The luminance origin of the pattern electroretinogram in man

Assessment of the uniform field electroretinogram for mouse retinal ganglion cell functional analysis

PURPOSE

The uniform field electroretinogram (UF-ERG) has been suggested as an alternative to the pattern electroretinogram (PERG) for non-invasive assessment of retinal ganglion cell (RGC) function in primates. We evaluated the validity of the UF-ERG to assess mouse RGC activity in vivo.

METHODS

Unilateral optic nerve crush (ONC) was performed on adult C57BL/6J mice. Contralateral eyes with uncrushed optic nerves and eyes from surgically naive mice served as experimental controls. Electrophysiological visual assessment was performed at 12 weeks post-ONC. Flash-mediated visual-evoked cortical potentials (VEPs) were measured to confirm the robustness of the ONC procedure. Full-field flash ERGs were used to interrogate photoreceptor and retinal bipolar cell function. RGC function was assessed with pattern ERGs. Summed onset and offset UF-ERG responses to alternating dark and light uniform field flash stimuli of different intensities and wavelengths were recorded from ONC and control eyes, and relative differences were compared to the PERG results. Following electrophysiological analysis, RGC loss was monitored by immunohistochemical staining of the RGC marker protein, RBPMS, in post-mortem retinal tissues.

RESULTS

ONC dramatically impacts RGC integrity and optic nerve function, demonstrated by reduced RGC counts and near complete elimination of VEPs. ONC did not affect scotopic ERG a-wave and b-wave amplitudes, while PERG amplitudes of eyes subjected to ONC were reduced by approximately 50% compared to controls. Summation of ON and OFF UF-ERG responses did not reveal statistically significant differences between ONC and control eyes, regardless of visual stimulus.

CONCLUSIONS

PERG responses are markedly impaired upon ONC, while UF-ERG responses are not significantly affected by surgical trauma to RGC axons in mice. The more closely related pattern and uniform field ERGs recorded in primates suggests species-specific differences in RGC features or subpopulations corresponding to PERG and UF-ERG response generators, limiting the utility of the UF-ERG for mouse RGC functional analysis.

Spatial properties of L- and M-cone driven incremental (On-) and decremental (Off-) electroretinograms: evidence for the involvement of multiple post-receptoral mechanisms

An overview of electroretinogram response components to incremental and decremental steps in L- and M-cone excitation was obtained in human observers, while varying the spatial properties of the stimulus. Responses to L- and M-cone stimuli of opposite polarity resembled each other, particularly within the central 35° of the retina. All amplitudes grew as stimulus size increased; however, earlier and later components of the On- and Off-responses showed differing degrees of dependency on stimulus size. Thus, they may reflect different proportions of responses originating in parvocellular (less stimulus size-dependent) and magnocellular (more stimulus size-dependent) streams.

On- and off-response ERGs elicited by sawtooth stimuli in normal subjects and glaucoma patients

The aim of this study is to measure the on- and off-responses and their response asymmetries elicited by sawtooth stimuli in normal subjects and glaucoma patients. Furthermore, the correlation between the ERGs and other functional and structural parameters are investigated. Full-field stimuli were produced using a Ganzfeld bowl with Light Emitting Diodes (LEDs) as light sources. On- and off-response ERGs were recorded from 17 healthy subjects, 12 pre-perimetric and 15 perimetric glaucoma patients using 4-Hz luminance rapid-on and rapid-off sawtooth stimuli (white light; mean luminance 55 cd/m(2)) at 100% contrast. The on- and off-responses were added to study response asymmetries. In addition, flash ERGs were elicited by red stimuli (200 cd/m(2)) on a blue background (10 cd/m(2)). The mean deviations (MD) of the visual field defects were obtained by standard automated perimetry. The retinal nerve fibre layer thickness (RNFLT) was measured with Spectral Domain Optical Coherence Tomography (SOCT). We studied the correlation between ERG response amplitudes, visual field mean deviation (MDs) and RNFLT values. The on-responses showed an initial negative (N-on) followed by a positive (P-on), a late positive (LP-on) and a late negative responses (LN-on). The off-responses showed an initial positive (P-off) a late positive (LP-off) and a late negative response (LN-off). The addition of on- and off-responses revealed an initial positive (P-add) and a late negative response (LN-add). The on-response components (N-on, P-on and LN-on) in the glaucoma patients were relatively similar to those of the control subjects. However, the LP-on was significantly elevated (p = 0.03) in perimetric patients. The LP-off was significantly elevated (p < 0.001), and the amplitude of LN-off was significantly reduced in perimetric patients (p = 0.02). The LN-add amplitude was significantly reduced (p < 0.001) and delayed (p = 0.03) in perimetric patients. The amplitudes of the LN-off and LN-add ERG components were significantly correlated with the PhNR in the flash ERG (LN-off: p = 0.01; LN-add: p < 0.001) and with RNFLT (LN-off: p = 0.006; LN-add: p = 0.001). On- and off-response ERGs and their response asymmetries, elicited by sawtooth stimuli, are altered in the glaucoma patients. The late components are affected. Changes in the late negative components are correlated with structural and other functional changes.

Social vision: sustained perceptual enhancement of affective facial cues in social anxiety

Heightened perception of facial cues is at the core of many theories of social behavior and its disorders. In the present study, we continuously measured electrocortical dynamics in human visual cortex, as evoked by happy, neutral, fearful, and angry faces. Thirty-seven participants endorsing high versus low generalized social anxiety (upper and lower tertiles of 2104 screened undergraduates) viewed naturalistic faces flickering at 17.5 Hz to evoke steady-state visual evoked potentials (ssVEPs), recorded from 129 scalp electrodes. Electrophysiological data were evaluated in the time-frequency domain after linear source space projection using the minimum norm method. Source estimation indicated an early visual cortical origin of the face-evoked ssVEP, which showed sustained amplitude enhancement for emotional expressions specifically in individuals with pervasive social anxiety. Participants in the low symptom group showed no such sensitivity, and a correlational analysis across the entire sample revealed a strong relationship between self-reported interpersonal anxiety/avoidance and enhanced visual cortical response amplitude for emotional, versus neutral expressions. This pattern was maintained across the 3500 ms viewing epoch, suggesting that temporally sustained, heightened perceptual bias towards affective facial cues is associated with generalized social anxiety.

Spatial distributions of on- and off-responses determined with the multifocal ERG

We studied the contribution of retinal on and off-mechanisms in the multifocal electroretinogram (mfERG) by measuring responses to saw tooth stimuli. Six healthy subjects participated in this study. Rapid-on and rapid-off sawtooth stimuli with a period of 427 ms were presented in a multifocal pattern composed of 19 hexagons. The stimuli were interleaved with a blank field of the mean luminance and chromaticity. On- and off-responses were added to extract response asymmetries. The amplitudes of on-, off-, and added-responses were determined for different eccentricities relative to a signal baseline that was defined as the average of the electrical level recorded in two different time windows in which no responses were present. Measurements were repeated with eight different stimulus stretch factors to account for changes in retinal cell density as a function of eccentricity. The amplitudes of all ERG components decreased with increasing eccentricity for all stretch factors. For stretch factors between 0 and 20, responses to the central and immediately adjacent hexagons were large in amplitude. For more peripheral hexagons, the responses were very small or absent. Three components were identified in the on-responses (N20(on), P46(on) and N100(on)). In the offresponses, we found one positive (P20(off)) and one negative (N90(off)) component, whereas in the addition, three components (N20(add), P46(add) and N100(add)) could be observed. The N20(on) and P46(on) amplitudes decreased less steeply with eccentricity than the N100(on) amplitude, whilst the P20(off) and N90(off) amplitudes exhibited a similar decrease with eccentricity. In the addition, the two negative components exhibited a similar decrease in amplitude as a function of eccentricity and decreased more steeply than the positive component. The number of stimulated cones and retinal ganglion cells was estimated from anatomical data and compared with the responses. The spatial properties of the amplitudes of N20(on), P46(on), P20(off), and N90(off) and P46(add) were similar to those of the stimulated cone numbers. The remaining components had spatial characteristics that resembled those of the retinal ganglion cells. It is proposed that the ERG asymmetries revealed in the summed responses have post-receptoral origins, some of them reflecting the activity of the ganglion cell population. The use of sawtooth stimuli provide, similar to the pattern ERG, a way to record the ERG asymmetries.

Brain computer interface using flash onset and offset visual evoked potentials

OBJECTIVE

This paper presents a brain computer interface (BCI) actuated by flash onset and offset visual evoked potentials (FVEPs). Flashing stimuli, such as digits or letters, are displayed on a LCD screen for inducing onset and offset FVEPs when one stares at one of them. Subjects can shift their gaze at target flashing digits or letters to generate a string for communication purposes.

METHODS

By designing the flickering sequences with mutually independent flash onsets (or offsets) and employing the inherent property that onset (or offset) FVEPs are time-locked and phase-locked to flash onsets (or offsets) of gazed stimuli, segmented epochs based on the flash onsets (or offsets) of gazed stimuli will be enhanced after averaging whereas those based on the onsets (or offsets) of non-gazed stimuli will be suppressed after averaging. The amplitude difference between the N2 and P2 peaks of averaged onset FVEPs, denoted by Amp(onset), and that between the N1 and P1 peaks of averaged offset FVEPs, denoted by Amp(offset), are detected during experiments. The stimulus inducing the largest value of the sum Amp(onset)+Amp(offset) is identified as the gazed target and the representative digit or letter is sent out.

RESULTS

Five subjects participated in two experiments. In the first experiment, subjects were asked to gaze at 25 flickering stimuli one by one with each for a duration of 1min. The mean accuracy with 10-epoch averages was 97.4%. In the second task, subjects were instructed to generate a string '0287513694E' by staring at stimuli on a pseudo keypad comprising ten digits '0-9' and two letters 'B' and 'E'. The mean accuracy and information transfer rates were 92.18% and 33.65bits/min, respectively.

CONCLUSIONS

The onset and offset FVEP-based BCI has shown that high information transfer rate has been achieved.

SIGNIFICANCE

A novel FVEP-based BCI system is proposed as an efficient and reliable tool for disabled people to communicate with external environments.

Contributions from lateral interaction mechanisms to the human ERG can be studied with a two-frequency method

We present a two-frequency method to investigate lateral interaction components (liERGs) in the human electroretinogram. Adjacent half cycles of sinusoidal gratings were modulated sinusoidally with different temporal frequencies f1 and f2. The liERGs were defined by the Fourier components at the intermodulation frequencies /f1 - f2/ and f1 + f2 which indicate nonlinear interactions between the half cycles. Significant liERGs were found in all subjects with a monotonic increase of the liERG magnitude in the spatial frequency range from f(s)=0.07 to 2.4 cpd. When /f1 - f2/ was below 5 Hz, liERGs were masked by noise intrusions. In a control experiment we demonstrated that the liERGs were not evoked by stray light artifacts. The liERGs may help to further differentiate the responses that are evoked by patterned stimuli within the retina.

Ocular optics, electroretinography and primary open angle glaucoma

Advancing the knowledge of ocular image quality is important in the development of new tests for the earlier detection of primary open angle glaucoma (POAG). Loss of blue sensitivity is an important feature of POAG; however, it is difficult to separate this loss from the effects of wavelength related intraocular light scatter and media absorption. A technique for recording the S-cone electroretinogram is described which may solve these problems. The effect of scattered light is avoided by a Ganzfeld stimulus and the absorption of short wavelength light is compensated by a photometric balancing of blue and green stimuli using the M-cone response. The modulation transfer function of the eye has enabled the true spatial tuning function of the retina to be determined by pattern electroretinography (PERG). This has demonstrated the role of retinal ganglion cells in the generation of the PERG. These techniques are shown to be highly sensitive tests in the early stages of POAG.

Pattern electroretinography (PERG) and an integrated approach to visual pathway diagnosis

The pattern electroretinogram (PERG) provides an objective measure of central retinal function, and has become an important element of the author's clinical visual electrophysiological practice. The PERG contains two main components, a positivity at approximately 50ms (P50) and a larger negativity at approximately 95ms (N95). The P50 component is affected by macular dysfunction with concomitant reduction in N95. The PERG therefore complements the Ganzfeld ERG in the assessment of patients with retinal disease. In contrast, the ganglion cell origins of the N95 component allow electrophysiological evaluation of ganglion cell function both in primary disease and in dysfunction secondary to optic nerve disease, where selective loss of N95 can be observed. Both macular dysfunction and optic nerve disease can give abnormalities in the visual evoked cortical potential (VEP), and the PERG thus facilitates more meaningful VEP interpretation. This review addresses the origins and recording of the PERG, and then draws on extensive clinical data from patients with genetically determined retinal and macular dystrophies, other retinal diseases and a variety of optic nerve disorders, to present an integrated approach to diagnosis.

The pattern electroretinogram in anterior visual pathway dysfunction and its relationship to the pattern visual evoked potential: a personal clinical review of 743 eyes

The pattern electroretinogram (PERG) has now been in routine clinical use for sufficiently long to allow a personal clinical review of its relationship to the cortically generated pattern visual evoked potential (PVEP). The PERG and PVEP findings are presented from 520 eyes with optic nerve demyelination (382 eyes), optic nerve compression (90 eyes) or heredofamilial optic atrophy (48 eyes), and these are compared with the findings obtained in 223 eyes with dysfunction anterior to the retinal ganglion cells. Dysfunction anterior to the retinal ganglion cells gives a reduction in the P50 component of the PERG, but this component is usually spared in optic nerve disease where selective loss of the N95 component is by far the most frequently occurring abnormality. A diagnostic strategy is presented.

Differentiation of VEP intermodulation and second harmonic components by dichoptic, monocular, and binocular stimulation

Modulation by two temporal frequencies differentiates visual processing at the fundamentals (1Fs), second harmonics (2Fs), and second-order intermodulation components (IMCs), the latter created neurally as the sum or difference of the two modulation frequencies. Steady-state VEPs were recorded while stereo-normal adults viewed luminance or grating stimuli modulated by up to three temporal frequencies under dichoptic, monocular, or ordinary (binocular) viewing conditions arranged using liquid crystal light shutters. In Experiment 1, modulation of luminance by a single temporal frequency produced strong 1F and 2F VEP components, but modulation of gratings produced only 2Fs. Modulation by two temporal frequencies resulted in IMCs, often in the absence of evoked activity in the EEG at the 1Fs. IMCs were generally larger during pattern as compared to luminance modulation. Amplitudes of 1Fs and IMCs were smaller, but 2Fs were larger, during dichoptic as compared to ordinary viewing. Although the 2F to a single modulation presented to one eye was not reduced when a second frequency was added to the opposite eye, monocular IMCs were diminished when a frequency was added to the opposite eye. We conclude that 2Fs and IMCs are associated with different neural substrates. Results are consistent with a two pathway model with one pathway having a nonlinear filter prior to binocular combination, the other pathway having a nonlinearity following binocular linear summation. Implications of these data for binocular function are discussed.

Effect of kainic acid and NMDA on the pattern electroretinogram, the scotopic threshold response, the oscillatory potentials and the electroretinogram in the urethane anaesthetized cat

Kainic acid (KA, 12.5-100 nmol) or N-methyl-D-aspartate (NMDA 25-250 nmol) was injected into the vitreous of one eye of urethane anaesthetized cats. Pattern electroretinograms (PERGs) were recorded to transient contrast reversing bars. Scotopic luminance electroretinograms (ERGs) were recorded to blue flashes. All doses of KA reduced the oscillatory potentials (OPs), PERG and focal ERG (FERG). At 50 nmol KA, the b-wave and scoptic threshold response (STR) were normal. At 100 nmol KA, the STR was absent and the b-wave reduced by over 50%. OPs and STRs were reduced in all NMDA injected eyes. NMDA at 25 nmol enhanced the FERG, PERG, and b-wave and high doses (above 150 nmol) reduced them. Light microscopic examination of retinas showed 25 nmol KA only damaged dendrites of ganglion cells. NMDA damage was slight with < 200 nmol. These data show that the cat PERG has a proximal component which is very sensitive to low doses of KA; the PERG and FERG are very similar; the STR and PERG are generated by different structures and that the OPs and the FERG and PERG are all generated close to the ganglion cell layer, proximal to the STR.

Pattern electroretinogram and visual evoked potential amplitudes are influenced by different stimulus field sizes and scotomata

The pattern electroretinogram and the visual evoked potential were recorded simultaneously with various stimulus fields and artificial scotomata of increasing sizes. In contrast to an earlier study, a smaller check size (20') and two stimulus field sizes (20 degrees x 20 degrees and 10 degrees x 10 degrees) for the scotomata were used. With a concentric decreasing stimulus field, a reduction of both the pattern electroretinogram and visual evoked potential was found. Both showed a simultaneous reduction of amplitudes, but, compared with the amplitude in the full field, the reduction was more extensive for the pattern electroretinogram at each test field size. This implies a greater contribution to the pattern electroretinogram from more eccentric retinal parts. An artificial central scotoma of increasing size in the 20 degrees x 20 degrees field had less influence on the pattern electroretinogram than on the visual evoked potential. The percentage amplitude loss of the visual evoked potential was more pronounced. The visual evoked potential was eventually abolished by a scotoma size from 10 degrees x 10 degrees upward, while the pattern electroretinogram was still registrable. When scotomata of similar size were introduced in a smaller (10 degrees x 10 degrees) field, percentage pattern electroretinogram and visual evoked potential amplitude losses were less separated than in a larger (20 degrees x 20 degrees) test field.

The second harmonic of the electroretinogram to sinusoidal flicker: spatiotemporal properties and clinical application

The photopic, focal electroretinogram elicited by sinusoidal modulation (8 Hz) of the luminance of a uniform field (flicker electroretinogram) presented in the macular region contains two main components: the first harmonic at the stimulus frequency (1F) and the second harmonic at twice the stimulus frequency (2F). Physiologic and clinical studies have demonstrated that the 1F originates in the outer retina, whereas the 2F has multiple postreceptoral sources, with a contribution of the innermost retina. The 2F, unlike 1F, is selectively altered in amplitude and phase in the early stages of optic neuritis and glaucoma. Both 1F and 2F are altered in maculopathies. An additional property of the focal electroretinogram is that the 2F, unlike 1F, has a very limited summation area (i.e., the function relating 2F amplitude with stimulus area is saturated for central stimulus sizes of the order of the macular region). This overrepresentation of the macular activity suggests that the 2F component of the flicker electroretinogram may provide a specific macular test even for full-field stimuli presented through opacities of the optical media.

The influence of pattern size on amplitude, latency and wave form of retinal and cortical potentials elicited by checkerboard pattern reversal and stimulus onset-offset

Transient pattern electroretinograms (PERGs) and visual evoked potentials (VEPs) were recorded with checkerboard pattern reversal and equiluminance stimulus onset-offset, elicited by a high quality moving mirror stimulator. Different sized checkerboard patterns (0.35-4.2 c/deg) were used as stimulus patterns. The wave forms of the equiluminance stimulus onset responses were similar to ERGs evoked with luminance decrease and the stimulus offset PERGs were like ERGs elicited by luminance increase. The PERG c wave and the VEP showed spatial frequency tuning with pattern reversal and stimulus offset. Spatial frequency tuning was not detectable with PERG a and b waves. Pattern reversal and stimulus onset evoked PERGs had no major spectral components above 40 Hz; stimulus offset evoked PERGs contained components up to 55.3 Hz. Retino-cortical time--measured as a latency difference of the PERG b wave to VEP P100--was identical with pattern reversal and stimulus onset and about 12 msec longer with stimulus offset. Our results suggest that the 3 stimulation modes, reversal, onset and offset induce different types of processing at the retinal and cortical levels. PERG a and b waves to our high luminance/contrast stimuli contain no pattern specific information and the c waves are the sum of luminance and pattern specific responses.

The different contributions of local luminance decreases and increases to the pattern electroretinogram (PERG)

The typical pattern-onset-offset stimulus (stimulus A) consisting of local luminance increases and decreases was broken down into stimuli presenting only local luminance increases (stimulus B) or only local luminance decreases (stimulus C). With stimulus B the onset ERGs are luminance responses. With stimulus C the onset ERGs are pattern-related responses showing a spatial band-pass function. With stimulus A the response is a linear addition of responses to stimuli B and C. The simultaneously recorded VEP is a pattern-related response with all three stimuli (A-C).

Electroretinograms (ERGs) and visual-evoked potentials (VEPs) elicited by pattern displacement

The relation between the amplitude of visual responses to a checkerboard stimulus and the degree of lateral displacement of the checks was examined across different check sizes with simultaneously recorded electroretinograms (ERGs) and visual-evoked potentials (VEPs). The amplitudes of both the b-wave and the after-potential of the ERG increase linearly with pattern displacement. However, the major components of the VEP (N70 and P100) were smaller than expected from linearity for both small checks with small displacements (thresholding) and for large checks with large displacements (saturation). These results suggest that the ERG is proportional to the number of receptors stimulated, but the VEP reflects neural processes influenced by the spatial structure of the stimulus.

Effect of spatial frequency on simultaneous recorded steady-state pattern electroretinograms and visual evoked potentials

Pattern electroretinograms (P-ERGs) and visual evoked potentials (VEPs) to 4 Hz alternating square-wave gratings were simultaneously recorded in 23 subjects. Responses were Fourier analyzed and amplitude and phase of the 2nd and 4th temporal harmonics were measured. The spatial frequency-amplitude function of the P-ERG 2nd harmonic component displayed either a bandpass tuning behavior, or a low-pass behavior. The peak amplitude for subjects with bandpass tuning was at 1.5 c/deg. The phase of the P-ERG 2nd harmonic decreased monotonically as spatial frequency increased. The VEP 2nd harmonic had a bimodal spatial frequency function with a peak at 3 c/deg and a second increase at spatial frequencies below 1 c/deg, regardless of the P-ERG characteristics. The phase of VEP 2nd and 4th harmonic had an inverted U-shaped function with peak at 3 c/deg and 1.5 c/deg respectively. Comparison of simultaneously recorded P-ERG and VEP spatial frequency functions demonstrated different tuning behavior for cortical and retinal responses. It is concluded that the proposed technique permits the separate analysis of retinal and cortical processing of visual information. The 2nd and 4th harmonic components of bEP behave independently of each other suggesting they may be generated by different subsystems.

Fundamental differences between the nonlinearities of pattern and focal electroretinograms

We directly compared nonlinear kernels of normal human pattern electroretinograms (PERGs) and corresponding localized flash ERGs (FERGs). The FERG was triphasic and resembled an adaptive process because it decayed slowly without changing shape over several kernel orders and interpulse intervals. The PERG was biphasic in the slice nearest the diagonal of the second-order kernel, similar to the FERG in slices farther from this diagonal, and without power in higher-order kernels. The unique PERG features were short-term effects that immediately followed a contrast transition. The appearance-disappearance PERG had a triphasic first-order kernel and a biphasic second-order kernel. The latter was similar to, but half the size of, that for the contrast-reversal PERG. When the first off-diagonal slices of the two PERG second-order kernels were analyzed in detail, we found in both that the first positive peak was larger than the FERG at intermediate spatial frequencies. Both PERG peaks in the slice had a low contrast threshold and were linear with contrast. The three FERG peaks of the corresponding FERG slice had a higher threshold and were saturated with increasing contrast. These observations show that the PERG contains substantial pattern specific nonlinear components and cannot be dismissed as merely the nonlinear subcomponents of the corresponding FERG.

Local pattern electroretinograms and ganglion cell activity in the turtle eye

1. Local electroretinograms and spike activity from ganglion cells were recorded from an eye cup preparation of the turtle retina. The responses were elicited with striped and plaid stimulus patterns. 2. The results obtained with the two forms of recording were highly similar. Both depended on the spatial phase of the pattern with respect to the recording electrode. Both had maximal response at the same stimulus spatial frequency. 3. The optimum spatial frequency (both for maximum electroretinogram amplitudes and spike discharge rates) shifted to lower values with the administration of the GABA antagonist, picrotoxin. 4. The low frequency falloff associated with this spatial tuning may point to a mechanism of lateral interactions common to the local electroretinogram and spike responses.

Lateral interaction component and local luminance nonlinearities in the human pattern reversal ERG

Two different mechanisms are presumed to contribute to pattern electroretinograms: non linearities of the local luminance response and nonlinear effects of lateral interactions. Previous attempts to discriminate the two components relied on the theoretical MTF of the optics. In this study, techniques of nonlinear systems analysis are used to extract the two components from the response to pattern reversal of different check sizes. The decomposition is based on the structure of the second order pattern reversal kernels alone. Detailed information about the two mechanisms can be gleaned from the kernel structure. The component properties are compared and discussed.

Electroretinograms to checkerboard pattern reversal in cats: physiological characteristics and effect of retrograde degeneration of ganglion cells

Pattern electroretinograms (P-ERGs) evoked by alternating checks were studied in cats. Spatial frequency functions for transient and steady-state P-ERG waves showed a bimodal distribution with a preferred frequency at 0.6-0.75 c/deg and a second amplitude increase at frequencies lower than 0.5 c/deg. Decreasing the pattern luminance by 0.5 log units produced a shift of the spatial tuning curve toward lower spatial frequencies. No temporal tuning was noted in the temporal frequency functions. The bimodal distribution suggests that at spatial frequencies higher than 0.5 c/deg, the tuning reflects a 'contrast response' originating in cells with center surround organization. At spatial frequencies below 0.5 c/deg, the 'luminance response' becomes predominant and is generated in cells sensitive to mean luminance changes. Transient and steady-state P-ERGs to medium and high spatial frequencies were abolished by section of the optic nerve, while low spatial frequency stimuli at or below 0.3 c/deg continued to evoke P-ERGs at 1, 5 and 10 months after surgery. Quantitative whole mount retina microscopic examination confirmed the retrograde degeneration of the ganglion cells. It is concluded that both transient and steady-state P-ERGs to small and medium spatial frequencies checks are predominantly related to ganglion cell activity, while P-ERGs to low spatial frequencies reflect preganglionic cell activity.

Distortion of spatial selectivity by pattern onset stimulation

Spatial selectivity of pattern evoked potentials has been thought to provide evidence of lateral inhibition. However, spatial tuning functions may be distorted by pattern onset stimulation that is applied repeatedly to the same area of retina so an after-image is formed. This only applies at low spatial frequencies because of the randomizing effects of eye movements. Low-frequency attenuation may therefore be exaggerated. Pattern reversal stimulation has the opposite effect and this is reflected in the literature by fewer reports of bandpass functions. A new method has therefore been devised to provide the true spatial response function. The spatial phase is reversed after every two consecutive presentations. By combining this paradigm with a correction for the optical transfer function of the eye, the true neural response function is obtained. Ten subjects participated in this study to evaluate the distortion of spatial selectivity in the pattern electroretinogram. The new stimulus paradigm reduced the low spatial frequency attenuation to a barely significant level giving an almost flat amplitude response for the + ve and - ve transients of the pattern electroretinogram for check sizes from 222' to 7' angular subtense. However, correction for optical degradation produces bandpass curves, which closely correspond to those predicted from recent data on receptive fields of primate retinal ganglion cells.

A review of the clinical applications of the pattern electroretinogram

The pattern electroretinogram (PERG) has recently been introduced as a clinical procedure. It has been thought by many to represent activity of the retinal ganglion cells, although this is still a matter of contention. The exciting prospect of a selective test of ganglion cell function led to the application of the PERG in a variety of ophthalmological conditions. In the course of these investigations the PERG was found to be diminished in cases of maculopathy, optic atrophy, optic neuritis, toxic optic neuropathy, neurotransmitter disorders, glaucoma and ocular hypertension and in retinal vascular disorders such as diabetes. It was also affected in some cases of amblyopia. This paper briefly describes the techniques used to record the PERG and reviews current literature pertaining to its clinical application.

Pattern ERG: effects of reference electrode site, stimulus mode and check size

We studied monocular pattern ERG (PERG) in 10 normal subjects and a patient with optic neuritis. No clinically significant PERG could be recorded from the occluded eye with any reference (ipsilateral ear or temple, or midfrontal), indicating that cross-contamination is not present with binocular testing. Ipsilateral temple reference minimized VEP (P100/N100) contribution to the PERG N95 which occurred with ipsilateral ear or midfrontal reference. The conclusions were confirmed by results from the patient, who had marked monocular delay of a normal amplitude P100. Twenty-four subjects were tested with monocular and binocular stimulation using an ipsilateral temple reference. There were differences in PERG latencies and amplitudes although the interside amplitude ratio showed smaller differences with binocular stimulation. Increasing check size (17, 35 and 70 min) decreased P50 and N95 latencies and increased P50 amplitude.

Temporal spacing of pattern alternation and human visual response

The electroretinogram and the visual evoked potential were recorded in response to checkerboard stimuli presented in phase alternation. The times in which the pattern was in its two positions were changed within a fixed total cycle time of two seconds. Both the amplitude of the electroretinogram and the evoked potential were affected by the length of the time interval between alternations, but in opposite directions. An electroretinogram that closely followed a preceding response was reduced in size while an evoked potential was larger. The results indicate that there is an interaction between adjacent pattern areas at the retinal level.

Isolation of scotopic human electroretinograms using color adaptation and pattern reversal stimuli

The human electroretinogram was recorded in response to alternations of a 500 nm checkerboard stimulus pattern presented against adaptation backgrounds of different wavelengths. A modification of Stiles' color adaptation paradigm was adapted to electroretinography. The results provided electrophysiological t.v.i. curves (threshold versus intensity curves) and permitted a determination of spectral sensitivity. The t.v.i. curves matched those seen in psychophysics. The spectral curve showed a close agreement with the scotopic CIE function.

Current source density analysis of linear and non-linear components of the primate electroretinogram

1. We have used the method of current source density analysis to locate the generators of harmonic electroretinogram (ERG) responses to contrast-modulated pattern and uniform-field stimuli in the primate retina. 2. Sinusoidal steady-state analysis was used, with a stimulus temporal frequency of 8 Hz. Fundamental and second-harmonic response components were measured for the uniform-field response. The second harmonic of the average of contrast-reversal pattern responses obtained at a series of spatial phases was also determined in the same experiments. In addition, retinal tissue resistance was measured. All of these measurements were obtained at a series of equally spaced depths in the retina. 3. Retinal resistivity was not observed to vary systematically with depth. In addition, any plausible undetected inhomogeneities of resistivity with depth were found to slightly affect the relative magnitudes of estimated current sources and sinks, but to have little effect on their localization. 4. In a given penetration, the phase lag of each harmonic component was relatively constant with depth in most cases; however the magnitude of this phase lag sometimes varied in different penetrations. To compare data from different penetrations, the constant phase lag for each harmonic was estimated, and the response data phase-shifted so as to bring all data into a standard (cosine) phase. 5. The resulting current source density analyses were found to be quite consistent in overall form for different penetrations and in different animals. These data were averaged to obtain a final estimate of the depth profiles for generators of different ERG components. 6. The uniform-field fundamental response was found to have a predominant source-sink pair in the distal half of the retina (receptor layer to outer plexiform layer). The pattern (second-harmonic) response generators had a quite different depth profile, consisting mainly of a source-sink pair in the proximal 20% of the retina (encompassing the nerve fibre layer to the middle of the inner plexiform layer). The uniform-field second-harmonic response showed a current source density (CSD) depth profile with multiple sources/sinks, as if it contained contributions from the other two.

Pattern electroretinogram can be more than the sum of local luminance responses

It is generally accepted that the pattern electroretinogram for very large spatial elements is the result of local luminance stimulation. Responses due to the luminance differences between elements may be assumed to be relatively unimportant because in the case of large elements only few retinal units are stimulated by gradients. With decreasing pattern element size one wonders to what extent the electroretinogram continues to be based on the local luminance stimulation. We investigated this question using 8 Hz checkerboard reversal and compared the pattern recordings with the recordings resulting from the same stimulus field modulated homogeneously (focal electroretinogram). A 100% modulated checkerboard at retinal level may be considerably less modulated because of imperfect optics of the eye. So the pattern electroretinogram should be compared with homogeneous field stimulation of correspondingly lower modulation depth. On the basis of the optical transfer properties of the eye we compared by subtracting the proper focal electroretinogram from the pattern electroretinogram. The difference response was virtually zero for check sizes larger than 120'. For checks from 60' down the difference response was of the same order of magnitude as the adjusted focal recording. This difference response for eyes with normal optics is largest around 30'; its wave form was found to be rather invariant with check size.

The pattern electroretinogram

Physiological experiments and the exploitation of clinical conditions have provided compelling evidence that retinal ganglion cells and other inner retinal structures generate the pattern ERG (PERG). As an increasing number of clinical reports have been published some contradictory findings have been reported. These may be ascribed to variation in recording and measuring techniques. The PERG consists of two major portions, the early positive and the following negative component which can be investigated separately if the stimulus conditions allow isolated (or "transient") responses to be recorded. Care has to be taken in positioning the reference electrode, maintaining accurate refraction, and the influence of pupil size must be considered. Furthermore the PERG is contaminated by a luminance component which may be generated in the outer retina. The size of this increases with low spatial frequency (large check-sizes) and high mean luminance. The PERG permits the examination of an additional level of the retina and helps the understanding of pathophysiology of various eye diseases, and is of clinical importance in routine diagnosis and assessment. In glaucoma the PERG amplitude is often reduced before it is possible to detect a scotoma and it is therefore an important prognostic indicator in patients with ocular hypertension. In diabetic retinopathy, retinal ischaemia sufficient to lead to the pre-proliferative state can be demonstrated. The PERG also has a major clinical role in examining localised retinal pathology. If combined with VECP recording, it greatly extends the interpretations possible, since not only can damage to the optic nerve be detected by both tests, but the normal PERG in the presence of an abnormal PVECP implies that the losses are confined to the central pathway.

The point-spread function of the eye from 0 degrees to 100 degrees and the pattern electroretinogram

In studies on the pattern electroretinogram the quality of the retinal image is a major concern. The use of contact lens electrodes was rejected since a good pattern could not be recorded. This is believed to be due to blurring of the retinal image. As indicator of image quality the patient's visual acuity is often used. We wondered whether this is a sufficient criterion. The retinal image is the product of the whole optical point-spread function of the eye whereas visual acuity refers only to the central portion of this function. On the basis of existing reports it can be estimated that for the young normal eye the outer edges of this function (straylight) causes considerable loss of contrast. The strength of the straylight can be much greater in older eyes. We studied the relation between the point-spread function including straylight and the pattern electroretinogram in normal eyes and some pathological cases. The measurements proved to follow the calculated contrasts on the basis of a local luminance model, with the exception of enhancement (tuning) around 60' checksize for the young normal eye. Because of the considerable differences in straylight in an older population one has to take into account that loss of pattern electroretinogram can be suffered in patients with otherwise good visual acuity.

Adaptation, stimulus color, and the human electroretinogram

(1) The human electroretinogram was recorded using two procedures that were designed to change the ratio of photopic to scotopic activity. (2) With the first procedure, responses were recorded with patterns that alternated with an asymmetrical time period to produce two average responses, one of which was more adapted than the other. (3) The second employed symmetrical alternation, but red and blue stimulation were used. (4) The electroretinogram was larger in responses to patterns of low spatial frequency for all conditions. The evoked potential, which was recorded simultaneously, was larger at intermediate frequencies. (5) The electroretinogram must be mediated before the antagonistic receptive field organization. (6) Sensitivity of electroretinogram with blue stimuli was higher than those with red at low spatial frequency. The possible scotopic contribution was discussed.

Computation of the luminance and pattern components of the bar pattern electroretinogram

Pattern onset electroretinograms (PERGs) were recorded from four normal subjects. Square-wave gratings of 75% contrast were presented in three approximately contiguous, concentric zones of outer angular radius, 5.1 degrees, 12.6 degrees, and 23.6 degrees. The zones were calculated to give equal numbers of ganglion cell receptive fields. The recorded PERGs were considered to include luminance and pattern components which have low and bandpass spatial tuning functions respectively. These components combine in the PERG to produce a broad spatial tuning characteristic. The amplitude of PERGs in response to low spatial frequency stimuli is widely reported to be linearly related to contrast. The retinal illuminance response at every spatial frequency was computed from the eye's modulation transfer function. This function characterizes the reduction in contrast that occurs because of optical degradation. The computed retinal illuminance response was subtracted from the PERG waveform and a pattern-specific response was revealed. The latter had a highly tuned bandpass function which peaked at higher spatial frequencies than the PERG at corresponding peripheral angles.

Pharmacological dissociation of the b-wave and pattern electroretinogram

Electroretinographic responses to modulation of either luminance (focal ERG) or spatial contrast (pattern ERG) were recorded from the pigeon eye before and after intravitreal injection of glutamate analogues DL alpha amino adipic acid (DL alpha AA) and 2-amino-4 phosphonobutyric acid (APB). Both toxins reversibly abolished the b-wave. The pattern ERG was still present, however, when the b-wave had been abolished by the toxins. This result demonstrates that the b-wave and the pattern ERG can be pharmacologically dissociated and suggests the possibility that in pigeons the b-wave and pattern ERG reflect the activity of different generators.

Effect of pattern luminance profile on the pattern ERG in man and pigeon

Pattern electroretinograms (PERGs) have been recorded in man and pigeon using phase reversing patterns. Studies were made in both species of the effect of grating spatial frequency, grating profile (sine or square) or the size of the squares of a checkerboard on the amplitude of the responses. The results obtained in the two species were similar. In general, the ratio of the amplitudes of the sine wave:square wave responses is that predicted by the Fourier fundamentals of the pattern, and is thus not determined by changes in local luminance or contrast as it is normally defined. Checks however gave larger responses than sine or square wave profile bars matched to the Fourier fundamental on the diagonal. The results are not explicable in terms of any single stimulus parameter and suggest that the PERG generator has a degree of center/surround organisation similar to retinal ganglion cells. It is however argued that, because of the known physiological properties, generators cannot be unequivocally localised by varying the stimulus or comparing retinal, cortical and perceptual response dynamics.

Macular electroretinograms and contrast sensitivity as sensitive detectors of early maculopathy

Eighteen patients with early maculopathies of various etiologies were tested with pattern and focal electroretinograms (macular ERGs), with high (400 cd/M2) and moderate (40 cd/M2) stimulus intensities and a four-alternative forced choice (4AFC) contrast sensitivity test in addition to intensive clinical examinations. High spatial frequency contrast sensitivity loss on the 4AFC test was the most striking and consistent feature of all cases. The only eyes not outside normal contrast sensitivity limits were three in which diagnosis was uncertain and the patients had not recognized any problem, including two marginal solar burns. Maculopathy also substantially reduced macular ERG amplitudes. Criterion scores on these tests separated patients from normals more effectively than other noninvasive procedures and only missed one eye detected by contrast sensitivity. Latencies were affected but the delays were of no clinical significance in the individual case. Stimulus intensity was not critical. The results indicate that contrast sensitivity testing and macular ERGs are very reliable indices of central visual dysfunction at a stage when visible macular changes are too subtle for confident diagnosis. Contrast sensitivity has appeal because of its reliability, objectivity, simplicity, and noninvasive nature. It is equally applicable to children and adults. Pattern and focal ERGs can establish that the visual deficit has a retinal origin and can provide the most reliable objective confirmation.

Pattern ERG and glaucomatous visual field defects

In the past five years numerous reports have suggested that ganglion cell function can be tested by means of a specialized form of electroretinography, the so-called pattern electroretinogram (PERG). Because of the important potentials of a ganglion cell test for clinical use this technique has been applied by several investigators to patients with (presumed) ganglion cell dysfunction, especially glaucoma. On grounds of principle we had reason to question whether the reported positive results should be attributed to ganglion cell dysfunction or to other factors such as optical disturbances. We investigated in this study the PERG as a function of visual field loss in glaucoma patients with careful control of optical factors. We did not find changes in PERG as a function of field loss. So either field loss is not related to the mass behaviour of ganglion cells, or ganglion cells are not the prime basis of the PERG. We believe the latter to be true.

The use of a system analysis approach to electrodiagnostic (ERG and VEP) assessment

To evaluate the integrity of a biological system and its constituent functional units, a systematic study of input-output relations adopted from engineering has proven appropriate. With such an approach, sequential analysis can be implemented to probe the various parameter extractions along, for example, the visual system. The a priori assumption in this approach is that the visual world is processed along functionally separate pathways yielding distinct percepts such as contrast and motion. This so-called channel approach has proven useful not only to basic vision research but also for clinical application. The present overview shows that on the basis of the ERG or VEP, a type of functional anatomy can be performed with the biological system of interest remaining intact. Finally, it will be demonstrated that electrophysiological output parameters of the visual system can also serve as a non-invasive entry to investigate general systemic disorders.