October 2003
Volume 44, Issue 10
Free
Visual Neuroscience  |   October 2003
Local Ganglion Cell Contributions to the Macaque Electroretinogram Revealed by Experimental Nerve Fiber Layer Bundle Defect
Author Affiliations
  • Brad Fortune
    From the Discoveries in Sight, Devers Eye Institute, Portland, Oregon.
  • Lin Wang
    From the Discoveries in Sight, Devers Eye Institute, Portland, Oregon.
  • Bang V. Bui
    From the Discoveries in Sight, Devers Eye Institute, Portland, Oregon.
  • Grant Cull
    From the Discoveries in Sight, Devers Eye Institute, Portland, Oregon.
  • Jin Dong
    From the Discoveries in Sight, Devers Eye Institute, Portland, Oregon.
  • George A. Cioffi
    From the Discoveries in Sight, Devers Eye Institute, Portland, Oregon.
Investigative Ophthalmology & Visual Science October 2003, Vol.44, 4567-4579. doi:10.1167/iovs.03-0200
  • Views
  • PDF
  • Share
  • Tools
    • Alerts
      ×
      This feature is available to Subscribers Only
      Sign In or Create an Account ×
    • Get Citation

      Brad Fortune, Lin Wang, Bang V. Bui, Grant Cull, Jin Dong, George A. Cioffi; Local Ganglion Cell Contributions to the Macaque Electroretinogram Revealed by Experimental Nerve Fiber Layer Bundle Defect. Invest. Ophthalmol. Vis. Sci. 2003;44(10):4567-4579. doi: 10.1167/iovs.03-0200.

      Download citation file:


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

      ×
  • Supplements
Abstract

purpose. To assess the structural and functional consequences of local ganglion cell (GC) loss in an experimental model of a retinal nerve fiber layer (NFL) bundle defect. To evaluate and compare three commonly used multifocal electroretinogram (mfERG) stimuli, as well as the standard transient pattern-reversal ERG (pERG) and the photopic full-field ERG, for detection of local GC damage.

methods. Intraretinal axotomy was achieved by multiple treatments with a diode laser adapted to a slit lamp biomicroscope. Retinal laser burns were applied along an arc, subtending approximately 60°, about one disc diameter superotemporal to the optic nerve. Functional measures were acquired before laser application and at numerous time points thereafter. These included mfERGs for three different stimuli: a standard fast m-sequence flicker, a global-field flash paradigm (MOFO), and a slowed m-sequence (with seven dark frames inserted to each m-step [7F]). pERGs were measured for a 24° × 32° checkerboard stimulus (0.56 cyc/deg, 90% contrast, 75 cd/m2, 5 reversals/s). Photopic full-field ERGs were measured for red flashes (0.42 log photopic cd-s/m2) on a blue rod-saturating background (30 scotopic cd/m2). Retinal photography, fluorescein angiography and postmortem histologic evaluation of the optic nerve, NFL, and retinal tissues were performed.

results. After six laser sessions, the NFL bundle defect appeared to be complete and contiguous and was visible both proximal to and distal to the site of the photoablation by clinical examination of the fundus and stereoscopic photographs. Histologic evaluation demonstrated localized loss of GC axons, confirmed at the level of the retrobulbar optic nerve. Retinal cross sections in the temporal retina (distal to the axotomy) showed loss of GC soma and NFL degeneration, whereas all other layers appeared intact. mfERGs showed loss of high-frequency components (HFCs) for responses located within the arcuate region corresponding to the NFL defect. Local GC damage was most easily detected using the slowed 7F m-sequence stimulus. This stimulus elicited relatively large HFCs that were significantly reduced from local responses after axotomy and that were tetrodotoxin (TTX)-sensitive in a control experiment. Low-frequency component loss with the 7F stimulus did not reach statistical significance. The photopic full-field ERG was not significantly affected. pERG amplitudes declined significantly from baseline but remained within normal limits.

conclusions. Focal loss of GC function in the macaque retina is most easily detected using the slowed-sequence mfERG. Local 7F HFCs depend on intact GC function.

Although several studies have demonstrated a diffuse component of damage in glaucoma, 1 2 3 4 the most salient and easily detected structural and functional abnormalities are localized. For example, the most characteristic visual field defects in glaucoma are spatially localized within the midperiphery 5 6 and are generally thought to reflect damage within corresponding bundles of ganglion cell (GC) axons. 7 8 9 10 These retinal nerve fiber layer (NFL) bundle defects, in turn, generally correspond to localized patterns of loss within the optic nerve head. 11 12 Evidence suggests that these observations may reflect a particular susceptibility of the upper and lower poles of the optic disc, perhaps conferred by local structural weakness within these areas of the laminar cribrosa. 13 14 15 Localized vascular events in and near this region may also play a role. 16 17  
Objective techniques such as electroretinography (ERG) can provide an important complimentary measure of function in glaucoma, especially for experimental animal models. Given the well-documented localized loss of GC axons in glaucoma, the multifocal technique 18 19 20 21 offers the distinct advantage of topographic functional assessment. Although the application of this technique to human glaucoma has thus far met with limited success, 22 23 24 25 26 27 several studies have demonstrated its promise in nonhuman primate models of glaucoma. 28 29 30 31 None of these animal studies, however, has examined localized structure-function relationships. To date, only experimental models with widespread effects throughout the entire retina have been adopted to validate the use of multifocal ERG (mfERG) for monitoring local GC function in glaucoma. For example, suppression of inner retinal neuronal responses by intravitreal injection of pharmacologically active compounds would be expected to affect the entire population of third-order neurons. 32 33 34 35 Similarly, optic nerve transection would be expected to destroy nearly all GCs (Kim CBY, et al. IOVS 2001;42:ARVO Abstract 3864). 
The purpose of this study was to examine local changes in mfERG responses after intraretinal photoablation of a limited set of GC axon bundles. Similar models have been used to study focal GC loss and NFL bundle defects for nearly two decades. 12 36 37 38 39 40 A secondary purpose of this study was to compare the ability of three different modes of mfERG stimulation, as well as the conventional pattern-reversal ERG (pERG) and the photopic full-field flash ERG (fERG), to detect localized GC damage. 
Methods
Animals
All experimental methods and animal care procedures adhered to the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research and were approved by the local Institutional Animal Care and Use Committee (IACUC). One adult female rhesus monkey (Macaca mulatta) aged 4 years was the primary subject of the study. Additional results are also presented for a second adult female macaque monkey who had mfERGs measured before and after intravitreal injection of tetrodotoxin (TTX). Normative mfERG data are presented for 11 other animals (all female, ages 7–11 years). Results for traditional pERGs and photopic full-field fERGs are compared with normative data collected during a separate study of 29 adult female rhesus monkeys, ranging in age from 9 to 14 years; full results of the latter studies are reported elsewhere. 41 42  
Anesthesia
All procedures were performed with animals under general anesthesia. In all cases, anesthesia was induced with 15 mg/kg intramuscular (IM) ketamine (Fort Dodge Animal Health, Fort Dodge, IA) and 1.5 mg/kg IM xylazine (Phoenix Scientific Inc., St. Joseph, MO), along with a single subcutaneous injection of atropine sulfate (0.05 mg/kg; Phoenix Scientific Inc.). Animals were intubated and breathed 100% oxygen. Body temperature was maintained with a warm-water heating pad set at 37°C. Pulse rate and oximetry were monitored (Propaq Encore unit, model 206EL; Protocol Systems, Inc., Beaverton, OR) and maintained between 85 to 125 per minute and 98% to 100%, respectively. Pupils were fully dilated (≥7 mm) with 1.0% tropicamide and 2.5% phenylephrine (Alcon Laboratories Inc., Fort Worth, TX). For all ERG sessions, anesthesia was maintained with ketamine (5 mg/kg per hour intravenously) and xylazine (0.8 mg/kg per hour IM). Topical corneal anesthesia was provided with 0.5% topical proparacaine (Alcon Laboratories Inc.) and an ocular lubricating agent (Celluvisc; Allergan, Irvine, CA) was periodically applied. Recording sessions lasted approximately 2 hours. During all other sessions (e.g., laser photocoagulation, stereo optic disc photography, fluorescein angiography), anesthesia was maintained with 2% to 3% isoflurane (Baxter, Deerfield, IL). 
Axotomy by Retinal Laser Photocoagulation
The left eye was studied in the experiment. Initially, burns were placed in a single row located approximately one disc diameter from the optic nerve, superotemporally, along an arc subtending approximately 60°. Burns were made using an infrared diode laser (810 nm; OcuLight SL; Iridex, Mountain View, CA) adapted for a slit lamp biomicroscope (model 900; Haag Streit AG, Köniz, Switzerland) and a fundus laser contact lens (Ocular Instruments Inc., Bellevue, WA). Based on prior work by Apple et al., 43 44 and Kormaramy et al. (IOVS 2001;42:ARVO Abstract 4430), multiple sessions were planned. The effect of photocoagulation (300-mW power, 200-ms duration, 75-μm diameter, 80–160 count per session) on the retinal NFL was observed during and between each treatment session by clinical fundus examination and simultaneous-stereoscopic optic disc and retinal photography (3-Dx; Nidek Co., Ltd., Gamagori, Japan). During each successive treatment, burns were placed along the hyperpigmented border of the preexisting site, and major blood vessels were avoided. 
After five treatments with the red laser (see Fig. 4 , bottom right, for timeline), there was a dense scar that appeared to involve the outer retina, retinal pigment epithelium, and choriocapillaris (confirmed by fluorescein angiography), yet only minor damage to the NFL was apparent. The striations of intact NFL were highlighted by dense pigment granules that collected between fiber bundles. The NFL could be seen easily, laying deep against markedly thinned tissue within the borders of the scar. Consequently, a sixth laser treatment was administered (in the 59th week): 96 burns were placed along the pigmented border and within the center of the scar with a green laser (532 nm; OcuLight GL; Iridex) that was set to a lower power (200 mW) and a larger spot size (125 μm), but the same 200-ms pulse duration. Subsequent examination of the clinical fundus and stereophotographs revealed a dense NFL defect of approximately 2 clock hours’ width and a marked notch in the neural rim tissue of the optic disc (see Figs. 1A 2A ). Final ERG records and photographs were obtained, and the animal was killed for histologic studies. 
Electroretinography
mfERGs were recorded using a visual evoked response recording system (VERIS, ver. 4; EDI, San Mateo, CA). Pupils were fully dilated (≥7 mm) and corneal anesthesia and lubrication were provided periodically throughout the session, as described earlier. Custom-designed Burian-Allen contact lens electrodes were used for all ERGs (10 mm diameter, +3.00 D; Hansen Ophthalmics, Iowa City, IA). The corneal ring on the stimulated eye served as the active electrode, and the corneal ring of the unstimulated (patched) contralateral eye served as the reference electrode. Both electrodes were referenced to a subcutaneous ground electrode placed in the thigh. Electrode impedance was accepted if less than 5 kΩ. 
Residual refractive error was measured by retinoscopy for the test distance (25 cm) and corrected to the nearest 0.50 D. The mfERG stimulus was presented on a 21-in. monochrome monitor with a 75-Hz refresh rate. Before the actual recording session, an initial set of brief recordings (2 minutes each) were used to center foveal responses within the response array and to position the blind-spot responses appropriately. Fixation stability was monitored both by direct visual inspection of the eye and of the on-line recording (i.e., baseline stability). On rare occasions, small, drifting eye movements and/or twitch-like partial blink artifacts (indicating insufficient depth of anesthesia) necessitated rerecording of particular ERG segments. 
The standard mfERG stimulus was a fast local luminance flicker of 103 unscaled hexagonal elements subtending a total field size of approximately 55° (Fig. 3A) . The luminance of each hexagon was independently modulated between dark (1 cd/m2) and light (200 cd/m2), according to a predetermined pseudorandom, binary m-sequence 18 19 with a base-interval of 13.3 ms, providing local contrasts of approximately 99%. Stimulus luminance was measured with a calibrated spot photometer (SpectraScan PR-650; Photo Research, Chatsworth, CA). Each recording was approximately 8 minutes in length (usually obtained in eight 60-second segments). Signals were amplified (gain = 100,000), band-pass filtered (10–300 Hz; with an additional 60-Hz line filter), sampled at 1.2 kHz (i.e., sampling interval = 0.83 msec), and digitally stored for subsequent off-line analyses. 
Two additional mfERG stimuli were also used. The first was introduced by Sutter et al. 45 and has been described in detail elsewhere 27 31 (for additional detail and theory, see Sutter 21 ). This stimulus differs from the standard fast-luminance flicker stimulus, in that during each m-frame, the amplitude of the pseudorandom local luminance modulation is smaller (hexagonal areas flicker between 1 and 100 cd/m2); and then after each m-frame, the entire stimulus area becomes dark for one video frame, flashes brightly (200 cd/m2) during the next video frame, and then returns to the dark luminance level for the duration of another frame. The sequence then repeats, beginning with another m-frame of pseudorandom local stimulation, a dark frame, a full-screen flash frame, a dark frame, and so forth. The schematic representation in Figure 3B illustrates two complete m-sequence cycles, with full-screen flashes interposed. This stimulus is referred to as MOFO throughout the article (M, m-sequence step; O, dark frame; F, full screen flash; O, dark frame). 
During the recording sessions of the final three time points, another slower m-sequence stimulus was also used that had seven blank frames inserted between each m-sequence step (called 7F in the present study). The summed responses under this condition have been shown to be similar to responses from single, brief Ganzfeld flashes on a photopic background. 46 47 Figure 3 shows a schematic representation of two successive m-sequence steps for the fast-luminance flicker (Fig. 3A) , the MOFO (Fig. 3B) , and the slowed flicker stimuli (Fig. 3C)
Specific mfERG records were exported for further analyses. Figure 1B shows the stimulus locations corresponding to these records. In Figure 1B the hexagonal stimulus pattern is overlaid onto the fundus photograph (the latter inverted to enable comparison with the field-projected response array). The responses from two concentric rings around the fundus were evaluated: locations numbered 1 to 12 around the first ring in black and numbered 1 to 11 around the partial second concentric ring (in red). The responses within these rings are of greatest interest, because the responses from the lower hemifield locations lie within the area of the NFL defect whereas those from the upper hemifield correspond to areas of intact retina. 
Local responses were further analyzed on computer, according to their frequency content (Excel; Microsoft, Redmond, WA). Each local mfERG response was band-passed filtered (90–300 Hz, −3 dB points) to extract the higher frequency components (HFCs). The lower frequency components (LFCs) of each response were represented as the raw response minus the HFC. The HFCs and LFCs for the 7F stimulus condition were quantified in the following manner. The amplitude of the HFCs were calculated as the root mean square (RMS) for the epoch between 15 and 55 ms of each filtered record. Noise amplitude was calculated as the RMS for a similar 40-ms epoch where there were no consistent response features (60–100 ms). The amplitude of the first negative feature (N1, analogous to the a-wave) was calculated as the maximum negative excursion from baseline in the epoch up to 40 ms. The amplitude of the positivity (P1, analogous to the b-wave) was calculated as the voltage difference between the maximum peak and the N1 trough, and the second negativity (N2, perhaps analogous to the photopic negative response [PhNR]) was calculated as the difference between the P2 peak and the minimum voltage between 40 and 60 ms. 
Immediately after mfERG recordings were completed, traditional large-field pERGs were recorded (Utas-E3000; LKC Technologies, Gaithersburg, MD), as previously described. 41 The pERG stimulus was a checkerboard pattern (0.56 cyc/deg), reversing at 2.5 Hz (5 reversals/s). The stimulus subtended 32° × 24° at the 50-cm test distance. Mean luminance was 75 cd/m2 and contrast was more than 90%. The position of the foveal projection determined during mfERG testing was used to centrally align the pERG stimulus. Residual refractive error was measured by retinoscopy for the test distance and corrected to the nearest 0.5 D. Fixation stability was monitored both by direct visual inspection of the eye and of the online recording (i.e., baseline stability) and was always steady with sufficient depth of anesthesia. 
Signals were band-pass filtered 1 to 500 Hz and sampled at 2 kHz. Two records were obtained for each eye and then averaged. Each single record was an average of 200 sweeps. Eye position was monitored continuously and remained stable with sufficient depth of anesthesia. 
Using the same recording system (UTAS-E3000; LKC Technologies), photopic full-field fERGs were obtained after 5 minutes of light adaptation to a rod-saturating blue background (30 scotopic cd/m2; Wratten no. 78 filter; Eastman Kodak Co., Rochester, NY). Red stimulus flashes (Wratten no. 29 filter; Eastman Kodak Co.) with an intensity of 0.42 log photopic cd-s/m2 were presented monocularly at 0.5 Hz by a Ganzfeld integrating sphere. Stimulus and background intensities were measured with a calibrated photometer (Spectra Pritchard PR-1980A; Photo Research). Signals were band-pass filtered 0.3 to 3000 Hz and sampled at 2 kHz. Two records were obtained and then averaged. Each single record was an average of 6 to 10 sweeps. Peak-to-trough amplitudes, as well as peak implicit times were measured as shown in Figure 4 . Oscillatory potential (OP) isolation was achieved by post hoc filtering using a Blackman filter (−3 dB at 70 and 280 Hz). The amplitude of the OP complex was quantified by calculating the RMS from the filtered waveform over a time epoch beginning at the a-wave trough and ending after OP4, as described previously. 42  
The normal ranges for mfERG responses to the 7F stimulus (n = 11 animals) were calculated using the t-distribution (α′ = 0.01, df = 10). Similarly, normal ranges for pERG and fERG parameters (n = 29 animals) were also determined using the t-distribution (α′ = 0.01, df = 28). This conservative estimate of normal ranges (i.e., α = 0.01) was adopted, because multiple comparisons were generally made for each ERG technique (e.g., three to four parameters each). 
Intravitreal Injection of TTX
mfERGs were measured in a single monkey before and 1 hour after a single injection of tetrodotoxin citrate (TTX; Sigma-Aldrich, St. Louis, MO), dissolved in balanced salt solution (Alcon Laboratories Inc., Fort Worth, TX). The 60 μL injection was made into the middle of the vitreal chamber by insertion of a sterile 30-gauge needle through the pars plana approximately 3.5 mm posterior to the temporal limbus. The final vitreal concentration (∼6–8 μM), was estimated assuming full dilution, no leakage and a vitreal volume of 2.1 mL. 48  
Histopathology
Tissue Allocation.
After 16 months of in vivo studies (e.g., ERGs, imaging) the animal was killed with an intravenous injection of 390 mg/mL sodium pentobarbital and 50 mg/mL sodium phenytoin (Euthasol; Diamond Animal Health, Inc., Des Moines, IA) after a bolus injection of heparin (∼5000 IU IV) and immediately perfused with approximately 1 L of 4% buffered paraformaldehyde through precannulated carotid arteries. The perfusion lasted approximately 30 to 45 minutes, and the eyes were enucleated. Two pieces of the left retrobulbar optic nerve, approximately 2 and 0.5 mm in length were sampled 2 to 3 mm posterior to the globe. The thin piece was processed for plastic embedding and sectioned (1 μm) for myelin stain. The thicker piece was prepared for paraffin sections for hematoxylin-eosin (H&E) and immunohistochemical stains. The left eye was then hemisectioned at the equator and the vitreous humor was removed by manual dissection. The anterior optic nerve was removed from the posterior eyecup using an 8-mm trephine and frozen fixed with dry ice (the trephine cut was shifted nasally to leave the temporal retina intact for wholemount study). The remaining eyecup was further cut into two pieces along a vertical line that ran just temporal to the fovea (approximately at the rightmost edge of the fundus photograph in Fig. 1A ). The piece containing the temporal retina was frozen fixed and later cross-sectioned for analysis of retinal structural changes in the area distal to the axotomy. The remaining retina from the posterior pole was wholemounted and processed for immunohistochemical stain of neurofilament and glial fibrillary acidic protein (GFAP; described later). 
Immunohistochemical Stain: Avidin-Biotin Method.
Paraffin sections were dewaxed and put in 0.3% H2O2 in methanol for 30 minutes for antigen retrieval. The sections were then incubated with 1% horse serum and 1% bovine serum albumin mixture for 30 minutes to reduce unspecific binding. Monoclonal mouse anti-human primary antibody of GFAP (Novocastra Laboratories Ltd., Newcastle-upon-Tyne, UK), was applied at a concentration of 1:200 and incubated at room temperature for 90 minutes. After three washes in 0.01 M PBS, 10 minutes each, biotinylated horse anti-mouse secondary antibody (1:100, Vectastain Elite ABC kit; Vector Laboratories, Inc., Burlingame, CA) was applied for 30 minutes, followed by avidin-biotin complex for another 30 minutes. 3,3′-Diaminobenzidine (DAB Kit; Vector Laboratories Inc.) was used for substrate chromogen staining for 2 to 5 minutes. The sections were counterstained with 0.1% Mayer’s hematoxylin (Sigma Diagnostics) and mounted. 
Immunohistochemical Stain: Double Fluorescent Stain of Wholemounted Retina.
The tissue was washed in PBS containing Triton X-100 0.2% for 1 hour and then incubated in blocking serum (1% horse serum and 1% goat serum in bovine serum albumin) for 12 to 24 hours at 4°C. The tissue was transferred into a mixed solution of 1:50 monoclonal mouse antibody to 200-kDa neurofilament and 1:100 rabbit antibody to GFAP diluted with 1% bovine serum albumin for 48 hours at 4°C. The tissue was then washed in 0.01 M PBS for three changes (1 hour each). Fluorescein isothiocyanate-conjugated secondary horse anti-mouse immunoglobulin and Texas red-conjugated goat anti-rabbit immunoglobulin at 1:100 (Vector Laboratories Inc.) were incubated for another 12 to 24 hours at 4°C. After a thorough wash with 0.01 M PBS, the tissue was mounted and viewed under a fluorescence microscope or a confocal microscope. 
Negative controls for immunohistochemical stains (both paraffin sections stains and wholemounted retina stains) were performed by omitting the corresponding antiserum from the primary antibody solution. 
H&E, Toluidine Blue, and Phenylenediamine Stains.
Dewaxed and hydrated paraffin sections were stained first with Mayer’s hematoxylin for 10 minutes and washed with Scott water for 3 to 5 minutes. After a brief dip in 1% acid alcohol (1% HCl in 70% alcohol) the sections were again rinsed with Scott’s water for approximately 3 minutes, stained with eosin for 10 minutes, washed in running tap water, dehydrated, cleared, and mounted. 
Toluidine blue was used to stain the 1-μm plastic-embedded cross sections of optic nerve. On a hot plate at 60°C to 70°C, the sections were stained with 1% toluidine blue solution for 3 minutes, followed by a few drops of phosphate buffer solution for another 2 minutes. Excess stain was washed off in running water, and the slide was air-dried and mounted. 
The plastic sections of the optic nerve were immersed in a staining solution (1% phenylenediamine in 1:1 methanol-isopropanol) for 20 minutes and rinsed lightly in isopropanol. The sections were air-dried and mounted. 
Results
Histopathology
Two weeks after the final laser treatment, a dense NFL defect had become clearly visible, along with a corresponding notch in the neural rim tissue of the optic disc. Figure 1A shows the wedge-shaped NFL defect, the characteristic loss of reflective striations, both distal and proximal to the burn site (note, the photograph has been inverted for comparison to the field-projected mfERG array; Fig. 1B ). Figure 1C shows a portion of the wholemounted retina distal to the burn (toward the GC soma) viewed from the vitreous side. Immunohistochemical staining for neurofilament demonstrates the sharp demarcation between intact nerve fiber bundles (bright green) and the dense NFL defect along the centralmost border of the defect (i.e., closest to the fovea). Figure 1E shows the same portion of retina with the excitation and barrier filters tuned to the GFAP fluorescent stain. The glial septa that normally run parallel to the axon bundles were intact and appeared similar on both sides of the defect’s border. Astrocytic and Müller cell processes could be seen wrapped around a blood vessel within the NFL. Figure 1G shows the overlay of Figures 1C and 1E and indicates that the normal striated pattern of the NFL in vivo depends on the presence of intact axon bundles but not the glial septa that border their channels. 
The outline of the laser scar is shown more clearly in the arterial phase of the fluorescein angiograph (Fig. 2A) taken on the same day as the fundus photograph in Figure 1A . Note that the retinal and choroidal circulation are both intact beyond the burn site, within the arcuate region of interest. The angiographic filling times were also normal (choroidal flush began 4 seconds after injection, arterial phase 5 seconds, and laminar venous phase 6 seconds after injection). Figure 2B shows a transverse section of the optic nerve, from approximately 2 mm behind the globe, stained for myelin with phenylenediamine. A distinct wedge-shaped sector of degenerating axons, corresponding to the retinal NFL defect was evident, contrasted against the otherwise darkly stained myelin sheaths of normal axons throughout the remainder of the optic nerve (note again that the section has been inverted; the notch at the bottom edge and the slightly lighter stained portion at the top right are both artifact). 
An adjacent transverse section, stained for GFAP and viewed at higher magnification, is shown in Figure 1F . The left panel shows intact axon bundles, normal architecture, and a normal level of GFAP expression in an unaffected portion of this cross section through the optic nerve. Cell bodies of astrocytes and oligodendrocytes are visible with normal distribution and density. The right panel demonstrates gliosis within the affected segment of the optic nerve, including increased density of glial cell bodies and increased expression of GFAP. Another adjacent transverse section, stained with toluidine blue is shown in Figure 1H . The left panel shows intact axons and normal bundle architecture within an unaffected portion of this section through the optic nerve. The right panel shows signs of active degeneration within the affected segment. Normal architecture is grossly disrupted, and numerous phagocytic cells (presumably activated microglia) are visible containing axonal debris, including lipid droplets and other byproducts of myelin breakdown. There were a few remaining intact axons that are likely to be migrant axons not adhering to the otherwise predictable retinotopic organization. 11 49 The longitudinal section through the anterior optic nerve (Fig. 1D) also revealed disruption of the normally columnar architecture of the axon bundles posterior to the lamina, as well as gliosis limited to these abnormal bundles within the laminar and retrolaminar regions of the ON. Thinning of the NFL was also apparent within the prelaminar ON head. 
A cross-sectional view of the retina temporal to the fovea (Fig. 2C) reveals marked thinning of the NFL and loss of GC bodies (arrowheads) within the region upstream from the axotomy. The black arrow indicates the approximate position of the horizontal raphe approximately 1 mm temporal to the fovea (the accuracy of this estimate is limited by the symmetry of the upper and lower edges of each eyecup section around the raphe). The normal complement of GCs and a thick NFL can be seen within the inferior retina below the horizontal raphe (i.e., to the left of the arrow). Damage appears to be limited to the GC and NF layers. The inner and outer plexiform layers, as well as the inner nuclear and photoreceptor layers are equally thick and dense on either side of the horizontal raphe. A similar pattern was observed in the next dozen serial sections examined. 
Electroretinography
Figure 4 plots retinal function versus time, as measured by the classic transient pERG. For reference, the normal average and range (mean ± 2 SD) for each parameter are indicated by the solid line and gray zone, respectively. The broken line in each panel indicates the lower limit for detection of significant change from the pretreatment baseline value in this animal (the latter calculation based on an estimate of intersession reliability (95% limits-of-agreement) for the normal group. 41 42  
There was a decrease in both the P50 and N95 components, compared with baseline, at all posttreatment time points (i.e., after the first four photoablation treatments). Although both parameters remained within the normal range (gray zones) throughout the experiment, they straddled the lower limit for detection of significant change until the green laser treatment was applied. Thereafter, the N95 component in particular showed a consistent, significant reduction from baseline. Implicit times of both components were unchanged throughout the experiment. 
The middle and right columns in Figure 4 show the photopic full-field fERG results over time. The amplitudes of the a-wave, b-wave, PhNR, and OPs varied slightly between sessions, also showing an initial decrease at the first posttreatment time point, but all parameters remained within normal limits throughout the experiment (including implicit times, not shown). At the final session, the values of all four parameters were essentially equal to the pretreatment baseline. None of these parameters showed a significant decrease below baseline at any posttreatment time points. 
Figure 5 shows the results for the three different modes of mfERG stimulation from the final mfERG session (66 weeks). Figure 5A contains the full response array for each stimulus. Good centration is evident, and the local consequences of the burn are apparent in the lower left quadrant of each full response array. Figure 5B shows individual records for each of the three stimuli, aligned in their respective columns according to the field position described by the legend to Figure 1B . These locations were chosen because the subset from the inferior hemifield lay within the area of the NFL defect, distal to the axotomy. Thus, responses from locations 7 to 11 in ring 3 can be compared with responses for 5 to 1, the intact areas in the opposite hemifield. Similarly, responses 8 to 10 in ring 2 can be compared with responses 6 to 4. Responses 11 and 12 in ring 2 should be considered with caution, as they are very close to the burn. Nevertheless, they are shown because location 11 lies within the NFL defect, whereas location 12 is superior to the predicted border of the scotoma. The solid black traces in each column are the records at 66 weeks after baseline. For comparison, the pretreatment baseline records are also shown for the fast m-sequence and MOFO stimuli (gray dashed traces). 
For standard fast m-sequence stimulation Figure 5B demonstrates that there were only subtle differences between responses from intact areas versus the areas within the NFL defect. Generally, the latter appeared to be slightly more smooth. The differences from pretreatment baseline responses (dashed gray) are also quite small. Similarly, responses to the MOFO stimulus from areas within the NFL defect appear to have smaller and fewer high-frequency oscillations compared with their counterparts from the opposite hemifield. In particular, the oscillations after the major deflection of what has been called the induced component 21 (here, the epoch from approximately 40 to 80 ms), as well as the third oscillation of what may be more of a direct response 21 (here, the epoch from 0 to 40 ms), are all less prominent. The differences from pretreatment baseline (dashed gray) follow a similar pattern: The oscillations in the induced component are smaller, most obviously within the area of the NFL defect. Having noted that the differences between responses within the NFL defect and either those from the opposite hemifield, those from the same locations before any treatment (i.e., baseline), or those from normal eyes, appeared to manifest most clearly as loss of higher frequency oscillations, the 7F mfERG stimulus was added to the protocol (from the 48th week on), because it is known to produce robust high-frequency oscillations in normal eyes. 
The third column in Figure 5B shows the responses for the 7F stimulus (slowed m-sequence). The differences between upper and lower hemifields are much more obvious for the 7F responses. There are very prominent oscillations in the responses from the intact areas, whereas the responses from the areas within the NFL defect are much smoother. It should also be noted that for responses from intact retina, the implicit times for some of the high-frequency oscillatory features increase as the distance between the stimulus area and the optic disc increases. This phenomenon is thought to be one of the primary characteristics of the optic nerve head component (ONHC), as proposed by Sutter and colleagues. 34 50 Of interest was that, as the stimulus location emerged from the NFL defect in ring 2, the oscillations again became prominent (see location 12). 
mfERG responses were further analyzed by their frequency content (see the Methods section). Figure 6 displays the HFCs in the left column and the LFCs in the right column. The filtered waveforms (Fig. 6A) from locations of greatest interest are presented for the slow 7F stimulus in an arrangement similar to that in Figure 5B . Again, note the dramatic differences in the HFCs between responses 1–5 and 11–7 for ring 3 and between 4–6 and 10–8 for ring 2. Small but consistent HFCs remained even for the group of responses corresponding to the NFL defect. The implicit times of these remnant HFCs did not vary with position around the ring as much as some of those from the opposite hemifield did. The differences between successive responses around the ring, for intact retinal areas, were less obvious for the LFCs shown in the right column. The LFC differences between records from areas within NFL defect and those from the opposite hemifield, however, include reduced amplitudes of the P2 and N2, as well as a small increase in the implicit times for both features. 
The differences between groups may be more easily appreciated by comparison of group average waveforms. In Figure 6B , the average of the records from locations within the NFL defect are shown as solid bold traces, and the average for the group from the intact hemifield are shown by thinner dashed traces. The HFC and LFC group average records are shown in the left and right columns, respectively. The same analysis is presented in Figures 6C and 6D for the MOFO and standard fast m-sequence responses, respectively. Whereas there were striking differences between the group averages for the slow 7F stimulus (Fig. 6B) , the differences between the average records for the standard fast-m-sequence stimulus (Fig. 6D) were negligible for the HFCs (left) and only very subtle for the LFCs (right). The group average differences for the MOFO stimulus (Fig. 6C) were intermediate (i.e., larger than those of the standard stimulus, but smaller than those of the slowed m-sequence stimulus). For the HFCs, the second and third oscillations in the MOFO direct responses (∼24–36 ms) were slightly smaller, as were the oscillations that follow the major deflection of the induced component (∼60–70 ms), in the NFL defect group compared with the intact group. The differences for the LFCs of MOFO responses included reduction in the depth of the trough between the direct and induced components (∼35–45 ms), as well as in the height of the primary peak of the induced component (∼70 ms). Thus, the 7F responses revealed the greatest differences between intact retinal areas and areas within the NFL defect. 
Because the 7F stimulus was not added to the protocol until week 48, comparisons to pretreatment baseline responses can be made only for the fast m-sequence and MOFO stimuli. Figures 6E and 6F show such a comparison for these two stimuli, respectively. The upper and lower group average records from 6D are shown again in 6E, now separated vertically into two rows. The baseline records (averages for identical groupings) are shown as thin gray traces. As expected from the results shown in Figure 5B (left column), the differences from baseline were quite small. Figure 6F presents the same comparison for the MOFO response group averages (the bold traces are taken from 6C, and the baseline group average records are shown as thin gray traces). The differences from baseline were small again, even for the lower hemifield HFCs (compare bold and gray traces, bottom left column). This result may be surprising, given the differences shown in Figure 5B , middle column, but can be explained by considering normal changes in HFC morphology and timing around each ring: the group average poorly characterizes differences between individual locations if they are relatively small, because such normal features become “smeared” in the average. 
Averaged data for upper and lower hemifield locations from the two peripheral rings of each response array (i.e., rings 4 and 5) were compared in a similar manner. The pattern of results was nearly identical with that shown in Figure 6 for rings 2 and 3, however, between-session variability was slightly larger compared with data for rings 2 and 3, especially for the MOFO stimulus (not shown). The peripheral extent of the NFL defect was less certain due to optical limitations of in vivo imaging and technical limitations of histologic preparations. Therefore peripheral mfERG responses were not analyzed further. 
The effects found using the 7F stimulus were large in comparison to the other two stimuli, but prelaser baseline responses were not available for the 7F stimulus. Therefore, the parameters of the 7F responses observed for this animal were compared against those for a group of normal control subjects to evaluate the possibility that the observed effects were due to random chance rather than to experimental intervention. 
The four panels in Figure 7 plot the average amplitude for the 7F response components, HFCs, N1, P1, and N2 for the group of responses within the area of the NFL defect (i.e., lower hemifield) and the group of responses from corresponding areas within the intact (upper) hemifield. Normal amplitude ranges (n = 11) are shown by the gray bars, and the mean normal amplitudes (±SE) are indicated by the hash marks (and dark bands). The data points plot the amplitudes for the experimental monkey at 48, 53, and 66 weeks after the experiment began. (These time points correspond to the second, third, and fourth points after the pretreatment baseline shown in Fig. 4 .) Figure 7 demonstrates that HFC amplitudes for the group of responses within the area of the NFL defect were below the lower limit of normal during all three sessions (P < 0.01, df = 10). Thus, it is not likely that the observed local loss of HFCs for 7F responses was due to random selection of a normal variant. Although all other parameters remained within the normal range until the end of the experiment, the N1, P1, and N2 amplitudes for the lower hemifield group declined substantially between the 53rd and 66th weeks. 
Because the 7F responses demonstrated such clear differences between the two groups of retinal locations within this experimental eye, while other functional measures showed little or no differences, the following experiment was designed to evaluate further the nature of the changes observed under the 7F paradigm. TTX is known to be a highly selective blocker of voltage-gated sodium currents (action potentials). 51 Action potentials are produced by GCs, and occasionally, but less vigorously, by some amacrine cells. 52 TTX has been used by other investigators to evaluate the contribution made by (primarily) GC spiking activity to mfERG responses for various stimuli. 33 34 35  
Figure 8 shows the results of TTX application for the slowed 7F stimulus. Preinjection (control) records for rings 2 and 3 are arranged as before in the left column of Figure 8A . Large HFCs were present and there was clear asymmetry as the position in the ring changed. Much of this asymmetry appeared to depend on the size and number of HFCs. The number of wavelets and amplitude of the HFCs increased in the temporal retina. The implicit times for some of the wavelets increased predictably with distance from the optic nerve head. The middle column in Figure 8A shows responses for the same locations 1 hour after TTX injection. The most obvious changes included a decrease in the number and amplitude of the HFCs. The nasal-temporal asymmetry was lost; the waveforms no longer varied with changes in ring position. The difference records (right column) contained both high- and low-frequency components, and showed distinct variation with changes in ring position, especially around ring 2. 
Figure 8B shows the band-pass-filtered HFCs for the corresponding records shown in Figure 8A . Comparison between the pre-TTX (left) and post-TTX (middle) waveforms demonstrates the predominant effects of TTX on the slowed 7F mfERG responses. The amplitude and number of HFC wavelets decreased, and the nasal-temporal asymmetry was lost after TTX. These effects were underscored by the filtered difference records (Fig. 8B , right) which showed that the HFCs removed by TTX were large and varied greatly around the rings, where they were largest and more numerous in the temporal retinal locations. Figure 8C shows the upper and lower group averages, as was shown in Figure 6 for the results after laser axotomy. The upper and lower retinal groups were similar to each other, both before and after TTX, but the post-TTX group averages were smaller, with fewer oscillations, similar to the findings after laser axotomy and GC death. 
As just outlined, the primary purpose of the TTX injection was to evaluate the origins of the HFCs in the 7F responses of normal monkeys. Space considerations prevent presentation here of TTX effects on other mfERG responses; the effects of TTX on the pERG and photopic full-field fERGs have been presented elsewhere. 41 42 The effects of TTX on the 7F responses are discussed in the following section within the context of prior published work. 
Discussion
The histologic results of this study confirm that this experimental model of a retinal NFL bundle defect produces localized damage to GCs, with little effect on outer retinal elements within the area of the NFL defect distal to the axotomy, nor to adjacent GCs and their axons. 36 37 38 39 The localized nature of the lesion was established anatomically through several lines of evidence: the appearance of the NFL bundle defect in the clinical fundus; the axonal degeneration proximal (central) to the laser axotomy site, both at the level of the laminar cribrosa and approximately 2 mm behind the globe; and the degeneration of GC bodies and their axons in the NFL distal to the axotomy in the temporal retina. 
The results of the functional analyses also suggest that the damage was localized and limited. There were no changes in either the full-field fERG, or in the mfERG responses from the anatomically intact hemifield, compared with preaxotomy baseline measures. Among the functional measures included in this study, the results revealed that the slowed (7F) mfERG stimulus was superior to the two other mfERG stimuli and to the standard transient pERG and the photopic full-field fERG, for detection of local GC loss in this model. 
After the initial group of four laser treatments, the amplitudes of both pERG components, the P50 and the N95, declined by approximately 50%. Some small proportion of this effect was probably due to destruction of the outer retina at the burn site, because the final burn size was approximately 15% of the pERG stimulus area. Throughout most of the remaining course of the experiment, these amplitudes remained nearly constant, and within the normal range, despite continued dramatic changes observed in the appearance of the NFL (over many months and after six laser treatments in total). Perhaps these early changes in pERG amplitude reflect dysfunction of GCs before substantial change in the clinical appearance of their axon bundles. In any case, because of the wide range of observed normal values, a cross-sectional analysis of pERG amplitude was unable to detect this moderate degree of local damage. This is consistent with recently published results for a study of pERG amplitudes in human glaucoma. 53 It should be noted that, in the current study, the normal group contained some animals that were much older than the experimental subject. This may have partially undermined the cross-sectional analysis. Longitudinal analysis of the pERG, however, was better, as the change from preaxotomy baseline was significant, particularly for the N95 component during the last two sessions. 
Although there was a larger effect on the N95 amplitude compared with the P50, it was surprising to find that the N95-to-P50 ratio did not change significantly from baseline (ratio data not shown). In a separate study, under the same pERG recording conditions used in our study, blockade of GC (and amacrine) spiking activity with TTX caused a greater reduction of the N95 component, with relative sparing of the P50 amplitude. 41 This latter result is consistent with a previous study, 54 and provides further support for the finding that loss of GCs causes a larger effect on the N95 component of the pERG. 55 Use of hemifield pERG stimuli with this model is likely to have been more effective, and perhaps, also to have shown a more selective effect on the N95 component within the damaged hemifield. 
The PhNR of the full-field ERG is also thought to reflect GC activity. 54 56 57 However, the relatively limited degree of GC damage created in the current experiment appears to be insufficient to have caused any change in the PhNR amplitude. Of note, the photopic OPs of the full-field ERG were similarly unaffected by this limited GC lesion. Both results may be surprising, given the changes observed for the slowed 7F mfERG responses, which included dramatic reduction of HFC amplitudes, as well as a decline of the N2 amplitude (e.g., Fig. 6 ). This discrepancy may be resolved by consideration of results from another recent study which suggest that the full-field photopic ERG resembles more closely the sum of only the most peripheral mfERG responses to slowed (14F) stimuli. 47 The NFL defect model, as applied in the present study, primarily affected more central retinal locations. Thus, it may be that the full-field ERG parameters were relatively unaffected, because the damage was localized, relatively central, and limited to few cells in the total GC population. Had the damage included a larger proportion of GCs, the full-field ERG may have manifested greater change. 
An emerging corollary is that the residual OPs in the slow 7F mfERG responses observed after GC degeneration (Fig. 6) and/or after TTX injection (Fig. 8) probably represent the dominant contributors to OPs in the photopic full-field fERG. For example, OPs of photopic full-field ERGs acquired under the same conditions as in the present study were reduced by only 10% to 25% after TTX injection. 42 Rangaswamy et al. 47 found that most, but not all of the oscillatory components of mfERG responses to a similarly slowed stimulus (14F) were TTX sensitive. Those that were TTX sensitive were distributed primarily within the central and near temporal retina. Taken together, these results support the hypothesis that TTX-sensitive oscillations of slow-sequence mfERGs contribute little to the photopic full-field ERG OPs. If the photopic full-field ERG OPs are dominated by non-TTX-sensitive generators, but perhaps rather by local amacrine cell inhibitory potentials, then they are less likely to be affected by GC loss. 
For the same reasons, decreased N1 and P1 amplitudes, as well as the small increase in the P1 implicit time of 7F responses, observed in this study, both after axotomy and TTX injection, may have been weakly reflected in their photopic full-field ERG counterparts. Generally, the effects of TTX injection, experimental glaucoma, and human glaucoma on photopic full-field ERG a- and b-waves are qualitatively similar to the 7F mfERG changes found in the present study but are smaller in magnitude, despite marked changes reported for the PhNR. 42 54 56 57 One study by Hare and Ton, 35 however, demonstrated more prominent effects of TTX on the monkey ERG b-wave and OPs for strobe flashes delivered without an adapting background. Perhaps their effect was larger because the stimulus was more concentrated on central retina and/or because the background adaptation level was different. 35 The results of these studies suggest that only relatively large GC lesions would be expected to cause significant alterations of the full-field, photopic ERG a-wave, b-wave, or OPs. 
Alternatively, it is possible that some of the decrease in local 7F response LFC amplitudes observed during the final weeks (e.g., N1 and P1) represent functional loss secondary to ensuing transsynaptic degenerative changes within the inner nuclear layer. Evidence against this, however, derives from the lack of significant changes observed in the LFCs of standard fast m-sequence responses (Figs. 5 6) . 58 Perhaps then, the differences between 7F response alterations observed after axotomy and those observed after TTX injection represent contributions of nonspiking GC activity. 
That the slowed 7F mfERG stimulus served to detect GC damage in this model better than the standard fast m-sequence and the MOFO stimulus may have to do with the relatively larger OPs produced by the slowed stimulus. Rangaswamy et al. 47 reported that the OPs were even larger for 14F than for the 7F stimulation. Both their study and the present findings demonstrate that most, but not all of the OPs produced under slowed multifocal stimulation are eliminated by TTX and that the TTX-sensitive OPs are the primary source of nasal-temporal waveform variation in these mfERGs. In the present study, GC axotomy also reduced the large central OPs and nasal-temporal asymmetry. Another group has suggested that such nasal-temporal asymmetry in the distribution of OPs for moderately slowed mfERG stimulation (3F) is produced mainly by interactions between local oscillatory wavelets and those belonging to an ONHC. 59 The results of the current study and in the one by Rangaswamy et al. 47 support the idea that the large TTX-sensitive OPs observed in central and temporal retinal mfERG responses to slowed stimuli reflect the presence of a relatively large ONHC. Thus, it may be that the slowed mfERG stimuli, which produce the largest OPs and the greatest nasal-temporal variation, show the best ability to detect GC damage, as the ONHC is thought to be directly dependent on GC spiking activity. 34 50 The same logic has been offered to explain why MOFO stimulation did better than standard fast m-sequence stimulation for detection of damage in human and experimental glaucoma. 27 31 Indeed, enhancement of the ONHC was the original motivation behind development of these alternative stimulus paradigms. 20 45  
As for why the OPs are so much larger in response to the slowed mfERG sequences, it may be that the lower time-averaged luminance level allows for enhanced rod-cone interactions, which are suggested to be an important factor in the generation of mfERG OPs. 60 Evaluation of this hypothesis and of the effects of different adaptation states on mfERG OPs awaits further study. The evidence presented in this study suggests that slowed sequence stimulation may be a better tool than the other mfERG stimuli for evaluation of GC function in human and experimental glaucoma. 
 
Figure 1.
 
(A) Fundus photograph showing arcuate NFL defect (between arrowheads). (B) Montage of color photograph and fluorescein angiograph, with mfERG stimulus array overlay. Stimulus-response locations analyzed are numbered, with the position of fovea and center of stimulus array indicated by the cross. Retinal wholemount with double-fluorescence stain for neurofilament (C) and GFAP (E) showing foveal-side border of NFL defect: green: axons; red: glia. (G) Overlay of (C) and (E). (D) Horizontal section through the optic nerve head stained with H&E. (F) Transverse section through the optic nerve approximately 2 mm behind the eye stained for GFAP. Local activation of glial cells was limited to bundles of degenerating axons (right) with normal appearance in the adjacent area (left). (H) Transverse section through the optic nerve approximately 2 mm behind the eye stained with toluidine blue, showing gliosis with phagocytosis of axonal debris (right) and normal appearance in the adjacent area (left).
Figure 1.
 
(A) Fundus photograph showing arcuate NFL defect (between arrowheads). (B) Montage of color photograph and fluorescein angiograph, with mfERG stimulus array overlay. Stimulus-response locations analyzed are numbered, with the position of fovea and center of stimulus array indicated by the cross. Retinal wholemount with double-fluorescence stain for neurofilament (C) and GFAP (E) showing foveal-side border of NFL defect: green: axons; red: glia. (G) Overlay of (C) and (E). (D) Horizontal section through the optic nerve head stained with H&E. (F) Transverse section through the optic nerve approximately 2 mm behind the eye stained for GFAP. Local activation of glial cells was limited to bundles of degenerating axons (right) with normal appearance in the adjacent area (left). (H) Transverse section through the optic nerve approximately 2 mm behind the eye stained with toluidine blue, showing gliosis with phagocytosis of axonal debris (right) and normal appearance in the adjacent area (left).
Figure 2.
 
(A) Fluorescein angiograph showing outline of laser burn, with intact retinal vascular pattern and choroidal filing within the areas distal to the burn in the temporal retina. (B) Transverse section through the optic nerve approximately 2 mm behind the eye stained with phenylenediamine. Bundles of degenerating axons localized to a wedge-shaped area (between arrows) with normal appearance in adjacent areas. (C) Vertical cross section through the retina sampled from the location approximated by the vertical line in (A). The position of the horizontal raphe approximately 1 mm temporal to the fovea is shown (dark arrow). Marked thinning of the NFL and loss of GC bodies occurred in the area corresponding to the NFL defect (arrowheads) with normal appearance of distal retinal layers.
Figure 2.
 
(A) Fluorescein angiograph showing outline of laser burn, with intact retinal vascular pattern and choroidal filing within the areas distal to the burn in the temporal retina. (B) Transverse section through the optic nerve approximately 2 mm behind the eye stained with phenylenediamine. Bundles of degenerating axons localized to a wedge-shaped area (between arrows) with normal appearance in adjacent areas. (C) Vertical cross section through the retina sampled from the location approximated by the vertical line in (A). The position of the horizontal raphe approximately 1 mm temporal to the fovea is shown (dark arrow). Marked thinning of the NFL and loss of GC bodies occurred in the area corresponding to the NFL defect (arrowheads) with normal appearance of distal retinal layers.
Figure 3.
 
Schematic of mfERG stimulus sequences. Two steps of the m-sequence are shown for each of the three stimuli: (A) standard fast m-sequence; (B) MOFO; and (C) slowed 7F sequence.
Figure 3.
 
Schematic of mfERG stimulus sequences. Two steps of the m-sequence are shown for each of the three stimuli: (A) standard fast m-sequence; (B) MOFO; and (C) slowed 7F sequence.
Figure 4.
 
Amplitude of pERG (left column) and photopic full-field ERG (middle and right columns) versus time. Arrows along the abscissa in the bottom right panel indicate time points of red (LTxr) and green (LTxg) laser treatment sessions. Baseline (BL) functional measures were acquired before any laser treatment. The normal average and range (mean ± 2 SD) for each parameter are indicated by the solid line and shaded zone, respectively. Dashed line: lower limit for detection of significant change from the baseline value (i.e., 95% limit of agreement for intersession reliability).
Figure 4.
 
Amplitude of pERG (left column) and photopic full-field ERG (middle and right columns) versus time. Arrows along the abscissa in the bottom right panel indicate time points of red (LTxr) and green (LTxg) laser treatment sessions. Baseline (BL) functional measures were acquired before any laser treatment. The normal average and range (mean ± 2 SD) for each parameter are indicated by the solid line and shaded zone, respectively. Dashed line: lower limit for detection of significant change from the baseline value (i.e., 95% limit of agreement for intersession reliability).
Figure 5.
 
(A) mfERG results from final recording session for all three stimuli; full response arrays. Gray ovals: approximate size and position of the blind spot. (B) Individual records from locations within rings 2 and 3 (see Fig. 1B ). Solid traces: records for the final mfERG session; dashed traces: pretreatment baseline records for the fast m-sequence and MOFO stimuli.
Figure 5.
 
(A) mfERG results from final recording session for all three stimuli; full response arrays. Gray ovals: approximate size and position of the blind spot. (B) Individual records from locations within rings 2 and 3 (see Fig. 1B ). Solid traces: records for the final mfERG session; dashed traces: pretreatment baseline records for the fast m-sequence and MOFO stimuli.
Figure 6.
 
(A) mfERGs for 7F stimulus separated into HFCs (left) and LFCs (right). (BD) Average HFC and LFC records for upper field group (thin dashed traces; locations 4 to 6, ring 2, and 1 to 5, ring 3) and lower field group (solid bold traces; locations 8 to 10, ring 2, and 7 to 11, ring 3) for 7F (B), MOFO (C), and standard fast stimulation (D). Comparison between records in (D) and prelaser baseline records (thin gray traces) for standard fast stimulation (E): upper field group (U), lower field group (L). Comparison between records from (C) and prelaser baseline records (thin gray traces) for MOFO stimulus (F).
Figure 6.
 
(A) mfERGs for 7F stimulus separated into HFCs (left) and LFCs (right). (BD) Average HFC and LFC records for upper field group (thin dashed traces; locations 4 to 6, ring 2, and 1 to 5, ring 3) and lower field group (solid bold traces; locations 8 to 10, ring 2, and 7 to 11, ring 3) for 7F (B), MOFO (C), and standard fast stimulation (D). Comparison between records in (D) and prelaser baseline records (thin gray traces) for standard fast stimulation (E): upper field group (U), lower field group (L). Comparison between records from (C) and prelaser baseline records (thin gray traces) for MOFO stimulus (F).
Figure 7.
 
Average amplitude of 7F response components, HFCs, N1, P1, and N2 versus time for upper and lower field locations (same groupings as Fig. 6 ). Data points (○) plot amplitudes at the 48th, 53rd, and 66th weeks (left to right). Shaded areas: normal amplitude ranges (•) statistical significance (P < 0.01). Top left, horizontal dotted line: mean normal amplitude of noise.
Figure 7.
 
Average amplitude of 7F response components, HFCs, N1, P1, and N2 versus time for upper and lower field locations (same groupings as Fig. 6 ). Data points (○) plot amplitudes at the 48th, 53rd, and 66th weeks (left to right). Shaded areas: normal amplitude ranges (•) statistical significance (P < 0.01). Top left, horizontal dotted line: mean normal amplitude of noise.
Figure 8.
 
(A) mfERGs for 7F stimulus before (control, left column) and after TTX injection (middle) and the difference traces (right). (B) Band-pass filtered records from (A) showing HFCs of control responses (left column), responses after TTX injection (middle), and difference records (right). (C) Average HFC records for upper and lower field locations (same groupings as Fig. 6 ) before (left) and after TTX injection (right).
Figure 8.
 
(A) mfERGs for 7F stimulus before (control, left column) and after TTX injection (middle) and the difference traces (right). (B) Band-pass filtered records from (A) showing HFCs of control responses (left column), responses after TTX injection (middle), and difference records (right). (C) Average HFC records for upper and lower field locations (same groupings as Fig. 6 ) before (left) and after TTX injection (right).
Airaksinen, PJ, Drance, SM, Douglas, GR, Mawson, DK, Nieminen, H. (1984) Diffuse and localized nerve fiber loss in glaucoma Am J Ophthalmol 98,566-571 [CrossRef] [PubMed]
Lachenmayr, BJ, Airaksinen, PJ, Drance, SM, Wijsman, K. (1991) Correlation of retinal nerve-fiber-layer loss, changes at the optic nerve head and various psychophysical criteria in glaucoma Graefes Arch Clin Exp Ophthalmol 229,133-138 [CrossRef] [PubMed]
Lachenmayr, BJ, Drance, SM, Airaksinen, PJ. (1992) Diffuse field loss and diffuse retinal nerve-fiber loss in glaucoma Ger J Ophthalmol 1,22-25 [PubMed]
Henson, DB, Artes, PH, Chauhan, BC. (1999) Diffuse loss of sensitivity in early glaucoma Invest Ophthalmol Vis Sci 40,3147-3151 [PubMed]
Heijl, A, Patella, VM. (2002) Essential Perimetry 3rd ed. Carl Zeiss Meditec, Inc. Dublin, CA.
Johnson, CA, Sample, PA, Cioffi, GA, Liebmann, JR, Weinreb, RN. (2002) Structure and function evaluation (SAFE): I. criteria for glaucomatous visual field loss using standard automated perimetry (SAP) and short wavelength automated perimetry (SWAP) Am J Ophthalmol 134,177-185 [CrossRef] [PubMed]
Hoyt, WF, Frisen, L, Newman, NM. (1973) Fundoscopy of nerve fiber layer defects in glaucoma Invest Ophthalmol 12,814-829 [PubMed]
Sommer, A, Katz, J, Quigley, HA, et al (1991) Clinically detectable nerve fiber atrophy precedes the onset of glaucomatous field loss Arch Ophthalmol 109,77-83 [CrossRef] [PubMed]
Quigley, HA, Enger, C, Katz, J, et al (1994) Risk factors for the development of glaucomatous visual field loss in ocular hypertension Arch Ophthalmol 112,644-649 [CrossRef] [PubMed]
Airaksinen, PJ, Drance, SM, Douglas, GR, Schulzer, M, Wijsman, K. (1985) Visual field and retinal nerve fiber layer comparisons in glaucoma Arch Ophthalmol 103,205-207 [CrossRef] [PubMed]
Minckler, DS. (1980) The organization of nerve fiber bundles in the primate optic nerve head Arch Ophthalmol 98,1630-1636 [CrossRef] [PubMed]
Radius, RL, Anderson, DR. (1979) The course of axons through the retina and optic nerve head Arch Ophthalmol 97,1154-1158 [CrossRef] [PubMed]
Quigley, HA, Green, WR. (1979) The histology of human glaucoma cupping and optic nerve damage: clinicopathologic correlation in 21 eyes Ophthalmology 86,1803-1830 [CrossRef] [PubMed]
Quigley, HA, Addicks, EM. (1981) Regional differences in the structure of the lamina cribrosa and their relation to glaucomatous optic nerve damage Arch Ophthalmol 99,137-143 [CrossRef] [PubMed]
Bellezza, AJ, Rintalan, CJ, Thompson, HW, et al (2003) Deformation of the lamina cribrosa and anterior scleral canal wall in early experimental glaucoma Invest Ophthalmol Vis Sci 44,623-637 [CrossRef] [PubMed]
Drance, SM, Fairclough, M, Butler, DM, Kottler, MS. (1977) The importance of disc hemorrhage in the prognosis of chronic open angle glaucoma Arch Ophthalmol 95,226-228 [CrossRef] [PubMed]
Sonnsjo, B, Dokmo, Y, Krakau, T. (2002) Disc haemorrhages, precursors of open angle glaucoma Prog Retinal Eye Res 21,35-56 [CrossRef]
Sutter, EE. (1991) The fast m-transform: a fast computation of cross-correlations with binary m-sequences SIAM J Comput 20,686-694 [CrossRef]
Sutter, EE, Tran, D. (1992) The field topography of ERG components in man: I. The photopic luminance response Vision Res 32,433-446 [CrossRef] [PubMed]
Sutter, EE, Bearse, MA, Jr, Stamper, RL, et al (2001) Monitoring retinal ganglion cell function with the mERG: recent advances Vision Science and Its Applications, OSA Technical Digest Series 1 ,10-13 Optical Society of America Washington, DC.
Sutter, EE. (2001) Imaging visual function with the multifocal m-sequence technique Vision Res 41,1241-1255 [CrossRef] [PubMed]
Hood, DC, Greenstein, VC, Holopigian, K, et al (2000) An attempt to detect glaucomatous damage to the inner retina with the multifocal ERG Invest Ophthalmol Vis Sci 41,1570-1579 [PubMed]
Palmowski, AM, Allgayer, R, Heinemann-Vemaleken, B. (2000) 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 101,35-49 [CrossRef] [PubMed]
Klistorner, AI, Graham, SL, Martins, A. (2000) Multifocal pattern electroretinogram does not demonstrate localised field defects in glaucoma Doc Ophthalmol 100,155-165 [CrossRef] [PubMed]
Hasegawa, S, Takagi, M, Usui, T, Takada, R, Abe, H. (2000) Waveform changes of the first-order multifocal electroretinogram in patients with glaucoma Invest Ophthalmol Vis Sci 41,1597-1603 [PubMed]
Fortune, B, Johnson, CA, Cioffi, GA. (2001) The topographic relationship between multifocal electroretinographic and behavioral perimetric measures of function in glaucoma Optom Vis Sci 78,206-214 [PubMed]
Fortune, B, Bearse, MA, Jr, Cioffi, GA, Johnson, CA. (2002) Selective loss of an oscillatory component from temporal retinal multifocal ERG responses in glaucoma Invest Ophthalmol Vis Sci 43,2638-2647 [PubMed]
Frishman, LJ, Saszik, S, Harwerth, RS, et al (2000) Effects of experimental glaucoma in macaques on the multifocal ERG in laser-induced glaucoma Doc Ophthalmol 100,231-251 [CrossRef] [PubMed]
Hare, WA, Ton, H, Ruiz, G, et al (2001) Characterization of retinal injury using ERG measures obtained with both conventional and multifocal methods in chronic ocular hypertensive primates Invest Ophthalmol Vis Sci 42,127-136 [PubMed]
Raz, D, Seeliger, MW, Geva, AB, et al (2002) The effect of contrast and luminance on mfERG responses in a monkey model of glaucoma Invest Ophthalmol Vis Sci 43,2027-2035 [PubMed]
Fortune, B, Cull, G, Wang, L, Van Buskirk, EM, Cioffi, GA. (2002) Factors affecting the use of multifocal electroretinography to monitor function in a primate model of glaucoma Doc Ophthalmol 105,151-178 [CrossRef] [PubMed]
Hood, DC, Greenstein, V, Frishman, L, et al (1999) Identifying inner retinal contributions to the human multifocal ERG Vision Res 39,2285-2291 [CrossRef] [PubMed]
Hood, DC, Frishman, LJ, Viswanathan, S, Robson, JG, Ahmed, J. (1999) Evidence for a ganglion cell contribution to the primate electroretinogram (ERG): effects of TTX on the multifocal ERG in macaque Vis Neurosci 16,411-416 [PubMed]
Hood, DC, Bearse, MA, Jr, Sutter, EE, Viswanathan, S, Frishman, LJ. (2001) The optic nerve head component of the monkey’s (Macaca mulatta) multifocal electroretinogram (mERG) Vision Res 41,2029-2041 [CrossRef] [PubMed]
Hare, WA, Ton, H. (2002) Effects of APB, PDA, and TTX on ERG responses recorded using both multifocal and conventional methods in monkey: effects of APB, PDA, and TTX on monkey ERG responses Doc Ophthalmol 105,189-222 [CrossRef] [PubMed]
Frisch, GD, Shawaluk, PD, Adams, DO. (1974) Remote nerve fibre bundle alterations in the retina as caused by argon laser photocoagulation Nature 248,433-435 [CrossRef] [PubMed]
Radius, RL, Anderson, DR. (1978) Retinal ganglion cell degeneration in experimental optic atrophy Am J Ophthalmol 86,673-679 [CrossRef] [PubMed]
Radius, RL, Anderson, DR. (1979) The histology of retinal nerve fiber layer bundles and bundle defects Arch Ophthalmol 97,948-950 [CrossRef] [PubMed]
Mae, Y, Matsui, M. (1982) Ophthalmoscopic model of nerve fiber bundle defects in monkey eyes produced by intraocular retinal coagulation Jpn J Ophthalmol 26,274-281 [PubMed]
Dawson, WW, Stratton, RD, Hope, GM, et al (1986) Tissue responses of the monkey retina: tuning and dependence on inner layer integrity Invest Ophthalmol Vis Sci 27,734-745 [PubMed]
Bui, BV, Fortune, B, Cull, G, Wang, L, Cioffi, GA. () Optimum recording conditions, inter-ocular variability and inter-session reliability of the monkey transient pattern electroretinogram Exp Eye Res In press
Fortune, B, Bui, BV, Cull, G, Wang, L, Cioffi, GA. () Inter-ocular and inter-session reliability of the electroretinogram photopic negative response (PhNR) in non-human primates Exp Eye Res Submitted
Apple, DJ, Wyhinny, GJ, Goldberg, MF, Polley, EH. (1976) Experimental argon laser photocoagulation. II. Effects on the optic disc Arch Ophthalmol 94,296-304 [CrossRef] [PubMed]
Apple, DJ, Wyhinny, GJ, Goldberg, MF, Polley, EH, Bizzell, JW. (1976) Experimental argon laser photocoagulation. I. Effects on retinal nerve fiber layer Arch Ophthalmol 94,137-144 [CrossRef] [PubMed]
Sutter, EE, Shimada, Y, Bearse, MA, Jr (1999) Mapping Inner Retinal Function through Enhancement of Adaptive Components in the m-ERG ,52-55 Optical Society of America Washington, DC.
Hood, DC, Seiple, W, Holopigian, K, Greenstein, V. (1997) A comparison of the components of the multifocal and full-field ERGs Vis Neurosci 14,533-544 [CrossRef] [PubMed]
Rangaswamy, NV, Hood, DC, Frishman, L. (2003) Regional variations in the local contributions to the primate photopic flash ERG revealed using the slow-sequence mfERG Invest Ophthalmol Vis Sci 44,3233-3247 [CrossRef] [PubMed]
Sieving, PA, Murayama, K, Naarendorp, F. (1994) Push-pull model of the primate photopic electroretinogram: a role for hyperpolarizing neurons in shaping the b-wave Vis Neurosci 11,519-532 [CrossRef] [PubMed]
Morgan, JE, Jeffery, G, Foss, AJ. (1998) Axon deviation in the human lamina cribrosa Br J Ophthalmol 82,680-683 [CrossRef] [PubMed]
Sutter, EE, Bearse, MA, Jr (1999) The optic nerve head component of the human ERG Vision Res 39,419-436 [CrossRef] [PubMed]
Hille, B. (1992) Ionic Channels of Excitable Membranes 2nd ed. Sinauer Associates, Inc. Sunderland, MA.
Dowling, JE. (1987) The Retina: An Approachable Part of the Brain Belknap-Harvard Cambridge, MA.
Garway-Heath, DF, Holder, GE, Fitzke, FW, Hitchings, RA. (2002) Relationship between electrophysiological, psychophysical, and anatomical measurements in glaucoma Invest Ophthalmol Vis Sci 43,2213-2220 [PubMed]
Viswanathan, S, Frishman, LJ, Robson, JG. (2000) The uniform field and pattern ERG in macaques with experimental glaucoma: removal of spiking activity Invest Ophthalmol Vis Sci 41,2797-2810 [PubMed]
Holder, GE. (2001) Pattern electroretinography (PERG) and an integrated approach to visual pathway diagnosis Prog Retinal Eye Res 20,531-561 [CrossRef]
Viswanathan, S, Frishman, LJ, Robson, JG, Harwerth, RS, Smith, EL, III (1999) The photopic negative response of the macaque electroretinogram: reduction by experimental glaucoma Invest Ophthalmol Vis Sci 40,1124-1136 [PubMed]
Viswanathan, S, Frishman, LJ, Robson, JG, Walters, JW. (2001) The photopic negative response of the flash electroretinogram in primary open angle glaucoma Invest Ophthalmol Vis Sci 42,514-522 [PubMed]
Hood, DC, Frishman, LJ, Saszik, S, Viswanathan, S. (2002) Retinal origins of the primate multifocal ERG: implications for the human response Invest Ophthalmol Vis Sci 43,1673-1685 [PubMed]
Bearse, MA, Jr, Shimada, Y, Sutter, EE. (2000) Distribution of oscillatory components in the central retina Doc Ophthalmol 100,185-205 [CrossRef] [PubMed]
Wu, S, Sutter, EE. (1995) A topographic study of oscillatory potentials in man Vis Neurosci 12,1013-1025 [CrossRef] [PubMed]
Figure 1.
 
(A) Fundus photograph showing arcuate NFL defect (between arrowheads). (B) Montage of color photograph and fluorescein angiograph, with mfERG stimulus array overlay. Stimulus-response locations analyzed are numbered, with the position of fovea and center of stimulus array indicated by the cross. Retinal wholemount with double-fluorescence stain for neurofilament (C) and GFAP (E) showing foveal-side border of NFL defect: green: axons; red: glia. (G) Overlay of (C) and (E). (D) Horizontal section through the optic nerve head stained with H&E. (F) Transverse section through the optic nerve approximately 2 mm behind the eye stained for GFAP. Local activation of glial cells was limited to bundles of degenerating axons (right) with normal appearance in the adjacent area (left). (H) Transverse section through the optic nerve approximately 2 mm behind the eye stained with toluidine blue, showing gliosis with phagocytosis of axonal debris (right) and normal appearance in the adjacent area (left).
Figure 1.
 
(A) Fundus photograph showing arcuate NFL defect (between arrowheads). (B) Montage of color photograph and fluorescein angiograph, with mfERG stimulus array overlay. Stimulus-response locations analyzed are numbered, with the position of fovea and center of stimulus array indicated by the cross. Retinal wholemount with double-fluorescence stain for neurofilament (C) and GFAP (E) showing foveal-side border of NFL defect: green: axons; red: glia. (G) Overlay of (C) and (E). (D) Horizontal section through the optic nerve head stained with H&E. (F) Transverse section through the optic nerve approximately 2 mm behind the eye stained for GFAP. Local activation of glial cells was limited to bundles of degenerating axons (right) with normal appearance in the adjacent area (left). (H) Transverse section through the optic nerve approximately 2 mm behind the eye stained with toluidine blue, showing gliosis with phagocytosis of axonal debris (right) and normal appearance in the adjacent area (left).
Figure 2.
 
(A) Fluorescein angiograph showing outline of laser burn, with intact retinal vascular pattern and choroidal filing within the areas distal to the burn in the temporal retina. (B) Transverse section through the optic nerve approximately 2 mm behind the eye stained with phenylenediamine. Bundles of degenerating axons localized to a wedge-shaped area (between arrows) with normal appearance in adjacent areas. (C) Vertical cross section through the retina sampled from the location approximated by the vertical line in (A). The position of the horizontal raphe approximately 1 mm temporal to the fovea is shown (dark arrow). Marked thinning of the NFL and loss of GC bodies occurred in the area corresponding to the NFL defect (arrowheads) with normal appearance of distal retinal layers.
Figure 2.
 
(A) Fluorescein angiograph showing outline of laser burn, with intact retinal vascular pattern and choroidal filing within the areas distal to the burn in the temporal retina. (B) Transverse section through the optic nerve approximately 2 mm behind the eye stained with phenylenediamine. Bundles of degenerating axons localized to a wedge-shaped area (between arrows) with normal appearance in adjacent areas. (C) Vertical cross section through the retina sampled from the location approximated by the vertical line in (A). The position of the horizontal raphe approximately 1 mm temporal to the fovea is shown (dark arrow). Marked thinning of the NFL and loss of GC bodies occurred in the area corresponding to the NFL defect (arrowheads) with normal appearance of distal retinal layers.
Figure 3.
 
Schematic of mfERG stimulus sequences. Two steps of the m-sequence are shown for each of the three stimuli: (A) standard fast m-sequence; (B) MOFO; and (C) slowed 7F sequence.
Figure 3.
 
Schematic of mfERG stimulus sequences. Two steps of the m-sequence are shown for each of the three stimuli: (A) standard fast m-sequence; (B) MOFO; and (C) slowed 7F sequence.
Figure 4.
 
Amplitude of pERG (left column) and photopic full-field ERG (middle and right columns) versus time. Arrows along the abscissa in the bottom right panel indicate time points of red (LTxr) and green (LTxg) laser treatment sessions. Baseline (BL) functional measures were acquired before any laser treatment. The normal average and range (mean ± 2 SD) for each parameter are indicated by the solid line and shaded zone, respectively. Dashed line: lower limit for detection of significant change from the baseline value (i.e., 95% limit of agreement for intersession reliability).
Figure 4.
 
Amplitude of pERG (left column) and photopic full-field ERG (middle and right columns) versus time. Arrows along the abscissa in the bottom right panel indicate time points of red (LTxr) and green (LTxg) laser treatment sessions. Baseline (BL) functional measures were acquired before any laser treatment. The normal average and range (mean ± 2 SD) for each parameter are indicated by the solid line and shaded zone, respectively. Dashed line: lower limit for detection of significant change from the baseline value (i.e., 95% limit of agreement for intersession reliability).
Figure 5.
 
(A) mfERG results from final recording session for all three stimuli; full response arrays. Gray ovals: approximate size and position of the blind spot. (B) Individual records from locations within rings 2 and 3 (see Fig. 1B ). Solid traces: records for the final mfERG session; dashed traces: pretreatment baseline records for the fast m-sequence and MOFO stimuli.
Figure 5.
 
(A) mfERG results from final recording session for all three stimuli; full response arrays. Gray ovals: approximate size and position of the blind spot. (B) Individual records from locations within rings 2 and 3 (see Fig. 1B ). Solid traces: records for the final mfERG session; dashed traces: pretreatment baseline records for the fast m-sequence and MOFO stimuli.
Figure 6.
 
(A) mfERGs for 7F stimulus separated into HFCs (left) and LFCs (right). (BD) Average HFC and LFC records for upper field group (thin dashed traces; locations 4 to 6, ring 2, and 1 to 5, ring 3) and lower field group (solid bold traces; locations 8 to 10, ring 2, and 7 to 11, ring 3) for 7F (B), MOFO (C), and standard fast stimulation (D). Comparison between records in (D) and prelaser baseline records (thin gray traces) for standard fast stimulation (E): upper field group (U), lower field group (L). Comparison between records from (C) and prelaser baseline records (thin gray traces) for MOFO stimulus (F).
Figure 6.
 
(A) mfERGs for 7F stimulus separated into HFCs (left) and LFCs (right). (BD) Average HFC and LFC records for upper field group (thin dashed traces; locations 4 to 6, ring 2, and 1 to 5, ring 3) and lower field group (solid bold traces; locations 8 to 10, ring 2, and 7 to 11, ring 3) for 7F (B), MOFO (C), and standard fast stimulation (D). Comparison between records in (D) and prelaser baseline records (thin gray traces) for standard fast stimulation (E): upper field group (U), lower field group (L). Comparison between records from (C) and prelaser baseline records (thin gray traces) for MOFO stimulus (F).
Figure 7.
 
Average amplitude of 7F response components, HFCs, N1, P1, and N2 versus time for upper and lower field locations (same groupings as Fig. 6 ). Data points (○) plot amplitudes at the 48th, 53rd, and 66th weeks (left to right). Shaded areas: normal amplitude ranges (•) statistical significance (P < 0.01). Top left, horizontal dotted line: mean normal amplitude of noise.
Figure 7.
 
Average amplitude of 7F response components, HFCs, N1, P1, and N2 versus time for upper and lower field locations (same groupings as Fig. 6 ). Data points (○) plot amplitudes at the 48th, 53rd, and 66th weeks (left to right). Shaded areas: normal amplitude ranges (•) statistical significance (P < 0.01). Top left, horizontal dotted line: mean normal amplitude of noise.
Figure 8.
 
(A) mfERGs for 7F stimulus before (control, left column) and after TTX injection (middle) and the difference traces (right). (B) Band-pass filtered records from (A) showing HFCs of control responses (left column), responses after TTX injection (middle), and difference records (right). (C) Average HFC records for upper and lower field locations (same groupings as Fig. 6 ) before (left) and after TTX injection (right).
Figure 8.
 
(A) mfERGs for 7F stimulus before (control, left column) and after TTX injection (middle) and the difference traces (right). (B) Band-pass filtered records from (A) showing HFCs of control responses (left column), responses after TTX injection (middle), and difference records (right). (C) Average HFC records for upper and lower field locations (same groupings as Fig. 6 ) before (left) and after TTX injection (right).
×
×

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.

×