MULTIFOCAL VISUAL EVOKED POTENTIAL IN OPTIC NEUROPATHIES AND HOMONYMOUS HEMIANOPIAS

MICHAEL WALL, KIMBERLY WOODWARD and TODD SLEEP

Departments of Neurology and Ophthalmology, Veterans Administration Medical Center, and University of Iowa, College of Medicine, Iowa City, IA, USA

Abstract

Purpose: To validate multifocal visual evoked potential (mfVEP) testing in patients with neuroophthalmological disorders. Methods: We tested 33 normals and 20 patients with documented lesions of the optic nerve and retrochiasmatic visual pathways. Fifteen of the patients had optic neuropathies and five had hemianopias. All patients had Humphrey SITA 24-2 testing and the mfVEP using the Opera mfVEP system. Visual field defects were defined for conventional automated perimetry as having three contiguous abnormal test locations at the p < 0.05 level or worse in a clinically suspicious area. Based on the testing of normals, all mfVEP results used the same criteria except abnormality at the p < 0.01 level was used. Amplitude information of three types was used for the mfVEP analysis: 1. signal of less than 90 nV; 2. probability plot data; and 3. intereye asymmetry comparison. Results: Using three contiguous abnormal test locations at p < 0.01, 12-18% of normals have defects with mfVEP. Conventional automated perimetry in normals showed 15% to have visual field defects. We found that mfVEP showed similar deficits to conventional automated perimetry in the non-demyelinating optic neuropathy patients. In recovered optic neuritis patients, mfVEP was superior. Three of five patients with hemianopias were missed by mfVEP. Conclusions: In non-demyelinating optic neuropathies, mfVEP results correlate very well with perimetry. In demyelinating optic neuropathies, mfVEP is a very sensitive test. MfVEP did not appear to be a good test for detecting hemianopias in our small sample. It appears that a larger database of normals is needed to better separate patients with visual field defects from normal subjects.

Introduction

Many attempts have been made to investigate visual field function using evoked potentials to visual stimuli. Evoked responses to small flashes of light at discrete test locations were attempted in the 1960s.1,2 The small responses found were of little clinical utility. Likewise, when the much lower amplitude responses to pattern shift stimulation were tried at discrete visual field loci, little useful information could be gleaned outside 10°. Half-field and quadrantic stimuli were used to record checkerboard pattern shift potentials, but the large stimulus sizes allowed only gross visual field defects to be documented thereby adding little to bedside visual field examina- tion.3-5

Address for correspondence: Michael Wall, MD, University of Iowa, College of Medicine, Department of Neurology, 200 Hawkins Drive #2007 RCP, Iowa City, IA 52242-1053, USA

Perimetry Update 2002/2003, pp. 265–274

Proceedings of the XVth International Perimetric Society Meeting, Stratford-upon-Avon, England, June 26–29, 2002

edited by David B. Henson and Michael Wall

© 2004 Kugler Publications, The Hague, The Netherlands

2 .6°

9 .8°

2 2 .2°

Fig. 1. The mfVEP stimulus is displayed on the left; the typical waveform results on the right.

A major breakthrough occurred when Sutter et al. used a 60-sector scaled checkerboard set of patches with the stimuli presented in a pseudorandom m-sequence.6,7 Using these sequences with cross-correlation techniques, electrical responses to pattern reversal stimuli can be extracted from occipital scalp recordings. Each of the 60 sectors in the central 22° of the visual field was examined.

Hood et al.8 and Klistorner and Graham9 have shown that mfVEPs generated by the two eyes of a single subject are usually nearly indistinguishable, except for small responses along the midline in the nasal field.8 This takes advantage of the fact that information from each eye synapses in the same homonymous location in visual cortex. When applied to patients, the interocular differences in mfVEP amplitude are greater in areas of visual field damage, thus allowing detection of monocular and nonhomonymous binocular visual field damage. However, detection of binocular homonymous damage can be problematic using this technique.

In 30 patients with established glaucoma, Klistorner and Graham noted good concordance of visual field defects found by conventional automated perimetry and with mfVEP.10 Their criteria for a defect with mfVEP were loss of signal amplitude to less than 120 nV in at least three adjacent points in the matching area. The defects were also detected using loss greater than 1.96 SD from the responses of healthy volunteers (empiric probability plots). Only one healthy volunteer had a visual field defect when the criteria were reapplied to the normals. Both the number and location of the defects were highly correlated between the two tests. This use of probability plots should allow detection of homonymous hemianopias and bilateral optic nerve damage. However, due to the large intersubject variability, clinical study is needed to determine whether this technique has merit for these clinical situations.

Little work has been done using mfVEP technique in patients with neuro-ophthal- mological disorders. Our aim was to test a series of these patients in order to determine the sensitivity and specificity of the technique.

Methods

Subjects

The visual testing protocol was approved by The University of Iowa Investigational Review Board. The tenets of the Declaration of Helsinki were followed. Twenty patients with various neuro-ophthalmological disorders and 33 normal subjects gave informed consent to participate in the study. Most of the normals were paid volunteers

who were hospital employees or friends or family members of eye clinic employees. The mean age of the neuro-ophthalmology patients was 42.4 ± 10.7 years and the normals, 43.8 ± 19.4 years.

Normals were included if they had: 1. no history of eye disease except refractive error (no more optical correction than six diopters of sphere or three diopters of cylinder); 2. no history of diabetes mellitus or systemic arterial hypertension; and 3. a normal ophthalmological examination. All normals were Caucasian; 17 were female and 13 male.

The neuro-ophthalmology patients were selected by a chart review or were asked to participate as part of a scheduled clinic visit. They all fulfilled the following inclusion criteria: a. unequivocal clinical diagnosis; b. reliable Humphrey full threshold visual fields. Patients were excluded if they had: a. any other disease causing visual field loss including background diabetic retinopathy; b. media opacity severe enough to make it difficult to visualize the optic disc. Ten patients had optic neuritis, two had idiopathic intracranial hypertension, three anterior ischemic optic neuropathy, and five homonymous hemianopias.

All subjects underwent perimetry with the Humphrey program SITA 24-2 using target locations fixed on a six-degree-spaced grid of the central 24°. We accepted all results of perimetry and mfVEP testing; normals were not retested if their results were abnormal.

Testing devices

Conventional automated perimetry was performed with the Humphrey Visual Field Analyzer Model 750 using the manufacturer’s recommendations. This test used a Goldmann size III, 4 mm2 spot of light on a 31.5 apostilb background; the size of the target was fixed, and the threshold to differential light intensity was found at each test site. We used the patients’ appropriate near correction and took care to prevent lens rim artifact.

We recorded mfVEP using the Opera (Accumap) system. This uses an alternating dartboard stimulus. Recordings are made from a four-channel EEG electrode channel set up using peak-to-trough amplitudes from 58 test locations. The stimulus of mfVEP is cortically scaled and uses a pseudo-randomly reversing pattern. The Opera system records EEG activity and a fast-Fourier analysis is performed on the data. EKG and alpha rhythm spikes are removed and the EEG is used to compute a scaling factor to adjust mfVEP amplitude so that interindividual differences are lessened. Each of the 58 separate checkerboard patches, with 16 checks each, is modulated in time according to a different m-sequence, such as progression. A cross-correlation technique is then used to extract the signal at each of the sites.

Subjects were seated in a comfortable chair, 30 cm from the stimulus display looking slightly up at the stimulus display in order to relax their neck muscles. The stimulus covered 26° surrounding fixation with two additional test location scaled patches covering 6° at the nasal horizontal. The signal was amplified 100,000 times and band-pass filtered between 3 and 30 Hz using a Grass four-channel amplifier. The sampling rate was 502 Hz.

The actual testing takes ten to 15 minutes per eye, but the setup time varies. Total testing time ranges from 35 to 60 minutes. The test is generally well accepted by the patients. They like not having to make some of the difficult decisions that threshold

268 M. Wall et al.

perimetry requires. Occasional patients prefer the more active involvement of conventional perimetry to staring at a flickering computer monitor. We have not made a formal assessment of patient acceptance.

We generated the visual stimulus on a 21-inch high-resolution monitor (Hitachi, Ltd.) In order to ensure good fixation, the numbers 3, 6, 8, or 9 were displayed in random sequence about every five seconds. The subjects were asked to press a button whenever they saw a number ‘3’. Subjects were not included if they could not perform this task without error.

The Opera system currently uses a database of 100 normal subjects to compute the ranges for normal values. An inter-eye asymmetry value for every stimulus location is calculated by dividing the difference in amplitude between right and left eyes by their sum (relative asymmetry coefficient). Probability plots for asymmetry are provided based on this same database of 100 normals.

Visual field defects

Visual field defects for the two tests were defined as follows. Total deviation pointwise probability plots (Statpac®) were used for the Humphrey SITA testing. We used both the amplitude deviation and relative asymmetry probability plots generated by the Opera system as two separate analyses. As a third analysis, we used the amplitudes of less than 90 nV as abnormal. For conventional automated perimetry three adjacent abnormal points at the p < 0.05 level or two adjacent abnormal points with at least one with a probability of occurrence in a normal population of less than 1% were needed. It was required that the abnormal test locations occurred in a clinically suspicious area and were non-edge points. For mfVEP we required three adjacent abnormal points at the p < 0.01 level. We also required that the abnormal test points occur in a clinically suspicious area (e.g., in a nerve fiber bundle-like pattern). We ignored the superior edge and two far nasal test locations because they were so frequently abnormal in our normal subjects.

Results

The results of our study are presented in Table 1. There were a substantial number of normals meeting the criteria for visual field defects with both conventional automated perimetry (15%) and mfVEP (12-21%). The reason for this high number of normals is unclear. As best we could determine, they did not have visual system disease.

Table 1. Results of conventional automated perimetry and mfVEP testing in 34 normals and 20 patients with neuro-ophthalmological disease

a.

b.

Fig. 3. A patient with optic neuritis has severe loss on an initial visual field examination. a. Fifteen years later, her examination has returned to normal. b. However, her mfVEP shows markers of the earlier loss as signal reduction throughout the visual field.

a.

b.

Fig. 4. A patient with a homonymous hemianopia due to a left temporal lobe oligodendroglioma has good agreement between: a. conventional automated perimetry; and b. mfVEP.

a.

b.

Fig. 5. Another patient with a homonymous hemianopia has a clearly present defect with conventional automated perimetry, which is not detected with mfVEP. a. Conventional automated perimetry; b. mfVEP.

mous hemianopias. We found that a substantial percentage of normals met criteria for defects with both conventional automated perimetry (15%) and mfVEP (12-21%).

In a large general population study of 140 subjects, 17 (or 8%) of those tested had visual field defects that were reproducible or explicable by ocular status or history on

Humphrey 30-2 testing.11 The total percentage of abnormal visual field examination is not clear from the study. In an earlier study, we found that 19% of 42 subjects from a general population had defects.12 A large random sample needs to be tested in order to determine the frequency of visual field defects with conventional automated perimetry using typical clinical criteria, but our two studies with different normative data sets suggest that this number is in the 15-20% range.

With mfVEP, using three contiguous abnormal test locations at p < 0.01, 18% of normals had defects. This dataset was used to determine the criteria for a defect. We suspect that a larger database of normals will be needed for traditional p < 0.05 criteria to be used.

In the patients tested with mfVEP, we found that, in non-demyelinating optic neuropathies, mfVEP results correlate very well with perimetry. In demyelinating optic neuropathies, mfVEP is a sensitive test and showed defects when conventional automated perimetry was normal. We did not evaluate latency measurements in this study, but we would expect the demyelinating optic neuropathies to be abnormal. We missed the patients clearly documented as homonymous hemianopia in three of five subjects tested. We have not yet tested enough patients to determine the mechanism for this failure. With the exception of patients with homonymous hemianopia, mfVEP results were very similar to those of conventional automated perimetry. While mfVEP performs well to detect patients with non-glaucomatous optic neuropathies, our results are limited by the large percentage of normal subjects having defects at the p < 0.01 level. Hopefully, this weakness of the technique should improve with the collection of a larger database of normals.

Acknowledgment

This study was supported by VA Merit Review, an unrestricted grant to the Department of Ophthalmology from Research to Prevent Blindness, New York, NY, USA

References

1.Regan D, Heron JR: Clinical investigation of lesions of the visual pathway: a new objective technique. J Neurol Neurosurg Psychiat 32(5):479-483, 1969

2.Schreinemachers HP, Henkes HP: Relation between localized retinal stimuli and the visual evoked response in man. Ophthalmologica 155:17-27, 1968

3.Maitland CG, Aminoff MJ, Kennard C, Hoyt WF: Evoked potentials in the evaluation of visual field defects due to chiasmal or retrochiasmal lesions. Neurology 32(9):986-991, 1982

4.Streletz LJ, Bae SH, Roeshman RM, Schatz NJ, Savino PJ: Visual evoked potentials in occipital lobe lesions. Arch Neurol 38(2):80-85, 1981

5.Celesia GG, Meredith JT, Pluff K: Perimetry, visual evoked potentials and visual evoked spectrum array in homonymous hemianopsia. Electroencephalogr Clin Neurophysiol 56(1):16-30, 1983

6.Sutter EE: The fast m-transform: a fast computation of cross-correlations with binary m-sequences. Soc Ind Appl Math 20:686-694, 1991

7.Baseler HA, Sutter EE, Klein SA, Carney T: The topography of visual evoked response properties across the visual field. Electroencephalogr Clin Neurophysiol 90(1):65-81, 1994

8.Hood DC, Zhang X, Greenstein VC, Kangovi S, Odel JG, Liebmann JM et al: An interocular comparison of the multifocal VEP: a possible technique for detecting local damage to the optic nerve. Invest Ophthalmol Vis Sci 41(6):1580-1587, 2000

9.Klistorner AI, Graham SL: Multifocal pattern VEP perimetry: analysis of sectoral waveforms. Doc Ophthalmol 98(2):183-196, 1999

10.Klistorner A, Graham SL: Objective perimetry in glaucoma. Ophthalmology 107(12):2283-2299, 2000

11.Heijl A, Lindgren G, Olsson J: Normal variability of static perimetric threshold values across the central visual field. Arch Ophthalmol 105:1544-1549, 1987

12.Wall M, Neahring RK, Woodward KR: Sensitivity and specificity of frequency doubling perimetry in neuro-ophthalmic disorders: a comparison with conventional automated perimetry. Invest Ophthalmol Vis Sci 43(4):1277-1283, 2002