August 2002
Volume 43, Issue 8
Free
Glaucoma  |   August 2002
Selective Loss of an Oscillatory Component from Temporal Retinal Multifocal ERG Responses in Glaucoma
Author Affiliations
  • Brad Fortune
    From the Discoveries in Sight, Devers Eye Institute, Portland, Oregon; and the
  • Marcus A. Bearse, Jr
    School of Optometry, University of California, Berkeley, California.
  • George A. Cioffi
    From the Discoveries in Sight, Devers Eye Institute, Portland, Oregon; and the
  • Chris A. Johnson
    From the Discoveries in Sight, Devers Eye Institute, Portland, Oregon; and the
Investigative Ophthalmology & Visual Science August 2002, Vol.43, 2638-2647. doi:https://doi.org/
  • Views
  • PDF
  • Share
  • Tools
    • Alerts
      ×
      This feature is available to authenticated users only.
      Sign In or Create an Account ×
    • Get Citation

      Brad Fortune, Marcus A. Bearse, George A. Cioffi, Chris A. Johnson; Selective Loss of an Oscillatory Component from Temporal Retinal Multifocal ERG Responses in Glaucoma. Invest. Ophthalmol. Vis. Sci. 2002;43(8):2638-2647. doi: https://doi.org/.

      Download citation file:


      © ARVO (1962-2015); The Authors (2016-present)

      ×
  • Supplements
Abstract

purpose. To evaluate electrophysiologic function in glaucoma by using a new stimulus designed to enhance ganglion cell and optic nerve head component (ONHC) contributions to multifocal electroretinogram (mfERG) responses.

methods. mfERGs of 16 individuals with glaucoma (POAG) and 18 normal control subjects were recorded and analyzed with a VER imaging system. The stimulus had three frames inserted between each m-sequence step: a full-field dark frame (1.0 cd/m2), a full-field flash (200 cd/m2), and another dark frame. Multifocal flashes were 100 cd/m2. The stimulus subtended approximately 40° total diameter and contained 103 scaled hexagonal elements. Signals were obtained using Burian-Allen bipolar electrodes, amplified ×106, band-pass filtered at 10 to 300 Hz, and sampled at 1200 Hz.

results. Local first-order responses (kernels) consisted of a direct component (DC) followed by an induced component (IC). Nasal–temporal response asymmetries in normal eyes were most easily observed in the IC. A small but distinct oscillation in the ICs of temporal retinal responses distinguished them from nasal IC waveforms. In individuals with glaucoma, there was less asymmetry between nasal and temporal responses, mostly because of the reduction of the oscillation in the temporal retinal ICs. The amplitude of this oscillation was 4.4 ± 2.1 nV/deg2 in the control group and 1.8 ± 1.2 nV/deg2 in the glaucoma group (P < 0.0001). Amplitude and latency measures of other response features were not significantly different from normal. Amplitude of the IC oscillation was not correlated with age in either the normal or glaucoma groups. In a group of normal subjects retested 3 months later, the average test–retest repeatability was ±12%.

conclusions. Selective loss of an oscillatory feature from IC responses in glaucoma may represent abnormalities in the inner plexiform layer of the temporal retina, where classic oscillatory potentials (OPs) are thought to arise. However, evidence suggests that this effect may also be due in part to loss of the ONHC.

Since the introduction of the multifocal electroretinogram (mfERG) technique, 1 2 3 it has become clear that pathologic changes within the outer retina can be detected with relatively high spatial resolution and sensitivity. 4 5 The clinical utility of the mfERG for evaluation of local outer retinal function has been validated by basic studies that show that mfERG responses are similar to scaled-down versions of photopic Ganzfeld ERGs. 5 6 The retinal mechanisms that generate the dominant features of each ERG type, namely cone photoreceptor and cone bipolar cell currents, are also similar. 5 7 8  
However, recent studies have provided strong evidence that mfERG responses also contain significant contributions generated by inner retinal mechanisms, including ganglion cell action potentials. 5 6 7 8 9 10 11 12 13 14 15 Response components originating with inner retinal activity may reflect both local events and events arising more proximally, such as at the optic nerve head. For example, the optic nerve head component (ONHC), first proposed by Sutter and Bearse, 13 is believed to represent an mfERG contribution produced by activity of retinal ganglion cell axons and generated in the vicinity of the optic nerve head. 13 14 In healthy eyes, this component would be delayed as a function of the distance that action potentials travel along unmyelinated nerve fibers, between the site of focal stimulation and the nerve head. 13 14  
The identification of mfERG components that may depend on normal electrophysiologic function of ganglion cells has given rise to interest in the mfERG as a potential tool for the study of glaucoma. 9 16 17 18 19 20 21 22 23 24 25 26 27 Glaucoma ultimately results in loss of vision because of damage to and death of ganglion cells (see Refs. 28 29 for review). It is often argued that the principal site of damage in glaucoma is within the anterior optic nerve, near the lamina cribrosa. 28 Yet, it is possible that pathologic mechanisms within the inner retina also contribute to altered vision function during the development and progression of the disease. 30 Thus, a relatively direct measure of local ganglion cell function, such as that which may be provided by the mfERG, would be beneficial to studies of human and experimental glaucoma. 
Initial studies have shown that mfERG responses are abnormal in both human glaucoma 9 16 17 18 19 20 21 22 23 24 25 26 27 and experimental glaucoma in nonhuman primates. 11 12 However, mfERG abnormalities, at least those determined by conventional measurements such as peak-to-trough amplitude, implicit time, and scalar-product amplitude density, generally do not correspond spatially with local sensitivity losses in standard psychophysical visual field (VF) tests. 16 19 22 23 24 One possible explanation is that loss of ganglion cells leads to loss of the ONHC, whose latency varies according to response location, while remaining local mfERG components maintain a more constant latency throughout the retina. Thus, the effect of a missing or diminished ONHC on the features of any local composite response is strongly dependent on its retinal locale. 13 16 Moreover, the luminance flicker stimulus common to most of the studies cited typically elicits only a relatively small ONHC in normal eyes. 27 31 Therefore, only subtle differences are revealed by the common mfERG flicker stimulus in eyes with early or moderately advanced glaucoma. 9 16 19 20 21 22 23  
Sutter et al. 31 32 and Shimada et al. 33 have developed an alternative mode of mfERG stimulation that is thought to elicit a relatively larger ONHC in normal eyes. 31 32 33 Hence, loss of the ONHC in glaucoma should become more apparent when such a stimulus is used. Initial studies of human 25 26 27 and experimental glaucoma 27 34 support this hypothesis. In this study, an mfERG stimulus thought to enhance ganglion cell signals and the ONHC, was used to test patients with glaucoma and aged-matched control subjects. The results have been presented previously in abstract form. 35  
Methods
Subjects
Sixteen patients with primary open-angle glaucoma (POAG) were recruited from the glaucoma service at the Devers Eye Institute (DEI). Mean age (±SD) was 56.6 ± 11.0 years. Each patient had previously performed 10 or more VF tests at the DEI (Humphrey 30-2 or 24-2 achromatic threshold standard automated perimetry, SAP; Humphrey Instruments, Dublin, CA) which demonstrated long-standing glaucomatous VF loss that was determined to be stable. Table 1 contains patient and VF information. As reflected by the pattern standard deviations (PSDs) relative to each patient’s mean defect (MD), most patients had relatively localized VF loss in the eye that was tested. The eye with the more localized VF defect was chosen for testing, or the test eye was chosen randomly if similar defects were present. All patients had controlled intraocular pressure (IOP) at the time of testing, and none had a history of any of the following conditions: exfoliation syndrome, pigment dispersion syndrome, corticosteroid use, iridocyclitis, ocular trauma, or other intraocular eye disease or other diseases known to affect the VFs (pituitary lesions, demyelinating diseases, HIV+ or AIDS, or diabetes). 
Eighteen healthy, visually normal volunteers with a mean age (±SD) of 54.0 ± 9.6 years served as the age-matched control group. Nine younger normal subjects (mean age, 31.6 ± 5.0 years) also participated in ancillary studies on the effects of normal aging and test–retest reliability, the results of which are reported herein. All normal subjects met the following inclusion criteria: good visual acuity (≥20/25), intraocular pressure of 22 mm Hg or less, symmetrical optic discs (asymmetry of vertical cup-to-disc ratio <0.2), normal standard achromatic increment threshold VFs (Humphrey 30-2 or 24-2 results within normal 95% confidence limits for both MD and PSD), and glaucoma hemifield test (GHT) results within normal limits. The exclusion criteria were history of ocular disease, surgery, or trauma, and systemic disease known to affect the visual system. 
All subjects were fully informed of the potential risks and benefits of the study and then provided voluntary written consent to participate. All procedures adhered to the tenets of the Declaration of Helsinki and were approved by the Legacy Health System’s institutional review board for the protection of human subjects. 
mfERG Stimulus and Data Acquisition
Multifocal ERGs were acquired using a VER imaging system (VERIS Science; EDI, San Mateo, CA). The stimulus was presented on a 21-in. monochrome monitor (Nortech, Plymouth, MN) with a 75-Hz refresh rate at a test distance of 40 cm (subtending ∼40° total visual angle, see Fig. 1A ). During each m-frame of the stimulus, the local luminance of each of the 103 individual hexagonal areas was determined to be either light (100 cd/m2) or dark (<1.5 cd/m2) by a pseudorandom binary m-sequence. 1 2 After each m-frame, the entire stimulus area was dark for one video frame, flashed brightly (200 cd/m2) during the next video frame, then returned to the dark luminance level for the duration of another frame. The sequence then repeated, beginning with another m-frame of pseudorandom local stimulation, a dark frame, a full-field flash frame, a dark frame, and so forth. The schematic representation in Figure 1B illustrates two complete m-sequence cycles, with full-field flashes interposed. This stimulus was chosen because it provides adequate separation of focal and full-field flashes while maintaining a reasonably short recording time and acceptable signal-to-noise characteristics. A similar mfERG stimulus paradigm was used by Shimada et al. 33 in a study of diabetes. 
The stimulus was viewed through natural pupils, primarily because these recordings were done as part of a larger study in which fine pattern-reversal stimulation was also used, thus requiring optimal clarity. At the time of testing, all subjects’ pupils measured between 2.5 and 4.0 mm. There was no significant difference in mean pupillary diameter between the glaucoma and control groups, and none of the patients were using miotic glaucoma medications. Burian-Allen bipolar contact lens electrodes (Hansen Ophthalmics, Iowa City, IA) were placed on the eye after topical anesthesia (0.5% proparacaine). Residual refractive error was corrected to the nearest 0.25 D. 
Each recording lasted approximately 8 minutes (m-sequence exponent = 13), which was broken into 16 shorter segments with brief rest periods between segments. Segments that were contaminated by two or more obvious voltage artifacts, due to blinks or eye movements, were discarded and rerecorded. Signals were amplified (gain, 106), band-pass filtered (10–300 Hz), sampled at 1200 Hz (i.e., sampling interval, 0.833 ms), and stored for subsequent off-line analyses. 
mfERG Data Analysis
Data were analyzed using with the system software (VERIS Science ver. 4.1; EDI). Before all subsequent analyses, all mfERG data were treated with one iteration of the system’s artifact-removal algorithm, which effectively eliminates voltage artifacts due to blinks and small eye movements (see Ref. 2 , Appendix B). One iteration of spatial smoothing was also used, so that each local ERG was averaged with one sixth of its six nearest neighbors. For some of the analyses, mfERG responses were pooled by retinal quadrant (see Fig. 1A for quadrant definition). 
Scalar-product amplitude densities are commonly used in clinical settings and are often reported in mfERG studies. The first part of the Results section presents the findings for scalar-product amplitudes. Mathematical details of the scalar-product amplitude density calculation can be found in Sutter and Tran (see Ref. 2 , Appendix C). In this study, scalar-product calculations for all local responses were based on 103 normal local templates, which were created by averaging the responses from the first 12 normal records. Our initial concern was that the template features might become inappropriately smooth if too many normal files were included in the average. On further inspection, we determined that this was not the case, because there was a negligible difference between these templates and a group comprising all normal files. Thus, the results of the scalar-product analyses would not be any different if the latter template group had been used. The response epochs were 5 to 50 ms (direct component [DC]) and 50 to 100 ms (induced component [IC]; described in the Results section; see Figs. 2 3 and 4 ). Narrower epochs were also evaluated but, in general, these resulted in unacceptably high intra- and intersubject variability. 
Results
Figure 2 shows a typical normal response to the mfERG stimulus used in this study. The trace at the top (Fig. 2A) is the average of the seven central macular responses from the arrays shown (Figs. 2B 2C) . This average response demonstrates the major features of the first-order kernel for one normal subject. The first-order kernel represents the mean difference between all retinal response epochs after local-light flashes and all response epochs after local dark stimulus events in the multifocal frame of the sequence. The initial part of the response is therefore dominated by retinal responses to local luminance flashes (100 cd/m2) and has been called the DC. 25 26 27 31 32 33 It is very similar in shape to dim, photopic, full-field flash ERG responses and most likely originates from similar generators—predominantly, hyperpolarizing and depolarizing cone bipolar cells—with additional contributions from cone photoreceptors and inner retinal neurons. 5 6 7 8  
The latter part of the first-order mfERG kernel recorded with this type of stimulus is the IC. 25 26 27 31 32 33 It represents the effect, at each location, of local luminance flicker stimulation on the response to subsequent full-screen bright flashes. It follows from the calculation of the first-order kernel, that unless the responses to the full-screen flashes were affected by prior stimulus-response history (i.e., local luminance flicker), there would be no IC. 27 31 32 33 36 That is, the IC represents the mean difference between successive bright, full-screen flash responses induced by the effect of prior local stimulation. In this regard, it represents adaptive effects acting within a train of stimulus responses. Such mechanisms are thought to increase in complexity as signals move from distal through proximal retinal layers—as gain control mechanisms increase in number and cumulative strength. Abnormalities of such mechanisms are believed to be a sensitive indication of early disease effects, perhaps specifically on the dynamics of inner retinal signal processing. Hence, in this study, we concentrated on the IC. 
The topographic distribution of the DCs and ICs in this normal subject are shown in Figures 2B and 2C , respectively. Of particular interest was the strong nasal–temporal asymmetry observed within the array of ICs in normal individuals. The most obvious feature of this asymmetry was the appearance of an oscillation at approximately 70 ms in the responses from the temporal retina (circled in Fig. 2C ). 
The scalar-product amplitude density distribution provided an overall representation of response topography for the major components. Figure 3 shows the response topographies for the DC (Fig. 3A) and the IC (Fig. 3B) for the normal subject shown in Figure 2 (left column) and for a patient with glaucoma (right column). In normal subjects, the amplitude density of the IC was generally larger, and the topographic profile steeper, than the DC. This may reflect differences in the topographic distribution of underlying retinal generators. 
The results of this particular patient represent the closest correspondence between the behavioral VF and the scalar-product density of all subjects in this study. The following analysis shows that there is generally poor correspondence between the commonly used scalar-product density parameter and VF sensitivity. 
The relationship between the overall response amplitude (scalar-product density) and behavioral sensitivity (SAP VF thresholds) was evaluated for all whole study participants by comparing the measurements averaged within each quadrant. Figure 4 shows the scalar-product amplitude density for the DC (Fig. 4A) and the IC (Fig. 4B) plotted against VF sensitivity in each patient. mfERG response amplitudes were averaged by spatial quadrant, as shown by the gray-shaded locations in Figure 1A . SAP-VF threshold values were also averaged by corresponding quadrant, but only the points from the 24-2 pattern were included, so that the field dimensions would be more similar, as well as consistent across patients (most patients, but not all, were tested with the 30-2 VF pattern). The normal range of mfERG amplitude density, for all quadrant averages, is shown by the box plot at the right of each graph. The normal ranges for all quadrants were similar; thus, the results for all four quadrants are shown pooled in a single distribution. Comparing the two distributions of normal values shows that the IC is approximately 65% larger than the DC, on average, under these stimulus conditions. 
Among the data of the patients with glaucoma, none of the four quadrants shows any correlation between mfERG amplitudes and SAP VF sensitivity, nor do the few ICs below the lower normal limit bear any obvious relationship to SAP VF sensitivity. Further, there are no significant differences between the glaucoma and normal group means of any of the individual quadrants or of the whole field combined. Thus, it is clear from the results shown in Fig. 4 that there is no spatial correspondence between SAP VF sensitivity and mfERG scalar-product amplitude of DCs or ICs, when averaged by quadrant. Furthermore, this overall response measurement does not reveal any significant difference between the glaucomatous and normal states. 
However, as mentioned earlier, a more subtle (localized) difference between the normal and the glaucomatous eye was observed in the IC. Reduction of an oscillation at approximately 70 ms in the responses from the temporal retina resulted in loss of nasal–temporal response asymmetry in patients with glaucoma. Figure 5 shows the IC portion (40–100 ms) of the responses from a concentric ring around the fovea. In the normal example (Fig. 5A) , there was obvious asymmetry when the responses near the blind spot (traces 12, 1, 2) were compared with those from the temporal retina (traces 6, 7, 8). The responses from the temporal retina contained an oscillation (marked with an asterisk) that was not readily visible in the responses from the nasal retina. The second major peak of these responses also changed systematically with distance from the blind spot (Fig. 5A , vertical hash marks). 
In contrast, the traces of the patient with glaucoma (Fig. 5B) show that the temporal retinal responses were much more similar to the nasal retinal responses, primarily because the temporal retinal oscillation was diminished. Although this patient’s behavioral VF showed a large difference in sensitivity between the superior and inferior hemifields, the mfERG responses did not show any significant differences between upper and lower locations at this eccentricity (∼6°). 
Among normal subjects, the oscillation in the IC was most apparent in the responses near the horizontal midline at approximately 5° to 15° in the temporal retina (see circled responses in Fig. 2C ). The trace at the top of Figure 6 shows the average of those seven local responses from the circled area of the array (Fig 2C) , with the major features labeled, in one normal subject. The box plots (Fig. 6B) show the peak-to-trough (P-to-T) amplitude distributions of the two major response features, and the oscillatory component, in the glaucoma and aged-matched normal groups for the response average of these seven locations in the temporal retina. The amplitude of the oscillatory feature was measured by the caliper method as shown—that is, between the voltage at the trough (Fig. 6A , n2) and the line spanning the two adjacent peaks (Fig. 6A , p2 and p3). There were no significant differences between the normal and glaucoma groups in any of the feature latencies, although on average, the oscillatory feature latency was slightly longer in the glaucoma group (Fig. 6A , n2 latency normal group: 70.9 ± 1.7 ms; glaucoma group: 72.2 ± 1.8 ms). There was no significant difference between the normal and glaucoma group means in either the amplitude of the DC or the amplitude of the main peak of the IC. However, at the IC oscillation, there was a significant difference between the group means (normal group mean ± SD, = 4.4 ± 2.1; glaucoma group, 1.8 ± 1.2 nV/deg2; P < 0.0001, two-tailed t-test). 
Figure 6C plots the amplitude of the oscillation versus the VF MD in each patient. Comparison with the global indicator MD is more appropriate than with local VF sensitivity for reasons to be discussed. Oscillation amplitude is not correlated with MD. The dashed line shows the optimal normal cutoff range derived from the analysis presented in the next section. 
Figure 7 shows the receiver operating characteristic (ROC) curve, which illustrates the sensitivity and specificity for discrimination of glaucoma from normal, based on various cutoff levels for the amplitude of the oscillatory component. The ROC curve shows that as sensitivity to detect glaucoma increased, the false-alarm rate increased, and thus, specificity declined. A measurement that perfectly discriminates between the normal and glaucomatous eye would plot along the ordinate to the top left corner, and along the top of the graph to the top right. That is, it would provide 100% specificity at all levels of sensitivity and vice versa. A test with performance equivalent to chance would plot along the diagonal (Fig. 7 , solid line from lower left to upper right corners). The area under the ROC curve quantifies the overall accuracy of the measurement. Perfect performance would result in an area under the ROC of 1.0, whereas chance performance would result in an area of 0.50. The area under this ROC curve is 0.88, which is significantly better than chance (P = 0.004). With the criterion value of 2.75 nV/deg2, sensitivity was 75%, with a specificity of 83% and 80% overall correct classification. 
The effect of age on the amplitude of the oscillation, as well as the test–retest reliability of this amplitude measurement, were evaluated in ancillary studies that included some younger normal subjects (see the Methods section). There was no significant correlation between amplitude and age in the age-matched normal group, the younger normal group, both nor mal groups combined, or the glaucoma group. Further, there was no significant difference in oscillation amplitude between the younger normal group and the older normal (control) group. The test–retest repeatability of the amplitude measurement was ±12%, on average, in 20 normal subjects who were retested within approximately 3 months of the original recording. 
The following analysis was performed to test the hypothesis that the origin of the oscillatory component and its topographic distribution in normal eyes (i.e., salient nasal–temporal asymmetry) may be due in part to dynamic phase relationships of underlying fundamental response components. A series of studies have proposed a model wherein two components, a retinal component (RC) and an ONHC, can account for much of the nasal–temporal asymmetry observed in mfERG responses. 13 14 27 31 32 In that model, the shape of the two components and the latency of the RC are constant in concentric rings around the foveal center. The latency of the ONHC, however, varies with distance from the optic nerve head. The method used to separate these two components has been described previously. 13 14 15  
Figure 8A shows the decomposition of the ICs, from responses around ring 2 in a normal subject, into an RC and ONHC. The original data are shown in the left-most column (solid traces). The sum of the model components is represented by the dashed traces overlying the original responses in the left column. The sum of the two model components fits the data well and accounts for most, if not all, of the nasal–temporal asymmetry, including variation in the timing of the second peak and the appearance of the oscillatory component in the temporal retinal responses. 
The RC and ONHC are shown overlaid in the right-most column of Figure 8A . For responses near the optic nerve head (traces 12, 1, 2), it can be seen how the primary troughs and peaks of the two fundamental components may combine to form one deep negativity followed by a peak without an oscillation. Whereas, in the temporal retina (traces 6, 7, 8), the ONHCs with longer latencies combine with the RCs to form a composite response with a relatively large oscillation. This type of cancellation and enhancement of induced oscillatory components has been observed previously in normal eyes, by using a sparse stimulation mfERG paradigm. 15  
It is possible that the reduced amplitude of the oscillation in the temporal retinal responses of patients with glaucoma was due to abnormalities of the ONHC, the RC, or both. Figure 8B shows the decomposition of the ICs from ring 2, into the RC and ONHC, in one of the patients with glaucoma. The original data are the solid traces in the leftmost column. Overlaid onto them is the sum of the two model components for each response location around ring 2 (dashed traces). The sum of the model components fit the data well. Loss of nasal–temporal asymmetry is apparent, because the responses around the ring are all very similar. The decomposition suggests that the ONHC is more reduced in the superior hemifield than in the inferior hemifield (Fig. 8 B , rightmost panel, locations 3, 4, 5), which in this case matches the pattern of sensitivity loss shown in the behavioral VF gray-scale plot. The RC also appeared to be reduced in some of these locations. Although the decomposition for this patient also suggests that the ONHC is not completely diminished, there is evidence that for two stimulus locations (Fig. 8C , rightmost panel, locations 1, 2), the decomposition of the mfERGs may have been incomplete. This will be discussed further later. 
Discussion
In the results of this study, there was an oscillation in mfERG responses from the temporal retina of normal subjects that was significantly reduced in subjects with POAG. Although the amplitude of this feature was shown to discriminate well between normal eyes and those in the early to moderate stages of glaucoma, its small size and complex origins warrant caution regarding any possible practical clinical utility. As discussed later in this section, this oscillatory feature may arise because of an interaction between local and nonlocal signal components. 
Although there are reports of conspicuous mfERG abnormalities in patients with glaucoma 18 and those with suspected glaucoma, 37 most studies, using the common luminance flicker stimulus at high or low contrast, have found only relatively subtle alterations. 9 16 17 18 19 20 21 22 23 It has been suggested that mfERG abnormalities in glaucoma are due, at least in part, to loss of the ONHC. 16 23 25 26 27 Luminance flicker mfERG responses in humans, however, contain a relatively small ONHC, which limits the usefulness of such stimuli for assessment of function in glaucoma. 25 26 27 31 Consequently, Sutter et al. 31 32 and Shimada et al. 33 developed a type of mfERG stimulus that enhances the relative size of the ONHC in normal subjects. 
The results of this study confirm their previous findings. First, control responses to this mfERG stimulus manifest a more salient nasal–temporal asymmetry than do control responses to common luminance flicker stimuli. 27 31 32 Second, the nasal–temporal asymmetry is attributable, largely, to the presence of a relatively large ONHC. 13 14 15 Third, it follows that differences between normal and glaucoma can be appreciated more readily using this type of global flash stimulus. 16 25 26 27 For instance, in a previous study in which high-contrast luminance flicker was used, response abnormalities were detected in this same group of patients with glaucoma, but only for the group as a whole. 23 That is, the difference between the glaucoma group average response and the control group average closely resembled the characteristics of the ONHC, but the differences in any given individual were probably too small to serve as reliable indicators of disease. 23 In contrast, the stimulus used in this study evoked responses with more nasal–temporal asymmetry, itself attributable to a greater contribution from the ONHC and thus allowed mfERG abnormalities to be more easily detected in this group of individuals with early to moderately advanced glaucoma. 
Specifically, the extra oscillation in the IC, found in responses from the temporal retina of control subjects, was the most salient feature of nasal–temporal asymmetry produced by this stimulus. The amplitude of this component provided good discrimination between individuals with glaucoma and control subjects. The area under the ROC curve was 0.88, which indicates that the performance of this measurement compares reasonably well with many other electrophysiologic and psychophysical measurements of function in patients with glaucoma with early to moderate VF damage (only a few of the patients in this study had advanced VF loss). For example, in one study which directly compared a large battery of functional measures in glaucoma, the areas under the ROC curves ranged from 0.556 to 0.898 and 0.522 to 0.898 in psychophysical and electrophysiologic tests, respectively. 38 Similarly, discriminability was on par with, or slightly better than, other mfERG studies that used other unique stimuli and/or methods, 19 20 21 although the nasal–temporal response asymmetry with this stimulus appears to be more salient. Yet the oscillation only appeared in the responses from the temporal retina, and its small absolute amplitude required some spatial pooling to achieve reliable signal-to-noise levels for analysis. As such, comparisons with local behavioral VF sensitivity would be somewhat compromised. However, the following discussion suggests that such comparisons are not likely to be meaningful in any event, because the origin of this feature, and the effects of glaucoma on it, may not be strictly associated with VF abnormalities. 
There are at least two interpretations that explain the origin of this oscillatory feature and its asymmetric distribution in the normal retina. Prior studies have found that the topographic distribution of oscillatory potentials (OPs) exhibits strong nasal–temporal asymmetry, such that OPs in the temporal retina are larger than those in the nasal retina. 39 40 Oscillatory components of the mfERG also have a strong nasal–temporal asymmetry. 13 15 34 41 42 OPs of full-field flash ERGs are thought to be generated within the inner plexiform layer (IPL), among the synaptic connections between bipolar, amacrine, and ganglion cells, and perhaps represent activity of inhibitory feedback circuits. 43 A recent report has shown that one of the earliest manifestations of experimental glaucoma in monkeys is retraction and swelling of ganglion cell (GC) dendrites in the IPL. 30 Complete loss of GCs within the IPL, as might occur in advanced glaucoma, would obviously have significant impact on the behavior of these circuits. Indeed, abnormalities of retinal OPs in advanced glaucoma have been demonstrated previously with full-field flash stimulation. 44 However, swelling and retraction of GC dendrites in early glaucoma may also be expected to result in abnormal properties of IPL electrophysiologic responses. Thus, it is possible that loss of the temporal retinal oscillatory component observed in this study represents abnormal activity of local IPL circuitry in the temporal retina. 
Alternatively, the appearance of the oscillation in normal temporal retinal response ICs can be explained by the presence of a relatively large ONHC, predictably delayed in the temporal retina, combined with the relatively latency-invariant retinal component (Fig 8A) . This hypothesis was initially driven by the observation that latency shifts of the second major peak of the IC were found to be proportional to distance from the optic nerve head (e.g., Fig 5A ). This behavior is a fundamental characteristic of the ONHC. 13 14 15 Further support for this hypothesis comes from another study that demonstrated that nasal–temporal asymmetries in the distribution of mfERG OPs can also be explained on the basis of latency shifts of the ONHC. 13 The consequences of changing the phase relationship between high-frequency oscillations present in RCs and ONHCs, led to enhancement of OPs in the composite responses of the temporal retina, but caused partial cancellation in the nasal retina. Thus, the reduction of the temporal retinal response oscillation observed in this study is more likely to be due to loss of the ONHC in glaucoma than to selective abnormalities in the temporal retina per se. 
In this light, it is interesting that the amplitude of the temporal retinal response oscillation was below the criterion value of 2.75 nV/deg2 in four of the six patients with glaucoma with a VF MD of −5 dB or less (Fig 6C ; five of these six did not trigger the P < 0.05 flag for VF MD, see Table 1 ). Yet, the correlation between the amplitude of this oscillatory feature in the temporal retina and VF sensitivity (either globally to MD, or locally to nasal VF thresholds) was not statistically significant. This suggests that electrophysiologic abnormalities such as loss of the ONHC, may begin at an early stage of glaucoma in many patients. 
This heterogeneity might be expected, given the complexity of pathophysiologic mechanisms thought to be involved in glaucoma and their relationship to the small signal measured here. For example, the ONHC is thought to be generated as ganglion cell action potentials arrive at or near the optic nerve head, where GC axons become myelinated, and/or abruptly change orientation as they join the optic nerve. 13 14 15 25 26 27 31 32 Because glial cell characteristics are also known to change at the optic nerve head, 43 mechanisms dependent on normal glial cell support, such as spatial buffering of ion concentration, may also contribute to generation of the ONHC. Thus, although the appearance of a normal ONHC in mfERG recordings may necessarily require output of spikes by GCs in the retina, it may also depend on normal physiologic function and/or anatomic integrity of the optic nerve head. Pathophysiologic changes within the optic nerve 45 46 47 may occur at a different rate than changes in the retina, including, glial cell activation, 48 ganglion cell dysfunction, and death (for example, Refs. 28 29 30 ) Moreover, the relative course of these events may vary among different patients. 
In summary, a small but reliable mfERG response feature was found to be selectively reduced in patients with glaucoma. The origin of this feature most likely depends on an interaction between local retinal signals mediating fast-light–adaptation mechanisms and a more proximal signal from the optic nerve head. The abnormalities seen in glaucoma probably reflect a combination of factors, including loss of normal GC function in the retina and pathophysiologic changes within the optic nerve head. Although the mfERG decomposition algorithm may provide a method for elucidating the relative contribution of these factors, currently, accurate measurement of local ONHC amplitudes may be hindered by incomplete isolation from the RC, especially in cases in which the ONHC waveform is locally altered by disease. Given that local RCs may also contain GC contributions originating from the cell bodies, the RC may itself be altered by glaucoma. 27 The decomposition algorithm of the mfERG may be enhanced in the future to account for this, improving the sensitivity of these component isolation techniques. Of course, the goal yet to be realized is the ability to detect significant physiologic abnormalities at a stage before GC death and permanent vision loss. 
 
Table 1.
 
Patient and Visual Field Information
Table 1.
 
Patient and Visual Field Information
Patient Age Eye VA C/D MD PSD Fovea
1 60 OD 20/25 0.90 −12.2 13.5 25
2 62 OS 20/20 0.60 −2.7* 8.2 36
3 40 OS 20/15 0.70 −2.0* 2.0* 37
4 60 OD 20/20 0.90 −14.3 16.0 36
5 52 OS 20/15 0.80 −2.7* 3.0* 35
6 66 OD 20/20 0.80 −9.6 16.0 38
7 57 OS 20/20 0.80 0.4* 5.5 36
8 37 OS 20/20 0.90 −17.4 13.2 28
9 75 OS 20/20 0.90 −4.7 9.0 33
10 62 OD 20/25 0.70 −9.6 14.3 35
11 49 OS 20/15 0.90 −8.0 15.5 37
12 72 OS 20/50 0.90 −18.4 14.0 21
13 52 OS 20/20 0.90 −5.4 5.1 36
14 57 OD 20/20 0.95 −22.7 18.7 30
15 57 OS 20/25 0.80 −1.8* 2.8* 33
16 39 OD 20/25 0.95 −18.0 12.1 30
Mean 56.1 0.84 −9.3 10.6 32.9
SD 11.0 0.10 7.2 5.5 4.8
Figure 1.
 
Schematic representation of multifocal stimulus geometry (A). Labeled rings: eccentricity (degrees radius). Gray oval: approximate projection of the optic nerve (i.e., blind spot). Solid gray hexagons: locations included within each quadrant average (supero-nasal, in this case). (B) Stimulus sequence. Frame transitions occur at 13.3-ms intervals (75 Hz).
Figure 1.
 
Schematic representation of multifocal stimulus geometry (A). Labeled rings: eccentricity (degrees radius). Gray oval: approximate projection of the optic nerve (i.e., blind spot). Solid gray hexagons: locations included within each quadrant average (supero-nasal, in this case). (B) Stimulus sequence. Frame transitions occur at 13.3-ms intervals (75 Hz).
Figure 2.
 
Normal mfERG response features for this stimulus. Average of macular responses (central response and first two concentric rings) in one control subject, with two components labeled (A). Response arrays for DC (B) and IC (C) in this normal subject are shown separately for clarity. Circled responses in the temporal retina (C) highlight the oscillatory component. Gray oval: blind spot.
Figure 2.
 
Normal mfERG response features for this stimulus. Average of macular responses (central response and first two concentric rings) in one control subject, with two components labeled (A). Response arrays for DC (B) and IC (C) in this normal subject are shown separately for clarity. Circled responses in the temporal retina (C) highlight the oscillatory component. Gray oval: blind spot.
Figure 3.
 
Topographic distribution of scalar-product amplitude densities for the DC (A) and IC (B) in a normal subject (left) and a patient with glaucoma (patient 4, right). The perimeter-generated VF gray-scale plot (Humphrey, San Leandro, CA) of patient 4 is shown at lower right. The VF and mfERG spatial dimensions are not equivalently scaled and should not be compared directly.
Figure 3.
 
Topographic distribution of scalar-product amplitude densities for the DC (A) and IC (B) in a normal subject (left) and a patient with glaucoma (patient 4, right). The perimeter-generated VF gray-scale plot (Humphrey, San Leandro, CA) of patient 4 is shown at lower right. The VF and mfERG spatial dimensions are not equivalently scaled and should not be compared directly.
Figure 4.
 
Scalar-product amplitude densities of the DC (A) and IC (B) averaged by quadrant and plotted versus the average VF sensitivity for the corresponding quadrant. Each symbol refers to a different quadrant (A, inset). Box plots: normal range (N) of all four quadrant averages in one distribution (middle hash: median; upper and lower box edges: 75th and 25th percentiles; upper and lower whiskers: 97.5th and 2.5th percentiles, respectively).
Figure 4.
 
Scalar-product amplitude densities of the DC (A) and IC (B) averaged by quadrant and plotted versus the average VF sensitivity for the corresponding quadrant. Each symbol refers to a different quadrant (A, inset). Box plots: normal range (N) of all four quadrant averages in one distribution (middle hash: median; upper and lower box edges: 75th and 25th percentiles; upper and lower whiskers: 97.5th and 2.5th percentiles, respectively).
Figure 5.
 
Normal nasal–temporal asymmetry of ICs (A). Traces show normal IC responses (IC, 40–105-ms epoch) from locations within the second concentric ring around the center (inset). Example for glaucoma patient 10 (B). The perimeter-generated VF gray-scale plot for this patient is shown at the bottom of the inset. VF and mfERG spatial dimensions are not equivalently scaled and should not be compared directly.
Figure 5.
 
Normal nasal–temporal asymmetry of ICs (A). Traces show normal IC responses (IC, 40–105-ms epoch) from locations within the second concentric ring around the center (inset). Example for glaucoma patient 10 (B). The perimeter-generated VF gray-scale plot for this patient is shown at the bottom of the inset. VF and mfERG spatial dimensions are not equivalently scaled and should not be compared directly.
Figure 6.
 
Normal response average from perimacular area of the temporal retina (A) in the normal subject in Figures 2 and 5A . The locations averaged are shown with gray shading in the inset and the corresponding responses are circled in Figure 2C . Major peaks of the DC and IC are labeled by polarity and order of appearance. (B) Trough-to-peak amplitude distributions for temporal retinal response average in patients with glaucoma (G) and normal (N) subjects. Box plots: percentiles as described in Figure 4 . *P < 0.0001. (C) Plot of signal amplitude versus VF mean defect.
Figure 6.
 
Normal response average from perimacular area of the temporal retina (A) in the normal subject in Figures 2 and 5A . The locations averaged are shown with gray shading in the inset and the corresponding responses are circled in Figure 2C . Major peaks of the DC and IC are labeled by polarity and order of appearance. (B) Trough-to-peak amplitude distributions for temporal retinal response average in patients with glaucoma (G) and normal (N) subjects. Box plots: percentiles as described in Figure 4 . *P < 0.0001. (C) Plot of signal amplitude versus VF mean defect.
Figure 7.
 
Receiver operating characteristic (ROC) plot based on amplitude of oscillatory component.
Figure 7.
 
Receiver operating characteristic (ROC) plot based on amplitude of oscillatory component.
Figure 8.
 
Example of decomposition of ring 2 responses (IC epoch only) for the normal subject (A) shown in Figure 5 . First (left) column: original responses (solid traces) and the sum of the two model components (dashed traces); second column: RCs; third column: ONHCs; fourth column: RCs and ONHCs superimposed. (B) Decomposition for glaucoma subject 4 . Description of the traces is as in (A). The perimeter-generated VF gray-scale plot for the patient is shown at right.
Figure 8.
 
Example of decomposition of ring 2 responses (IC epoch only) for the normal subject (A) shown in Figure 5 . First (left) column: original responses (solid traces) and the sum of the two model components (dashed traces); second column: RCs; third column: ONHCs; fourth column: RCs and ONHCs superimposed. (B) Decomposition for glaucoma subject 4 . Description of the traces is as in (A). The perimeter-generated VF gray-scale plot for the patient is shown at right.
The authors thank Erich E. Sutter, Smith-Kettlewell Eye Research Institute, San Francisco, California, for providing assistance with the extraction of the ONHCs. 
Sutter EE. The fast m-transform: fast computation of cross-correlations with binary m-sequences. Society of Industrial and Applied Mathematics Journal on Computing. 1991;20:686–694.
Sutter EE, Tran D. The field topography of ERG components in man. I: the photopic luminance response. Vision Res. 1992;32:433–446. [CrossRef] [PubMed]
Bearse MA, Sutter EE. Imaging localized retinal dysfunction with the multifocal electroretinogram. J Opt Soc Am A. 1996;13:634–640. [CrossRef]
Kretschmann U, Bock M, Gockeln R, Zrenner E. Clinical applications of multifocal electroretinography. Doc Ophthalmol. 2000;100:99–113. [CrossRef] [PubMed]
Hood DC. Assessing retinal function with the multifocal technique. Prog Retinal Eye Res. 2000;19:607–646. [CrossRef]
Hood DC, Seiple W, Holopigian K, Greenstein V. A comparison of the components of the multifocal and full-field ERGs. Vis Neurosci. 1997;14:533–544. [CrossRef] [PubMed]
Hare WA, Ton H. Effects of APB, PDA, and TTX on ERG responses recorded using both multifocal and conventional methods in monkey. Doc Ophthalmol. .In press
Horiguchi M, Suzuki S, Kondo M, Tanikawa A, Miyake Y. Effect of glutamate analogues and inhibitory neurotransmitters on the electroretinograms elicited by random sequence stimuli in rabbits. Invest Ophthalmol Vis Sci. 1998;39:2171–2176. [PubMed]
Hood DC, Greenstein V, Frishman L, et al. Identifying inner retinal contributions to the human multifocal ERG. Vision Res. 1999;39:2285–2291. [CrossRef] [PubMed]
Hood DC, Frishman LJ, Viswanathan S, Robson JG, Ahmed J. Evidence for a ganglion cell contribution to the primate electroretinogram (ERG): effects of TTX on the multifocal ERG in macaque. Vis Neurosci. 1999;16:411–416. [PubMed]
Frishman LJ, Saszik S, Harwerth RS, et al. Effects of experimental glaucoma in macaques on the multifocal ERG: multifocal ERG in laser-induced glaucoma. Doc Ophthalmol. 2000;100:231–251. [CrossRef] [PubMed]
Hare WA, Ton H, Ruiz G, et al. Characterization of retinal injury using ERG measures obtained with both conventional and multifocal methods in chronic ocular hypertensive primates. Invest Ophthalmol Vis Sci. 2001;42:127–136. [PubMed]
Sutter EE, Bearse MA. The optic nerve head component of the human ERG. Vision Res. 1999;39:419–436. [CrossRef] [PubMed]
Hood DC, Bearse MA, Sutter EE, Viswanathan S, Frishman LJ. The optic nerve head component of the monkey’s (Macaca mulatta) multifocal electroretinogram (mERG). Vision Res. 2001;41:2029–2041. [CrossRef] [PubMed]
Bearse MA, Shimada Y, Sutter EE. Distribution of oscillatory components in the central retina. Doc Ophthalmol. 2000;100:185–205. [CrossRef] [PubMed]
Bearse MA, Sutter EE, Sim D, Stamper R. Glaucomatous dysfunction revealed in higher order components of the electroretinogram. Vision Science and Its Applications, OSA Technical Digest Series. 1996;1:104–107. Optical Society of America Washington, DC.
Vaegan , Buckland L. The spatial distribution of ERG losses across the posterior pole of glaucomatous eyes in multifocal recordings. Aust NZ J Ophthalmol. 1995;24(suppl 2)28–31.
Chan HL, Brown B. Multifocal ERG changes in glaucoma. Ophthalmic Physiol Opt. 1999;19:306–316. [CrossRef] [PubMed]
Hood DC, Greenstein VC, Holopigian K, et al. An attempt to detect glaucomatous damage to the inner retina with the multifocal ERG. Invest Ophthalmol Vis Sci. 2000;41:1570–1579. [PubMed]
Hasegawa S, Takagi M, Usui T, Takada R, Abe H. Waveform changes of the first-order multifocal electroretinogram in patients with glaucoma. Invest Ophthalmol Vis Sci. 2000;41:1597–1603. [PubMed]
Palmowski AM, Allgayer R, Heinemann-Vernaleken B. The multifocal ERG in open angle glaucoma: a comparison of high and low contrast recordings in high- and low-tension open angle glaucoma. Doc Ophthalmol. 2000;101:35–49. [CrossRef] [PubMed]
Fortune B, Cioffi GA, Johnson CA, Kondo Y, Mochizuki K, Kitazwa Y. The relationship between multifocal electroretinogram and standard automated perimetry findings in normal tension glaucoma. Weinreb RN Krieglstein GK Kitazawa Y eds. Glaucoma in the 21st Century. 2000;73–78. Harcourt Publishers London.
Fortune B, Johnson CA, Cioffi GA. The topographic relationship between multifocal electroretinographic (MERG) and behavioral perimetric measures of function in glaucoma. Optom Vis Sci. 2001;78:206–214. [PubMed]
Klistorner AI, Graham SL, Martins A. Multifocal pattern electroretinogram does not demonstrate localised field defects in glaucoma. Doc Ophthalmol. 2000;100:155–165. [CrossRef] [PubMed]
Bearse MA, Jr, Stamper RL, Sutter EE. Detection of functional abnormalities in glaucoma using a new multifocal ERG paradigm [ARVO Abstract]. Invest Ophthalmol Vis Sci. 2000;41(4)S103.Abstract nr 536
Bearse MA, Sutter EE, Stamper RL. Detection of glaucomatous dysfunction using a global flash multifocal electroretinogram (mERG) paradigm. OSA Technical Digest Series. 2001;14–17. Optical Society of America Washington, DC. Vision Science and Its Applications. Vol. 1
Sutter EE, Bearse MA, Stamper RL, et al. Monitoring retinal ganglion cell function with the mERG: recent advances. OSA Technical Digest Series. 2001;10–13. Optical Society of America Washington, DC. Vision Science and Its Applications. Vol. 1
Quigley HA. Neuronal death in glaucoma. Prog Retinal Eye Res. 1999;18:39–57. [CrossRef]
Osborne NN, Wood JP, Chidlow G, Bae JH, Melena J, Nash MS. Ganglion cell death in glaucoma: what do we really know?. Br J Ophthalmol. 1999;83:980–986. [CrossRef] [PubMed]
Weber AJ, Kaufman PL, Hubbard WC. Morphology of single ganglion cells in the glaucomatous primate retina. Invest Ophthalmol Vis Sci. 1998;39:2304–2320. [PubMed]
Sutter EE, Shimada Y, Li Y, Bearse MA. Mapping inner retinal function through enhancement of adaptive components in the M-ERG. OSA Technical Digest Series. 1999;1:52–55. Optical Society of America Washington, DC. Vision Science and Its Applications. Vol. 1
Sutter EE, Bearse MA, Shimada Y, Li Y. A multifocal ERG protocol for testing retinal ganglion cell function [ARVO Abstract]. Invest Ophthalmol Vis Sci. 1999;40(4)S15.Abstract nr 79
Shimada Y, Li Y, Bearse MA, Sutter EE, Fung W. Assessment of early retinal changes in diabetes using a new multifocal ERG protocol. Br J Ophthalmol. 2001;85:414–419. [CrossRef] [PubMed]
Fortune B, Cull G, Wang L, Van Buskirk EM, Cioffi GA. Factors affecting the use of multifocal electroretinography to monitor function in a primate model of glaucoma. Doc Ophthalmol. .In press
Fortune B, Cioffi GA, Johnson CA. Glaucoma causes selective reduction of an oscillatory component in multifocal ERG responses from the temporal retina [ARVO Abstract]. Invest Ophthalmol Vis Sci. 2001;42(4)S147.Abstract nr 783
Sutter EE, Bearse MA. The retinal topography of local and lateral gain control mechanisms. OSA Technical Digest Series. 1998;1:20–23. Optical Society of America Washington, DC. Vision Science and Its Applications. Vol. 1
Chan HH, Brown B. Pilot study of the multifocal electroretinogram in ocular hypertension. Br J Ophthalmol. 2000;84:1147–1153. [CrossRef] [PubMed]
Graham SL, Drance SM, Chauhan BC, et al. Comparison of psychophysical and electrophysiological testing in early glaucoma. Invest Ophthalmol Vis Sci. 1996;37:2651–2662. [PubMed]
Miyake Y. Macular oscillatory potentials in humans: macular OPs. Doc Ophthalmol. 1990;75:111–124. [CrossRef] [PubMed]
Miyake Y, Shiroyama N, Horiguchi M, Ota I. Asymmetry of focal ERG in human macular region. Invest Ophthalmol Vis Sci. 1989;30:1743–1749. [PubMed]
Hood DC, Frishman LJ, Robson JG, et al. A frequency analysis of the regional variation in the contribution from action potentials to the primate multifocal ERG. OSA Technical Digest Series. 1999;56–59. Optical Society of America Washington, DC. Vision Science and Its Applications. Vol. 1
Wu S, Sutter EE. A topographic study of oscillatory potentials in man. Vis Neurosci. 1995;12:1013–1025. [CrossRef] [PubMed]
Wachtmeister L. Oscillatory potentials in the retina: what do they reveal. Prog Retinal Eye Res. 1998;17:485–521. [CrossRef]
Gur M, Zeevi YY, Bielik M, Neumann E. Changes in the oscillatory potentials of the electroretinogram in glaucoma. Curr Eye Res. 1987;6:457–466. [CrossRef] [PubMed]
Hernandez MR. The optic nerve head in glaucoma: role of astrocytes in tissue remodeling. Prog Retinal Eye Res. 2000;19:297–321. [CrossRef]
Neufeld AH. Microglia in the optic nerve head and the region of parapapillary chorioretinal atrophy in glaucoma. Arch Ophthalmol. 1999;117:1050–1056. [CrossRef] [PubMed]
Yuan L, Neufeld AH. Tumor necrosis factor-alpha: a potentially neurodestructive cytokine produced by glia in the human glaucomatous optic nerve head. Glia. 2000;32:42–50. [CrossRef] [PubMed]
Wang L, Cioffi GA, Cull G, Dong J, Fortune B. Immunohistologic evidence of retinal glial cell changes in human glaucoma. Invest Ophthalmol Vis Sci. 2002;43:1088–1094. [PubMed]
Figure 1.
 
Schematic representation of multifocal stimulus geometry (A). Labeled rings: eccentricity (degrees radius). Gray oval: approximate projection of the optic nerve (i.e., blind spot). Solid gray hexagons: locations included within each quadrant average (supero-nasal, in this case). (B) Stimulus sequence. Frame transitions occur at 13.3-ms intervals (75 Hz).
Figure 1.
 
Schematic representation of multifocal stimulus geometry (A). Labeled rings: eccentricity (degrees radius). Gray oval: approximate projection of the optic nerve (i.e., blind spot). Solid gray hexagons: locations included within each quadrant average (supero-nasal, in this case). (B) Stimulus sequence. Frame transitions occur at 13.3-ms intervals (75 Hz).
Figure 2.
 
Normal mfERG response features for this stimulus. Average of macular responses (central response and first two concentric rings) in one control subject, with two components labeled (A). Response arrays for DC (B) and IC (C) in this normal subject are shown separately for clarity. Circled responses in the temporal retina (C) highlight the oscillatory component. Gray oval: blind spot.
Figure 2.
 
Normal mfERG response features for this stimulus. Average of macular responses (central response and first two concentric rings) in one control subject, with two components labeled (A). Response arrays for DC (B) and IC (C) in this normal subject are shown separately for clarity. Circled responses in the temporal retina (C) highlight the oscillatory component. Gray oval: blind spot.
Figure 3.
 
Topographic distribution of scalar-product amplitude densities for the DC (A) and IC (B) in a normal subject (left) and a patient with glaucoma (patient 4, right). The perimeter-generated VF gray-scale plot (Humphrey, San Leandro, CA) of patient 4 is shown at lower right. The VF and mfERG spatial dimensions are not equivalently scaled and should not be compared directly.
Figure 3.
 
Topographic distribution of scalar-product amplitude densities for the DC (A) and IC (B) in a normal subject (left) and a patient with glaucoma (patient 4, right). The perimeter-generated VF gray-scale plot (Humphrey, San Leandro, CA) of patient 4 is shown at lower right. The VF and mfERG spatial dimensions are not equivalently scaled and should not be compared directly.
Figure 4.
 
Scalar-product amplitude densities of the DC (A) and IC (B) averaged by quadrant and plotted versus the average VF sensitivity for the corresponding quadrant. Each symbol refers to a different quadrant (A, inset). Box plots: normal range (N) of all four quadrant averages in one distribution (middle hash: median; upper and lower box edges: 75th and 25th percentiles; upper and lower whiskers: 97.5th and 2.5th percentiles, respectively).
Figure 4.
 
Scalar-product amplitude densities of the DC (A) and IC (B) averaged by quadrant and plotted versus the average VF sensitivity for the corresponding quadrant. Each symbol refers to a different quadrant (A, inset). Box plots: normal range (N) of all four quadrant averages in one distribution (middle hash: median; upper and lower box edges: 75th and 25th percentiles; upper and lower whiskers: 97.5th and 2.5th percentiles, respectively).
Figure 5.
 
Normal nasal–temporal asymmetry of ICs (A). Traces show normal IC responses (IC, 40–105-ms epoch) from locations within the second concentric ring around the center (inset). Example for glaucoma patient 10 (B). The perimeter-generated VF gray-scale plot for this patient is shown at the bottom of the inset. VF and mfERG spatial dimensions are not equivalently scaled and should not be compared directly.
Figure 5.
 
Normal nasal–temporal asymmetry of ICs (A). Traces show normal IC responses (IC, 40–105-ms epoch) from locations within the second concentric ring around the center (inset). Example for glaucoma patient 10 (B). The perimeter-generated VF gray-scale plot for this patient is shown at the bottom of the inset. VF and mfERG spatial dimensions are not equivalently scaled and should not be compared directly.
Figure 6.
 
Normal response average from perimacular area of the temporal retina (A) in the normal subject in Figures 2 and 5A . The locations averaged are shown with gray shading in the inset and the corresponding responses are circled in Figure 2C . Major peaks of the DC and IC are labeled by polarity and order of appearance. (B) Trough-to-peak amplitude distributions for temporal retinal response average in patients with glaucoma (G) and normal (N) subjects. Box plots: percentiles as described in Figure 4 . *P < 0.0001. (C) Plot of signal amplitude versus VF mean defect.
Figure 6.
 
Normal response average from perimacular area of the temporal retina (A) in the normal subject in Figures 2 and 5A . The locations averaged are shown with gray shading in the inset and the corresponding responses are circled in Figure 2C . Major peaks of the DC and IC are labeled by polarity and order of appearance. (B) Trough-to-peak amplitude distributions for temporal retinal response average in patients with glaucoma (G) and normal (N) subjects. Box plots: percentiles as described in Figure 4 . *P < 0.0001. (C) Plot of signal amplitude versus VF mean defect.
Figure 7.
 
Receiver operating characteristic (ROC) plot based on amplitude of oscillatory component.
Figure 7.
 
Receiver operating characteristic (ROC) plot based on amplitude of oscillatory component.
Figure 8.
 
Example of decomposition of ring 2 responses (IC epoch only) for the normal subject (A) shown in Figure 5 . First (left) column: original responses (solid traces) and the sum of the two model components (dashed traces); second column: RCs; third column: ONHCs; fourth column: RCs and ONHCs superimposed. (B) Decomposition for glaucoma subject 4 . Description of the traces is as in (A). The perimeter-generated VF gray-scale plot for the patient is shown at right.
Figure 8.
 
Example of decomposition of ring 2 responses (IC epoch only) for the normal subject (A) shown in Figure 5 . First (left) column: original responses (solid traces) and the sum of the two model components (dashed traces); second column: RCs; third column: ONHCs; fourth column: RCs and ONHCs superimposed. (B) Decomposition for glaucoma subject 4 . Description of the traces is as in (A). The perimeter-generated VF gray-scale plot for the patient is shown at right.
Table 1.
 
Patient and Visual Field Information
Table 1.
 
Patient and Visual Field Information
Patient Age Eye VA C/D MD PSD Fovea
1 60 OD 20/25 0.90 −12.2 13.5 25
2 62 OS 20/20 0.60 −2.7* 8.2 36
3 40 OS 20/15 0.70 −2.0* 2.0* 37
4 60 OD 20/20 0.90 −14.3 16.0 36
5 52 OS 20/15 0.80 −2.7* 3.0* 35
6 66 OD 20/20 0.80 −9.6 16.0 38
7 57 OS 20/20 0.80 0.4* 5.5 36
8 37 OS 20/20 0.90 −17.4 13.2 28
9 75 OS 20/20 0.90 −4.7 9.0 33
10 62 OD 20/25 0.70 −9.6 14.3 35
11 49 OS 20/15 0.90 −8.0 15.5 37
12 72 OS 20/50 0.90 −18.4 14.0 21
13 52 OS 20/20 0.90 −5.4 5.1 36
14 57 OD 20/20 0.95 −22.7 18.7 30
15 57 OS 20/25 0.80 −1.8* 2.8* 33
16 39 OD 20/25 0.95 −18.0 12.1 30
Mean 56.1 0.84 −9.3 10.6 32.9
SD 11.0 0.10 7.2 5.5 4.8
×
×

This PDF is available to Subscribers Only

Sign in or purchase a subscription to access this content. ×

You must be signed into an individual account to use this feature.

×