Electroretinogram responses in myopia: a review

The prevalence of myopia is on the rise worldwide since a few decades [1]. Various meta-analysis studies have predicted that approximately 5 billion of the global population may develop myopia by the year 2050, with around 1 billion of them having high myopia [2, 3]. The increasing prevalence of myopia and its associated sight-threatening risks [3‐9] make myopia a major public health concern [10‐12] and demands investigation into the fundamentals of eye growth regulation. The ocular stretching in myopia is associated with several structural and functional changes in posterior segment of the eye [13, 14]. Recent research highlights the role of retinal signaling in ocular growth, and various studies have investigated the electrophysiological responses in different types of refractive errors. This review is aimed to provide a summary of research work on electroretinogram (ERG) responses in myopia and their implications in understanding the nature of retinal functioning in myopic eyes.

The development of neural retina usually begins on day 26 of gestation [15], where the inner neural ectoderm divides into 3–4 layers of cells [16]. By week 12, the retinal layers start to form, with the inner neuroblastic layer giving rise to the ganglion, amacrine, and Muller cells creating the inner retina [17]. Similarly, the outer neuroblastic layer gives rise to photoreceptors (rods and cones), bipolar and horizontal cells forming the outer retina [17]. By the end of week 14, the ganglion cells migrate away from the fovea toward retinal periphery and cone photoreceptors migrate toward the fovea [16, 18].

Retina being the only photo-sensitive neural layer in the eye [19], incorporates about 55 types of structurally and functionally specific neurons [20, 21] including photoreceptors, bipolar cells, ganglion cells, horizontal cells, and amacrine cells [22, 23]. The distinct arrangement of these neurons from outer to inner retina forms a complex circuit to capture the photons of light from an object [24‐26]. These photons are converted into electrical/neuronal signals by the photoreceptors with the help of visual pigments present in them by a process called “photo-transduction” [26, 27]. The retinal photoreceptors, through synapses with retinal bipolar cells, transmit signals to the retinal ganglion cells. Upon activation, the axons of retinal ganglion cells carry neuronal signals to the brain via optic nerve for visual perception [28]. Evidence from animal studies suggests that both inner and outer retina may influence the detection of optical defocus and signaling for the corresponding development of ocular growth [19, 29‐34].

Several animal species including chicks [35‐44], squid [45], tree shrews [33, 46], monkeys [47‐55], marmosets [56], guinea pigs [57], kittens [58‐60], mice [61‐63], and also humans [64‐68] are capable of identifying the sign and magnitude of retinal image defocus and make compensatory alterations in ocular growth [69‐74]. Evidence from the experiments conducted on animal models indicates that the absence of input from the accommodative system (cycloplegia, ciliary nerve section, or damage to the Edinger–Westphal nucleus) [31, 39] or higher visual center (optic nerve section) [75] does not influence the ocular response to imposed form-deprivation [38, 75], or optical defocus [31, 39, 42, 75, 76], suggesting that the visual mechanisms regulating the refractive development are primarily localized at the retina.

Given that the fovea provides the best visual acuity (largely attributed to cone signaling) [77, 78], it was traditionally assumed that cone pathways may have a greater influence on visual signaling for refractive development [70, 79]. However, as the foveal area corresponds to only a small part of the visual field, it is reasonable to assume that the peripheral retinal areas might also be important in driving refractive status. There is growing evidence involving animal models indicating the presence of ocular growth pathways mediated by signals from the peripheral retina. The normal response to a) form-deprivation in monkeys treated with laser ablation at the cone-rich fovea [52], b) similar myopic responses in monkeys with form-deprivation [48, 53] and hyperopic defocus [51, 54] imposed on the rod-dominated peripheral regions or the entire visual field, and c) recent work on Gnat1^−/− mice with rod dysfunction [63] indicate that the peripheral rod pathways may be equally critical for visually guided eye growth. Blocking the functions of photoreceptors [19, 80], ON and OFF pathways [60, 81‐83] by pharmacological means (neurotoxins) [30‐32, 34] or genetic means (such as in mouse models) [84‐88] is known to affect both normal refractive development and response to form-deprivation myopia (FDM) showing the importance of various retinal neurons, neuronal pathways, and neurotransmitters in the refractive development of eye [89, 90]. Overall, these studies support the hypothesis that refractive development occurs through a cascade of local and regionally selective mechanisms in the retina [55, 70, 73, 74, 79].

The retinal function can be assessed by electrophysiological tests that study the electrical properties of the biological cells and tissues, driven by the flow of ions (ion current) [26, 91‐93]. Of various electrophysiological tests, the electroretinogram (ERG) with the standard protocol by the International Society for Clinical Electrophysiology of Vision (ISCEV) is widely used to determine the global and localized retinal responses [94‐97]. When a bright flash of light illuminates the retina, changes in membrane potentials across the neuronal and non-neuronal retinal cells simultaneously and instantaneously with a high temporal resolution (milliseconds) [94‐97] give rise to an extracellular current, which forms the basis of ERG [98‐100]. Hence, the ERG test provides a unique opportunity to investigate changes in retinal electrical activity to several inherited and acquired retinal diseases [94‐97] and several disorders or ocular conditions including refractive errors [101]. The most commonly used ERG techniques are full-field flash ERG (ffERG), multifocal ERG (mfERG), and pattern ERG (PERG).

A flash ERG measures the average response of retinal cells from a relatively broad retinal region to a full-field luminance stimulation [94]. By varying the background illumination, the light- or dark-adapted state of the eye, and the intensity of stimulus flash, one can elicit and isolate responses from different retinal cells. A standard ERG waveform is usually biphasic, with an initial cornea-negative response (a-wave), followed by a cornea-positive response (b-wave), and an additional slower positive wave or c-wave (Fig. 1a: scotopic and Fig. 1c: photopic) [97, 102]. In general, electrical activity in the photoreceptors, ON-bipolar cells, and retinal pigment epithelium initiate the a-wave [103], b-wave [104], and c-wave [102], respectively. The oscillatory potentials (OPs) that indicate the activity of amacrine cells in inner retina are represented by a small high-frequency wavelet component on the ascending limb of b-waves [94].

The results from several studies that investigated ffERG responses in myopia are given in Table 1. The overall incidence of abnormal electrophysiological findings in myopes younger than 18 years of age was reported to be 29%, with a higher proportion of ERG abnormalities reported in higher ametropias (spherical equivalent refractive error, SER worse than  ± 6.00 D; 52%) compared to individuals with emmetropia (SER:  − 0.75 D to + 1.50 D; 26%) or low ametropia (SER lower than ± 6.00 D) [101].

Since the first report of conventional ERG in myopes by Karpe in 1945 [105], various studies have reported impairment of retinal function in myopia. Several studies reported a significant reduction of b-wave amplitude in myopia that closely correlated with the degree of myopia and the axial length of eye [106‐115]. For every 1-mm increase in axial length, the dark-adapted 3.0 ERG showed a reduction of 15.7 μV and 23.4 μV in a-wave and b-wave amplitude, respectively, in an absence of a myopic retinal degeneration [109]. Significant differences in both a- and b-wave amplitudes of ffERG have been reported in high myopia (SER:  − 6.00 D to − 14.50 D), moderate myopia ( − 3.00 D to − 5.00), and low refractive error (+ 0.75 D to − 2.75 D) [108, 110] with a significant reduction in the b-wave amplitude under both scotopic (Fig. 1b) and photopic (Fig. 1d) conditions for individuals with high myopia (Fig. 2) [106, 107, 110, 111]. Scotopic responses (dark-adapted 3.0 ERG) were, however, reported to be more significantly affected than the photopic responses (light-adapted 3.0 and 30 Hz flicker ERG) [109, 110].

Studies have shown significantly lower short (S), long (L) and middle (M) wavelength-sensitive cone ERG b-wave amplitudes in high myopic eyes than the non-myopic eyes [116]. Significant reduction of cone and rod responses (mostly cones) in individuals with early-onset of high myopia (onset age, ≤ 5 years) than those with late-onset of high myopia (onset age, 12.4 ± 2.5 years) suggests that cone-rod dysfunction may be a sign for early onset of high myopia [117, 118]. The ffERG [119] and focal macular ERG [120] findings in pathologic myopes (presence of myopic retinal degeneration caused by progressive stretching and thinning of posterior segment of eye due to excessive axial length elongation which can result in reduced best-corrected visual acuity [12, 121]) showed a significant reduction in the amplitude of a- and b-wave in high myopic eyes with tigroid fundus appearance when compared to emmetropic eyes, whereas the implicit time was within the normal range. Similarly, there was a significant reduction in the amplitude of a-wave, b-wave, and OPs and delay in implicit time in high myopia with posterior staphyloma involving the macula compared to early myopia with tigroid fundus [120]. The reduced amplitude with normal implicit time in high myopia with tigroid fundus was related to a significant reduction in the macular cone density (focal macular ERG), which is considered to be an early macular change in high myopia [119, 120]. Furthermore, it was suggested that the reduced amplitude with delayed latency in high myopia associated with macular pathologies such as posterior staphyloma involving the macula could further reduce macular cone photoreceptors [120]. Likewise, chorioretinal vascular changes, retinal pigment epithelium degeneration, and receptor changes found in degenerative myopia may also play a major role in altering the ERG responses [119].

In contrast, Wan et al. (2020) recently reported an increase in the amplitude of a- and b-wave of the scotopic/dark-adapted 3.0 ERG (combined responses arising from the photoreceptors and ON-bipolar cells of both the rod and cone systems; rod-dominated) with the degree of myopia [122]. In addition, the average peak frequency of the rod-driven dark-adapted OPs arising from amacrine cells and the inner plexiform layer also showed a significant positive correlation with the magnitude of myopia [122]. The authors argued that these inconsistencies in comparison with other studies reflected the composition of the participants in their study, being young adults without any sign of pathological myopia (i.e., myopic retinal degenerations). In addition, the responses obtained from previous studies reflected combined contributions of the rod- and cone-driven OPs, interfering with each other [123‐126], while Wan et al. isolated the rod-driven OPs by subtracting the light-adapted ERG from the dark-adapted ERG [127, 128]. Their findings indicate an alteration in the rod and ON-bipolar cell function in myopia, with minimal effect on the cone system. It is hypothesized that changes in the rod-mediated retinal function may be related to the changes in the retinal dopaminergic pathways (dopamine D2 receptors) [122, 129]. The variation in the dark-adapted 3.0 ERG OP amplitudes in myopes indicates an imbalance of the 'ON' and 'OFF' retinal activity, which may be associated with the development of myopia and its progression [130].

A conventional ffERG measures the global electrical response of the entire retina, but it does not provide a localized response [131]. A mfERG is applicable for objectively studying local retinal health as well as characterizing and monitoring focal retinal lesions in various pathological conditions [132‐134]. The mfERG uses a specific hexagonal stimulus pattern to obtain a topographic map of retinal electrophysiological activity over a restricted retinal region (~ 40–50°), unlike ffERG, that reflects light-induced electrical activity from almost the entire retina [95]. This specific hexagonal pattern stimulus illuminates the retina using a pseudo-random binary m-sequence algorithm and gives rise to a continuously recorded signal from individual retinal locations [95]. All the localized responses can be averaged to compare quadrants, hemi-retinal areas, normal and abnormal regions of the two eyes, or successive rings from center to periphery [95]. Routinely, the stimulus pattern (array) with 61 or 103 hexagons is used within a field diameter of 40–50° (20–25° radius from the point of fixation to the edge of display) [95]. In the case of 61 hexagons, they are grouped from center to periphery into five rings (R1–R5), where R1 is the central ring and R5 is the peripheral ring. The approximate eccentricity from R1 to R5 is < 2°, 2–5°, 5–10°, 10–15°, and > 15° (~ 23°), respectively [135]. A similar grouping for 103 hexagons display would have a total of six rings within the same field diameter.

A typical mfERG waveform (also called the first-order response, or first-order kernel) is analogous to the conventional ffERG response as it is biphasic, with an initial negative component (N1), followed by a positive peak (P1) (Fig. 3a) [95]. There is another second negative deflection (N2) after the positive peak (P1). In humans, the N1 component primarily originates from the cone photoreceptors with minimal contribution from ON- and OFF-cone bipolar cells, the P1 component arises from the activity of ON- and OFF-cone bipolar cells, and the N2 component is derived from inner retinal cells (amacrine and ganglion cells) [94, 95, 136, 137].

The results from several studies that investigated mfERG responses in myopia are given in Table 2. Previous investigations on early changes in retinal function using the first-order kernel responses in low, medium, and high myopes showed a significant reduction of N1, P1 [106, 109‐111, 138‐143], and N2 amplitude density [139] with a greater effect on P1 amplitudes than N1 amplitudes (Fig. 3b and Fig. 4) [109, 143, 144]. The reduction in N1 and P1 amplitude density and significant delay in corresponding latencies [145, 146] in all rings, as well as four retinal quadrants, was significantly correlated with axial length and the degree of myopic refractive error [106, 109‐111, 138‐144]. The delayed implicit time was attributed to the possible altered synaptic transmission between the ON- and OFF-bipolar cells or structural changes in the inner plexiform layer of retina [147].

Previous studies have found that the mfERG amplitude density are maximum, and the latencies of P1 and N2 waves are the longest at the central R1 (fovea), which progressively decreased with the increase in retinal eccentricity [135, 138, 142, 148]. However, P1 amplitude density reduction in case of children with progressive myopia (< − 1.00D/2 years) was reported to be significantly smaller than stable myopes at central 5 degrees (R1) [149]. Progressive myopes also showed significantly shorter implicit times for OPs arising from inner retina compared to emmetropes and stable myopes, with similar implicit times for stable myopes and emmetropes [150]. These findings collectively indicate that progression of myopia may lead to inner retinal changes and alter the electrophysiological responses at both center and periphery of the retina [150].

There are various other factors such as optical, electrical, and retinal factors that contribute to the reduced mfERG responses. For optical factors, the degraded mfERG responses are associated with the decreased retinal image size and retinal illuminance due to elongated axial length in myopes. Electrical factors such as increased electrical path and ocular resistance from the electrical sources (at retinal plane) and the ERG electrodes [151] and lower retinal cell responsivity are proposed to be linked with reduced mfERG responses in myopes [152]. Retinal factors attributing to the reduced mfERG responses in myopes are increased sub-retinal space and morphological alterations in retinal cells due to axial elongation [153, 154]. Morphological changes include decreased retinal thickness, decreased retinal photoreceptor density [155], structural changes in photoreceptor outer segment [19], and photoreceptor dysfunction [156]. The prolonged implicit times/delayed latency may also be due to altered synaptic transmission from retinal photoreceptors (primarily rods) to ON and OFF pathways [157].

The PERG is a contrast-based response, driven by macular photoreceptors and originating from retinal ganglion cells [94]. It is a measure of both central retinal function and retinal ganglion cell function [163]. In macaque monkeys, both ON and OFF pathways equally contribute to the transient PERG amplitudes [164]. Clinically, transient PERG response has two main components: P50 (positive peak at 50 ms from stimulus onset) is an inner retinal component driven by macular photoreceptors and N95 (negative peak at 95 ms from stimulus onset) is the second component which is contrast-related and is generated by the retinal ganglion cells [96].

The P50 and N95 wave amplitudes of the transient PERG response were reduced in high myopes with longer axial length compared to that of emmetropes and low myopes [101, 165‐167]. The amount of loss in P50 amplitude was proportional to the degree of myopia, i.e., 8% in low myopes (− 1.00 to − 3.00 D), 16% in moderate myopes (− 3.25 to − 6.00 D), and 36% in high myopes (− 6.25 to − 10.00 D), when compared with emmetropes or myopes up to − 0.75 D [166]. Similarly, the amount of loss in N95 amplitude was also proportional to the degree of myopia, i.e., 7% in low myopes, 21% in moderate myopes, and 43% in high myopes, when compared with emmetropes [166]. Although P50 wave latencies show no difference, N95 wave latencies were reported to significantly increase in high myopia [166]. The reduced P50 and N95 amplitudes in higher degrees of myopia may indicate early macular and ganglion cell dysfunction even in eyes with normal vision and a healthy appearance of the macula [165‐167].

A further refinement of mfERG is the gmfERG, in which a successive insertion of a full-field or global-flash stimulus between consecutive focal flashes of a standard mfERG stimulus enhances the adaptive response, isolating the outer and inner retinal responses into two major components [168, 169]. The direct component (DC) is predominantly derived from the outer retinal cells (photoreceptors and bipolar cells), whereas the induced component (IC) is derived from the inner retinal cells (ganglion and amacrine cells) [168, 170‐172].

The results from several studies that investigated gmfERG in myopia are given in Table 3. Evaluation of neural response from outer to inner retina in emmetropes and myopes using the gmfERG showed that both the DC and IC responses gradually decreased from R1 to R5. The IC responses were more affected as compared to the DC responses, indicating that the inner retina was greatly affected in myopes [173]. Both the DC and IC amplitude densities were significantly correlated for retinal mid-peripheral regions corresponding to R2 to R3 in myopic refractive error [173]. It is hypothesized that these gmfERG responses are mediated by light-adapted changes in the retinal dopaminergic system.

The gmfERG on myopic children with different contrast levels exhibited a significant reduction in central macular (R1) DC amplitude density at 96% contrast, while the IC amplitude density was unaffected [174]. But myopic adults showed a significant reduction in the paracentral DC amplitude density for 29% and 49% contrasts [175]. The IC amplitude density in myopic adults is reduced for all measured contrast levels in both central and peripheral retinas [175]. There were no significant changes for both DC and IC implicit times in children and adults [174, 175]. Overall, these findings suggest that gmfERG-derived inner and outer retinal function in myopes vary significantly with age and retinal eccentricity.

A similar contrast-based gmfERG setup was used to determine whether myopia progression measured over 1 year was associated with changes in retinal function. At 49% contrast, both the DC and IC amplitude densities at the macula (central R1) were significantly reduced with the progression of myopia [176‐178]. The DC and IC implicit times were also reduced considerably in the paracentral retinal region [176, 177]. However, the high-contrast responses remained unaffected by the myopia progression [176‐178]. The findings indicate that myopia progression in children alters the inner retinal function at central retina, along with partial involvement of paracentral retina [176, 177]. The retinal electrophysiological functions seem not only differentially affected in children and adult myopes, but also in outer and inner retina that differentially process the spatial details [179].

Given an alarming rise in the prevalence of myopia worldwide, various optical, pharmacological, and environmental strategies are being incorporated to prevent the development of myopia [180] as well as to slow down the rate of myopia progression in children [181, 182]. One of the popular optical anti-myopia strategies includes orthokeratology [183, 184], which decreases the hyperopic defocus at the peripheral retina in myopic eyes [185‐187]. A recent investigation on the effect of 60 days of orthokeratology treatment on PERG reported significantly delayed implicit time of P50 and N95 wave, but no effect on PERG amplitudes [188]. Because the blur induced by orthokeratology and other peripheral defocus inducing anti-myopia strategies is not the same across all retinal eccentricities, it would be useful to use the mfERG to investigate how ERG responses vary in different retinal regions.

Besides optical anti-myopia strategies, pharmacological management of myopia progression with atropine eye drops has also been one of the most effective strategies to control myopia progression in children [181, 189‐193]. The majority of previous ERG studies that investigated different concentrations of atropine eye drops (0.01%, 0.1%, 0.5%, and 1%) on retinal signals reported no significant effect of atropine on retinal function as demonstrated by ffERG [194‐196], mfERG [194, 197], or PERG [198] in young myopic children (< 14 years of age). However, there are a few studies that report contradictory results. Firstly, Khanal et al. [199] reported that 0.1% atropine eye drops resulted in a 14% reduction of dark-adapted 3.0 OP amplitudes and 4% delay in the a-wave implicit time of dark-adapted 10.0 ERG (stronger ffERG), indicating that atropine could alter neural activity in inner retina and activity of photoreceptors, respectively [199]. Secondly, Kothari et al. [194] reported a reduction in the P50 amplitude of PERG with 0.01% atropine eye drops, indicating that the induced optical blur due to cycloplegia and mydriasis may alter signal transmission in inner retina (amacrine cells) [194]. Lastly, it was reported that gmfERG responses increased with 0.1% atropine eye drops in the presence of optically induced myopic defocus, suggesting that the atropine may enhance the effects of myopic defocus in the inner layers of the peripheral retina in controlling the eye growth for potential anti-myopia effects [200]. Overall, the literature related to the influence of atropine eye drops on altering retinal signals and regulating ocular growth is sparse and warrants further in-depth investigations to improve understanding of this important relationship and mechanism.

To summarize, there are significant changes in retinal function, as assessed from ERG testing in myopes, and these changes strongly correlate with axial length and the degree of myopia. Although some investigations with the ffERG show significantly reduced dark-adapted and light-adapted a- and b-wave amplitudes with increasing degree of myopic refractive error, there is some evidence that dark-adapted responses are further attenuated than the light-adapted responses. These findings suggest that myopia may be associated with reduced photoreceptor (mainly rod response) and ON-bipolar cell activity. Several studies with the mfERG show reduced P1 amplitude density in myopes, suggesting alterations in retinal ON-and-OFF cone bipolar cells in myopia. The mfERG amplitude density associated with refractive error varies significantly with retinal eccentricity (larger reduction in peripheral retina than in the fovea), axial length, and the degree of myopia. Finally, studies have reported reduced PERG amplitudes and gmfERG amplitude density in both the central and paracentral retina in high myopia. While these studies illustrate important associations between myopic refractive error and changes in retinal electrical activity, there has been limited work to understand the longitudinal changes in the ERG and how they relate to myopia progression in younger eyes. Future work investigating electrophysiological responses in combination with the measurements of retinal structural changes (using optical coherence tomography) will provide valuable insights into how retinal electrical changes may influence ocular growth and refractive error development in humans. Given the availability and wide use of optical and pharmacological anti-myopia management strategies that are known to act at retinal level (such as orthokeratology, multifocal contact lenses, and atropine), it would also be interesting for future studies to examine how these anti-myopia interventions interact with retinal signals to prevent myopia.