0042-6989/91 $3.00 + 0.00 Copyright © 1991 Pergamon Press pic
Vision Res. Vol. 31, No. II, pp. 1859-1864, 1991 Printed in Great Bntain. All rights reserved
VISUAL EVOKED MAGNETIC FIELDS TO FLASH AND PATTERN IN 100 NORMAL SUBJECTS R.
A. ARMSTRONG, A. SLAVEN
and G. F. A.
HARDING
Clinical Neurophysiology Unit, Department of Vision Sciences, Aston University, Aston Triangle, Birmingham B4 7ET, England (Received 20 June 1990; in revised form 3 January 1991) Abstract-The practicality of recording visual evoked magnetic fields in 100 subjects 15-87 yr of age using a single channel d.c. SQUID second order gradiometer in an unshielded environment was investigated. The pattern reversal response showed a major positive component between 90 and 120 msec (P1OOM) while the response to flash produced a major positive component between 90 and 140 msec (P2M). Latency norms of the P100M were more variable than the corresponding P100 and P2 visual evoked potentials. The latency of the P100M may show a steep increase with age in most subjects after about 55 yr whereas only a small trend of latency with age was detected for the flash P2M. Neuromagnetism
Norms
Flash
Pattern reversal
INTRODUCTION
There have been no studies of the clinical value of visual evoked magnetic fields (VEF). However, there may be advantages to making magnetic recordings which can complement the visual evoked potential (VEP) in clinical diagnosis. First, no electrodes have to be attached to the subject so that it is relatively easy to gather large amounts of data from different subjects. Second, the fields are less smeared by the low conductive skull (Wikswo, 1983; Okada, 1983). The other advantages often quoted are more problematical. The observation that radial dipoles may not provide a field outside the head is not purely an advantage of magnetic recording (Cohen & Cuffin, 1983; Melcher & Cohen, 1988); combinations of both radial and tangential sources can contribute to both the EEG and MEG. In addition, the lack of a reference in magnetic recording may not be an advantage since the position of the compensation coil can influence the field measurements and improved techniques of analysing the EEG such as source derivation removes the effect of the reference. Hence, the particular clinical advantage of MEG may be in a more sensitive signal to pathology due to less smearing. Since VEPs can be recorded in some clinical conditions, with one active electrode and one reference electrode. it would seem useful to establish whether one SQUID could be used in the same way. A single channel magnetometer can be used in unshielded environ-
Age
ment and so could be established easily for routine use in many clinical environments. In order to investigate the clinical utility of the VEF it is necessary to establish normative data across the age span. However, one of the limitations of a single channel magnetometer is that recordings from different scalp positions have to be made sequentially. Hence, a diagnostic protocol involving full field and half field stimulation to a variety of visual stimuli at different recording positions would increase the recording time substantially for the patient and is probably not feasible with a single channel system. However, a problem in choosing a single recording position is that magnetic signals are more focally distributed over the scalp than the corresponding electrical signals (Cohen & Cuffin, 1983). In addition, the exact topography of the visual signals may vary from subject to subject (Stok, 1986). It may, therefore, be difficult to locate the region of the maxima for a magnetic component in many patients without full knowledge of the topographic distribution. If recordings were made over a less active region of the scalp, the corresponding decrease in signal amplitude worsens the signal to noise ratio and could influence the component latency. In addition, there is evidence that some visual signals tend to cancel in the midline to full field stimulation because of the orientation of the sources in the visual cortex (Okada, 1983). This suggests that more lateral recording positions may give more consistent responses. In our studies, we
1859
R.
1860
A. ARMSTRONG
wished to use flash and pattern reversal stimuli to complement our YEP studies and their topographic distributions have not been well described. To circumvent this problem, we made a series of preliminary measurements recording from nine positions over the occipital scalp, which indicated that a region of scalp between 6-9 cm above the inion on the midline and 2 cm to the left of the midline was a good recording position for our stimuli in most subjects. Another problem is that the VEF may be more variable than the YEP. It is common practice to superimpose two successive traces of the latter to ensure consistency in the measurement of component latency. With increased variability in sequential recordings in the VEF, several tracings may need to be recorded to ensure consistency. The purpose of this study is to assess the feasibility of establishing normative data for evoked magnetic components using a single channel magnetometer in an unshielded environment. This has been carried out using a simple protocol on 100 normal subjects 15-87 yr of age.
MATERIALS AND METHODS
et al.
Stimulus parameters
For clinical use norms need to be established for a variety of check sizes. In this initial survey fairly large checks (70') were used because they gave higher amplitude magnetic signals. It should be noted that large checks may have a significant luminance component (Celesia, Kaufman & Cone, 1987; Novak, Wiznitzer, Kurtzberg, Gresser & Vaughan, 1988). The patterned stimuli were back projected onto a screen (16° x 14° field) with the checkerboard reversing at 2/sec. The mean luminance of the screen was 1050 cd/ m 2 while the contrast was maintained at 0.76. Diffuse light flashes at 2/sec were produced by a Grass PS22 stroboscope set at intensity 8 (3939 cd/m 2 ) and subtending a visual angle of 17°. All stimuli were viewed binocularly. Signal processing
Signals from the magnetometer were recorded at sensitivity x 1000 and were filtered between 0.1 and 30 Hz (24 dB/octave roll off) and also passed through a 50 Hz comb filter. Between 64 and 128 signals were averaged on a 4 channel Datalab signal processor with a time base of 500msec.
Subjects
Recording protocol
One hundred normal subjects 15-87 yr of age participated in the study. Before the recording all subjects were interviewed and carefully instructed as to the protocol and consent obtained. Seven subjects were eliminated because either they could not be recorded with corrected vision since their spectacles produced excessive magnetic artefact or they had an ophthalmological problem, e.g. cataract in elderly subjects. All remaining subjects had visual acuities better than 6/6. The study was approved by the University Ethical Committe.
Subjects were comfortably seated under the magnetometer with the Dewar Tail positioned normal to the scalp. We commenced recording 6-9 cm above the inion in the midline as determined by our preliminary measurements. If no signals with recognisable components were recorded then the Dewar Tail may be located in a transition zone between extrema. The Dewar Tail was therefore repositioned 1-2 cm to the left of the midline. Responses to pattern stimulation were recorded first. The subject was asked to fixate the screen and count the number of reversals to ensure attention to the stimulus. Recording was continued until two consistent responses were recorded. After a brief interval, flash responses were recorded. In a number of the subjects the recording was repeated on subsequent occasions to check longer term consistency of the responses.
Magnetometer
Magnetic fields were measured using a BTl magnetometer (Model 601). The device was a single channel d.c. SQUID second order gradiometer with a 5 cm baseline and 2 cm pick up coil located 12 mm from the outer surface of the Dewar Tail. The system had a white noise level of 16.6 fT/ at 5 Hz. The signal measured by the SQUID is proportional to the field generated by the source and not to the gradient of the field when the distance from the source to the pick up coil is small compared with the baseline (Carelli, Modena & Romani, 1983).
JHz
RESULTS
Magnetic waveforms to flash and pattern were successfully recorded in the majority of our subjects. Most of these recordings were made at or near the midline. If recordings failed to
1861
Visual evoked magnetic fields Table I. Normal values for the latency of the magnetic visual evoked field and visual evoked potential to a reversing checkerboard pattern (mean ± ISD) Age group (yr) 10-19 20-29 30-39 40-49 50-59
YEP PlOO (msec) 108.5 ± 10.9 101.7 ± 7.6 106.7 ± 4.9 104.6 ± 5.2 102.9 ± 6.75 109.2±1O.45 110.0 ± 8.7
60-69
70-79 80-89
P2m
VEF PlOOM (msec) 105 ± 4.27 106 ± 9.3 106 ± 9.5 109 ± 13.1 108 ± 7.9 114±15.1 119 ± 16.6 153 ± 5
produce recognisable components in the midline then movement to the lateral position was usually successful However, we failed to record evoked fields with recognisable components in 8 subjects of whom 3 were elderly with very low amplitude responses. The pattern reversal stimulus produced a major positive deflection (outgoing magnetic field) between 90 and 120 msec (PlOOM) in the majority of subjects. Figure 1 shows 5 successive recordings of the PI00M taken 1 min apart on the same subject and indicates that component latency may vary by about 8 msec from minute to minute. The PI00M may be preceded by a negative deflection (ingoing field) of latency of 75-96 msec and succeeded by a negative component of latency 125-155 msec. The presence of later peaks was more variable but many subjects exhibited positive components at about 200 and/or 300 msec.
P100m
Latency
500 msec
Fig. 2. Evoked magnetic waveforms to a flash stimulus (3939 cd/m2 ) recorded I min apart in one subject (AS).
The response to flash stimulation was more complex with several positive components being recorded, the most consistent of which was a major positive peak between 90 and 140 msec in the majority of subjects (Fig. 2). We have called this peak the P2M since in many subjects it was preceded by a smaller positive peak and we have followed the nomenclature for flash VEPS of Harding (1974). Normative values of the latency of the PIOOM and P2M were calculated for each decade and compared with YEP norms currently used in our laboratories (Harding & Wright, 1986). Although VEF and YEP norms were recorded under different conditions and on different populations of subjects, it is evident that VEF norms have a greater variability than the electrical norms, particularly in some of the older age groups. In addition, a plot of latency of the PIOOM against age (Fig. 3) revealed little change in latency until about 55 yr of age and then evidence Table 2. Normal values for the latency of the magnetic visual evoked field and visual evoked potential to a flashing light stimulus (mean ± I SD) Age group (yr)
Latency
500msec
Fig. I. Evoked magnetic waveforms to the pattern reversal stimulus (70' checks) recorded I min apart in one subject (RAA). The timebase is 500 msec. Positive, that is outgoing magnetic field is indicated upwards.
10-19 20-29 30-39 40-49 50-59 60-96 70-79 80-89
YEP P2 (msec)
VEF P2M (msec)
114.5 ± 9.8 120.7 + 10.99 121.7 + 7.8 126.8 + 11.26 122.5 5.45 127.2+11.28 134.2+12.72 -
104 ± 10.8 109 ± 14 118 ± 12.2 109 ± 14.2 106 ± 9.7 115±15.5 119±9.1 125 ± 19.7
±
R.
1862
A. ARMSTRONG
250
180
UQ)
y=1474717-19109X+00211x 2 R=048 A
'/ ..;;.
160
III
•• • • ••• /
E E 0 0
'0 >u c: Q)
.
°
140 0
... Q..
et al.
ac ° °° ~ ° ° D~ °° ePa. ••• ° 0
120
0
a 100
0
...J
0 DCDJ DO
0 0
20
40
80 0
0
DO
D.
D·.
60
...... ..
0
200
.§ E
N Q..
...
150
0
>c:
.
0
u
Q)
~
80
~III
0
100
°
0
...J
100
Age (yr)
Fig. 3. The relationship between the latency of the PlooM reversal component and age in normal subjects. Fitted regressions are (a) linear 15-54 yr of age (r = 0.1, P > 0.05), (b) linear> 55 yr of age (r = 0.57, P < 0.01).
of a more rapid increase. A I-way analysis of variance of the data showed significant differences between the latency means in the 60, 70 and 80s age groups when compared with the 20s age group. However, the data after 55 yr show greatly increased variability and hence the delay in the PI00M may be apparent in only some of the elderly subjects. By contrast, the flash P2M (Fig. 4) may show an initial decline in latency and then a more gradual increase in latency after about 50 yr of age. Analysis of variance showed a significant difference between the 70 and 80s age groups and the 40s age group. DISCUSSION
This study, one of the first attempts to measure evoked magnetic fields on a large number of subjects over the age span, has been partially successful in that recordings could be made from the majority of subjects tested. However, we failed to record any VEFs in 8 subjects, who may have had very low amplitude responses which could not be detected in our present noise environment. In addition, variation in the topography of the VEF may have resulted in recordings being made over an inactive region of the scalp. A survey of the topography of the PIOOM and P2M across the age span will be necessary to determine the extent of this problem. The normative values of the VEF were more variable than their electrical counterparts. The increased variability between subjects may be due to more superimposed noise on the traces due to the measurements being made in an unshielded environment. In preliminary experiments, more repetitions of the response did not decrease
0
0
50 0
20
40
60
80
100
Age (yr)
Fig. 4. The latency of the P2M flash component and age in normal subjects. Data were best fitted by a second order polynomial (r = 0.48, P < 0.001).
the intersubject variability. The numbers of repetitions required to substantially reduce the variability in an unshielded environment needs to be critically investigated. It is less likely that the increased variability is due to spontaneous background MEG since the physical mechanisms of EEG and MEG generation may be similar. Another problem is that recording cannot be made from a single recording position which can be defined for all subjects. We currently record over a region of scalp extending from the midline to a lateral position. It is possible that only at the positive extrema of the PlOOM and P2M would more consistent reponses be recorded. In addition, the VEF may be sensitive to poor fixation or lack of attention on the part of the subject (Snyder, Dustman & Shearer, 1981) and although these factors were controlled as far as possible, they may have contributed to the increased variability of the response, particularly in the elderly group. The increased variability may also be intrinsic to the VEF itself. The YEP may be derived from the volume current emanating from a mixture of radial and tangential sources while the VEF may be determined by activity in the dendritic trees of a more select group of tangentially oriented neurons (Wikswo, 1983; Okada, 1983). Hence, the VEF may be more sensitive to the pattern of activation of the calcarine fissure and the individual location of fissures in different subjects (Steinmetz, Furst & Meyer, 1989). Further improvements could also be made to our recording protocol. We currently record from seated subjects with their head unrestrained whereas more consistent responses may be recorded from prone subjects where lateral head movement is restricted. In addition, this study
Visual evoked magnetic fields
describes full field responses which may be necessary to examine pre-chiasmal lesions in diagnostic tests. However, full field responses may be more variable in topography than half field responses (Okada, 1983; Stok, 1986) and hence, the latter recorded from lateral positions may give more successful and consistent norms. Clearly, more needs to be known about the VEF before it can be used in clinical tests. Little is known of how stimulus conditions influence the VEF. A variety of factors influence the YEP including age and gender (Snyder et al., 1981; Allison, Wood & Goff, 1983; Cohn, Dustman & Shearer, 1985; Wright, Williams, Drasdo & Harding, 1985; Celesia et al., 1987), head size (Guthkelch, Bursick & Sclabassi, 1987), size of stimulus field and check size (Harding, 1982; Yiannikas & Walsh, 1983; Ristanovi6 & Hajdukovi6, 1981) and contrast and blur (Sokol, Moskowitz & Towle, 1981; Harding, 1982, 1988; Botak, Bodis-Wollner & Guillory, 1987). Many of these factors may also influence the VEF but not necessarily in a predictable or identical way. Our data suggest a much steeper rise with increasing age in latency of the magnetic pattern reversal PIOOM compared with the PIOO component of the YEP (Asselman, Chadwick & Marsden, 1975; Snyder et al., 1981). Changes in the PIOOM may reflect aging processes in the eye such as changes in cornea, lens and pupil which may reduce the level of light reaching the retina but such changes would only effect the latency of the response by around 5 msec (Corso, 1971; Verma & Kooi, 1984; Wright et al., 1985; Celesia et al., 1987). However, changes in the PIOOM may reflect aging both of the optic nerve tract and the visual cortex particularly in the calcarine fissure. However, the latency of the PIOOM may be difficult to identify and was more variable in some elderly subjects and more age series need to be recorded before this trend in the VEF can be confirmed. In addition, little is known of the topographic distribution of visual magnetic components and of the variation in distribution between subjects. A topographic survey across the age span to establish the position of the maxima more accurately will be essential to establish the optimum regions of the scalp for single channel recording. Response to hemifield or quadrantic stimuli may also need to be recorded to aid interpretation of magnetic waveforms in a variety of clinical conditions (Blumhardt, 1987). Once these investigations have been made and a variety of magnetic norms of sufficient quality
1863
established for a variety of stimulus conditions, such as different check sizes, left and right hemifield stimuli, and for the normal limits of interocular variation, then a step forward will have been made in establishing MEG as a clinical tool. The differing effects of age on the YEP and VEF responses to flash and pattern stimulation tempts us to believe that the MEG may well contain independent data that may complement the YEP information rather than simply being a difficult way of recording the same information. Ultimately, however, the recording of clinical cases will be essential to determine whether the VEMR will give more information than the YEP. REFERENCES Allison. T.. Wood. C. C. & Goff, W. R. (1983). Brainstem auditory, pattern reversal visual and short latency somatosensory evoked potentials: Latencies in relation to age. sex and brain and body size. Electroencephalography and Clinical Neurophysiology, 55, 619-639. Asselman, P., Chadwick, D. W. & Marsden, C. D. (1975). Visual evoked responses in the diagnosis and management of patients suspected of multiple sclerosis. Brain. 98. 261-282. Blumhardt. L (1987). The abnormal pattern visual evoked response III neurology. In Halliday, A. M., Butler, S. R & Paul, R. (Eds), A textbook of clinical neurophysiology (pp. 307-342). New York: Wiley. Botak, P., Bodis-Wollner, I. & Guillory, S. (1987). The effect of blur and contrast on VEP latency: Comparison between check and sinusoidal grating patterns. Electroencephalography and Clinical Neurophysiology, 68, 247-255. Carelli, P., Modena, I. & Romani, G. L. (1983). Detection coils. In Williamson, S. J., Romani, G., Kaufman, L. & Modena, I. Biomagnetism. An introductory approach (NATO ASI Senes, pp. 85-99). New York: Plenum Press Celesia, C G., Kaufman, D. & Cone, S. (1987). The effect of age and sex on pattern electroretinograms and visual evoked potentials. Electroencephalography and Clinical Neurophysiology, 68, 161-171. Cohen, D & Cuffin, R. N. (1983). Demonstration of useful differences between magnetoencephalogram and electroencephalogram. Electroencephalography and Climcal Neurophysiology, 56, 38-51. Cohn, N. B., Dustman, R. E. & Shearer, D. E. (1985). The effect of age, sex and interstimulus interval on augmenting and reducing of occipital VEPs. Electroencephalography and Chmcal Neurophysiology, 62, 177-183. Corso, J. F. (1971). Sensory processes and age effects in normal adults. Journal of Gerontology, 26, 90-105. Guthkelch, A. N., Bursick, D. & Sclabassi, R.J. (1987). The relatIOnship of the latency of the Visual Ploo wave to gender and head size. Electroencephalography and Clinical Neurophyswlogy, 68, 219-222. Harding, G F. A. (1974). The visual evoked response. In Roper-Hall, M. 1., el al. (Eds), Advances in ophthalmology (pp. 2-28). Basel: Karger. Hardmg, G. F. A. (1982). The flash evoked visual response and its use in ocular conditions. Journal of Electrophysiology and Technology, 8, 63-78.
1864
R. A.
ARMSTRONG
Harding, G. F. A. (1988). Neurophysiology of vision and its clinical application. In Edwards, K. & Llewellyn, R. (Eds), Optometry (Chap. 3, pp. 44-60). London: Butterworths. Harding, G. F. A. & Wright, e. E. (1986). Visual evoked potentials in acute optic neuritis. In Hess, R. F. & Plant, e. T. (Eds), Optic neuritis (pp. 230-254). Cambridge University Press. Melcher, J. R. & Cohen, D. (1988). Dependence of the magnetoencephalogram on dipole orientation in the rabbit head. Electroencephalography and Clinical Neurophysiology, 70,460-472. Novak, G. P., Wiznitzer, M., Kurtzberg, D., Gresser, B. S. & Vaughan, H. G. (1988). The utility of visual evoked potentials using hemifield stimulation and several check sizes in the evaluation of suspected multiple sclerosis. Electroencephalography and Clinical Neurophysiology, 71, 1-9.
Okada, Y. (1983). Neurogenesis of evoked magnetic fields. In Williamson, S. J., Romani, G., Kaufman, L. & Modena, I. (Eds), Biomagnetism. An interdisciplinary approach (Nato ASI Series, pp. 399-408). New York: Plenum Press. Ristanovic, D. & Hajdukovic, R. (1981) Effects of spatially structured stimulus fields on pattern reversal visual evoked potentials. Electroencephalography and Clinical Neurophysiology, 51, 599-610. Snyder, E. W., Dustman, R. E. & Shearer, D. E. (1981). Pattern reversal evoked potential amplitudes: Life span
et al.
changes. Electroencephalography and Clinical Neurophysiology, 52, 429-434. Sokol, S., Moskowitz, A. & Towle, V. L. (1981). Age related changes in the latency of the visual evoked potential: Influence of check size. Electroencephalography and Clinical Neurophysiology, 51, 559-562. Steinmetz, H., Furst, G. & Meyer, B. U. (1989). Craniocerebral topography within the international 10-20 system. Electroencephalography and Clinical Neurophysiology, 72, 499-506.
Stok, C. J. (1986). The inverse problem in EEG and MEG with application to visual evoked responses. Ph.D. thesis, Leiden University. Verma, N. P. & Kooi, K. A. (1984). Gender factor in longer PlOO latency of elderly persons. Electroencephalography and Clinical Neurophysiology, 59, 361-365. Wikswo, J. P. (1983). Cellular action currents. In Williamson, S. J., Romani, G., Kaufman, L. & Modena, I. (Eds), Biomagnetism: An interdisciplinary approach (NATO ASI Series, pp. 173-207). New York: Plenum Press. Wright, e. E., Williams, D. E., Drasdo, N. & Harding G. F. A. (1985). The influence of age on the electroretinogram and visual evoked potential. Documenta Ophthalmologica, 59, 365-384. Yiannikas, C. & Walsh, J. C. (1983). The variation of the pattern shift visual evoked response with the size of the stimulus field. Electroencephalography and Clinical Neurophysiology, 55, 427-435.
Vision Res. Vol. 31, No. II, pp. 1859-1864, 1991 Printed in Great Bntain. All rights reserved
VISUAL EVOKED MAGNETIC FIELDS TO FLASH AND PATTERN IN 100 NORMAL SUBJECTS R.
A. ARMSTRONG, A. SLAVEN
and G. F. A.
HARDING
Clinical Neurophysiology Unit, Department of Vision Sciences, Aston University, Aston Triangle, Birmingham B4 7ET, England (Received 20 June 1990; in revised form 3 January 1991) Abstract-The practicality of recording visual evoked magnetic fields in 100 subjects 15-87 yr of age using a single channel d.c. SQUID second order gradiometer in an unshielded environment was investigated. The pattern reversal response showed a major positive component between 90 and 120 msec (P1OOM) while the response to flash produced a major positive component between 90 and 140 msec (P2M). Latency norms of the P100M were more variable than the corresponding P100 and P2 visual evoked potentials. The latency of the P100M may show a steep increase with age in most subjects after about 55 yr whereas only a small trend of latency with age was detected for the flash P2M. Neuromagnetism
Norms
Flash
Pattern reversal
INTRODUCTION
There have been no studies of the clinical value of visual evoked magnetic fields (VEF). However, there may be advantages to making magnetic recordings which can complement the visual evoked potential (VEP) in clinical diagnosis. First, no electrodes have to be attached to the subject so that it is relatively easy to gather large amounts of data from different subjects. Second, the fields are less smeared by the low conductive skull (Wikswo, 1983; Okada, 1983). The other advantages often quoted are more problematical. The observation that radial dipoles may not provide a field outside the head is not purely an advantage of magnetic recording (Cohen & Cuffin, 1983; Melcher & Cohen, 1988); combinations of both radial and tangential sources can contribute to both the EEG and MEG. In addition, the lack of a reference in magnetic recording may not be an advantage since the position of the compensation coil can influence the field measurements and improved techniques of analysing the EEG such as source derivation removes the effect of the reference. Hence, the particular clinical advantage of MEG may be in a more sensitive signal to pathology due to less smearing. Since VEPs can be recorded in some clinical conditions, with one active electrode and one reference electrode. it would seem useful to establish whether one SQUID could be used in the same way. A single channel magnetometer can be used in unshielded environ-
Age
ment and so could be established easily for routine use in many clinical environments. In order to investigate the clinical utility of the VEF it is necessary to establish normative data across the age span. However, one of the limitations of a single channel magnetometer is that recordings from different scalp positions have to be made sequentially. Hence, a diagnostic protocol involving full field and half field stimulation to a variety of visual stimuli at different recording positions would increase the recording time substantially for the patient and is probably not feasible with a single channel system. However, a problem in choosing a single recording position is that magnetic signals are more focally distributed over the scalp than the corresponding electrical signals (Cohen & Cuffin, 1983). In addition, the exact topography of the visual signals may vary from subject to subject (Stok, 1986). It may, therefore, be difficult to locate the region of the maxima for a magnetic component in many patients without full knowledge of the topographic distribution. If recordings were made over a less active region of the scalp, the corresponding decrease in signal amplitude worsens the signal to noise ratio and could influence the component latency. In addition, there is evidence that some visual signals tend to cancel in the midline to full field stimulation because of the orientation of the sources in the visual cortex (Okada, 1983). This suggests that more lateral recording positions may give more consistent responses. In our studies, we
1859
R.
1860
A. ARMSTRONG
wished to use flash and pattern reversal stimuli to complement our YEP studies and their topographic distributions have not been well described. To circumvent this problem, we made a series of preliminary measurements recording from nine positions over the occipital scalp, which indicated that a region of scalp between 6-9 cm above the inion on the midline and 2 cm to the left of the midline was a good recording position for our stimuli in most subjects. Another problem is that the VEF may be more variable than the YEP. It is common practice to superimpose two successive traces of the latter to ensure consistency in the measurement of component latency. With increased variability in sequential recordings in the VEF, several tracings may need to be recorded to ensure consistency. The purpose of this study is to assess the feasibility of establishing normative data for evoked magnetic components using a single channel magnetometer in an unshielded environment. This has been carried out using a simple protocol on 100 normal subjects 15-87 yr of age.
MATERIALS AND METHODS
et al.
Stimulus parameters
For clinical use norms need to be established for a variety of check sizes. In this initial survey fairly large checks (70') were used because they gave higher amplitude magnetic signals. It should be noted that large checks may have a significant luminance component (Celesia, Kaufman & Cone, 1987; Novak, Wiznitzer, Kurtzberg, Gresser & Vaughan, 1988). The patterned stimuli were back projected onto a screen (16° x 14° field) with the checkerboard reversing at 2/sec. The mean luminance of the screen was 1050 cd/ m 2 while the contrast was maintained at 0.76. Diffuse light flashes at 2/sec were produced by a Grass PS22 stroboscope set at intensity 8 (3939 cd/m 2 ) and subtending a visual angle of 17°. All stimuli were viewed binocularly. Signal processing
Signals from the magnetometer were recorded at sensitivity x 1000 and were filtered between 0.1 and 30 Hz (24 dB/octave roll off) and also passed through a 50 Hz comb filter. Between 64 and 128 signals were averaged on a 4 channel Datalab signal processor with a time base of 500msec.
Subjects
Recording protocol
One hundred normal subjects 15-87 yr of age participated in the study. Before the recording all subjects were interviewed and carefully instructed as to the protocol and consent obtained. Seven subjects were eliminated because either they could not be recorded with corrected vision since their spectacles produced excessive magnetic artefact or they had an ophthalmological problem, e.g. cataract in elderly subjects. All remaining subjects had visual acuities better than 6/6. The study was approved by the University Ethical Committe.
Subjects were comfortably seated under the magnetometer with the Dewar Tail positioned normal to the scalp. We commenced recording 6-9 cm above the inion in the midline as determined by our preliminary measurements. If no signals with recognisable components were recorded then the Dewar Tail may be located in a transition zone between extrema. The Dewar Tail was therefore repositioned 1-2 cm to the left of the midline. Responses to pattern stimulation were recorded first. The subject was asked to fixate the screen and count the number of reversals to ensure attention to the stimulus. Recording was continued until two consistent responses were recorded. After a brief interval, flash responses were recorded. In a number of the subjects the recording was repeated on subsequent occasions to check longer term consistency of the responses.
Magnetometer
Magnetic fields were measured using a BTl magnetometer (Model 601). The device was a single channel d.c. SQUID second order gradiometer with a 5 cm baseline and 2 cm pick up coil located 12 mm from the outer surface of the Dewar Tail. The system had a white noise level of 16.6 fT/ at 5 Hz. The signal measured by the SQUID is proportional to the field generated by the source and not to the gradient of the field when the distance from the source to the pick up coil is small compared with the baseline (Carelli, Modena & Romani, 1983).
JHz
RESULTS
Magnetic waveforms to flash and pattern were successfully recorded in the majority of our subjects. Most of these recordings were made at or near the midline. If recordings failed to
1861
Visual evoked magnetic fields Table I. Normal values for the latency of the magnetic visual evoked field and visual evoked potential to a reversing checkerboard pattern (mean ± ISD) Age group (yr) 10-19 20-29 30-39 40-49 50-59
YEP PlOO (msec) 108.5 ± 10.9 101.7 ± 7.6 106.7 ± 4.9 104.6 ± 5.2 102.9 ± 6.75 109.2±1O.45 110.0 ± 8.7
60-69
70-79 80-89
P2m
VEF PlOOM (msec) 105 ± 4.27 106 ± 9.3 106 ± 9.5 109 ± 13.1 108 ± 7.9 114±15.1 119 ± 16.6 153 ± 5
produce recognisable components in the midline then movement to the lateral position was usually successful However, we failed to record evoked fields with recognisable components in 8 subjects of whom 3 were elderly with very low amplitude responses. The pattern reversal stimulus produced a major positive deflection (outgoing magnetic field) between 90 and 120 msec (PlOOM) in the majority of subjects. Figure 1 shows 5 successive recordings of the PI00M taken 1 min apart on the same subject and indicates that component latency may vary by about 8 msec from minute to minute. The PI00M may be preceded by a negative deflection (ingoing field) of latency of 75-96 msec and succeeded by a negative component of latency 125-155 msec. The presence of later peaks was more variable but many subjects exhibited positive components at about 200 and/or 300 msec.
P100m
Latency
500 msec
Fig. 2. Evoked magnetic waveforms to a flash stimulus (3939 cd/m2 ) recorded I min apart in one subject (AS).
The response to flash stimulation was more complex with several positive components being recorded, the most consistent of which was a major positive peak between 90 and 140 msec in the majority of subjects (Fig. 2). We have called this peak the P2M since in many subjects it was preceded by a smaller positive peak and we have followed the nomenclature for flash VEPS of Harding (1974). Normative values of the latency of the PIOOM and P2M were calculated for each decade and compared with YEP norms currently used in our laboratories (Harding & Wright, 1986). Although VEF and YEP norms were recorded under different conditions and on different populations of subjects, it is evident that VEF norms have a greater variability than the electrical norms, particularly in some of the older age groups. In addition, a plot of latency of the PIOOM against age (Fig. 3) revealed little change in latency until about 55 yr of age and then evidence Table 2. Normal values for the latency of the magnetic visual evoked field and visual evoked potential to a flashing light stimulus (mean ± I SD) Age group (yr)
Latency
500msec
Fig. I. Evoked magnetic waveforms to the pattern reversal stimulus (70' checks) recorded I min apart in one subject (RAA). The timebase is 500 msec. Positive, that is outgoing magnetic field is indicated upwards.
10-19 20-29 30-39 40-49 50-59 60-96 70-79 80-89
YEP P2 (msec)
VEF P2M (msec)
114.5 ± 9.8 120.7 + 10.99 121.7 + 7.8 126.8 + 11.26 122.5 5.45 127.2+11.28 134.2+12.72 -
104 ± 10.8 109 ± 14 118 ± 12.2 109 ± 14.2 106 ± 9.7 115±15.5 119±9.1 125 ± 19.7
±
R.
1862
A. ARMSTRONG
250
180
UQ)
y=1474717-19109X+00211x 2 R=048 A
'/ ..;;.
160
III
•• • • ••• /
E E 0 0
'0 >u c: Q)
.
°
140 0
... Q..
et al.
ac ° °° ~ ° ° D~ °° ePa. ••• ° 0
120
0
a 100
0
...J
0 DCDJ DO
0 0
20
40
80 0
0
DO
D.
D·.
60
...... ..
0
200
.§ E
N Q..
...
150
0
>c:
.
0
u
Q)
~
80
~III
0
100
°
0
...J
100
Age (yr)
Fig. 3. The relationship between the latency of the PlooM reversal component and age in normal subjects. Fitted regressions are (a) linear 15-54 yr of age (r = 0.1, P > 0.05), (b) linear> 55 yr of age (r = 0.57, P < 0.01).
of a more rapid increase. A I-way analysis of variance of the data showed significant differences between the latency means in the 60, 70 and 80s age groups when compared with the 20s age group. However, the data after 55 yr show greatly increased variability and hence the delay in the PI00M may be apparent in only some of the elderly subjects. By contrast, the flash P2M (Fig. 4) may show an initial decline in latency and then a more gradual increase in latency after about 50 yr of age. Analysis of variance showed a significant difference between the 70 and 80s age groups and the 40s age group. DISCUSSION
This study, one of the first attempts to measure evoked magnetic fields on a large number of subjects over the age span, has been partially successful in that recordings could be made from the majority of subjects tested. However, we failed to record any VEFs in 8 subjects, who may have had very low amplitude responses which could not be detected in our present noise environment. In addition, variation in the topography of the VEF may have resulted in recordings being made over an inactive region of the scalp. A survey of the topography of the PIOOM and P2M across the age span will be necessary to determine the extent of this problem. The normative values of the VEF were more variable than their electrical counterparts. The increased variability between subjects may be due to more superimposed noise on the traces due to the measurements being made in an unshielded environment. In preliminary experiments, more repetitions of the response did not decrease
0
0
50 0
20
40
60
80
100
Age (yr)
Fig. 4. The latency of the P2M flash component and age in normal subjects. Data were best fitted by a second order polynomial (r = 0.48, P < 0.001).
the intersubject variability. The numbers of repetitions required to substantially reduce the variability in an unshielded environment needs to be critically investigated. It is less likely that the increased variability is due to spontaneous background MEG since the physical mechanisms of EEG and MEG generation may be similar. Another problem is that recording cannot be made from a single recording position which can be defined for all subjects. We currently record over a region of scalp extending from the midline to a lateral position. It is possible that only at the positive extrema of the PlOOM and P2M would more consistent reponses be recorded. In addition, the VEF may be sensitive to poor fixation or lack of attention on the part of the subject (Snyder, Dustman & Shearer, 1981) and although these factors were controlled as far as possible, they may have contributed to the increased variability of the response, particularly in the elderly group. The increased variability may also be intrinsic to the VEF itself. The YEP may be derived from the volume current emanating from a mixture of radial and tangential sources while the VEF may be determined by activity in the dendritic trees of a more select group of tangentially oriented neurons (Wikswo, 1983; Okada, 1983). Hence, the VEF may be more sensitive to the pattern of activation of the calcarine fissure and the individual location of fissures in different subjects (Steinmetz, Furst & Meyer, 1989). Further improvements could also be made to our recording protocol. We currently record from seated subjects with their head unrestrained whereas more consistent responses may be recorded from prone subjects where lateral head movement is restricted. In addition, this study
Visual evoked magnetic fields
describes full field responses which may be necessary to examine pre-chiasmal lesions in diagnostic tests. However, full field responses may be more variable in topography than half field responses (Okada, 1983; Stok, 1986) and hence, the latter recorded from lateral positions may give more successful and consistent norms. Clearly, more needs to be known about the VEF before it can be used in clinical tests. Little is known of how stimulus conditions influence the VEF. A variety of factors influence the YEP including age and gender (Snyder et al., 1981; Allison, Wood & Goff, 1983; Cohn, Dustman & Shearer, 1985; Wright, Williams, Drasdo & Harding, 1985; Celesia et al., 1987), head size (Guthkelch, Bursick & Sclabassi, 1987), size of stimulus field and check size (Harding, 1982; Yiannikas & Walsh, 1983; Ristanovi6 & Hajdukovi6, 1981) and contrast and blur (Sokol, Moskowitz & Towle, 1981; Harding, 1982, 1988; Botak, Bodis-Wollner & Guillory, 1987). Many of these factors may also influence the VEF but not necessarily in a predictable or identical way. Our data suggest a much steeper rise with increasing age in latency of the magnetic pattern reversal PIOOM compared with the PIOO component of the YEP (Asselman, Chadwick & Marsden, 1975; Snyder et al., 1981). Changes in the PIOOM may reflect aging processes in the eye such as changes in cornea, lens and pupil which may reduce the level of light reaching the retina but such changes would only effect the latency of the response by around 5 msec (Corso, 1971; Verma & Kooi, 1984; Wright et al., 1985; Celesia et al., 1987). However, changes in the PIOOM may reflect aging both of the optic nerve tract and the visual cortex particularly in the calcarine fissure. However, the latency of the PIOOM may be difficult to identify and was more variable in some elderly subjects and more age series need to be recorded before this trend in the VEF can be confirmed. In addition, little is known of the topographic distribution of visual magnetic components and of the variation in distribution between subjects. A topographic survey across the age span to establish the position of the maxima more accurately will be essential to establish the optimum regions of the scalp for single channel recording. Response to hemifield or quadrantic stimuli may also need to be recorded to aid interpretation of magnetic waveforms in a variety of clinical conditions (Blumhardt, 1987). Once these investigations have been made and a variety of magnetic norms of sufficient quality
1863
established for a variety of stimulus conditions, such as different check sizes, left and right hemifield stimuli, and for the normal limits of interocular variation, then a step forward will have been made in establishing MEG as a clinical tool. The differing effects of age on the YEP and VEF responses to flash and pattern stimulation tempts us to believe that the MEG may well contain independent data that may complement the YEP information rather than simply being a difficult way of recording the same information. Ultimately, however, the recording of clinical cases will be essential to determine whether the VEMR will give more information than the YEP. REFERENCES Allison. T.. Wood. C. C. & Goff, W. R. (1983). Brainstem auditory, pattern reversal visual and short latency somatosensory evoked potentials: Latencies in relation to age. sex and brain and body size. Electroencephalography and Clinical Neurophysiology, 55, 619-639. Asselman, P., Chadwick, D. W. & Marsden, C. D. (1975). Visual evoked responses in the diagnosis and management of patients suspected of multiple sclerosis. Brain. 98. 261-282. Blumhardt. L (1987). The abnormal pattern visual evoked response III neurology. In Halliday, A. M., Butler, S. R & Paul, R. (Eds), A textbook of clinical neurophysiology (pp. 307-342). New York: Wiley. Botak, P., Bodis-Wollner, I. & Guillory, S. (1987). The effect of blur and contrast on VEP latency: Comparison between check and sinusoidal grating patterns. Electroencephalography and Clinical Neurophysiology, 68, 247-255. Carelli, P., Modena, I. & Romani, G. L. (1983). Detection coils. In Williamson, S. J., Romani, G., Kaufman, L. & Modena, I. Biomagnetism. An introductory approach (NATO ASI Senes, pp. 85-99). New York: Plenum Press Celesia, C G., Kaufman, D. & Cone, S. (1987). The effect of age and sex on pattern electroretinograms and visual evoked potentials. Electroencephalography and Clinical Neurophysiology, 68, 161-171. Cohen, D & Cuffin, R. N. (1983). Demonstration of useful differences between magnetoencephalogram and electroencephalogram. Electroencephalography and Climcal Neurophysiology, 56, 38-51. Cohn, N. B., Dustman, R. E. & Shearer, D. E. (1985). The effect of age, sex and interstimulus interval on augmenting and reducing of occipital VEPs. Electroencephalography and Chmcal Neurophysiology, 62, 177-183. Corso, J. F. (1971). Sensory processes and age effects in normal adults. Journal of Gerontology, 26, 90-105. Guthkelch, A. N., Bursick, D. & Sclabassi, R.J. (1987). The relatIOnship of the latency of the Visual Ploo wave to gender and head size. Electroencephalography and Clinical Neurophyswlogy, 68, 219-222. Harding, G F. A. (1974). The visual evoked response. In Roper-Hall, M. 1., el al. (Eds), Advances in ophthalmology (pp. 2-28). Basel: Karger. Hardmg, G. F. A. (1982). The flash evoked visual response and its use in ocular conditions. Journal of Electrophysiology and Technology, 8, 63-78.
1864
R. A.
ARMSTRONG
Harding, G. F. A. (1988). Neurophysiology of vision and its clinical application. In Edwards, K. & Llewellyn, R. (Eds), Optometry (Chap. 3, pp. 44-60). London: Butterworths. Harding, G. F. A. & Wright, e. E. (1986). Visual evoked potentials in acute optic neuritis. In Hess, R. F. & Plant, e. T. (Eds), Optic neuritis (pp. 230-254). Cambridge University Press. Melcher, J. R. & Cohen, D. (1988). Dependence of the magnetoencephalogram on dipole orientation in the rabbit head. Electroencephalography and Clinical Neurophysiology, 70,460-472. Novak, G. P., Wiznitzer, M., Kurtzberg, D., Gresser, B. S. & Vaughan, H. G. (1988). The utility of visual evoked potentials using hemifield stimulation and several check sizes in the evaluation of suspected multiple sclerosis. Electroencephalography and Clinical Neurophysiology, 71, 1-9.
Okada, Y. (1983). Neurogenesis of evoked magnetic fields. In Williamson, S. J., Romani, G., Kaufman, L. & Modena, I. (Eds), Biomagnetism. An interdisciplinary approach (Nato ASI Series, pp. 399-408). New York: Plenum Press. Ristanovic, D. & Hajdukovic, R. (1981) Effects of spatially structured stimulus fields on pattern reversal visual evoked potentials. Electroencephalography and Clinical Neurophysiology, 51, 599-610. Snyder, E. W., Dustman, R. E. & Shearer, D. E. (1981). Pattern reversal evoked potential amplitudes: Life span
et al.
changes. Electroencephalography and Clinical Neurophysiology, 52, 429-434. Sokol, S., Moskowitz, A. & Towle, V. L. (1981). Age related changes in the latency of the visual evoked potential: Influence of check size. Electroencephalography and Clinical Neurophysiology, 51, 559-562. Steinmetz, H., Furst, G. & Meyer, B. U. (1989). Craniocerebral topography within the international 10-20 system. Electroencephalography and Clinical Neurophysiology, 72, 499-506.
Stok, C. J. (1986). The inverse problem in EEG and MEG with application to visual evoked responses. Ph.D. thesis, Leiden University. Verma, N. P. & Kooi, K. A. (1984). Gender factor in longer PlOO latency of elderly persons. Electroencephalography and Clinical Neurophysiology, 59, 361-365. Wikswo, J. P. (1983). Cellular action currents. In Williamson, S. J., Romani, G., Kaufman, L. & Modena, I. (Eds), Biomagnetism: An interdisciplinary approach (NATO ASI Series, pp. 173-207). New York: Plenum Press. Wright, e. E., Williams, D. E., Drasdo, N. & Harding G. F. A. (1985). The influence of age on the electroretinogram and visual evoked potential. Documenta Ophthalmologica, 59, 365-384. Yiannikas, C. & Walsh, J. C. (1983). The variation of the pattern shift visual evoked response with the size of the stimulus field. Electroencephalography and Clinical Neurophysiology, 55, 427-435.
