologic blockade of postreceptoral responses and specifically those of ON bipolar cells or mutations that prevent signaling by ON bipolar cells can eliminate the b-wave (e.g., Fig. 24.6 ).

There is also good evidence that for most of its dynamic range, the dark-adapted b-wave is generated by rod bipolar cells in the primary rod circuit. Only for very strong stimuli will rod signals pass via gap junctions to cones and then to cone bipolar cells, which will also then contribute to the dark-adapted flash response. For the dark-adapted, scotopic ERGs it has been possible to isolate Granit’s PII from other ERG components and to compare its characteristics with those of rod bipolar cells recorded in retinal slices. One approach is to pharmacologically isolate PII.

Intravitreal injection of the inhibitory neurotransmitter γ-aminobutyric acid (GABA) has been used to suppress inner retinal activity in order to isolate PII in the dark-adapted retina of the C57BL/6 mouse ( Fig. 24.10A ). GABA receptors are present on bipolar cell terminals, as well as on amacrine and ganglion cells of inner retina. Following GABA inhibition, inner retinal responses involved in generating the very sensitive pSTR and nSTR are blocked, leaving a pharmacologically isolated PII, likely generated by rod bipolar cells. Loss of the sensitive STRs was not due to a general loss of retinal sensitivity; a-wave amplitude and kinetics, reflecting photoreceptor function, did not change after GABA injection ( Fig. 24.10C ).

Jun 29, 2024 | Posted by in OPHTHALMOLOGY | Comments Off on Electroretinogram

Full access? Get Clinical Tree

Get Clinical Tree app for offline access