- Email: [email protected]
Cortical generators of the CI component of the pattern-onset visual evoked potential
256
Electroencephalography and clinical Neurophysiology, 1987, 68: 256- 267 Elsevier Scientific Publishers Ireland, Ltd.
EEG 01890
Cortical generators of the CI component of the pattern-onset visual evoked potential S.R. Butler, G.A. Georgiou, A. Glass, R.J. Hancox, J.M. Hopper and K.R.H. Smith Anatomy Department, The Medical School, Birmingham University, Birmingham B15 2TJ ( U.K.) (Accepted for publication: 8 November, 1986)
Summary Thirteen-channel visual evoked potentials (VEPs) to pattern-onset were recorded with stimuli restricted to individual octants of the peripheral field, to halves and to quadrants of the fovea. The voltage of the CI component was measured in each channel to define its topography for each stimulated sector. The potential fields so obtained were then analysed to find the orientation and location of a dipole that would produce a corresponding pattern of voltages at the scalp. The locations of the computed dipoles are consistent with the hypothesis that CI is generated in striate cortex. The computed locations and orientations are not compatible with alternative arrangements of sources in extrastriate cortex. A significant problem remains. If CI is indeed generated by the striate cortex then the orientation of the dipoles excited by stimulation of the peripheral field indicates that the cortex is surface negative. This leads to the prediction that foveal stimuli will elicit a CI which is negative at posterior electrodes. The experiments reported here confirm that CI is positive with such stimuli, and its source is calculated as a horizontal dipole with its positive pole oriented posterolaterally. Two possible explanations are considered for the reversed polarity of foveal CI: (a) that macropotentials associated with stimulation of the fovea are opposite in polarity from those associated with stimulation of the peripheral field; (b) that the foveal area of the retinotopic map extends into the lateral calcarine fissure with the effect that much of it faces in the reverse direction from the cortex at the pole. Key words: Dipole; Visual evoked potentials; Source location; Visual cortex; Pattern-onset
T h e visual evoked p o t e n t i a l (VEP) to a low spatial frequency checkerboard stimulus s u b t e n d ing a large visual angle c o n t a i n s a p r o m i n e n t early positive wave. This c o m p o n e n t is referred to as CI w h e n elicited by p a t t e r n - o n s e t a n d P100 w h e n evoked b y pattern-reversal. Both CI a n d P100 r e m a i n positive if the stimulus is restricted to the lower hemifield b u t reverse their polarity over certain parts of the scalp when the stimulus is in the u p p e r hemifield (Halliday a n d Michael 1970; Michael a n d H a l l i d a y 1971; Jeffreys a n d Axford 1972; Jeffreys a n d Smith 1979). This suggests that the cortical n e u r o n e s which generate b o t h CI a n d P100 are retinotopically organised, a n d that the Correspondence to: Dr. S.R. Butler, Department of Anatomy, The Medical School, University of Birmingham, Vincent Drive, Birmingham B15 2TJ (U.K.). 0168-5597/87/$03.50 © Elsevier Scientific Publishers Ireland Ltd.
areas on which u p p e r a n d lower fields are m a p p e d have different orientations. A l t h o u g h both CI a n d P100 display a form of polarity reversal with stimuli above a n d below the horizontal meridian, they p r o b a b l y have somewhat different cortical generators because their scalp d i s t r i b u t i o n s are dissimilar in several respects. F o r example, when pattern-reversal stimulation is restricted to a lateral hemifield, the m a i n positivity is ' p a r a d o x i c a l l y ' ipsilateral to the stimulus (Barrett et al. 1976; H a r d i n g et al. 1980) b u t not for p a t t e r n - o n s e t stimulation (Shagass et al. 1976). The response recorded over the contralateral (i.e., stimulated) hemisphere is a negative deflection with pattern-reversal ( B l u m h a r d t et al. 1978) but positive with p a t t e r n - o n s e t (Kriss a n d Halliday 1980). The physiological characteristics of the response also differ. Change in ampli-
CORTICAl. G E N E R A T O R S OF PATI'ERN-ONSET VEP
tude as a function of luminance is much greater for P100 than CI and the latency of P100 appears to change with check size whereas that of CI does not (Kriss and Barrett 1985). Finally, Estevez and Spekreijse (1974) showed that P100 is equivalent to a component of similar latency in the patternoffset response which can be distinguished from CI. The scalp distributions of these two components have been used to deduce the anatomical location of their generators. Early investigators proposed that CI arises in striate cortex (Jeffreys and Axford 1972; Jeffreys 1977) whereas P100 originates in the extrastriate areas of the occipital convexity (Halliday and Michael 1970), possibly extending onto its medial surface (Halliday et al. 1977). These proposals require that CI is generated by surface negative, and P100 by surface positive cortex. However, the view that CI is generated in the striate cortex has been questioned because it fails to predict the effects of foveal stimulation. This part of the visual field is believed to be mapped around the posterior pole of the occipital lobe whereas the peripheral field is mapped on the medial surface and in the calcarine sulcus (Holmes 1944; Spalding 1952). Electrical stimulation of the striate cortex in man suggests that relatively little of the medial surface of the hemisphere is devoted to the fovea (Brindley 1973; Dobelle et al. 1979), in contrast to anatomical findings (Brouwer 1932). Fig. 1A and 1B show how the peripheral field is believed to be mapped within the calcarine sulcus. The striate cortex is shown surface negative as Jeffreys proposes for the generation of CI. He argued that this arrangement accounts for the observation that a superiorly located electrode detects a negative CI when the upper field is stimulated and a positive CI when the lower field is stimulated. It follows that stimulation of the central area of the visual field, F, should evoke a surface negative response in the cortex around the occipital pole which serves the fovea (Fig. 1C). Posterior electrodes should therefore detect a negative CI to foveal stimulation. In fact, pattern-onset restricted to the central 2 ° of the visual field usually evokes a positive CI (Drasdo 1980; Les~vre and Joseph 1980). This and other
257
¥
R
A
B
ii
c Fig. 1. The octants of the peripheral field (A) and their representation on the banks of the calcarine fissure (B). The polarity of CI and its reversal can be accounted for if it is generated by surface negative striate cortex. Then an electrode over the midline would record a positive CI with stimuli in the lower half of the peripheral field (activating cortex at T, U, V, W). The same electrode would record a negative CI with stimuli in the periphery of the upper half-field (activating cortex at S, R, Y, X). The fovea is believed to be mapped around the external convexity of the occipital pole (C). It follows that stimulation of the fovea (F) alone should evoke a negative C1 at posterior electrodes.
features of its scalp distribution have led to the counter-proposal that CI is generated in surface positive cortex of extrastriate regions (Lesrvre and Joseph 1980; Lesrvre 1982).
)
258
We have studied the retinotopic organisation of the generators of CI of the pattern onset VEP to gain further information about their anatomical location. We stimulated sectors of the fovea and peripheral field individually and computed the location and orientation of the neural generators of CI using the method of Henderson et al. (1975), modified to incorporate the effects of skull thickness. A similar source location technique has been used by others to localise the generators of the pattern-onset evoked response to stimuli occupying individual quadrants of the visual field (Darcey et al. 1980a,b; Wood 1982).
S.R. B U T L E R ET AL. 17°
b
4°
5°
fo
13° 16° c
Method
Subjects and stimulation Twenty-three normal subjects took part in various sections of this investigation. The majority were male or female undergraduates aged 18-23. They sat in a sound-attenuated, electrically screened chamber and their E E G was recorded while they observed stimuli displayed on a high resolution, monochrome (black-and-white) television monitor. The stimuli were generated by an Acorn BBC microcomputer and took the form of checkerboard patterns (contrast ratio 0.9) whose onset replaced an isoluminant grey background. Subjects saw one of two different sets of stimuli: (a) Peripheral octants and fooeal half-fields. Fig. 2a shows each hemifield divided into 5 sectors. Evoked potentials were recorded from all 5 sectors of either the right or left half-field in each subject. Each sector of the peripheral field began 4 ° above or below the fixation point (Fig. 2b) and 5 ° lateral to it, and extended to 15 ° above or below the horizontal meridian and to 17 ° along it. Fig. 2c shows the form of stimulus used to stimulate half of the fovea, extending 1.6 ° laterally and 1.5 ° above and below the horizontal meridian. In each sector the individual checks subtended 50' for the majority of subjects. In a few individuals slightly smaller checks were used, down to 20'. Any effect this may have had on the amplitude of CI was small compared with differences between subjects. The stimulus was presented 192 times (3 groups
d
Fig. 2. a: schematic representation of the visual field, showing its division into peripheral octants and foveal half-fields, b and c: dimensions of the foveal and peripheral sectors, and fixation point F. d: dimensions of the foveal quadrant stimuli.
of 64 stimuli with short rest periods in between) in each sector of the field. Stimulus duration was 170 msec and the interstimulus interval randomised between 600 and 1200 msec. Pattern-onset was frame-locked to the TV raster to avoid the introduction of spurious spatial frequencies. Stimulation of such small areas of the field evokes potentials of relatively low amplitude and the signal-to-noise ratio is correspondingly degraded. Recordings were made from the 5 sectors of one hemifield in each subject until clear EPs were obtained for all the sectors of each hemifield in at least 6 subjects. For this reason the number of subjects used to obtain VEPs for left and right half-fields was unequal (10 and 7 respectively). (b) Fooeal quadrants. In a further 6 subjects VEPs were recorded with stimuli restricted to individual quadrants of the fovea. Each quadrant comprised a sector of a circular stimulus field, radius 2 ° and filled with 30' checks (Fig. 2d). In other respects the conditions of stimulation were the same as for peripheral octants.
EEG recording and averaging The EEG was recorded from Ag-AgC1 electrodes, previously shorted in normal saline to drain DC sources that might have generated thermal
CORTICAL GENERATORS OF PATTERN-ONSETVEP drift. For VEPs to peripheral octants and hemifovea, the recordings were made from 13 electrodes simultaneously, with A1 and A2 as an average reference: T3, T4, C3, Cz, C4, O1, 02, Oz and from modified positions T5, T6, P3, Pz, P4. The latter were shifted 5% of the nasion-inion distance posterior to their conventional locations to cluster more closely around the visual cortex. Note that A1 and A2 were not linked as this would disturb the potentials at the scalp, but averaged through the Grass average reference box. For the purpose of dipole localisation using the mathematical technique described below, the choice of reference is immaterial (so long as it is incorporated in the algorithm) - - although of course it affects the wave form in conventional write-outs. For recordings made with stimuli in foveal quadrants the reference was Fz, and F7 and F8 replaced T3 and T4. The signals were amplified on a 14-channel Grass Model 8-18C EEG machine (gain, 7/iV/cm; time constant, 1 sec; H F filter, 70 Hz; 50 Hz notch filter normally out) and saved together with trigger pulses from the stimulator on a RACAL Store 14 FM tape-recorder for subsequent offline analysis. The recordings were digitised by a 12-bit analogue-to-digital converter (ADC) and averaged on a North Star Advantage computer. The amplitude of the trace in each channel was measured at the time of the peak of CI in channel Pz. We have not investigated the source location of these EPs at other latencies. For this measurement, the baseline was taken as the average of the first 20 msec of the trace in each channel. The procedure resulted in a set of 13 numerical values defining CI over the electrode array for each sector of the field in each subject. When the investigation of peripheral sectors was carried out, only 8 channels of analogueto-digital conversion were available and the 13 channels were therefore obtained in two passes, with P3 channel common to both. Minor differences in the two passes of P3 are due to flutter in the tape-recorder. Dipole source location The pattern of 13 potential differences served as data from which to compute the locus and
259 orientation of the generators of CI by the following method. The potential at a point on the surface of a conducting sphere due to a dipole within it may be calculated given the orientation, positional coordinates and magnitude of the dipole; the potential is referenced to the centre of the dipole (for example, Wilson and Bayley 1950). In this way, the potential difference between any two points on the surface of the sphere can be derived. However, an inverse solution to such equations does not exist. That is to say, one cannot directly calculate the location and orientation of a source given the scalp potentials it creates. Instead the surface potentials are calculated for many values of orientation and location until a voltage pattern is found which matches the observed potentials (Henderson et al. 1975). An iterative minimisation algorithm is used to search for the best match between observed and calculated potentials without scanning all possible locations and orientations. This method provides a result which is meaningful only when the generator comprises a single, restricted area of cortex. The consequences of applying the method (inappropriately) to the case of a more complex arrangement of sources will be considered below. The skull has important effects on the electrotonic spread of voltage to the scalp. It both attenuates the signal and smoothes the spatial distribution of its peaks and troughs (Ary et al. 1981; Nunez 1981; Wood 1982). To correct for this, the algorithm employed in the present study is that described by Kavanagh et al. (1978) which models the head as 3 concentric spheres. The inner and outer shells contain high conductivity brain and scalp; the middle shell represents the skull. In our implementation, the radius of inner and outer tables of the skull were taken as 0.87 and 0.94 of the radius of the head. The absolute radius of the head is immaterial since the algorithm works on normalised values of the voltage distribution, ignoring the magnitude of the dipole. Values adopted for the volume resistivity of the brain and scalp (222 I2- cm) and the ratio of conductivity of brain to bone (80 : 1) were those obtained by Rush and Driscoll (1968). The effect of the bone un-
260
S.R. B U T L E R ET AL.
@@@ From left
From ebove
From behind
Fig. 3. Computed location and orientation of the dipoles to stimulation of the upper left octant next to the vertical meridian. Five of them form a cluster. In the sixth case (filled circle), the system was unable to find a close match between the field of a single dipole and the recorded voltages. The best match was obtained with a dipole of location and orientation which places it outside the cluster.
doubtedly outweighs the effect of lesser resistive anisotropies such as the CSF in the subarachnoid space and ventricles, not modelled here. The computation began with the dipole at an arbitrary location and orientation. The search for a minimum difference between observed and calculated potentials was executed several times for each VEP using different starting points as a check on the success of the minimisation technique in avoiding false minima. It is not normally possible to find a close match between observed and calculated potentials if the real source does not approximate to a dipole (i.e., is not singular, small in area, and uniformly polarised). To confine the method to cases where the source behaved like a dipole, the mean square difference between normalised calculated and normalised observed fields was obtained. The data were rejected if the ratio of this mean square difference to the maximum value in the normalised potential distribution exceeded 1:4. This is an arbitrary cut-off, derived from inspection of the fit of calculated and observed potential fields. In the study of foveal hemifields and peripheral octants, the computed locations of the source of the VEP for each sector were plotted on diagrams like that in Fig. 3. Here the majority of dipoles form a cluster even though they have been obtained from the VEPs of different subjects. However, one is clearly deviant. Such deviant dipoles were rejected before the cosine coordinates of the remainder were averaged to obtain a typical loca-
tion and orientation of the generator for each sector of the visual field. Accordingly, of the 6 clear EPs for each sector submitted to the localisation analysis, some were excluded at this stage. Thus for sectors R 1, S, T, U, V, W, X, Y and the left and right foveal sectors, nl (number of dipoles excluded) = 0, 0, 1, 0, 0, 1, 0, 1, 1, 2, respectively, and n2 (number of dipoles initially computed) = 2, 6, 7, 7, 8, 6, 9, 6, 5, 6, respectively. In the case of the study of foveal quadrants grand averages were obtained over all subjects to characterise the typical wave form and the amplitudes from this used as input for the localisation program.
Results
Fig. 4A shows the 13-channel evoked potential to pattern-onset in each of the 5 sectors of the left half of the visual field in one subject. Fig. 4B shows the corresponding evoked potentials for the right half of the field in another subject. The amplitude of CI varied considerably between subjects and between sectors of the field, being least well defined for the top octant of the right visual field (Fig. 4B). The phase reversal of CI for stimuli above and below the horizontal meridian is most marked in parietal and occipital channels contralateral to the half of the field in which the stimuli appeared. The latency of CI varied between subjects (90-110 msec). Numerical values of the voltage in all channels at the time of the peak in Pz (normally about 100 msec) were taken from all channels and used as data for localisation of the source active at that latency. The baseline for measurement of amplitude was taken as the average of the first 20 msec of each trace. Figs. 5A, B and 6 show the computed locations and orientations of the generators excited by stimulation of the peripheral octants, foveal halffields and foveal quadrants. The location of the dipole is referred to the sphere whose outer limit is depicted. In the model, the nasion-inion line cor-
i Note: no average dipole obtained for this sector.
CORTICAL GENERATORS OF PATTERN-ONSET VEP responds to a horizontal diameter at the equator of the sphere. The brain is shown in its approximate registration to this line, with the frontal pole just above the nasion and the occipital pole deep to the inion. In Fig. 5A and B the coordinates of the dipole have been obtained by averaging the cosine vectors across subjects for all cases in which a good match was found and where the dipoles from different subjects clustered. Stimulation of the upper right octant next to the vertical meridian failed to evoke a well defined CI in many subjects, its source was poorly matched as a dipole and where matches meeting the criterion were obtained the dipoles did not cluster across subjects. Consequently no dipole is shown for this sector in Fig. 5B. There is a consistent reversal in the orientation of the sources excited by stimulation of the pe-
A
261 ripheral octants above and below the horizontal meridian. This occurs irrespective of whether the octant lies next to the vertical or horizontal meridian. The sources lie in the hemisphere contralateral to the field stimulated in all cases, and usually somewhat closer to the midline than those excited by foveal stimulation. In general, the computed locations and orientations display mirror symmetry across the midline. Figs. 5 and 6 show that foveal stimuli excite sources far back in the occipital lobe and somewhat lateral to the midline in the hemisphere contralateral to stimulation. In the case of stimuli confined to individual quadrants of the fovea there is a tendency for dipoles to be located lower within the lobe for stimuli in the upper field than those for the lower field. However, there is no consistent swing in orientation as a function of the
B
Fig. 5. Dipoles computed for CI for each sector of the left (A) and right (B) half-fields. The circle marks the location of the dipole and its negative pole. The length and direction of the line in each plane shows its orientation. The coronal section shows the head viewed from behind.
262
S.R. BUTLER ET AL.
rr r
Tr
~
Tr
.,
T3 T4
T4
01
C3
c:
C4
c.~
02
TS T6
Oz
T5
P3
P3
P2
Tt
Tr
T3
P4
Tr T3 ~ T~ ~
T4
T5
Pz
~
Ts
~
1"6
o~ ~
T6
~
P3
P3
Tr
Tr
T3 ~ T4
~
cz
C4
G
02 ~
C.;
02
C3
Oz
~ Tr
Tr
~ C~
~
C4
02 Cz
T5
Pz
P3
P3
~
C~.
p:,---~P'~ i
Tr P4 T4 ~
~
.
T~
~
01
e
C:.
C4 Pz
T6.
Oz
P3 ~
P3
Tr!
Tr
L T~ ! T
4
;
T5
,
~ c.~ - ~ _ _ ~ , - - , . ~ w
C3 C4
Tri
'~
Cz Pz
- .
cz_.-.-.._./-x~.j
Oz P2 u, ÷
A
B
Fig. 4. Thirteen-channel evoked potentials to pattern onset in each sector of the left (A) and right (B) halves of the visual field. C o m m o n reference: A1 and A2. The VEPs to each half of the field were recorded in different subjects. The 13 channels were digitised in two passes of an 8-channel analogue to digital converter. P3 appears in both passes. Tr marks the period for which the stimulus was on (upward deflection). Ticks on the abscissa mark 100 msec intervals.
CORTICAL GENERATORS OF PATTERN-ONSET VEP
Fig. 6. Dipoles computed for CI with stimuli confined to foveal quadrants. The conventions are those used in Fig. 5.
location of the stimulus above or below the fixation point: the positive pole of the source faces posterolaterally in all foveal quadrants. There was little variation between subjects in the orientation of the source for the foveal EPs whose fields met our criteria for dipole matching.
Discussion
The evoked potentials to checkerboard stimuli (Fig. 4) replicate many important features of CI observed by previous workers using stimuli in restricted regions of the visual field. In particular they show the polarity inversion of CI obtained with stimuli above and below the horizontal meridian (Jeffreys and Axford 1972; Jeffreys 1977; Jeffreys and Smith 1979) and the absence of the paradoxical lateralisation characteristic of the pattern-reversal response (Barrett et al. 1976). The computed locations of the generators of CI varied systematically with stimulus location (Figs. 5 and 6). The equivalent dipoles were found to lie in the occipital lobe contralateral to the hemifield within which the stimulus occurred. Stimulation in
263
the peripheral octants excited sources which lay closer to the medial surface of the occipital lobe than to its lateral convexity. Their orientation changes according to whether the stimulus was above or below the horizontal meridian, the positive pole was directed upwards for stimuli in the lower field, downwards for stimuli in the upper field. It was not possible to resolve the upper right octant as a single dipole with consistent orientation across subjects. It is conceivable that the area generating CI for this sector spans a curvature in the surface of the cortex such as the lip of a gyrus or base of a sulcus. The generators would be distributed at various angles and the field would be particularly susceptible to individual differences in convolution. Stimulation of the fovea excites sources in a more posterior location, oriented radially to the posterolateral convexity of the occipital lobe, positive pole outward. Stimuli in the upper part of the fovea excited sources somewhat lower in the occipital lobe than did stimuli in the lower part of the fovea on the same side. This is in accord with the inverted representation of this part of the visual field on the occipital cortex, irrespective of whether the generators lie in striate or peristriate cortex (Drasdo 1977; 1980). Darcey et al. (1980a, b) describe the location and orientation of sources excited by pattern-onset, using a red checkerboard. They examined the first peak in the VEP (92 msec) when individual quadrants of the field were stimulated and analysed the distribution of scalp potentials using the algorithm of Kavanagh et al. (1978). Their stimuli included foveal quadrants and extended 10 ° into the peripheral field, i.e., only partially overlapping the area of the peripheral field stimulated by us. The location and orientation of their dipoles correspond closely with those found for foveal stimulation in this study. Both investigations disclose dipoles for quadrants of the left field oriented obliquely above and below the horizontal with stimuli on opposite sides of the horizontal meridian; no such change of orientation occurred with stimuli in the right field (Fig. 6). Jeffreys and Axford (1972) also recorded VEPs
264
with pattern onset in separate quadrants of the field. With the stimuli employed in that study, the amplitude of CI was found to be chiefly dependent on stimulation of the peripheral part of each quadrant. The location and orientation of the sources excited by such stimuli correspond well with those we report here for peripheral octants. The generators for both upper and lower fields appear to lie close to one another, in a position congruent with that of the striate cortex. Their orientations reverse through 180 ° as would be expected if the generators faced one another across the calcarine fissure. Jeffreys envisaged a cruciform representation of the peripheral field: octants close to the vertical meridian were mapped on the medial face of the hemisphere, in the cuneus and lingual gyrus while those close to the horizontal meridian were supposed to lie in the depths of the calcarine fissure as proposed by Holmes (1944). This model leads to the expectation that sources excited by stimulation of octants close to the vertical meridian will be orientated horizontally, i.e., normal to the medial surface of the hemisphere. However, our d a t a provide no support for this prediction (Fig. 5). Although there may be variations between individual subjects, one would expect at least a tendency toward a more horizontal orientation of the dipole for octants adjacent to the vertical meridian. This is not the case consistently. It might be argued that the cortex of the medial surface makes a relatively minor contribution to the evoked response because most of the striate cortex lies within the calcarine fissure (Filimonoff 1932; Brindley 1972; Stensaas et al. 1974). If so the cruciform model would need revision to accommodate this. A more serious objection to the idea that CI is generated by striate cortex concerns the orientation of the source excited by foveal stimuli. Assuming that the fovea is mapped around the occipital pole and if CI arises in surface negative cortex, the dipoles excited by stimulation of the central field should be oriented with the negative pole posteriorly (Fig. 1). Our results show that their orientation is the opposite. In this respect our observations accord with those of Les6vre and Joseph (1980) and Les+vre (1982) and appear to
S.R. BUTLER ET AL.
rule out an explanation of the origin of CI in terms of generally accepted concepts of the organisation of the striate cortex. However, the data cannot by interpreted in terms of sources in prestriate cortex, instead. The observed changes in orientation could be produced only if the potentials for sectors above and below the horizontal meridian were generated in superior and inferior occipital territories. These sources would be far apart; the data reveal no such separation as a function of the sector stimulated. The lack of separation is not an artefact caused by the use of inappropriate values for bone conductivity. We have examined the effects of assuming brain-bone conductivity ratios between 1:1 and 1:800, thus bracketing the values obtained by Rush and Driscoll (1968) of 1 : 80. These procedures failed to produce any separation of dipoles for upper and lower fields (Butler et al. 1984; Hancox et al. 1984). Nor is it likely that the generators lie outside the cortex (Blumhardt et al. 1982; Celesia et al. 1982; Fava et al. 1985). Nor can it be postulated that the invariance of the computed locations is due to the simultaneous activation of many supplementary visual areas in prestriate cortex distributed round the occipital pole, because in this case, no inversion of the equivalent dipole should occur with stimulation of different parts of the field. Moreover, this interpretation would conflict with the view of Jeffreys (Jeffreys 1977; Jeffreys and Musselwhite 1983) that CI originates in a single anatomical area. Furthermore, the superior and inferior cortical territories seem distinguished by their roles in spatial perception and object recognition rather than their retinotopic relationships (Ungerleider and Mishkin 1982). We have seen that neither single nor multiple sources outside the striate cortex provide a satisfactory explanation for the origin of CI. We must therefore reconsider the question of its origin in that cortex. Its appearance as the first major component of the onset VEP, its computed location, and the orientation of its sources when stimuli are in the peripheral field, all fit with its generation by surface negative striate cortex. Only its polarity when the fovea is stimulated poses a problem. The problem exists because it is believed that the fovea
C O R T I C A L G E N E R A T O R S OF P A T T E R N - O N S E T VEP
is represented only around the occipital pole, and because we assume that the macropotentials evoked by our stimuli will be equivalent for both foveal and peripheral parts of the field. The projection of the fovea is distinguished from that of the peripheral retina in several ways, including the distribution of colour opponent channels (Zrenner and Gouras 1983), the absence of input from rods (Polyak 1941), the size of the receptive fields and the proportion of X and Y cells (de Monasterio 1978). Moreover, the cytoarchitecture of the striate cortex is not uniform (Ngowyang 1934) and it has been suggested that the major subdivisions reflect a different cellular organisation for the projection of the fovea from that serving the peripheral field (Smith 1930a,b). Accordingly it is indeed possible that the macropotentials of foveal and extrafoveal striate areas may differ - - even to the extent that their responses to the afferent volley are of opposite polarity for stimuli of the same spatial frequency. The second assumption concerns the representation of the fovea at the occipital pole. Serial sections reveal that the calcarine sulcus normally forks in its depth to become Y or T shaped in cross-section. It is seldom a simple straight groove. The top of the T often emerges as the lateral calcarine sulcus (LCS) on the medial surface of the lobe at the posterior end of the calcarine fissure, just anterior to the tip of the pole (Bolton --}~"
I
Super,o~
Anterior
Inl
Poslerioc
X b
c
Fig. 7. a: medial view of right occipital lobe. Lateral calcarine sulcus (lcs), calcarine fissure (c) and parieto-occipital fissure (po). b: coronal section through lobe in plane X - X showing the typically T-shaped fundus of the calcarine sulcus. Stippled area marks the typical distribution of striate cortex, c: horizontal section through lobe in plane Y - Y showing the manner in which cortex of the pole curves into the lcs where it continues as the fundus of the calcarine sulcus. Typical distribution of striate cortex is marked as b.
265
1900; Filimonoff 1932; Polyak 1957) (Fig. 7a,b,c). In an anatomical study (in preparation), we have found LCS to be present, though highly variable in length, in 26 of 28 hemispheres examined. In such cases the cortex of the pole does not wrap directly onto the medial surface of the lobe but onto the posterior wall of the LCS where its orientation is exactly the reverse of that on the pole. The variability of this territory is in accord with the wide individual differences in the topography of the VEP observed using stimuli confined to the fovea (Jeffreys 1977; Halliday 1982; Brecelj and Cunningham 1985). How far the foveal map extends into the LCS, when present, is unknown. If the majority of the fovea were mapped on its posterior bank as an extension of the cortex at the pole, surface negative cortex would give rise to a positive CI at the occiput. Such a location for the foveal projection remains compatible with the data of Holmes (1944) and Spalding (1952) since the fovea would still be mapped within the occipital pole. Extrapolation of the cortical magnification factor from subhuman primates indicates that a disproportionate area of striate cortex is allocated to the fovea (Daniel and Whitteridge 1961; Cowey and Rolls 1974; Drasdo 1977). At present we cannot determine whether the factor is such that in man the foveal striate cortex extends beyond the pole and into the LCS. If the representation of the central field were folded back on itself in this way, then in some individuals stimulation of the central part of the fovea would excite sources both on the pole and within the sulcus. The resulting local fields would tend to cancel, and CI would have a disproportionately small amplitude. Several authors (Jeffreys 1977; Spekreijse et al. 1977; Drasdo 1980) have commented on the low amplitude of CI with foveal stimuli, though the effect might equally be a consequence of other variation in the morphology of the pole. The present study has not enabled us to distinguish between a striate and an extrastriate origin for the CI component of the VEP. A source, or multiple sources, outside the striate cortex seem to fit the data less well than a striate origin. However, any simple explanation, in terms of its origin in the striate cortex, may necessitate either refining the accepted view of retinotopic organisation,
266 or assuming that the polarity of macropotentials i n t h e s t r i a t e c o r t e x is n o t u n i f o r m f o r all p a r t s o f t h e v i s u a l field. This investigation was supported by the Medical Research Council. We are grateful to Dr. R.M. Flinn for assistance with computing.
References Ary, J.P., Klein, S.A. and Fender, D.H. Location of sources of evoked scalp potentials: corrections for skull and scalp thicknesses. IEEE Trans. biomed. Engng, 1981, BME-28: 447-452. Barrett, G., Blumhardt, L., Halliday, A.M., Halliday, E. and Kriss, A. A paradox in the lateralisation of the visual evoked response. Nature (Lond.), 1976, 261: 253-255. Blumhardt, L.D., Barrett, G., Halliday, A.M. and K.riss, A. The effect of experimental 'scotomata' on the ipsilateral and contralateral responses to pattern-reversal in one half-field. Electroenceph. clin. Neurophysiol., 1978, 45: 376-392. Blumhardt, L.D., Barrett, G., Kriss, A. and Halliday, A.M. The pattern-evoked potential in lesions of the posterior visual pathways. Ann. N.Y. Acad. Sci., 1982, 388: 264-289. Bolton, J.S. The exact histological localisation of the visual area of the human cerebral cortex. Phil. Trans. roy. Soc. B, 1900, 193: 165-222. Brecelj, J. and Cunningham, K. Occipital distribution of foveal half-field responses. Docum. ophthal. (The Hague), 1985, 59: 157-165. Brindley, G.S. The variability of human striate cortex. J. Physiol. (Lond.), 1972, 225: lp-3p. Brindley, G.S. Sensory effects of electrical stimulation of the visual and paravisual cortex in man. In: R. Jung (Ed.), Handbook of Sensory Physiology. 7/3. Central Processing of Visual Information. B. Springer, Berlin, 1973: 583-594. Brouwer, B. Projection of the retina on the cortex in man. Res. Publ. Ass. nerv. ment. Dis., 1932, 13: 529-534. Butler, S.R., Hancox, R.J., Georgiou, G.A. and Glass, A. The 90-100 msec component of the pattern onset visual evoked response does not originate in striate cortex. Behav. Brain Res., 1984, 12: 177. Celesia, G.G., Polcyn, R.D., Holden, J.E., Nickles, R.J., Gatley, J.S. and Koeppe, R.A. Visual evoked potentials and positron emission tomographic mapping of regional cerebral blood flow and cerebral metabolism: can the neuronal potential generators be visualized? Electroenceph. clin. Neurophysiol., 1982, 54: 243-256. Cowey, A. and Rolls, E.T. Human cortical magnification factor and its relation to visual acuity. Exp. Brain Res., 1974, 21: 447-454. Daniel, P.M. and Whitteridge, D. The representation of the visual field on the cerebral cortex in monkeys. J. Physiol. (Lond.), 1961, 159: 203-221. Darcey, T.M., Ary, J.P. and Fender, D.H. Methods for the localization of electrical sources in the human brain. In:
S.R. BUTLER ET AL. H.H. Kornhuber and L. Deecke (Eds.), Prog. Brain Res., Vol. 54. Elsevier, Amsterdam, 1980a: 128-133. Darcey, T.M., Ary, J.P. and Fender, D.H. Spatio-temporal visually evoked scalp potentials in response to partial-field patterned stimulation. Electroenceph. clin. Neurophysiol., 1980b, 50: 348-355. Dobelle, W.H., Turkel, J., Henderson, D.C. and Evans, J.R. Mapping the representation of the visual field by electrical stimulation of human visual cortex. Amer. J. Ophthal., 1979, 88: 727-735. Drasdo, N. The neural representation of visual space. Nature (Lond.), 1977, 266: 554-556. Drasdo, N. Cortical potentials evoked by pattern presentation in the foveal region. In: C. Barber (Ed.), Evoked Potentials. MTP Press, Lancaster, 1980: 167-174. Estevez, O. and Spekreijse, H. Relationship between pattern appearance-disappearance and pattern reversal responses. Exp. Brain Res., 1974, 19: 233-238. Fava, E., Ducati, A., Motti, E.D.F. and Marossero, F. Intracerebral recordings of PR-VEPs in three awake patients. Electroenceph. clin. Neurophysiol., 1985, 61: S102. Filimonoff, I.N. 0ber die Variabilit~it der Grosshirnrindenstruktur. Mitteilung II. Regio occipitalis beim erwachsenen Menschen. J. Psychol. Neurol. (Lpz.), 1932, 44: 1-96. Halliday, A.M. Evoked Potentials in Clinical Testing. Churchill-Livingstone, Edinburgh, 1982. Halliday, A.M. and Michael, W.F. Changes in pattern-evoked responses in man associated with the vertical and horizontal meridians of the visual field. J. Physiol. (Lond.), 1970, 208: 499-513. Halliday, A.M., Barrett, G., Halliday, E. and Michael, W.F. The topography of the pattern evoked potential. In: J.E. Desmedt (Ed.), Visual Evoked Potentials in Man: New Developments. Clarendon Press, Oxford, 1977: 121-133. Hancox, R.J., Butler, S.R., Glass, A. and Georgiou, G. Source location of the 100 msec component of the pattern onset visual evoked response. Electroenceph. clin. Neurophysiol., 1984, 58: 14P. Harding, G.F.A., Smith, G.F. and Smith, P.A. The effect of various stimulus parameters on the lateralization of the VEP. In: C. Barber (Ed.), Evoked Potentials. MTP Press, Lancaster, 1980: 213-218. Henderson, C.J., Butler, S.R. and Glass, A. The localization of equivalent dipoles of EEG sources by the application of electrical field theory. Electroenceph. clin. Neurophysiol., 1975, 39: 117-130. Holmes, G. The organisation of the visual cortex in man. Proc. roy. Soc. B, 1944, 132: 348-361. Jeffreys, D.A. The physiological significance of pattern visual evoked potentials. In: J.E. Desmedt (Ed.), Visual Evoked Potentials in Man: New Developments. Clarendon Press, Oxford, 1977: 134-167. Jeffreys, D.A. and Axford, J.G. Source locations of patternspecific components of human visual evoked potentials. I. Component of striate cortical origin. Exp. Brain Res., 1972, 16: 1-21.
CORTICAL GENERATORS OF PATI'ERN-ONSET VEP Jeffreys, D.A. and Musselwhite, M.J. Evoked potential evidence of parallel processing in human striate cortex. Electroenceph, clin. Neurophysiol., 1983, 56: 17P. Jeffreys, D.A. and Smith, A.T. The polarity inversion of scalp potentials evoked by upper and lower half-field stimulus patterns: latency or surface distribution differences? Electroenceph, clin. Neurophysiol., 1979, 46: 409-415. Kavanagh, R.N., Darcey, T.M., Lehmann, D. and Fender, D.H. Evaluation of methods for three-dimensional localization of electrical sources in the human brain. IEEE Trans. biomed. Engng, 1978, BME-25: 421-429. Kriss, A. and Barrett, G. Checkerboard VEPs: a review of stimulus response relationships, In: D. Papakostopoulos, S.R. Butler and I. Martin (Eds.), Clinical and Experimental Neuropsychophysiology. Croom Helm, London, 1985: 1-29. Kriss, A. and Halliday, A.M. A comparison of occipital potentials evoked by pattern onset, offset and reversal by movement. In: C. Barber (Ed.), Evoked Potentials. MTP Press, Lancaster, 1980: 205-212. Lesrvre, N. Chronotopographical analysis of the human evoked potential in relation to the visual field (data from normal individuals and hemianopic patients). Ann. N.Y. Acad. Sci., 1982, 388: 156-182. Les+vre, M. and Joseph, J.P. Hypothesis concerning the most probable sites of origin of the various components of the pattern evoked potential. In: C. Barber (Ed.), Evoked Potentials. MTP Press, Lancaster, 1980: 159-166. Michael, W.F. and Halliday, A.M. Differences between the occipital distribution of upper and lower field pattern evoked responses in man. Brain Res., 1971, 32: 311-324. Monasterio, F.M. de. Properties of concentrically organised X and Y ganglion cells of macaque retina. J. Neurophysiol., 1978, 41: 1394-1417. Ngowyang, G. Die cytoarchitektonischen Unterfelder der Area striata beim Menschen. J. Psychol. Neurol. (Lpz.), 1934, 46: 351-363. Nunez, P.L. Electric Fields of the Brain: the Neurophysics of EEG with Contributions by Ron D. Katznelson. Oxford University Press, New York, 1981.
267 Polyak, S.L. The Retina. University of Chicago Press, Chicago, IL, 1941. Polyak, S.L. The Vertebrate Visual System. University of Chicago Press, Chicago, IL, 1957. Rush, S. and Driscoll, D.A. Current distribution in the brain from surface electrodes. Anesth. Analg., 1968, 47: 717-723. Shagass, C., Amadeo, M. and Roemer, R.A. Spatial distribution of potentials evoked by half field pattern-reversal and pattern-onset stimuli. Electroenceph. clin. Neurophysiol., 1976, 41: 609-622. Smith, G.E. New light on vision. Nature (Lond.), 1930a, 125: 820-824. Smith, G.E. The cortical representation of the macula. J. Anat. (Lond.), 1930b, 64: 477-478. Spalding, J.M.K. Wounds of the visual pathways. Part I. The visual radiation. J. Neurol. Neurosurg. Psychiat., 1952, 15: 99-109. Spekreijse, H., Estevez, O. and Reits, D. Visual evoked potentials and the physiological analysis of visual processes in man. In: J.E. Desmedt (Ed.), Visual Evoked Potentials in Man: New Developments. Clarendon Press, Oxford, 1977: 16-89. Stensaas, S.S., Eddington, D.K. and Dobelle, W.H. The topography and variability of the primary visual cortex in man. J. Neurosurg., 1974, 40: 747-755. Ungerleider, L.G. and Mishkin, M. Two cortical visual systems. In: D.J. Ingle, M.A. Goodale and R.J.W. Mansfield (Eds.), Analysis of Visual Behavior. MIT Press, Cambridge, MA, 1982: 549-586. Wilson, F.N. and Bayley, R.H. The electric field of an eccentric dipole in a homogeneous spherical conducting medium. Circulation, 1950, 1: 84-92. Wood, C.C. Application of dipole localization metl~ods to source identification of human evoked potentials. Ann. N.Y. Acad. Sci., 1982, 388: 139-155. Zrenner, E. and Gouras, P. Cone opponency in tonic ganglion cells and its variation with eccentricity in rhesus monkey retina. In: J.T. Mollon and L.T. Sharpe (Eds.), Colour Vision: Psychophysiology and Psychophysics. Academic Press, London, 1983: 211-223.
Electroencephalography and clinical Neurophysiology, 1987, 68: 256- 267 Elsevier Scientific Publishers Ireland, Ltd.
EEG 01890
Cortical generators of the CI component of the pattern-onset visual evoked potential S.R. Butler, G.A. Georgiou, A. Glass, R.J. Hancox, J.M. Hopper and K.R.H. Smith Anatomy Department, The Medical School, Birmingham University, Birmingham B15 2TJ ( U.K.) (Accepted for publication: 8 November, 1986)
Summary Thirteen-channel visual evoked potentials (VEPs) to pattern-onset were recorded with stimuli restricted to individual octants of the peripheral field, to halves and to quadrants of the fovea. The voltage of the CI component was measured in each channel to define its topography for each stimulated sector. The potential fields so obtained were then analysed to find the orientation and location of a dipole that would produce a corresponding pattern of voltages at the scalp. The locations of the computed dipoles are consistent with the hypothesis that CI is generated in striate cortex. The computed locations and orientations are not compatible with alternative arrangements of sources in extrastriate cortex. A significant problem remains. If CI is indeed generated by the striate cortex then the orientation of the dipoles excited by stimulation of the peripheral field indicates that the cortex is surface negative. This leads to the prediction that foveal stimuli will elicit a CI which is negative at posterior electrodes. The experiments reported here confirm that CI is positive with such stimuli, and its source is calculated as a horizontal dipole with its positive pole oriented posterolaterally. Two possible explanations are considered for the reversed polarity of foveal CI: (a) that macropotentials associated with stimulation of the fovea are opposite in polarity from those associated with stimulation of the peripheral field; (b) that the foveal area of the retinotopic map extends into the lateral calcarine fissure with the effect that much of it faces in the reverse direction from the cortex at the pole. Key words: Dipole; Visual evoked potentials; Source location; Visual cortex; Pattern-onset
T h e visual evoked p o t e n t i a l (VEP) to a low spatial frequency checkerboard stimulus s u b t e n d ing a large visual angle c o n t a i n s a p r o m i n e n t early positive wave. This c o m p o n e n t is referred to as CI w h e n elicited by p a t t e r n - o n s e t a n d P100 w h e n evoked b y pattern-reversal. Both CI a n d P100 r e m a i n positive if the stimulus is restricted to the lower hemifield b u t reverse their polarity over certain parts of the scalp when the stimulus is in the u p p e r hemifield (Halliday a n d Michael 1970; Michael a n d H a l l i d a y 1971; Jeffreys a n d Axford 1972; Jeffreys a n d Smith 1979). This suggests that the cortical n e u r o n e s which generate b o t h CI a n d P100 are retinotopically organised, a n d that the Correspondence to: Dr. S.R. Butler, Department of Anatomy, The Medical School, University of Birmingham, Vincent Drive, Birmingham B15 2TJ (U.K.). 0168-5597/87/$03.50 © Elsevier Scientific Publishers Ireland Ltd.
areas on which u p p e r a n d lower fields are m a p p e d have different orientations. A l t h o u g h both CI a n d P100 display a form of polarity reversal with stimuli above a n d below the horizontal meridian, they p r o b a b l y have somewhat different cortical generators because their scalp d i s t r i b u t i o n s are dissimilar in several respects. F o r example, when pattern-reversal stimulation is restricted to a lateral hemifield, the m a i n positivity is ' p a r a d o x i c a l l y ' ipsilateral to the stimulus (Barrett et al. 1976; H a r d i n g et al. 1980) b u t not for p a t t e r n - o n s e t stimulation (Shagass et al. 1976). The response recorded over the contralateral (i.e., stimulated) hemisphere is a negative deflection with pattern-reversal ( B l u m h a r d t et al. 1978) but positive with p a t t e r n - o n s e t (Kriss a n d Halliday 1980). The physiological characteristics of the response also differ. Change in ampli-
CORTICAl. G E N E R A T O R S OF PATI'ERN-ONSET VEP
tude as a function of luminance is much greater for P100 than CI and the latency of P100 appears to change with check size whereas that of CI does not (Kriss and Barrett 1985). Finally, Estevez and Spekreijse (1974) showed that P100 is equivalent to a component of similar latency in the patternoffset response which can be distinguished from CI. The scalp distributions of these two components have been used to deduce the anatomical location of their generators. Early investigators proposed that CI arises in striate cortex (Jeffreys and Axford 1972; Jeffreys 1977) whereas P100 originates in the extrastriate areas of the occipital convexity (Halliday and Michael 1970), possibly extending onto its medial surface (Halliday et al. 1977). These proposals require that CI is generated by surface negative, and P100 by surface positive cortex. However, the view that CI is generated in the striate cortex has been questioned because it fails to predict the effects of foveal stimulation. This part of the visual field is believed to be mapped around the posterior pole of the occipital lobe whereas the peripheral field is mapped on the medial surface and in the calcarine sulcus (Holmes 1944; Spalding 1952). Electrical stimulation of the striate cortex in man suggests that relatively little of the medial surface of the hemisphere is devoted to the fovea (Brindley 1973; Dobelle et al. 1979), in contrast to anatomical findings (Brouwer 1932). Fig. 1A and 1B show how the peripheral field is believed to be mapped within the calcarine sulcus. The striate cortex is shown surface negative as Jeffreys proposes for the generation of CI. He argued that this arrangement accounts for the observation that a superiorly located electrode detects a negative CI when the upper field is stimulated and a positive CI when the lower field is stimulated. It follows that stimulation of the central area of the visual field, F, should evoke a surface negative response in the cortex around the occipital pole which serves the fovea (Fig. 1C). Posterior electrodes should therefore detect a negative CI to foveal stimulation. In fact, pattern-onset restricted to the central 2 ° of the visual field usually evokes a positive CI (Drasdo 1980; Les~vre and Joseph 1980). This and other
257
¥
R
A
B
ii
c Fig. 1. The octants of the peripheral field (A) and their representation on the banks of the calcarine fissure (B). The polarity of CI and its reversal can be accounted for if it is generated by surface negative striate cortex. Then an electrode over the midline would record a positive CI with stimuli in the lower half of the peripheral field (activating cortex at T, U, V, W). The same electrode would record a negative CI with stimuli in the periphery of the upper half-field (activating cortex at S, R, Y, X). The fovea is believed to be mapped around the external convexity of the occipital pole (C). It follows that stimulation of the fovea (F) alone should evoke a negative C1 at posterior electrodes.
features of its scalp distribution have led to the counter-proposal that CI is generated in surface positive cortex of extrastriate regions (Lesrvre and Joseph 1980; Lesrvre 1982).
)
258
We have studied the retinotopic organisation of the generators of CI of the pattern onset VEP to gain further information about their anatomical location. We stimulated sectors of the fovea and peripheral field individually and computed the location and orientation of the neural generators of CI using the method of Henderson et al. (1975), modified to incorporate the effects of skull thickness. A similar source location technique has been used by others to localise the generators of the pattern-onset evoked response to stimuli occupying individual quadrants of the visual field (Darcey et al. 1980a,b; Wood 1982).
S.R. B U T L E R ET AL. 17°
b
4°
5°
fo
13° 16° c
Method
Subjects and stimulation Twenty-three normal subjects took part in various sections of this investigation. The majority were male or female undergraduates aged 18-23. They sat in a sound-attenuated, electrically screened chamber and their E E G was recorded while they observed stimuli displayed on a high resolution, monochrome (black-and-white) television monitor. The stimuli were generated by an Acorn BBC microcomputer and took the form of checkerboard patterns (contrast ratio 0.9) whose onset replaced an isoluminant grey background. Subjects saw one of two different sets of stimuli: (a) Peripheral octants and fooeal half-fields. Fig. 2a shows each hemifield divided into 5 sectors. Evoked potentials were recorded from all 5 sectors of either the right or left half-field in each subject. Each sector of the peripheral field began 4 ° above or below the fixation point (Fig. 2b) and 5 ° lateral to it, and extended to 15 ° above or below the horizontal meridian and to 17 ° along it. Fig. 2c shows the form of stimulus used to stimulate half of the fovea, extending 1.6 ° laterally and 1.5 ° above and below the horizontal meridian. In each sector the individual checks subtended 50' for the majority of subjects. In a few individuals slightly smaller checks were used, down to 20'. Any effect this may have had on the amplitude of CI was small compared with differences between subjects. The stimulus was presented 192 times (3 groups
d
Fig. 2. a: schematic representation of the visual field, showing its division into peripheral octants and foveal half-fields, b and c: dimensions of the foveal and peripheral sectors, and fixation point F. d: dimensions of the foveal quadrant stimuli.
of 64 stimuli with short rest periods in between) in each sector of the field. Stimulus duration was 170 msec and the interstimulus interval randomised between 600 and 1200 msec. Pattern-onset was frame-locked to the TV raster to avoid the introduction of spurious spatial frequencies. Stimulation of such small areas of the field evokes potentials of relatively low amplitude and the signal-to-noise ratio is correspondingly degraded. Recordings were made from the 5 sectors of one hemifield in each subject until clear EPs were obtained for all the sectors of each hemifield in at least 6 subjects. For this reason the number of subjects used to obtain VEPs for left and right half-fields was unequal (10 and 7 respectively). (b) Fooeal quadrants. In a further 6 subjects VEPs were recorded with stimuli restricted to individual quadrants of the fovea. Each quadrant comprised a sector of a circular stimulus field, radius 2 ° and filled with 30' checks (Fig. 2d). In other respects the conditions of stimulation were the same as for peripheral octants.
EEG recording and averaging The EEG was recorded from Ag-AgC1 electrodes, previously shorted in normal saline to drain DC sources that might have generated thermal
CORTICAL GENERATORS OF PATTERN-ONSETVEP drift. For VEPs to peripheral octants and hemifovea, the recordings were made from 13 electrodes simultaneously, with A1 and A2 as an average reference: T3, T4, C3, Cz, C4, O1, 02, Oz and from modified positions T5, T6, P3, Pz, P4. The latter were shifted 5% of the nasion-inion distance posterior to their conventional locations to cluster more closely around the visual cortex. Note that A1 and A2 were not linked as this would disturb the potentials at the scalp, but averaged through the Grass average reference box. For the purpose of dipole localisation using the mathematical technique described below, the choice of reference is immaterial (so long as it is incorporated in the algorithm) - - although of course it affects the wave form in conventional write-outs. For recordings made with stimuli in foveal quadrants the reference was Fz, and F7 and F8 replaced T3 and T4. The signals were amplified on a 14-channel Grass Model 8-18C EEG machine (gain, 7/iV/cm; time constant, 1 sec; H F filter, 70 Hz; 50 Hz notch filter normally out) and saved together with trigger pulses from the stimulator on a RACAL Store 14 FM tape-recorder for subsequent offline analysis. The recordings were digitised by a 12-bit analogue-to-digital converter (ADC) and averaged on a North Star Advantage computer. The amplitude of the trace in each channel was measured at the time of the peak of CI in channel Pz. We have not investigated the source location of these EPs at other latencies. For this measurement, the baseline was taken as the average of the first 20 msec of the trace in each channel. The procedure resulted in a set of 13 numerical values defining CI over the electrode array for each sector of the field in each subject. When the investigation of peripheral sectors was carried out, only 8 channels of analogueto-digital conversion were available and the 13 channels were therefore obtained in two passes, with P3 channel common to both. Minor differences in the two passes of P3 are due to flutter in the tape-recorder. Dipole source location The pattern of 13 potential differences served as data from which to compute the locus and
259 orientation of the generators of CI by the following method. The potential at a point on the surface of a conducting sphere due to a dipole within it may be calculated given the orientation, positional coordinates and magnitude of the dipole; the potential is referenced to the centre of the dipole (for example, Wilson and Bayley 1950). In this way, the potential difference between any two points on the surface of the sphere can be derived. However, an inverse solution to such equations does not exist. That is to say, one cannot directly calculate the location and orientation of a source given the scalp potentials it creates. Instead the surface potentials are calculated for many values of orientation and location until a voltage pattern is found which matches the observed potentials (Henderson et al. 1975). An iterative minimisation algorithm is used to search for the best match between observed and calculated potentials without scanning all possible locations and orientations. This method provides a result which is meaningful only when the generator comprises a single, restricted area of cortex. The consequences of applying the method (inappropriately) to the case of a more complex arrangement of sources will be considered below. The skull has important effects on the electrotonic spread of voltage to the scalp. It both attenuates the signal and smoothes the spatial distribution of its peaks and troughs (Ary et al. 1981; Nunez 1981; Wood 1982). To correct for this, the algorithm employed in the present study is that described by Kavanagh et al. (1978) which models the head as 3 concentric spheres. The inner and outer shells contain high conductivity brain and scalp; the middle shell represents the skull. In our implementation, the radius of inner and outer tables of the skull were taken as 0.87 and 0.94 of the radius of the head. The absolute radius of the head is immaterial since the algorithm works on normalised values of the voltage distribution, ignoring the magnitude of the dipole. Values adopted for the volume resistivity of the brain and scalp (222 I2- cm) and the ratio of conductivity of brain to bone (80 : 1) were those obtained by Rush and Driscoll (1968). The effect of the bone un-
260
S.R. B U T L E R ET AL.
@@@ From left
From ebove
From behind
Fig. 3. Computed location and orientation of the dipoles to stimulation of the upper left octant next to the vertical meridian. Five of them form a cluster. In the sixth case (filled circle), the system was unable to find a close match between the field of a single dipole and the recorded voltages. The best match was obtained with a dipole of location and orientation which places it outside the cluster.
doubtedly outweighs the effect of lesser resistive anisotropies such as the CSF in the subarachnoid space and ventricles, not modelled here. The computation began with the dipole at an arbitrary location and orientation. The search for a minimum difference between observed and calculated potentials was executed several times for each VEP using different starting points as a check on the success of the minimisation technique in avoiding false minima. It is not normally possible to find a close match between observed and calculated potentials if the real source does not approximate to a dipole (i.e., is not singular, small in area, and uniformly polarised). To confine the method to cases where the source behaved like a dipole, the mean square difference between normalised calculated and normalised observed fields was obtained. The data were rejected if the ratio of this mean square difference to the maximum value in the normalised potential distribution exceeded 1:4. This is an arbitrary cut-off, derived from inspection of the fit of calculated and observed potential fields. In the study of foveal hemifields and peripheral octants, the computed locations of the source of the VEP for each sector were plotted on diagrams like that in Fig. 3. Here the majority of dipoles form a cluster even though they have been obtained from the VEPs of different subjects. However, one is clearly deviant. Such deviant dipoles were rejected before the cosine coordinates of the remainder were averaged to obtain a typical loca-
tion and orientation of the generator for each sector of the visual field. Accordingly, of the 6 clear EPs for each sector submitted to the localisation analysis, some were excluded at this stage. Thus for sectors R 1, S, T, U, V, W, X, Y and the left and right foveal sectors, nl (number of dipoles excluded) = 0, 0, 1, 0, 0, 1, 0, 1, 1, 2, respectively, and n2 (number of dipoles initially computed) = 2, 6, 7, 7, 8, 6, 9, 6, 5, 6, respectively. In the case of the study of foveal quadrants grand averages were obtained over all subjects to characterise the typical wave form and the amplitudes from this used as input for the localisation program.
Results
Fig. 4A shows the 13-channel evoked potential to pattern-onset in each of the 5 sectors of the left half of the visual field in one subject. Fig. 4B shows the corresponding evoked potentials for the right half of the field in another subject. The amplitude of CI varied considerably between subjects and between sectors of the field, being least well defined for the top octant of the right visual field (Fig. 4B). The phase reversal of CI for stimuli above and below the horizontal meridian is most marked in parietal and occipital channels contralateral to the half of the field in which the stimuli appeared. The latency of CI varied between subjects (90-110 msec). Numerical values of the voltage in all channels at the time of the peak in Pz (normally about 100 msec) were taken from all channels and used as data for localisation of the source active at that latency. The baseline for measurement of amplitude was taken as the average of the first 20 msec of each trace. Figs. 5A, B and 6 show the computed locations and orientations of the generators excited by stimulation of the peripheral octants, foveal halffields and foveal quadrants. The location of the dipole is referred to the sphere whose outer limit is depicted. In the model, the nasion-inion line cor-
i Note: no average dipole obtained for this sector.
CORTICAL GENERATORS OF PATTERN-ONSET VEP responds to a horizontal diameter at the equator of the sphere. The brain is shown in its approximate registration to this line, with the frontal pole just above the nasion and the occipital pole deep to the inion. In Fig. 5A and B the coordinates of the dipole have been obtained by averaging the cosine vectors across subjects for all cases in which a good match was found and where the dipoles from different subjects clustered. Stimulation of the upper right octant next to the vertical meridian failed to evoke a well defined CI in many subjects, its source was poorly matched as a dipole and where matches meeting the criterion were obtained the dipoles did not cluster across subjects. Consequently no dipole is shown for this sector in Fig. 5B. There is a consistent reversal in the orientation of the sources excited by stimulation of the pe-
A
261 ripheral octants above and below the horizontal meridian. This occurs irrespective of whether the octant lies next to the vertical or horizontal meridian. The sources lie in the hemisphere contralateral to the field stimulated in all cases, and usually somewhat closer to the midline than those excited by foveal stimulation. In general, the computed locations and orientations display mirror symmetry across the midline. Figs. 5 and 6 show that foveal stimuli excite sources far back in the occipital lobe and somewhat lateral to the midline in the hemisphere contralateral to stimulation. In the case of stimuli confined to individual quadrants of the fovea there is a tendency for dipoles to be located lower within the lobe for stimuli in the upper field than those for the lower field. However, there is no consistent swing in orientation as a function of the
B
Fig. 5. Dipoles computed for CI for each sector of the left (A) and right (B) half-fields. The circle marks the location of the dipole and its negative pole. The length and direction of the line in each plane shows its orientation. The coronal section shows the head viewed from behind.
262
S.R. BUTLER ET AL.
rr r
Tr
~
Tr
.,
T3 T4
T4
01
C3
c:
C4
c.~
02
TS T6
Oz
T5
P3
P3
P2
Tt
Tr
T3
P4
Tr T3 ~ T~ ~
T4
T5
Pz
~
Ts
~
1"6
o~ ~
T6
~
P3
P3
Tr
Tr
T3 ~ T4
~
cz
C4
G
02 ~
C.;
02
C3
Oz
~ Tr
Tr
~ C~
~
C4
02 Cz
T5
Pz
P3
P3
~
C~.
p:,---~P'~ i
Tr P4 T4 ~
~
.
T~
~
01
e
C:.
C4 Pz
T6.
Oz
P3 ~
P3
Tr!
Tr
L T~ ! T
4
;
T5
,
~ c.~ - ~ _ _ ~ , - - , . ~ w
C3 C4
Tri
'~
Cz Pz
- .
cz_.-.-.._./-x~.j
Oz P2 u, ÷
A
B
Fig. 4. Thirteen-channel evoked potentials to pattern onset in each sector of the left (A) and right (B) halves of the visual field. C o m m o n reference: A1 and A2. The VEPs to each half of the field were recorded in different subjects. The 13 channels were digitised in two passes of an 8-channel analogue to digital converter. P3 appears in both passes. Tr marks the period for which the stimulus was on (upward deflection). Ticks on the abscissa mark 100 msec intervals.
CORTICAL GENERATORS OF PATTERN-ONSET VEP
Fig. 6. Dipoles computed for CI with stimuli confined to foveal quadrants. The conventions are those used in Fig. 5.
location of the stimulus above or below the fixation point: the positive pole of the source faces posterolaterally in all foveal quadrants. There was little variation between subjects in the orientation of the source for the foveal EPs whose fields met our criteria for dipole matching.
Discussion
The evoked potentials to checkerboard stimuli (Fig. 4) replicate many important features of CI observed by previous workers using stimuli in restricted regions of the visual field. In particular they show the polarity inversion of CI obtained with stimuli above and below the horizontal meridian (Jeffreys and Axford 1972; Jeffreys 1977; Jeffreys and Smith 1979) and the absence of the paradoxical lateralisation characteristic of the pattern-reversal response (Barrett et al. 1976). The computed locations of the generators of CI varied systematically with stimulus location (Figs. 5 and 6). The equivalent dipoles were found to lie in the occipital lobe contralateral to the hemifield within which the stimulus occurred. Stimulation in
263
the peripheral octants excited sources which lay closer to the medial surface of the occipital lobe than to its lateral convexity. Their orientation changes according to whether the stimulus was above or below the horizontal meridian, the positive pole was directed upwards for stimuli in the lower field, downwards for stimuli in the upper field. It was not possible to resolve the upper right octant as a single dipole with consistent orientation across subjects. It is conceivable that the area generating CI for this sector spans a curvature in the surface of the cortex such as the lip of a gyrus or base of a sulcus. The generators would be distributed at various angles and the field would be particularly susceptible to individual differences in convolution. Stimulation of the fovea excites sources in a more posterior location, oriented radially to the posterolateral convexity of the occipital lobe, positive pole outward. Stimuli in the upper part of the fovea excited sources somewhat lower in the occipital lobe than did stimuli in the lower part of the fovea on the same side. This is in accord with the inverted representation of this part of the visual field on the occipital cortex, irrespective of whether the generators lie in striate or peristriate cortex (Drasdo 1977; 1980). Darcey et al. (1980a, b) describe the location and orientation of sources excited by pattern-onset, using a red checkerboard. They examined the first peak in the VEP (92 msec) when individual quadrants of the field were stimulated and analysed the distribution of scalp potentials using the algorithm of Kavanagh et al. (1978). Their stimuli included foveal quadrants and extended 10 ° into the peripheral field, i.e., only partially overlapping the area of the peripheral field stimulated by us. The location and orientation of their dipoles correspond closely with those found for foveal stimulation in this study. Both investigations disclose dipoles for quadrants of the left field oriented obliquely above and below the horizontal with stimuli on opposite sides of the horizontal meridian; no such change of orientation occurred with stimuli in the right field (Fig. 6). Jeffreys and Axford (1972) also recorded VEPs
264
with pattern onset in separate quadrants of the field. With the stimuli employed in that study, the amplitude of CI was found to be chiefly dependent on stimulation of the peripheral part of each quadrant. The location and orientation of the sources excited by such stimuli correspond well with those we report here for peripheral octants. The generators for both upper and lower fields appear to lie close to one another, in a position congruent with that of the striate cortex. Their orientations reverse through 180 ° as would be expected if the generators faced one another across the calcarine fissure. Jeffreys envisaged a cruciform representation of the peripheral field: octants close to the vertical meridian were mapped on the medial face of the hemisphere, in the cuneus and lingual gyrus while those close to the horizontal meridian were supposed to lie in the depths of the calcarine fissure as proposed by Holmes (1944). This model leads to the expectation that sources excited by stimulation of octants close to the vertical meridian will be orientated horizontally, i.e., normal to the medial surface of the hemisphere. However, our d a t a provide no support for this prediction (Fig. 5). Although there may be variations between individual subjects, one would expect at least a tendency toward a more horizontal orientation of the dipole for octants adjacent to the vertical meridian. This is not the case consistently. It might be argued that the cortex of the medial surface makes a relatively minor contribution to the evoked response because most of the striate cortex lies within the calcarine fissure (Filimonoff 1932; Brindley 1972; Stensaas et al. 1974). If so the cruciform model would need revision to accommodate this. A more serious objection to the idea that CI is generated by striate cortex concerns the orientation of the source excited by foveal stimuli. Assuming that the fovea is mapped around the occipital pole and if CI arises in surface negative cortex, the dipoles excited by stimulation of the central field should be oriented with the negative pole posteriorly (Fig. 1). Our results show that their orientation is the opposite. In this respect our observations accord with those of Les6vre and Joseph (1980) and Les+vre (1982) and appear to
S.R. BUTLER ET AL.
rule out an explanation of the origin of CI in terms of generally accepted concepts of the organisation of the striate cortex. However, the data cannot by interpreted in terms of sources in prestriate cortex, instead. The observed changes in orientation could be produced only if the potentials for sectors above and below the horizontal meridian were generated in superior and inferior occipital territories. These sources would be far apart; the data reveal no such separation as a function of the sector stimulated. The lack of separation is not an artefact caused by the use of inappropriate values for bone conductivity. We have examined the effects of assuming brain-bone conductivity ratios between 1:1 and 1:800, thus bracketing the values obtained by Rush and Driscoll (1968) of 1 : 80. These procedures failed to produce any separation of dipoles for upper and lower fields (Butler et al. 1984; Hancox et al. 1984). Nor is it likely that the generators lie outside the cortex (Blumhardt et al. 1982; Celesia et al. 1982; Fava et al. 1985). Nor can it be postulated that the invariance of the computed locations is due to the simultaneous activation of many supplementary visual areas in prestriate cortex distributed round the occipital pole, because in this case, no inversion of the equivalent dipole should occur with stimulation of different parts of the field. Moreover, this interpretation would conflict with the view of Jeffreys (Jeffreys 1977; Jeffreys and Musselwhite 1983) that CI originates in a single anatomical area. Furthermore, the superior and inferior cortical territories seem distinguished by their roles in spatial perception and object recognition rather than their retinotopic relationships (Ungerleider and Mishkin 1982). We have seen that neither single nor multiple sources outside the striate cortex provide a satisfactory explanation for the origin of CI. We must therefore reconsider the question of its origin in that cortex. Its appearance as the first major component of the onset VEP, its computed location, and the orientation of its sources when stimuli are in the peripheral field, all fit with its generation by surface negative striate cortex. Only its polarity when the fovea is stimulated poses a problem. The problem exists because it is believed that the fovea
C O R T I C A L G E N E R A T O R S OF P A T T E R N - O N S E T VEP
is represented only around the occipital pole, and because we assume that the macropotentials evoked by our stimuli will be equivalent for both foveal and peripheral parts of the field. The projection of the fovea is distinguished from that of the peripheral retina in several ways, including the distribution of colour opponent channels (Zrenner and Gouras 1983), the absence of input from rods (Polyak 1941), the size of the receptive fields and the proportion of X and Y cells (de Monasterio 1978). Moreover, the cytoarchitecture of the striate cortex is not uniform (Ngowyang 1934) and it has been suggested that the major subdivisions reflect a different cellular organisation for the projection of the fovea from that serving the peripheral field (Smith 1930a,b). Accordingly it is indeed possible that the macropotentials of foveal and extrafoveal striate areas may differ - - even to the extent that their responses to the afferent volley are of opposite polarity for stimuli of the same spatial frequency. The second assumption concerns the representation of the fovea at the occipital pole. Serial sections reveal that the calcarine sulcus normally forks in its depth to become Y or T shaped in cross-section. It is seldom a simple straight groove. The top of the T often emerges as the lateral calcarine sulcus (LCS) on the medial surface of the lobe at the posterior end of the calcarine fissure, just anterior to the tip of the pole (Bolton --}~"
I
Super,o~
Anterior
Inl
Poslerioc
X b
c
Fig. 7. a: medial view of right occipital lobe. Lateral calcarine sulcus (lcs), calcarine fissure (c) and parieto-occipital fissure (po). b: coronal section through lobe in plane X - X showing the typically T-shaped fundus of the calcarine sulcus. Stippled area marks the typical distribution of striate cortex, c: horizontal section through lobe in plane Y - Y showing the manner in which cortex of the pole curves into the lcs where it continues as the fundus of the calcarine sulcus. Typical distribution of striate cortex is marked as b.
265
1900; Filimonoff 1932; Polyak 1957) (Fig. 7a,b,c). In an anatomical study (in preparation), we have found LCS to be present, though highly variable in length, in 26 of 28 hemispheres examined. In such cases the cortex of the pole does not wrap directly onto the medial surface of the lobe but onto the posterior wall of the LCS where its orientation is exactly the reverse of that on the pole. The variability of this territory is in accord with the wide individual differences in the topography of the VEP observed using stimuli confined to the fovea (Jeffreys 1977; Halliday 1982; Brecelj and Cunningham 1985). How far the foveal map extends into the LCS, when present, is unknown. If the majority of the fovea were mapped on its posterior bank as an extension of the cortex at the pole, surface negative cortex would give rise to a positive CI at the occiput. Such a location for the foveal projection remains compatible with the data of Holmes (1944) and Spalding (1952) since the fovea would still be mapped within the occipital pole. Extrapolation of the cortical magnification factor from subhuman primates indicates that a disproportionate area of striate cortex is allocated to the fovea (Daniel and Whitteridge 1961; Cowey and Rolls 1974; Drasdo 1977). At present we cannot determine whether the factor is such that in man the foveal striate cortex extends beyond the pole and into the LCS. If the representation of the central field were folded back on itself in this way, then in some individuals stimulation of the central part of the fovea would excite sources both on the pole and within the sulcus. The resulting local fields would tend to cancel, and CI would have a disproportionately small amplitude. Several authors (Jeffreys 1977; Spekreijse et al. 1977; Drasdo 1980) have commented on the low amplitude of CI with foveal stimuli, though the effect might equally be a consequence of other variation in the morphology of the pole. The present study has not enabled us to distinguish between a striate and an extrastriate origin for the CI component of the VEP. A source, or multiple sources, outside the striate cortex seem to fit the data less well than a striate origin. However, any simple explanation, in terms of its origin in the striate cortex, may necessitate either refining the accepted view of retinotopic organisation,
266 or assuming that the polarity of macropotentials i n t h e s t r i a t e c o r t e x is n o t u n i f o r m f o r all p a r t s o f t h e v i s u a l field. This investigation was supported by the Medical Research Council. We are grateful to Dr. R.M. Flinn for assistance with computing.
References Ary, J.P., Klein, S.A. and Fender, D.H. Location of sources of evoked scalp potentials: corrections for skull and scalp thicknesses. IEEE Trans. biomed. Engng, 1981, BME-28: 447-452. Barrett, G., Blumhardt, L., Halliday, A.M., Halliday, E. and Kriss, A. A paradox in the lateralisation of the visual evoked response. Nature (Lond.), 1976, 261: 253-255. Blumhardt, L.D., Barrett, G., Halliday, A.M. and K.riss, A. The effect of experimental 'scotomata' on the ipsilateral and contralateral responses to pattern-reversal in one half-field. Electroenceph. clin. Neurophysiol., 1978, 45: 376-392. Blumhardt, L.D., Barrett, G., Kriss, A. and Halliday, A.M. The pattern-evoked potential in lesions of the posterior visual pathways. Ann. N.Y. Acad. Sci., 1982, 388: 264-289. Bolton, J.S. The exact histological localisation of the visual area of the human cerebral cortex. Phil. Trans. roy. Soc. B, 1900, 193: 165-222. Brecelj, J. and Cunningham, K. Occipital distribution of foveal half-field responses. Docum. ophthal. (The Hague), 1985, 59: 157-165. Brindley, G.S. The variability of human striate cortex. J. Physiol. (Lond.), 1972, 225: lp-3p. Brindley, G.S. Sensory effects of electrical stimulation of the visual and paravisual cortex in man. In: R. Jung (Ed.), Handbook of Sensory Physiology. 7/3. Central Processing of Visual Information. B. Springer, Berlin, 1973: 583-594. Brouwer, B. Projection of the retina on the cortex in man. Res. Publ. Ass. nerv. ment. Dis., 1932, 13: 529-534. Butler, S.R., Hancox, R.J., Georgiou, G.A. and Glass, A. The 90-100 msec component of the pattern onset visual evoked response does not originate in striate cortex. Behav. Brain Res., 1984, 12: 177. Celesia, G.G., Polcyn, R.D., Holden, J.E., Nickles, R.J., Gatley, J.S. and Koeppe, R.A. Visual evoked potentials and positron emission tomographic mapping of regional cerebral blood flow and cerebral metabolism: can the neuronal potential generators be visualized? Electroenceph. clin. Neurophysiol., 1982, 54: 243-256. Cowey, A. and Rolls, E.T. Human cortical magnification factor and its relation to visual acuity. Exp. Brain Res., 1974, 21: 447-454. Daniel, P.M. and Whitteridge, D. The representation of the visual field on the cerebral cortex in monkeys. J. Physiol. (Lond.), 1961, 159: 203-221. Darcey, T.M., Ary, J.P. and Fender, D.H. Methods for the localization of electrical sources in the human brain. In:
S.R. BUTLER ET AL. H.H. Kornhuber and L. Deecke (Eds.), Prog. Brain Res., Vol. 54. Elsevier, Amsterdam, 1980a: 128-133. Darcey, T.M., Ary, J.P. and Fender, D.H. Spatio-temporal visually evoked scalp potentials in response to partial-field patterned stimulation. Electroenceph. clin. Neurophysiol., 1980b, 50: 348-355. Dobelle, W.H., Turkel, J., Henderson, D.C. and Evans, J.R. Mapping the representation of the visual field by electrical stimulation of human visual cortex. Amer. J. Ophthal., 1979, 88: 727-735. Drasdo, N. The neural representation of visual space. Nature (Lond.), 1977, 266: 554-556. Drasdo, N. Cortical potentials evoked by pattern presentation in the foveal region. In: C. Barber (Ed.), Evoked Potentials. MTP Press, Lancaster, 1980: 167-174. Estevez, O. and Spekreijse, H. Relationship between pattern appearance-disappearance and pattern reversal responses. Exp. Brain Res., 1974, 19: 233-238. Fava, E., Ducati, A., Motti, E.D.F. and Marossero, F. Intracerebral recordings of PR-VEPs in three awake patients. Electroenceph. clin. Neurophysiol., 1985, 61: S102. Filimonoff, I.N. 0ber die Variabilit~it der Grosshirnrindenstruktur. Mitteilung II. Regio occipitalis beim erwachsenen Menschen. J. Psychol. Neurol. (Lpz.), 1932, 44: 1-96. Halliday, A.M. Evoked Potentials in Clinical Testing. Churchill-Livingstone, Edinburgh, 1982. Halliday, A.M. and Michael, W.F. Changes in pattern-evoked responses in man associated with the vertical and horizontal meridians of the visual field. J. Physiol. (Lond.), 1970, 208: 499-513. Halliday, A.M., Barrett, G., Halliday, E. and Michael, W.F. The topography of the pattern evoked potential. In: J.E. Desmedt (Ed.), Visual Evoked Potentials in Man: New Developments. Clarendon Press, Oxford, 1977: 121-133. Hancox, R.J., Butler, S.R., Glass, A. and Georgiou, G. Source location of the 100 msec component of the pattern onset visual evoked response. Electroenceph. clin. Neurophysiol., 1984, 58: 14P. Harding, G.F.A., Smith, G.F. and Smith, P.A. The effect of various stimulus parameters on the lateralization of the VEP. In: C. Barber (Ed.), Evoked Potentials. MTP Press, Lancaster, 1980: 213-218. Henderson, C.J., Butler, S.R. and Glass, A. The localization of equivalent dipoles of EEG sources by the application of electrical field theory. Electroenceph. clin. Neurophysiol., 1975, 39: 117-130. Holmes, G. The organisation of the visual cortex in man. Proc. roy. Soc. B, 1944, 132: 348-361. Jeffreys, D.A. The physiological significance of pattern visual evoked potentials. In: J.E. Desmedt (Ed.), Visual Evoked Potentials in Man: New Developments. Clarendon Press, Oxford, 1977: 134-167. Jeffreys, D.A. and Axford, J.G. Source locations of patternspecific components of human visual evoked potentials. I. Component of striate cortical origin. Exp. Brain Res., 1972, 16: 1-21.
CORTICAL GENERATORS OF PATI'ERN-ONSET VEP Jeffreys, D.A. and Musselwhite, M.J. Evoked potential evidence of parallel processing in human striate cortex. Electroenceph, clin. Neurophysiol., 1983, 56: 17P. Jeffreys, D.A. and Smith, A.T. The polarity inversion of scalp potentials evoked by upper and lower half-field stimulus patterns: latency or surface distribution differences? Electroenceph, clin. Neurophysiol., 1979, 46: 409-415. Kavanagh, R.N., Darcey, T.M., Lehmann, D. and Fender, D.H. Evaluation of methods for three-dimensional localization of electrical sources in the human brain. IEEE Trans. biomed. Engng, 1978, BME-25: 421-429. Kriss, A. and Barrett, G. Checkerboard VEPs: a review of stimulus response relationships, In: D. Papakostopoulos, S.R. Butler and I. Martin (Eds.), Clinical and Experimental Neuropsychophysiology. Croom Helm, London, 1985: 1-29. Kriss, A. and Halliday, A.M. A comparison of occipital potentials evoked by pattern onset, offset and reversal by movement. In: C. Barber (Ed.), Evoked Potentials. MTP Press, Lancaster, 1980: 205-212. Lesrvre, N. Chronotopographical analysis of the human evoked potential in relation to the visual field (data from normal individuals and hemianopic patients). Ann. N.Y. Acad. Sci., 1982, 388: 156-182. Les+vre, M. and Joseph, J.P. Hypothesis concerning the most probable sites of origin of the various components of the pattern evoked potential. In: C. Barber (Ed.), Evoked Potentials. MTP Press, Lancaster, 1980: 159-166. Michael, W.F. and Halliday, A.M. Differences between the occipital distribution of upper and lower field pattern evoked responses in man. Brain Res., 1971, 32: 311-324. Monasterio, F.M. de. Properties of concentrically organised X and Y ganglion cells of macaque retina. J. Neurophysiol., 1978, 41: 1394-1417. Ngowyang, G. Die cytoarchitektonischen Unterfelder der Area striata beim Menschen. J. Psychol. Neurol. (Lpz.), 1934, 46: 351-363. Nunez, P.L. Electric Fields of the Brain: the Neurophysics of EEG with Contributions by Ron D. Katznelson. Oxford University Press, New York, 1981.
267 Polyak, S.L. The Retina. University of Chicago Press, Chicago, IL, 1941. Polyak, S.L. The Vertebrate Visual System. University of Chicago Press, Chicago, IL, 1957. Rush, S. and Driscoll, D.A. Current distribution in the brain from surface electrodes. Anesth. Analg., 1968, 47: 717-723. Shagass, C., Amadeo, M. and Roemer, R.A. Spatial distribution of potentials evoked by half field pattern-reversal and pattern-onset stimuli. Electroenceph. clin. Neurophysiol., 1976, 41: 609-622. Smith, G.E. New light on vision. Nature (Lond.), 1930a, 125: 820-824. Smith, G.E. The cortical representation of the macula. J. Anat. (Lond.), 1930b, 64: 477-478. Spalding, J.M.K. Wounds of the visual pathways. Part I. The visual radiation. J. Neurol. Neurosurg. Psychiat., 1952, 15: 99-109. Spekreijse, H., Estevez, O. and Reits, D. Visual evoked potentials and the physiological analysis of visual processes in man. In: J.E. Desmedt (Ed.), Visual Evoked Potentials in Man: New Developments. Clarendon Press, Oxford, 1977: 16-89. Stensaas, S.S., Eddington, D.K. and Dobelle, W.H. The topography and variability of the primary visual cortex in man. J. Neurosurg., 1974, 40: 747-755. Ungerleider, L.G. and Mishkin, M. Two cortical visual systems. In: D.J. Ingle, M.A. Goodale and R.J.W. Mansfield (Eds.), Analysis of Visual Behavior. MIT Press, Cambridge, MA, 1982: 549-586. Wilson, F.N. and Bayley, R.H. The electric field of an eccentric dipole in a homogeneous spherical conducting medium. Circulation, 1950, 1: 84-92. Wood, C.C. Application of dipole localization metl~ods to source identification of human evoked potentials. Ann. N.Y. Acad. Sci., 1982, 388: 139-155. Zrenner, E. and Gouras, P. Cone opponency in tonic ganglion cells and its variation with eccentricity in rhesus monkey retina. In: J.T. Mollon and L.T. Sharpe (Eds.), Colour Vision: Psychophysiology and Psychophysics. Academic Press, London, 1983: 211-223.