The record of the changes in the electrical potential of the retina following stimulation by a flash of light is called an electroretinogram. In this section, the historic development of electroretinography, the origin of the response to light, and the components and types of normal ERGs are described.
HISTORIC DEVELOPMENT OF ELECTRORETINOGRAPHY
The first work in this field was concerned with the corneoretinal potential, the resting potential, which is defined as the difference in potential between the cornea and the posterior pole of the eye. This potential was first described by Emile Dubois; Reymond, who showed in 1849 that the cornea is electrically positive with respect to the posterior pole. In 1865, Holmgren reported that the resting potential can be modified by the action of light shining on the retina. Shortly after this, Dewar and McKendrick independently discovered this light response. They were able to show that the changes in potential on impact of light amounted to 3 to 10% of the normal resting potential and were independent of the anterior portion of the eye. Initially, their measurements were made by placing electrodes on the cornea and the posterior pole of the eye, but they subsequently showed that the response to light could also he recorded between the exposed brain and the cornea, allowing the eye to be left in situ. They then found that the same electrical changes could be recorded by placing electrodes on the cornea and an adjacent area of skin. Having made this discovery, they were able to attempt to produce a human ERG; this was achieved by using a clay trough filled with saline as the corneal electrode.
These early attempts at human electroretinography were far from satisfactory, and it was not until the turn of the century that advances in recording techniques allowed more accurate records to be made. At this point, the stage had been reached when a waveform could be accurately recorded and measured (Fig. 11.1), and it was known beyond doubt that this waveform was produced by the retina, even though it was recorded through electrodes placed at some distance from the eye. By the early 1930s, attempts were being made to record a human ERG using the valve amplifier.
Figure 11.1. A biphasic ERG. This is the typical response recorded from the human eye, using the simplest equipment and a relatively weak flash stimulus.
At the same time, an important milestone was reached in the study or responses from animals in the classic work of Granit. He developed the idea, put forward by previous workers, that the ERG represents the sum of three waveforms, which he termed "processes." These he enumerated as P1, P11, and P111. He showed that if the ERG is recorded from a cat subjected to deepening levels of ether anesthesia, the waveform changes in a characteristic manner. This change in the waveform was thought to be caused by the selective inhibition of each of the three processes—P1, P11, and P111—in turn. Although they have been elaborated to some extent, these original ideas about the nature of the ERG are still held to be true today. As soon as the knowledge of the basic components of the ERG had become well established, much interest was centered on the relative contribution of photopic and scotopic mechanisms to the response. For example, in 1940 it was noted that the flicker fusion frequency was not the same under photopic and scotopic conditions. This difference is now being used in many electrodiagnostic clinics to assess cone function.
These various advances in our understanding of the electrical responses from the eye were based on work in animals. Meanwhile, investigations on human subjects were hampered by the technical problem of fixing the electrodes. A great step forward was made in 1941 when Riggs introduced the contact lens electrode. Until that time, clinical electroretinography did not really exist, and little was known about alterations in disease. Use of the contact lens electrode was limited until the pioneering work of Karpe began to be published from Stockholm in 1945. It soon became apparent that the contact lens electrode eliminated much of the interference caused by background noise. The late 1970s saw a further development in this respect: the introduction of a flexible electrode that hangs over the lower lid and allows good recordings to be made but at the same time enables the subject to view the stimulus directly rather than through a contact lens. This electrode has certain advantages when a formed stimulus is required rather than a simple flash.
Using the method described by Karpe, the human ERG appeared as a biphasic response. However, it had previously been shown in other animals that a series of small wavelets could sometimes be seen on the 'b' wave. In 1954, Cobb and Morton described the same phenomenon in man and named it the oscillatory potential. They counted four to six wavelets using a brief flash stimulus, since then it has been shown that these wavelets may be selectively abolished by disease. The early receptor potential (ERP)—another component of the human ERG—can be seen at the very beginning of the response, immediately before the 'a' wave. Brown and Murakami first described it in 1964 and showed that it could be elicited only by an intense light stimulus. The latent period of the ERP is very short, less than 60 msec, and this response appears as a small positive peak followed by a larger negative one. The ERP is thought to be an electrical manifestation of the bleaching of the photo pigment in the retina. These newer components of the human ERG can usually be seen if a photoflash stimulus is used; a typical ERG is shown in Figure 11.2.
Figure 11.2. ERG produced by a photoflash stimulus showing the early receptor potential (ERP) and oscillatory potential.
ORIGIN OF THE ELECTRORETINOGRAM
If the clinical value of the ERG is to be fully realized in the future, a full understanding of the mode of production of the waveform will be essential. A superficial lne spection of the problem might lead us to search for an origin of the 'a' wave in one layer of the retina, the 'b' wave in another, and so on. However, the research carried out so far indicates that the source of the recorded response must be considered in two stages. First, we must find out what component waves are added together to produce the final response; then, having isolated these components, we must determine their anatomic site of origin.
A further question arises when we consider that the ERG elicited by a diffuse flash of light is a mass response. Is this response the sum of different kinds of response from different parts of the retinal sphere? Local responses from different parts of the retina can now be obtained, and it is perhaps surprise ing that the response from a small area of retina is similar to the mass response. Differences can be seen in the waveforms of records from the fovea and the peripheral retina in the cynomolgus monkey. At the fovea, the 'b' wave is small, whereas the 'a' wave is large. These differences are partly due to the fact that the rods and cones do not produce exactly the same kind of response and partly because of the different anatomic configuration of the nerve elements at the fovea and the peripheral retina.
ANALYSIS OF THE RESPONSE
In the electrodiagnostic clinic, the human ERG is seen as a biphasic response, a negative 'a' wave followed by a positive 'b' wave. The 'b' wave is modified by the oscillatory wavelets in its ascending part. The other components of the ERG can be elicited only by using special recording techniques. It is now widely accepted that the 'a' wave represents the leading edge of P111, which is a negative wave derived front the inner segments of the receptors. P111 has a faster cut-off from cones than it does from rods, and this difference can explain differences between ERGs from vertebrates with rod-dominated or cone-dominated retinas (Fig. 11.3).
Figure 11.3. Comparison of ERGs from the rod-dominated eye of a cat (i) and the cone-dominated eye of a frog (ii). Note the absence of the c wave and the presence of the d wave in the cone-dominated retina.
Although the negative P111 continues for the duration of the ERG, its leading edge is all that is seen because P11 appears, and this positive wave is superimposed upon it. P11 is thought to arise from the middle layer of the retina, either from one type of bipolar cell or from the Milner cells. The rate of depolarization of the Muller cell is a little slow compared with that of the 'b' wave. P11 is sensitive to ischemia in the retinal circulation.
The 'c' wave, which is a positive wave following the familiar biphasic 'a'/' b' wave pattern, is not usually seen in the clinic because recording conditions do not allow it. The 'c' wave is thought to arise from the pigment epithelium. Figure 11.4 shows how P1, P11, and P111 summate to produce an ERG.
Figure 11.4 Diagram to show how PI, PII, and PIII summate to produce the ERG. Changes in the shape of PIII can account for the difference between rod- and cone-dominated retinas.
THE OSCILLATORY POTENTIAL. As mentioned, under suitable stimulus conditions, a number of small wavelets are seen on the 'b' wave. In the monkey, these wavelets are abolished in a striking manner by clamping the retinal circulation; they are also abolished in the human eye after central retinal artery occlusion. The dependence of these wavelets on the integrity of the retinal circulation suggests that they may arise in the inner part of the retina, which receives its nourishment from this source. It is interesting, however, that the wavelets seem to be more susceptible to ischemic change than is the 'b' wave. Intraretinal microelectrodes have been used to record oscillatory responses from the inner nuclear layer of the frog's retina, and the response can be produced only if a wide area of retina is stimulated. There is also some evidence that the wavelets are produced by tangentially oriented structures, and they are particularly well seen in vertebrates, whose retinas have a thick and well-developed inner nuclear layer.
It has been suggested that the wavelets are related to observed cyclical changes in the amplitude of the spike discharges in the Optic nerve, thus explaining their origin from the ganglion cells. However, they are still present in optic atrophy, and antidromic stimulation of optic nerve fibers does not reset the rhythm of the wavelets. The wavelets can best be produced by exposing the eye to double flashes spaced about 15 seconds apart. The second flash tends to produce a more well-defined response, and this requirement for preadaptation becomes more marked when the eye is dark adapted. The maximum chromatic sensitivity of the wavelets has been shown to be at the red end of the spectrum, and it has been claimed that the wavelets are abolished in patients with congenital achromatopsia. Their exact site of Origin is therefore still in doubt, although the evidence at present seems to point to the inner nuclear layer. In view of the fact that the wavelets show certain features in common with a response recordable from the amacrine cells, it has been suggested that they could represent a feedback mechanism from the amacrines to the bipolars.
More recently, it has been found that the last of the oscillatory potentials appears to behave differently from the others. Its timing alters with increase in stimulus frequency in a different manner. In fact, the last wavelet appears to be time locked to the stimulus offset, whereas the others are not. The suggestion is that the last wavelet is part of the off response, being generated by the retinal off elements described in single-cell recordings.
THE EARLY RECEPTOR POTENTIAL (ERP). Although the latent period of the 'a' wave becomes much shorter with stronger stimuli, it is never less than about 2 msec. For some years before 1964, it was suspected that a response might exist that bridged the gap between the moment of excitation and the onset of P111. In 1964, Brown and Murakami found that an electrical response of no detectable latency could be recorded from a microelectrode inserted into the inner segment of the receptors. It was then shown that this rapid biphasic response—termed the early receptor potential—could be recorded with large electrodes outside the retina. The ERP can now be recorded as a clinical procedure in the human.
The action spectrum of this response agrees with that for bleaching of visual pigment. The ERP was elusive in the past because it is easily obscured by artifacts and a strong flash is required to elicit it. The ERP comprises a small positive component, known as R1, followed by a larger component, known as R2, which leads directly into the 'a' wave. Because the latency is virtually zero, both R, and R2 are thought to arise from the outer segments of the receptors, and it has been suggested that they are caused by movements of charge in visual pigment molecules.
The ERP is more resistant to disease than are the other components of the ERG. When recorded from the isolated retina, it is not much altered by formaldehyde or metal-chelating agents. When the retina is heated, the ERP disappears at the same temperature at which the regular orientation of the pigment molecules is lost.
VARIATIONS IN THE NORMAL ERG. The clinician who interprets ERG traces should fully understand all the different factors that may influence a normal ERG. Without such knowledge, the reporting may be extremely misleading, and conclusions based on it unjustified. The factors that may influence a normal response can be summarized as follows:
- Physiologic factors
State of dark adaptation
Age and sex
- Type or adjustment of equipment
Amplifier (setting of gain and time con-stant)
Type of recorder
Stimulus color, duration, and intensity
Bubbles in contact lens
Photoelectric and electric artifacts
TYPES OF NORMAL ELECTRORETINOGRAM
PHOTOPIC AND SCOTOPIC ERG. The electrical response of the dark-adapted retina to a white flash reflects both rod and cone activity, but in the clinic, it is often useful to be able to separate these. This separation may be achieved in the following way:
- If a flickering light is used with a frequency of 30 cycles per second, then a pure cone response results because the rod system cannot respond at this rate.
- A pure cone ERG can also be produced if the stimulus is superimposed on a steady background illumination. The background illumination serves to saturate the rods so that they cannot respond to brief flashes.
- A rod response can be produced by stimulating the dark-adapted retina with a dim blue light, which is below the cone threshold.
FLASH ERG. When an intense photoflash is used to produce an ERG, certain special features appear in the response. First, din ERP can be seen immediately before the 'It! wave. The 'a' wave is abnormally large and the 'b' wave is small when measured front the resting potential. However, the potential difference from the peak of the 'a' wave to the peak of the wave is not very different from that obtained using the more classic technique of electroretinography described by Karpe. A further feature of this type of response is the prominence of the oscillatory potential. Finally, there is a pronounced refractory period after each response. If the stimulus is repeated half a minute after the first one, then the resulting response is half the size of the first; a repeat flash within 2 or 3 seconds of the first one produces no response whatsoever. Subjectively, a flash of this sort produces a dense afterimage, which changes color over a period of one-half to three-quarters of a minute and then disappears.
RESPONSE TO FLICKER. One method of obtaining a photopic ERG is to present the eye with a rapid series of flashes. If a stimulus flash is repeated every few seconds, using a weak stimulus, then the second response resembles a normal ERG. If the flash rate is increased to two per second, then the second response and successive responses have a photopic character and are reduced in amplitude. As the frequency is increased, the amplitudes of 'a' and 'b' waves approach one another. Beyond a certain frequency, the trace becomes sinusoidal, and finally it flattens altogether when the critical fusion frequency is reached. The critical fusion frequency varies with the intensity of the stimulus. If a graph is made by plotting intensity against critical fusion frequency, the resulting curve has a kink in it at about 20 per second. This kink corresponds with the rod/ cone break in the dark-adaptation curve, and it suggests that the rods do not respond above the level of about 20 per second. With high intensities, a fusion frequency of 70 per second can be reached.
"OFF EFFECT." As a rule, the clinical ERG is recorded using a brief stimulus flash whose duration is limited to less than 20 msec. If a more prolonged stimulus is used, however, a change in electrical activity is evident when the stimulus is discontinued. In animal experiments, this has been termed the "off effect," or wave. The human "off effect" is a negative-going wave with a weak stimulus, but it becomes a positive wave as the stimulus is increased in intensity.
The clinical applications of the "off effect" have been limited. It has been shown that the negative-going wave elicited by a dim stimulus is absent in congenital stationary night blindness but present in rod monochromatopsia. It has also been shown that a series of wavelets may be found on the "off effect" that bear a resemblance to the oscilatory potential.
Nilsson has developed a DC registration technique that has allowed more detailed study of the "off effect." After a very fast positive wave, there occurs a fast negative change (the 'f' wave), a slower positive wave with a maximum at 0.9 to 1.5 seconds after "off" (the 'g' wave), and a slow negative change with a maximum of 4 to 6 seconds (the wave). The wave seems to be the "off" equivalent of the wave.
PATTERN ERG. Until recently it was assumed that the ERG produced by a pattern stimulus was similar to that produced by a flash stimulus if matched for luminance. Considerable interest has been aroused by reports that the pattern ERG, unlike the flash ERG, is reduced after section of the optic nerve. The response to a diffuse flash is normal after optic nerve section because it arises from the unaffected distal components of the retina. Furthermore, during an attack of acute optic neuritis, both flash and pattern ERGs may be normal at first, but over the ensuing weeks the pattern ERG may decrease in amplitude, whereas the flash ERG, as expected, remains normal. Such findings have stimulated considerable research interest, and it now looks as though routine measurement of the pattern ERG in the clinic can be useful.
The pattern ERG is now recognized to have two major components: P1 (or P-50, indicating the latency in milliseconds) and N1 (or N-95, a negative wave following P-50 at about 95 msec). Evidence is accumulating that the early P1 component may be generated in the bipolar region, whereas the N1 component is generated in the ganglion cell layer, The pattern ERG has been shown to be a sensitive indicator of diabetic retinopathy and particularly of early glaucomatous damage to the retina.
FOCAL ERG AND MULTIFOCAL ERG. In theory, it is possible to obtain an ERG from a small area of retina if a suitable background stimulus is used to preadapt the surrounding area and prevent spurious responses from stray light. However, development of a reliable and widely accepted means of producing a focal ERG from a localized region of abnormal retina has proved difficult. One promising technique has involved the use of a stimulator ophthalmoscope with which the examiner can place the stimulus on the desired region of retina under direct observation. Such techniques are capable of detecting quite small foveal lesions. The response obtained from local stimulation is very small indeed and can only be identified by repeating the stimulus many times and averaging the results. This is done electronically so that the patient views a flickering light for a few minutes and a readout is obtained at the end. The size of such a response is in the region of 1 or 2
Recently this approach has been developed in a very interesting manner; a multi-input stimulus has been developed. The patient views a TV screen on which numerous illuminated squares or hexagons are presented, each being illuminated at independent moments or simultaneously with a probability of 1/2. The response from each consecutive moment of this pseudorandom sequence is recorded and analyzed. The analysis assumes that the local flash ERG shows spatial summation in a linear manner.
The end result of this technique is a map of the responses that resembles a plot of the
visual field obtained by standard clinical means. The method has the advantage of being objective, and it can be performed on both eyes simultaneously (Fig. 11.5).
Figure 11.5. The type of graphical recording obtained by the multifocal ERG.
STANDARDIZATION OF THE ELECTRORETINOGRAM
The present standardized ERG protocol was developed by a committee of the International Society for the Clinical Electrophysiology of Vision (ISCEV) and published in 1989. It was updated in 1994. This standard is now widely used. An example of responses obtained using this protocol is shown in Figure 11.6. In principle, the standardized technique involves:
- Initial dark adaptation after dilating the pupil, followed by measurement of the rod response to a dim white flash of known intensity.
- After this, a bright flash stimulus is applied to elicit a maximal response and the oscillatory potential. The latter can be portrayed separately by filtering the response.
- A single-flash cone response is then measured by suppressing the rods with a con. stant background luminance.
- The flicker response from the cones is measured.
Figure 11.6. The various traces that are obtained using the standardized ERG.
The important features of the normal ERG can be summarized as follows:
- Although the ERG can be recorded through electrodes placed outside and even at a distance from the eye, there 18 no doubt that it is produced by the retina; other structures in and around the eye probably make no contribution to it.
- The ERG is made up of the following components: the ERP, the 'a' wave, the 'b' wave, and the 'c' wave. There is also an "off effect," whose position depends on the timing of the stimulus flash.
- The ERP is thought to arise from the outer segments of the receptors. The 'a' wave is
part of Granit's P111 component and is thought to arise from the inner segments of the receptors. The 'b' wave corresponds to Granit's P11 component and is thought to arise from the inner nuclear layer. The 'c' wave, which corresponds to P1, probably arises from Under suitable stimulus conditions, the 'b' wave is modified by the appearance of three or four small wavelets, which probably arise in the inner nuclear layer but not from the same source as the 'b' wave itself. These wavelets are particularly sensitive to pathologic changes in the retina.
- Repetitive focal stimuli to different parts of the retina produce responses that can now be analyzed, giving a form of objective visual field measurement.
So far we have been considering electrical responses that are produced by exposing the eye to a brief flash of light. Although a variety of light stimuli are used in electroretinography, they are all relatively short flashes, lasting for milliseconds rather than seconds. In electro-oculography, a slightly different technique, the electrical responses of the eye to a prolonged light stimulus last-ing several minutes are measured.
The difference in potential between the cornea and the posterior pole of the eye, known as the corneoretinal potential (the resting potential) normally amounts to several millivolts. When the eye is exposed to a brief flash of light, the corneoretinal potential changes; the tracings of these changes constitute the ERG. Unfortunately, it is not easy to measure the corneoretinal potential over long periods of time because in practice the response is obscured by blinks, random eye movements, and other artifacts, which make the baseline potential unsteady. This problem can be alleviated during electroretinography by the use of an AC-coupled amplifier; this type of amplifier only responds to relatively rapid changes in potential, and in this way, a steady baseline is more easily maintained. When the eye is exposed to a continuous light stimulus, a slow change in the corneoretinal potential occurs. This change would not normally be seen using an AC-coupled amplifier, and the baseline would be too unsteady to obtain accurate measurements if a directly coupled amplifier were used.
Electro-oculography is a recording technique that allows an AC amplifier to be used to record these slow changes in the corneoretinal potential; rapid changes of potential are produced by moving the eyes to and fro and are fed through an AC amplifier to a pen recorder. These changes in potential have been shown to be related to the size of the corneoretinal potential if the size of the eye movements is kept constant.
To perform the test, the electrodes are placed on the skin on either side of the eye at the medial and the lateral canthi, and one indifferent electrode is usually placed on the forehead. The subject is seated, facing a screen that can be illuminated. In addition, two small red fixation lights are mounted on either side of the screen. The subject is then asked to look briskly from one fixation light to the other, thus making horizontal eye movements of a constant size.
The eye can be regarded as an electrical dipole, the cornea being positive with respect to the posterior pole. Figure 11.7 illustrates how eye movements in a horizontal direction can produce a modified square wave and how the vertical limbs of this waveform can increase and decrease in amplitude, depending on the size of the corneoretinal potential. The same method can, of course, be used to measure eye movements, but here we are concerned with changes in the corneoretinal potential, and the eye movements are kept at a constant value.
Figure 11.7. Diagram illustrating electro-oculography and calculation of the Arden index. Values of the amplitude of the square Wave in the dark and light are recorded. See Figure 11.17 for an example of an actual EOG.
As long ago as 1929, it was shown that eye movements produce electrical changes that can be measured by skin electrodes, but at that time it was assumed that these electrical changes were related to muscle action potentials. It was later conclusively proved that the changes in potential were due solely to the existence of the standing potential.
Electro-oculography thus provides a means of monitoring long-term changes in the corneoretinal potential and, in particular, a means of assessing the changes induced by light. In the clinic, the tests can be performed in an automated fashion, so that a nurse is needed only to fix on the electrodes. A value known as the Arden index is calculated from the raw data. This value should be above 180 in a normal subject, although some myopes have slightly lower values.
In its most commonly used form, the test involves seating the subject in the dark; during this time, the corneoretinal potential tends to fall. In continued darkness, the potential remains at a low level but tends to wander up and down slightly. After 12 minutes, the light stimulus is applied. There is an initial electroretinographic response, and then the potential falls for about 2 minutes, after which it begins to rise steadily over a period of about 7 minutes. The initial fall is termed the fast oscillation, or transient, and the rise is known as the light rise. After about 7 minutes, the corneoretinal potential reaches a peak value and then begins to fall in spite of the fact that the light stimulus is still being applied. This fall in potential is followed by a further rise, and it becomes apparent that the response is a form of damped oscillation.
For clinical purposes, a Standard for Clinical Electro-oculography has now been developed, and as is the case for electroretinography, it is recommended that the standards be carefully followed.
ORIGIN OF THE ELECTRO-OCULOGRAM AND ITS RELATION TO THE ELECTRORETINOGRAM
The electro-oculogram (EOG) recorded with skin electrodes is probably the resultant of several potentials. Potentials in the skin or other parts do not change as the eye rotates and therefore do not affect the results. Nonretinal potentials arising out of the eye itself, on the other hand, could be more important. For example, the cornea is thought to be polarized so that its anterior surface is negative; that is, it acts against the resting potential. If the corneal potential were to be abolished, then one might expect a corresponding increase in the resting potential. Both the standing potential and the dark trough of the EOG are markedly reduced by the administration of azide, which produces selective damage to the pigment epithelium. Hence, this may be the anatomic origin of at least part of the EOG response. Laboratory studies have shown that a delayed hyperpolarization of the basal membrane of the retinal pigment epithelium correlates with the fast oscillation of the EOG. Longer exposure to light produces a gradual depolarization of the basal membrane, which accounts for the light rise of the standing potential. These electrical changes in the retinal pigment epithelium presumably depend on the integrity of the receptors, and the EOG could not therefore be expected to be a specific indicator of retinal pigment epithelial function. In certain diseases the ERG may be normal when the EOG is grossly abnormal or vice versa, and this may be of diagnostic value. Patients with vitelliform macular degeneration, for example, may have a diminished light rise and a normal ERG (see case 11.3 at the end of the chapter). In congenital retinal dysgenesis, the EOG is normal but the ERG is abnormal, whereas in a congenital retinal abiotrophy, both the ERG and the EOG are affected. Despite a few examples in which changes in the EOG do not reflect changes in the ERG, our knowledge of the meaning of these differences is still inadequate.
In practice, the important differences between electroretinography and electro-oculography can be summarized as follows:
- Electroretinography measures rapid changes in the resting potential in response to light, whereas electro-oculography measures slow changes.
- A contact lens is not required for electro-oculography.
- Less skill is required for electro-oculography, and the equipment is more portable.
- Electro-oculography cannot easily be performed on patients who cannot fixate, and it is not suitable for testing retinal function in blind patients.
- Children under the age of 5 or 6 years cannot usually cooperate sufficiently to allow accurate electro-oculography.