Brain Research, 101 (1976) 11-22 © Elsevier Scientific Publishing Company, Amsterdam - Printed in The Netherlands
ll
S P I N D L E WAVE S Y N C H R O N Y IN T H E SOMATOSENSORY CORTEX OF T H E CAT
T R O N D GANES
The Institute of Neurophysiology, University of Oslo, Oslo (Norway) (Accepted June 6th, 1975)
SUMMARY
(1) Spontaneous barbiturate spindles were recorded from the primary and secondary somatosensory cortex. The recordings were concentrated to areas surrounding several reference loci. The recording sites producing the maximal response evoked by stimulation of an exposed nerve in a contralateral limb were used as reference loci. (2) Spindles recorded at various distances from the respective reference loci were cross-correlated to spindles developing simultaneously in the latter. High correlation coefficients, indicating a considerable degree of wave synchrony, were obtained between spindles in the reference locus and spindles recorded a few millimetres from this site. The correlation coefficients decreased with increasing interelectrode distance. A relatively sharp fall in the correlation coefficients was generally found 2-3 mm from the reference locus. Small amounts of sodium pentobarbital, given intravenously at intervals of 5 min, had no effect upon this pattern. (3) The change in the correlation coefficients was followed by a parallel change in the amplitude of the evoked potentials. The iso-correlation lines of spindle wave synchrony and the iso-amplitude lines of the evoked potentials had a similar distribution and extension for each particular reference locus. (4) Lateral spread o f spindle waves in the cortex seems to be of minor importance, since a vertical lesion cutting the cortico-cortical fibres did not reduce the wave synchrony of the spindles recorded from either side of the lesion. (5) The majority of the spindles recorded in the close vicinity of a reference locus started simultaneously within ~: 0.1 sec. This pattern changed with increasing distance from the reference locus and 5-6 mm away only a fraction of the spindles started simultaneously. However, within the entire primary somatosensory cortex a small but significant coupling existed between onset of the spindles.
12 INTRODUCTION
The sensory input to the cortex is organized in a strictly topographical fashion. This applies to both the primary (SI) and to the secondary (SII) somatosensory receiving areasl,",l~,~4,17 ~a:'3,26,~L There is also evidence that the thalamo-cortical CIC) systems, which take part in barbiturate spindle activity, are organized in a similar way3,6,9,10,1;',~6. Several investigations2,6,1z,13, '~4,9~ have suggested that barbiturate spindle activity in the somatosensory cortex is generated by rhythmic activity of the same TC connections and intracortical synapses which are involved in the prinlary somatosensory responses. One consequence of this hypothesis is that various cortical areas ought to show different rhythmic patterns because they receive their rhythmic input from dissimilar thalamic nuclei and subnuclei. The purpose of this study was to examine to what extent the barbiturate spindle activity in SI and SII areas is topographically organized and to determine some of the factors which influence the distribution of such spindle waves in the cortex. METHODS
The experiments were performed on cats anaesthetized with 30 mg/kg sodium pentobarbital given intraperitoneally. Light anaesthesia characterized by withdrawal of the forelimb and hindlimb to light pinching was maintained during the experiments by additional intravenous injections of small amounts of barbiturate. Operation and stimulation. The cortical surface was exposed by gentle removal of the overlying skull and covered by a pool of liquid paraffin kept at 37 °C during the experiment. Three peripheral forelimb nerves, the superficial radial, the superficial branch of the ulnar and the median nerve were exposed distal to the elbow joint and mounted in separate stimulation cuffs: In the hindlimb, the posterior tibial nerve was exposed for stimulation distal to the knee joint. The stimulus threshold for evoking a short latency cortical potential was determined for each nerve. Later in the experiment voltages from 1.5 to 2 times the threshold were used. The pulse duration was 0.1 msec. Recording technique. The cortical potentials were recorded monopolarly with silver or platinum ball electrodes (diameter 0.2-0.3 mm) placed on the pial surface under direct visual guidance through a stereoscopic microscope. Three to four electrodes with a permanent interelectrode distance of I mm were mounted in a manipulator for rapid mapping of cortical potentials. The cortical site giving maximal short latency response to stimulation of one of the peripheral nerves was defined as the cortical reference point for that particular nerve projection, To map the distribution of evoked potentials and barbiturate spindles with respect to this site, one electrode remained in this position while the other electrodes were moved systematically around in adjacent cortical areas. The recorded potentials were fed through preamplifiers and displayed on a storage oscilloscope for direct examination and on a normal oscilloscope for filming. In addition, the records were taped on a multichannelled tape recorder for later analysis.
13
Computer and calculation methods'. Analyses of the evoked responses, as well as the barbiturate spindle waves, were done by a NORD-I computer, either on-line or by off-line analysis. The computer methods and programmes have been presented earlier t6. The program used for cortical evoked potentials had no amplitude limit, measuring all peaks above or below the averaged baseline. When plotted on a cortical map, however, the smallest amplitude used was taken as 30 % of the maximal amplitude. In plotting the spindle correlation values between the records from two electrodes, only the initial value (with zero displacement of the records) of each pair of spindles was plotted. The lowest correlation value (r) used was selected to match the lowest relative value of the evoked potential. This corresponds to a cross-correlation coefficient of 0.5 and a coefficient of determination (r e) of 0.25, indicating that 25 % of the variability in the relative amplitudes of the spindle waves was due to a linear relationship between the two records. RESULTS
Spindle wave coherences and interelectrode distance Spontaneous barbiturate spindles recorded from two sites on the cortical surface 0.3-0.5 mm apart started simultaneously and were almost identical in form. When such spindles were cross-correlated, the maximal cross-correlation values were close to + 1. Since two surface electrodes in this position obviously are influenced by much the same cortical activity, this result was expected. In fact, in this case, cross-correla-
A
O.Smm
ULN
C
B
STIM
o~05
--/'L/"
0
-IO
_0.2 IOtas
"
. . . .
" • . . . . - I 0 0 ms
"
®
•
,
.
I 0 0 ms
0 ,
, is
~ -05
.
.
.
.
.
.
.
.
o - I 0
.
.
.
.
.
.
.
- I 0 0 ms
.
.
.
.
.
0
°
•
.
IO0 ms
DISPLACEMENT
Fig. I. W a v e s y n c h r o n y o f cortical spindles. A : two electrodes on the pial surface in the p r i m a r y s o m a t o s e n s o r y cortex (SI). Distance between the electrodes is 0.3 ram. Cortical responses evoked by contralateral s t i m u l a t i o n are nearly similar in the two sites. B: s p o n t a n e o u s spindles f r o m the s a m e sites. C: cross-correlogram o f the spindles in B. D : one o f the electrodes h a s been m o v e d 4 m m medially. T h e evoked response f r o m this locus h a s nearly vanished. E: spindles with a similar onset in the two recording sites. F: c r o s s - c o r r e l o g r a m o f the spindles in E.
14 tion does not differ significantly from auto-correlation. In Fig. t spindles were recorded with two electrodes 0.3 mm apart in the ulnar projection area o|" SI (A). The spindles (B) were similar, consisting of individual spindle waves at regular intervals. Cross-correlation of the records (C) resulted in a regular sinus-shaped cross-correlogram with a maximal correlation factor of 0.98. Peaks of high correlation values appeared at record displacements of about 100 msec, corresponding to an intraspindle frequency of l0 Hz. When. the intraspindle frequency was less regular, the cross-correl ograms appeared more damped with a less regular sinusoidal form. In Fig. I D, one o f the electrodes was moved 4 mm medially while the other remained unchanged. The cortical evoked potential revealed that the new electrode locus was outside the ulnar projection area. Spindles were, however, still present at both electrode sites and sometimes occurred simultaneously (E). They differed, however, both in duration and shape of the spindle waves. The cross-correlogram (f~), although still retaining some of the sinusoidal features, invariably gave a low correlation, the factors never exceeding ~ 0.3. Based on this pilot experiment it was clear that spindle wave coherence, expressed in terms of the cross-correlation factor, was related to the distance between the recording electrodes and also in some way to the organization of the cortex. In order to get more precise information on these points, the distribution of spindle wave coherences and evoked potentials in SI and SII were mapped systematically. In the primary sensory cortex the forelimb and hindlimb projection areas were mapped. In the former, the subdivisions represented by the projection areas of three peripheral forelimb nerves, the ulnar (U), median (M) and superficial radial nerve (SR) were determined. In the hindlimb area, stimulation of contralateral posterior tibial nerve (PT) was used. No attempts were made to separate between the different sensory modalities. The cortical site giving a maximal short latency response to contralateral nerve stimulation was defined as the cortical reference site (REF) of the particular nerve. Mapping was concentrated to the cortex surrounding the REF sites. One electrode was left in this locus and the other moved along radial tracks in steps of 1 mm. At each new recording site several spindles were recorded and cross-correlated to spindles simultaneously recorded by the REF electrode. The amplitudes of evoked potentials were recorded simultaneously. Generally, the electrode was moved along four different rays through the REF site (dots in Fig. 2A). For convenience only two of the rows shown in Fig. 2Aa and b are plotted in B and C_ Each plot of spindle wave coherence and amplitude of evoked potentials was based upon 10 cross-correlated spindles and the same number of evoked potentials. Fig. 2D shows similar data from SII. Only one row of recording points in the SII forelimb area (b) and one row in the hindlimb area (a) were plotted in E and F, respectively. The spindle wave synchrony was directly related to the distance from the REF sites. The cross-correlation coefficient decreased rapidly with increasing interelectrode
15
A
C
B
a~ ~ v
SPINOLE CORRELATION .... EV~tOPOTENTIAL *1 100%
*1 tO0% s' ~',
01150 I* *- °*
4
D
3
Z
0
2
I
S
E °l, 100%
4
4
F
,100%
3
2
r
F
I
O
I
5
~'
~'~SUPRASYLV. F
a
ECTOSYLV.
~0.5, S
,
ooeoOeeo MED
b ",,/O,H,
"•
~ 4
L .-"
., 3
2
I
0
I
Z
INTERELECTRODE DISTANCE
3 4 rm~
3
2
L
L
2
3
".
I
.0
I
4
Fig. 2. Distribution of cortical spindle wave synchrony. A: the cortical recording sites are illustrated by dots. The two open circles indicate the reference locus in the SI forepaw (lower) and hindlimb (upper) projection area. The results from the two recording tracks a and b are plotted in the diagrams of B and C respectively. The maximal cross-correlation factors of the spindles (average of 5-10 analyzed spindle pairs) are plotted as filled circles• The relative amplitude of the evoked responses (in per cent of the maximal response) are plotted as crosses. The abscissae of the diagrams give the distance from the reference electrode. The vertical lines through the correlation plots indicate dispersion of the average wave synchrony from one spindle pair to the next expressed in S.D. D : mapping in the second somatosensory area (SII). The results of two recording tracks a and b are plotted in E and F, respectively.
distance. In m o s t tracks it reached a value o f + 0.3 or less w h e n spindles were recorded m o r e than 3-4 m m f r o m the R E F site. A l o n g all tracks this decrease in wave coherence was followed by a parallel decrease in amplitude o f the e v o k e d potentials. B o t h the spindle wave synchrony and the evoked potentials were reduced m o r e rapidly in SII than in SI when the interelectrode distance was increased. A l o n g s o m e tracks, h o w ever, a correlation factor as high as + 0 . 5 to + 0 . 6 persisted up to distances o f 4 - 5 m m f r o m the R E F site. This feature was observed when spindles were recorded f r o m the cortical area between SI and SII, with the R E F site either in SI or in SII. Examples o f this are seen in Fig. 2C and E.
Topographical organization of coherent spindles In order to compare the distribution o f coherent spindle activity with the sensory projection areas, maps were m a d e by drawing lines between cortical points with identical correlation values and similarly, between points with evoked potentials o f the same amplitude (Fig. 3). Linear intrapolation between adjacent recording points was done to obtain isocorrelation lines representing correlation factors o f + 0 . 9 , + 0 . 7 , + 0 . 5 and iso-amplitude lines representing 90, 70, and 30 ~ o f the m a x i m a l e v o k e d potential amplitude.
16
D4p
G //
iI
SPINDLEWAVE SYNCHRONY
.
.
.
H
.
SII
EVOKED
POTENTIAL
F
c
/
2mm
J ~ S
'..
i::
EP= 70
"
EP°
5o
"
EP = 5 0
"
......
EVOKED
POTENTIAL
, SPINDLEWAVE SYNCHRONY
Fig. 3. Extent of cortical areas with a high spindle wave synchrony. A: the enclosed area indicates the areas shown in detail in B and C. The black dots in B and C represent the reference sites. The surrounding lines are iso-correlation lines for spindle wave synchrony (B) and iso-amplitude lines for the sensory evoked potentials (C). PT, SR, U and M are the cortical reference loci for the contralateral posterior tibial, superficial radial, ulnar and median nerves, respectively. E and F: similar maps of iso-correlation lines and iso-amplitude lines for the second somatosensory area, as indicated by the enclosed area in D. G : the iso-correlation lines and iso-amplitude lines relative to a reference focus in SI are superimposed. H: similar to G but with the reference focus in SII.
Fig. 3B a n d E illustrates the d i s t r i b u t i o n o f spindle wave s y n c h r o n y s u r r o u n d i n g six different R E F sites. A l l spindles r e c o r d e d inside the o u t e r line (r = 0.5) have a crosscorrelation coefficient o f + 0 . 5 or m o r e when c o m p a r e d to spindles simultaneously rec o r d e d f r o m the R E F site. A certain overlap between the three f o r e l i m b areas investig a t e d was observed, b u t the size a n d shape o f their c o r r e s p o n d i n g spindle coherent areas differed significantly (Fig. 3B). A c o m p a r i s o n between Fig. 3B a n d C a n d between E a n d F shows t h a t the spindle wave coherence was i n t i m a t e l y related to the extent o f the various sensory p r o j e c t i o n areas b o t h in the SI a n d SII. The i s o - c o r r e l a t i o n lines a n d i s o - a m p l i t u d e lines relative to a reference focus in SI are s u p e r i m p o s e d in Fig. 3G, a n d a similar s u p e r i m p o s i t i o n o f d a t a for a reference p o i n t in SII is seen in Fig. 3H.
17
A
W
:~ 0.9
B
._1
.3
08
0-= 0.1
~ 07 3mm
Q9'
C M- 0.77
M=0.5 No= 8
0.8.
0"= 0 0 7
0.7
o 0.6
0.6
"' 0.5 .J
0.5
z 0.4
0.4
No- 9
a
ffl
BEFORE LESION
AFTER LESION
Fig. 4. Lack of lateral spread of spindle wave synchrony. A : recording situation. The jagged line between the recording sites illustrates the extent of the cortical lesion. SSM, middle suprasylvian gyrus; LAT, lateral gyrus. B and C: dots give the maximal cross-correlation coefficients of 8 spindle pairs recorded before (B) and 9 spindle pairs recorded after (C) the cortical lesion. M, mean; ~, the S.D. No, number of spindle pairs analyzed.
Spread of spindle waves in the cortex Although highly coherent spindles were found within definite and relatively small cortical zones corresponding to the main sensory receiving areas, some of this activity could be due to spread from neighbouring regions by intracortical or corticocortical connections. Therefore, experiments were performed in which the corticocortical connections between two recording sites in the cortex were cut (Fig. 4). Cortical spindles were recorded in SI from two loci 2 mm apart (A). Cross-correlation of the spindles resulted in correlation factors between -+-0.7 and +0.4, the mean value being +0.5 (B). A vertical incision into the cortex was made between the recording electrodes with the electrodes left in situ. Care was taken not to displace the electrodes. The lesion, verified histologically, extended through all cortical layers and between 1 and 2 mm into the underlying white matter. Cortical spindle activity returned immediately after the lesion in both recording sites. Contrary to what might be expected, cross-correlation of the spindles (C) revealed a significant increase in spindle wave coherences. The experiment, repeated several times, both in SI and in the lateral and suprasylvian gyri, always gave the same result. Start intervals of cortical spindles When recorded from closely adjacent cortical areas the majority of spindles started simultaneously. With increasing interelectrode distance the amount of strictly coincident spindles decreased and the amount of local spindle activity, defined as isolated appearance of spindles, increased correspondingly. In order to get an impression of the size of the cortical area with strictly coincident spindle activity, spindles were recorded continuously for 11 rain from three different sites in SI (Fig. 5). The incidence of spindle activity in this period was 7/min and the duration of each spindle varied between 1 and 2 sec. The intervals between the start of spindles developing at the different recording sites were measured. The criterion for onset of a spindle was that the first spindle wave to be accepted should have an amplitude of at least twice the random baseline variation and be followed by a regular sequence of at least three
18
IOO
D
C
B
A
I00'
U/M IO~ INTERM
%, FZ W
~ O,I
s
[]
~ 0.4.
s
[]
~2~4s
85 INTERV.
50 ¸
50
START INTERVAL
U/TR
J
n~ w 0-
START INTERVAL
START INTERVAL
Fig. 5. The degree of spindle coincidence in SI. A: recording situation. The electrodes U and M are 3 m m apart. B: intervals between onset of spindles in these two recording sites. C: the interelectrode distance has been increased to 6 ram. D : intervals between onset of spindles recorded from the cortical loci shown in C. For details see text.
spindle waves. The records were arbitrarily divided in periods of 4 sec, and the time difference between the spindle onsets within this period was measured with an accuracy of 100 msec. Thus, assuming that one spindle from each recording electrode occurs within the 4 sec, spindles from either recording site could appear first in one of the 40 bins, thus giving 40 possible start intervals. I f a consecutive series of spindles from two cortical recording sites were mutually unrelated, there would be a chance of 1/40 that one pair would show one particular start interval. Applying this method to the spindle records of the experiment illustrated in Fig. 5, it appeared that 18 % of the spindles recorded 3 mm apart started simultaneously (within 4- 100 msec), 48 % started within a time lag or lead of 400 msec while the amount of local spindles (interval between 2 and 4 sec) was 19 %. In Fig. 5C the electrodes were 6 mm apart, one recording locus being in the forelimb area and the other in the hindlimb area of SI. In this recording situation 6 % of the spindles started simultaneously (within -~: 100 smec), 25 % started within 4- 400 msec. Forty-one per cent were local spindles.
Effect of barbiturate on spindle coherence In lightly anaesthetized preparations the EEG was complex, consisting of well defined periods of typical spindle activity interspersed between a mixture of fast and slow background activity. Small amounts of barbiturate reduced the background activity temporarily, leaving the spindle waves more conspicuous. Consequently, cross-correlation of cortical spindles recorded from adjacent areas revealed a small overall increase in the correlation values after barbiturate administration. When no additional barbiturate was given, these changes were temporary. This apparent increase in the cortical spindle wave coherence evidently was related to the mentioned 'smoothing effect' of barbiturate. Experimental smoothing either by frequency filters in the preamplifiers or by computer smoothing subroutines had the same effect,
19
LL
r
"'U
g
,
f
T
-e.
T
l
?
[
I
e
-?
I
,~05 J 0 {.)
5 rnin
5'
5'
q
I'
I'
I'
1'
5`
5`
I'
i'
N
2
2
2
2
2
2
2
2
2
4
BARBITURATE
(mg)
]Fig. 6. Effect of small amounts of barbiturate on the cortical spindle wave synchrony. The recording situation is illustrated in the inset. Spindles from these loci were recorded and cross-correlated after sinai amounts of sodium pentobarbital (2 rag) were administered intravenously. Ten spindle pairs were analyzed alter each administration; the average of their maximal correlation coefficients is plotted in the diagram. The vertical lines through the plots are the S.D. The abscissa of the diagram indicates the amount of barbiturate given.
thus supporting the proposed explanation. Apart from this initial effect, increasing doses of barbiturates did not increase the cross-correlation coefficients between records taken from adjacent cortical points. Art example of this is shown in Fig. 6. A series of doses of sodium pentobarbital, each of 0.7 mg/kg, was given to a lightly anaesthetized preparation at intervals of 5 rain. Spindles were recorded and crosscorrelated between each injection. Each circle in Fig. 6 represents the average of the maximal correlation factor of I0 cross-correlated spindles. The vertical bars represent ± 1 S.D. Except for a moderate reduction in spindle wave coherence following the second dose the cortical spindle wave coherence did not change by barbiturate administration. Cortical points which did not show any spindle wave coherence also failed to reveal such coherence after barbiturate administration. DISCUSSION
Localized topographical organization of cortical rhythmic activity The thalamo-cortical (TC) systems taking part in rhythmic 10/sec activity appear to be topographically organized. Thus, Andersson 7 and Andersson and Wolpow 1°, by transection of the dorsal columns, induced continuous slow wave activity in the denervated zone of the cortex with a sharp delineation to activity in the neighbouring non-denervated areas. Similar results were later obtained from the thalamus 9. Furthermore, Andersen et al. 6, Ganes 15 and Ganes and Andersen 16 recorded spontaneous barbiturate spindles from a focus in the ventro-basal part of the thalamus and its corresponding cortical projection area. Maximal thalamo-cortical spindle wave synchrony was found when the recording electrodes were functionally on-line. The wave synchrony decreased rapidly when the cortical electrode was displaced from the on-line position. Although the TC spindle wave synchrony was not reduced uniformly in the
20 different directions from the on-line locus, the reduction was always paralleled by a similar reduction of the responses evoked in the cortex by contralateral cuta~lecms stimulation. Therefore, the extent of the TC synchronous areas seems determined by the projection from the thalamic recording locus to the cortex. The thalamic ceil group apparently serves as pacemaker for the cortical rhythmic activity. Based on these results one would expect that full wave synchrony between spindles recorded lrom two different cortical points would also be limited to small areas ot' the cortex and further that the size and shape of these areas would be determined by the inpul f'rom the corresponding thalamic pacemaker. Mapping experiments within SI and SII supported this hypothesis. However, the absolute size of these cortical areas showing synchronous spindle waves evidently depends on the criterion for spindle wave synchrony. Andersen et al. 6 found a i'ull wave synchrony only when the spindles were recorded at a maximal interelectrode distance of 1--2 ram. The applied criterion was that the peak latency of the waves in both spindles corresponded to each other. Up to an interelectrode distance of 3-4 mm a considerable fraction of the waves in a pair of spindle records were still synchronous while the wave synchrony fell sharply with larger interelectrode distances. Similar results were obtained in the present study using the cross-correlation coefficient between two records as a quantitative index of their wave synchrony. High cross-correlation values indicate that each wave in one spindle corresponds to a similar wave in the other; full spindle wave synchrony thus implies a correlation coefficient close to i-I or - I . Such a high wave correlation was only found with interelectrode distances of less than 1-2 ram, the iso-correlation lines of + 0 . 9 thus circumscribed a small cortical area in the vicinity of the reference electrode. A certain wave synchrony was, however, also found between spindles recorded with an interelectrode distance up to 3 4 mm. In these cases the cross-correlation factors were significantly lower indicating that only a fraction of the spindle waves were synchronous. The lowest correlation factor used in the present study was 0.5. This value indicates that only 257~, of the digitized values of one spindle record are linearly related to the corresponding digitized values of the other record. The cortical area enclosed by the 0.5 iso-correlation line was still relatively small and corresponded fairly well to the 30 ?/o amplitude line of the evoked potential. Lack o f cortico-cortical influence on spindle synchrony In addition to the size of the cortical projection from a given thalamic pacemaker, the size of the cortical entities for rhythmic activity might conceivably depend upon cortico-cortical triggering of rhythmic spindle activity. If, however, active or passive spread is important for the distribution of cortical spindle over larger distances, one would have expected a more diffuse topographical organization of the thalamocortical spindle systems and thereby larger cortical areas to show synchronous spindle waves. Furthermore, cutting of association fibres between two cortical sites should have reduced the spindle wave synchrony between these sites, a feature not observed, neither in previous studies 6 nor in the present investigation. In conclusion, therefore, spindle wave synchrony in the cortex seems limited to relatively small areas, the size and shape of which are to a great extent defined by the
21 c o r t i c o - p e t a l c o n n e c t i o n s f r o m the r h y t h m i c a l l y active t h a l a m i c p a c e m a k e r s a n d n o t significantly by i n t r a c o r t i c a l spread. In a g r e e m e n t with this view a series o f small intravenous doses o f b a r b i t u r a t e h a d no significant effect o n any o f the p a r a m e t e r s for cortical spindle wave s y n c h r o n y used in the present study. Cortical areas with simultaneous spindle onset Lehtinen 22 r e p o r t e d t h a t b a r b i t u r a t e spindles in the m o t o r cortex o f the cat started s i m u l t a n e o u s l y in a relatively large area, a n d he suggested the term m a c r o u n i t s to describe such areas. In the present study a similar m a p p i n g o f spindle starts was m a d e in the sensory cortex. A l t h o u g h a certain c o u p l i n g existed between onset o f spindles r e c o r d e d f r o m areas 5-6 m m a p a r t , full coincidence o f all spindles was observed only with a n interelectrode distance o f less t h a n 2 ram. Since b a r b i t u r a t e anaesthesia tends to m a k e spindle activity in different cortical, as well as thalamic, areas m o r e independent3,z,6, 8, the c o u p l i n g o f the spindles will evidently vary a c c o r d i n g to the d e p t h o f anaesthesia. D e t e r m i n a t i o n o f so-called m a c r o u n i t s a n d c o m p a r i s o n o f their size, can, therefore, be misleading. In a d d i t i o n , as also m e n t i o n e d by Lehtinen 2z, s y n c h r o n o u s l y starting spindles in two cortical zones do not imply t h a t their waves are in synchrony. A similar onset does n o t necessarily require a c o m m o n t h a l a m i c pacem a k e r b u t m a y rather reflect the fact that spindle activity in one p a r t o f the t h a l a m u s c a n quickly induce spindle activity in another.
REFERENCES l ADRIAN, E. D., Double representation of the feet in the sensory cortex of the cat, J. Physiol. (Lond.), 98 (1940) 16-18P. 2 ADRIAN,E. D., Afferent discharges to the cerebral cortex from peripheral sense organs, J. Physiol. (Lond.), 100 (1941) 159-191. 3 ANDERSEN,P., AND ANDERSSON, S. A., Physiological Basis of the Alpha Rhythm, Appleton-CenturyCrofts, New York, 235 p. 4 ANDERSEN,P., ANDERSSON,S. A., AND LOMO, T., Thalamocortical relations during spontaneous barbiturate spindle potentials, J. Physiol. (Lond.), 186 (1966) 37-38P. 5 ANDERSEN, P., ANDERSSON, S. A., AND LOMO, T., Patterns of spontaneous rhythmic activity within
various thalamic nuclei, Nature (Lond.), 211 (1966) 881-889. 6 ANDERSEN, P., ANDERSSON,S. A., AND LOMO, T., Nature of thalamocortical relations during spontaneous barbiturate spindle activity, J. Physiol. (Lond.), 192 (1967) 283-307. 7 AND~RSSON,S. A., Projections of different spinal pathways to the second somatic sensory area in cat, Acta physiol, scand., 56, Suppl. 194 (1962) 1-74. 8 ANDERSSON,S. A., HOLMGREN, E., AND MANSON, J. R., Synchronization and desynchronization in
the thalamus of the unanaesthetized decorticate cat, Electroenceph. clhz. Neurophysiol., 31 (1971) 335-345. 9 ANDERSSON,S. A., HOLMGREN, E., AND MANSON, J. R., Localized thalamic rhythmicity induced by spinal and cortical lesions, Electroenceph. clin. Neurophysiol., 31 (1971) 347-356. 10 ANDERSSON,S. A., ANDWOLPOW,E. R., Localized slow wave activity in the somatosensory cortex of the cat, Acta physiol, scand., 61 (1964) 130--140. 11 BENJAMIN, R. M., AND WELKER, W. 1., Somatic receiving areas of cerebral cortex of squirrel
monkey, J. Neurophysiol., 20 (1957) 286-299. 12 CREUTZFELDT,O. D., WA'rANABE,S., AND Lux, H. D., Relations between EEG phenomena and potentials of single cortical cells. I. Evoked responses after thalamic and epicortical stimulaton, Electroenceph. clin. Neurophysiol., 20 (1966) 1-18. 13 CREtJTZFELDT,O. D., WATANABE,S., AND LUX, H. D., Relations between EEG phenomena and
22 potentials of single cortical cells. II. Spontaneous and convulsoid activity, Etectroenceph~ clh~. Neurophysiol., 20 (1966 ) t 9- 37. 14 DARIAN-SMI'rH, I., I~ISTER, J., MOK, H., AND YOKOTA, T., Somatic sensory cortical projectio~ areas excited by tactile stimulation of the cat, a triple representation, J. Ph~'shd. (Lond.). 182 (1966) 671 689. 15 GANES, T., Barbiturate spindle activity in the thalamic lateral ventroposterior nucleus and lhe second somatosensory area of the cortex, Brain Research, 98 (1975) 473-483. 16 GANES, T., AND ANDF,RSEN, P., Barbiturate spindle activity in functionally corresponding thalamic and cortical somatosensory areas in the cat, Brain Research, 98 (1975) 457-472. 17 GUILLERY. R. W., ADRIAN, H. G., WOOLSEY, C. N., AND ROSE, J. E., Activation of somatosensor3, areas I and 11 of cat's cerebral cortex by focal stimulation of the ventrobasal complex. In D. P. PURPURA AND M. D. YAHR (Eds.), The Thalamus, Columbia Univ. Press, New York, 1966, pp. 197-207. 18 HAND, P. J., AND MORRISON, A. R., Thalamocortical relationships in somatic sensory system as revealed by silver impregnation techniques, Brain Behav. Evol., 5 (1972) 273-302. 19 HAND, P. J., AND MORRISON, A. R., Thalamocortical projections from the ventrobasal complex to somatic sensory areas I and lI of the cat, Exp. Neurol., 26 (1970) 291-308. 20 HAWS, G., AND WOOLSEY, C. N., The pattern of organization within the primary tactile area of the cerebral cortex of the cat, Fed. Proc., 3 (1944) 18. 21 JONES, E. G., AND POWELL, T. P. S., The cortical projection of the ventroposterior nucleus of lhe thalamus in the cat, Brain Research, 13 (1969) 298-318. 22 LErITINEN, 1., Temporal distribution of barbiturate spindle activity and recruiting responses in tl~e neocortex and pyramidal tract of cat, Ann. Aead. Sci. fenn. As, 149 (1972) 1--40. 23 LIVINGSTON,A., AND PHILLIPS, C. G., Maps and thresholds for the sensorimotor cortex of the cat, Quart. J. exp. PhysioL, 42 (1957) 190-205. 24 SPENCER, A., AND BROOKHART, J. M., Electrical patterns of augmenting and recruiting waves in depths of sensory motor cortex of cat, J. Neurophysiol., 24 (1961) 26-50. 25 SPENCER, A., AND BROOKHART, J. M., A study of spontaneous spindle waves in sensory motor cortex of cat, J, Nearophyskd., 24 (1961) 50-65. 26 WOOLSEY, C. N., AND FAIRMAN, D., Contralateral, ipsilateral and bilateral representation of cutaneous receptors in somatic areas I and 11 of the cerebral cortex of pig, sheep and other mammals, Surgery, 19 (1946) 684-702. 27 WOOLSEY, C. N., AND WANG, G. H., Somatic sensory areas 1 and II of the cerebral cortex of the rabbit, Fed. Proc., 4 (1945) 79.
ll
S P I N D L E WAVE S Y N C H R O N Y IN T H E SOMATOSENSORY CORTEX OF T H E CAT
T R O N D GANES
The Institute of Neurophysiology, University of Oslo, Oslo (Norway) (Accepted June 6th, 1975)
SUMMARY
(1) Spontaneous barbiturate spindles were recorded from the primary and secondary somatosensory cortex. The recordings were concentrated to areas surrounding several reference loci. The recording sites producing the maximal response evoked by stimulation of an exposed nerve in a contralateral limb were used as reference loci. (2) Spindles recorded at various distances from the respective reference loci were cross-correlated to spindles developing simultaneously in the latter. High correlation coefficients, indicating a considerable degree of wave synchrony, were obtained between spindles in the reference locus and spindles recorded a few millimetres from this site. The correlation coefficients decreased with increasing interelectrode distance. A relatively sharp fall in the correlation coefficients was generally found 2-3 mm from the reference locus. Small amounts of sodium pentobarbital, given intravenously at intervals of 5 min, had no effect upon this pattern. (3) The change in the correlation coefficients was followed by a parallel change in the amplitude of the evoked potentials. The iso-correlation lines of spindle wave synchrony and the iso-amplitude lines of the evoked potentials had a similar distribution and extension for each particular reference locus. (4) Lateral spread o f spindle waves in the cortex seems to be of minor importance, since a vertical lesion cutting the cortico-cortical fibres did not reduce the wave synchrony of the spindles recorded from either side of the lesion. (5) The majority of the spindles recorded in the close vicinity of a reference locus started simultaneously within ~: 0.1 sec. This pattern changed with increasing distance from the reference locus and 5-6 mm away only a fraction of the spindles started simultaneously. However, within the entire primary somatosensory cortex a small but significant coupling existed between onset of the spindles.
12 INTRODUCTION
The sensory input to the cortex is organized in a strictly topographical fashion. This applies to both the primary (SI) and to the secondary (SII) somatosensory receiving areasl,",l~,~4,17 ~a:'3,26,~L There is also evidence that the thalamo-cortical CIC) systems, which take part in barbiturate spindle activity, are organized in a similar way3,6,9,10,1;',~6. Several investigations2,6,1z,13, '~4,9~ have suggested that barbiturate spindle activity in the somatosensory cortex is generated by rhythmic activity of the same TC connections and intracortical synapses which are involved in the prinlary somatosensory responses. One consequence of this hypothesis is that various cortical areas ought to show different rhythmic patterns because they receive their rhythmic input from dissimilar thalamic nuclei and subnuclei. The purpose of this study was to examine to what extent the barbiturate spindle activity in SI and SII areas is topographically organized and to determine some of the factors which influence the distribution of such spindle waves in the cortex. METHODS
The experiments were performed on cats anaesthetized with 30 mg/kg sodium pentobarbital given intraperitoneally. Light anaesthesia characterized by withdrawal of the forelimb and hindlimb to light pinching was maintained during the experiments by additional intravenous injections of small amounts of barbiturate. Operation and stimulation. The cortical surface was exposed by gentle removal of the overlying skull and covered by a pool of liquid paraffin kept at 37 °C during the experiment. Three peripheral forelimb nerves, the superficial radial, the superficial branch of the ulnar and the median nerve were exposed distal to the elbow joint and mounted in separate stimulation cuffs: In the hindlimb, the posterior tibial nerve was exposed for stimulation distal to the knee joint. The stimulus threshold for evoking a short latency cortical potential was determined for each nerve. Later in the experiment voltages from 1.5 to 2 times the threshold were used. The pulse duration was 0.1 msec. Recording technique. The cortical potentials were recorded monopolarly with silver or platinum ball electrodes (diameter 0.2-0.3 mm) placed on the pial surface under direct visual guidance through a stereoscopic microscope. Three to four electrodes with a permanent interelectrode distance of I mm were mounted in a manipulator for rapid mapping of cortical potentials. The cortical site giving maximal short latency response to stimulation of one of the peripheral nerves was defined as the cortical reference point for that particular nerve projection, To map the distribution of evoked potentials and barbiturate spindles with respect to this site, one electrode remained in this position while the other electrodes were moved systematically around in adjacent cortical areas. The recorded potentials were fed through preamplifiers and displayed on a storage oscilloscope for direct examination and on a normal oscilloscope for filming. In addition, the records were taped on a multichannelled tape recorder for later analysis.
13
Computer and calculation methods'. Analyses of the evoked responses, as well as the barbiturate spindle waves, were done by a NORD-I computer, either on-line or by off-line analysis. The computer methods and programmes have been presented earlier t6. The program used for cortical evoked potentials had no amplitude limit, measuring all peaks above or below the averaged baseline. When plotted on a cortical map, however, the smallest amplitude used was taken as 30 % of the maximal amplitude. In plotting the spindle correlation values between the records from two electrodes, only the initial value (with zero displacement of the records) of each pair of spindles was plotted. The lowest correlation value (r) used was selected to match the lowest relative value of the evoked potential. This corresponds to a cross-correlation coefficient of 0.5 and a coefficient of determination (r e) of 0.25, indicating that 25 % of the variability in the relative amplitudes of the spindle waves was due to a linear relationship between the two records. RESULTS
Spindle wave coherences and interelectrode distance Spontaneous barbiturate spindles recorded from two sites on the cortical surface 0.3-0.5 mm apart started simultaneously and were almost identical in form. When such spindles were cross-correlated, the maximal cross-correlation values were close to + 1. Since two surface electrodes in this position obviously are influenced by much the same cortical activity, this result was expected. In fact, in this case, cross-correla-
A
O.Smm
ULN
C
B
STIM
o~05
--/'L/"
0
-IO
_0.2 IOtas
"
. . . .
" • . . . . - I 0 0 ms
"
®
•
,
.
I 0 0 ms
0 ,
, is
~ -05
.
.
.
.
.
.
.
.
o - I 0
.
.
.
.
.
.
.
- I 0 0 ms
.
.
.
.
.
0
°
•
.
IO0 ms
DISPLACEMENT
Fig. I. W a v e s y n c h r o n y o f cortical spindles. A : two electrodes on the pial surface in the p r i m a r y s o m a t o s e n s o r y cortex (SI). Distance between the electrodes is 0.3 ram. Cortical responses evoked by contralateral s t i m u l a t i o n are nearly similar in the two sites. B: s p o n t a n e o u s spindles f r o m the s a m e sites. C: cross-correlogram o f the spindles in B. D : one o f the electrodes h a s been m o v e d 4 m m medially. T h e evoked response f r o m this locus h a s nearly vanished. E: spindles with a similar onset in the two recording sites. F: c r o s s - c o r r e l o g r a m o f the spindles in E.
14 tion does not differ significantly from auto-correlation. In Fig. t spindles were recorded with two electrodes 0.3 mm apart in the ulnar projection area o|" SI (A). The spindles (B) were similar, consisting of individual spindle waves at regular intervals. Cross-correlation of the records (C) resulted in a regular sinus-shaped cross-correlogram with a maximal correlation factor of 0.98. Peaks of high correlation values appeared at record displacements of about 100 msec, corresponding to an intraspindle frequency of l0 Hz. When. the intraspindle frequency was less regular, the cross-correl ograms appeared more damped with a less regular sinusoidal form. In Fig. I D, one o f the electrodes was moved 4 mm medially while the other remained unchanged. The cortical evoked potential revealed that the new electrode locus was outside the ulnar projection area. Spindles were, however, still present at both electrode sites and sometimes occurred simultaneously (E). They differed, however, both in duration and shape of the spindle waves. The cross-correlogram (f~), although still retaining some of the sinusoidal features, invariably gave a low correlation, the factors never exceeding ~ 0.3. Based on this pilot experiment it was clear that spindle wave coherence, expressed in terms of the cross-correlation factor, was related to the distance between the recording electrodes and also in some way to the organization of the cortex. In order to get more precise information on these points, the distribution of spindle wave coherences and evoked potentials in SI and SII were mapped systematically. In the primary sensory cortex the forelimb and hindlimb projection areas were mapped. In the former, the subdivisions represented by the projection areas of three peripheral forelimb nerves, the ulnar (U), median (M) and superficial radial nerve (SR) were determined. In the hindlimb area, stimulation of contralateral posterior tibial nerve (PT) was used. No attempts were made to separate between the different sensory modalities. The cortical site giving a maximal short latency response to contralateral nerve stimulation was defined as the cortical reference site (REF) of the particular nerve. Mapping was concentrated to the cortex surrounding the REF sites. One electrode was left in this locus and the other moved along radial tracks in steps of 1 mm. At each new recording site several spindles were recorded and cross-correlated to spindles simultaneously recorded by the REF electrode. The amplitudes of evoked potentials were recorded simultaneously. Generally, the electrode was moved along four different rays through the REF site (dots in Fig. 2A). For convenience only two of the rows shown in Fig. 2Aa and b are plotted in B and C_ Each plot of spindle wave coherence and amplitude of evoked potentials was based upon 10 cross-correlated spindles and the same number of evoked potentials. Fig. 2D shows similar data from SII. Only one row of recording points in the SII forelimb area (b) and one row in the hindlimb area (a) were plotted in E and F, respectively. The spindle wave synchrony was directly related to the distance from the REF sites. The cross-correlation coefficient decreased rapidly with increasing interelectrode
15
A
C
B
a~ ~ v
SPINOLE CORRELATION .... EV~tOPOTENTIAL *1 100%
*1 tO0% s' ~',
01150 I* *- °*
4
D
3
Z
0
2
I
S
E °l, 100%
4
4
F
,100%
3
2
r
F
I
O
I
5
~'
~'~SUPRASYLV. F
a
ECTOSYLV.
~0.5, S
,
ooeoOeeo MED
b ",,/O,H,
"•
~ 4
L .-"
., 3
2
I
0
I
Z
INTERELECTRODE DISTANCE
3 4 rm~
3
2
L
L
2
3
".
I
.0
I
4
Fig. 2. Distribution of cortical spindle wave synchrony. A: the cortical recording sites are illustrated by dots. The two open circles indicate the reference locus in the SI forepaw (lower) and hindlimb (upper) projection area. The results from the two recording tracks a and b are plotted in the diagrams of B and C respectively. The maximal cross-correlation factors of the spindles (average of 5-10 analyzed spindle pairs) are plotted as filled circles• The relative amplitude of the evoked responses (in per cent of the maximal response) are plotted as crosses. The abscissae of the diagrams give the distance from the reference electrode. The vertical lines through the correlation plots indicate dispersion of the average wave synchrony from one spindle pair to the next expressed in S.D. D : mapping in the second somatosensory area (SII). The results of two recording tracks a and b are plotted in E and F, respectively.
distance. In m o s t tracks it reached a value o f + 0.3 or less w h e n spindles were recorded m o r e than 3-4 m m f r o m the R E F site. A l o n g all tracks this decrease in wave coherence was followed by a parallel decrease in amplitude o f the e v o k e d potentials. B o t h the spindle wave synchrony and the evoked potentials were reduced m o r e rapidly in SII than in SI when the interelectrode distance was increased. A l o n g s o m e tracks, h o w ever, a correlation factor as high as + 0 . 5 to + 0 . 6 persisted up to distances o f 4 - 5 m m f r o m the R E F site. This feature was observed when spindles were recorded f r o m the cortical area between SI and SII, with the R E F site either in SI or in SII. Examples o f this are seen in Fig. 2C and E.
Topographical organization of coherent spindles In order to compare the distribution o f coherent spindle activity with the sensory projection areas, maps were m a d e by drawing lines between cortical points with identical correlation values and similarly, between points with evoked potentials o f the same amplitude (Fig. 3). Linear intrapolation between adjacent recording points was done to obtain isocorrelation lines representing correlation factors o f + 0 . 9 , + 0 . 7 , + 0 . 5 and iso-amplitude lines representing 90, 70, and 30 ~ o f the m a x i m a l e v o k e d potential amplitude.
16
D4p
G //
iI
SPINDLEWAVE SYNCHRONY
.
.
.
H
.
SII
EVOKED
POTENTIAL
F
c
/
2mm
J ~ S
'..
i::
EP= 70
"
EP°
5o
"
EP = 5 0
"
......
EVOKED
POTENTIAL
, SPINDLEWAVE SYNCHRONY
Fig. 3. Extent of cortical areas with a high spindle wave synchrony. A: the enclosed area indicates the areas shown in detail in B and C. The black dots in B and C represent the reference sites. The surrounding lines are iso-correlation lines for spindle wave synchrony (B) and iso-amplitude lines for the sensory evoked potentials (C). PT, SR, U and M are the cortical reference loci for the contralateral posterior tibial, superficial radial, ulnar and median nerves, respectively. E and F: similar maps of iso-correlation lines and iso-amplitude lines for the second somatosensory area, as indicated by the enclosed area in D. G : the iso-correlation lines and iso-amplitude lines relative to a reference focus in SI are superimposed. H: similar to G but with the reference focus in SII.
Fig. 3B a n d E illustrates the d i s t r i b u t i o n o f spindle wave s y n c h r o n y s u r r o u n d i n g six different R E F sites. A l l spindles r e c o r d e d inside the o u t e r line (r = 0.5) have a crosscorrelation coefficient o f + 0 . 5 or m o r e when c o m p a r e d to spindles simultaneously rec o r d e d f r o m the R E F site. A certain overlap between the three f o r e l i m b areas investig a t e d was observed, b u t the size a n d shape o f their c o r r e s p o n d i n g spindle coherent areas differed significantly (Fig. 3B). A c o m p a r i s o n between Fig. 3B a n d C a n d between E a n d F shows t h a t the spindle wave coherence was i n t i m a t e l y related to the extent o f the various sensory p r o j e c t i o n areas b o t h in the SI a n d SII. The i s o - c o r r e l a t i o n lines a n d i s o - a m p l i t u d e lines relative to a reference focus in SI are s u p e r i m p o s e d in Fig. 3G, a n d a similar s u p e r i m p o s i t i o n o f d a t a for a reference p o i n t in SII is seen in Fig. 3H.
17
A
W
:~ 0.9
B
._1
.3
08
0-= 0.1
~ 07 3mm
Q9'
C M- 0.77
M=0.5 No= 8
0.8.
0"= 0 0 7
0.7
o 0.6
0.6
"' 0.5 .J
0.5
z 0.4
0.4
No- 9
a
ffl
BEFORE LESION
AFTER LESION
Fig. 4. Lack of lateral spread of spindle wave synchrony. A : recording situation. The jagged line between the recording sites illustrates the extent of the cortical lesion. SSM, middle suprasylvian gyrus; LAT, lateral gyrus. B and C: dots give the maximal cross-correlation coefficients of 8 spindle pairs recorded before (B) and 9 spindle pairs recorded after (C) the cortical lesion. M, mean; ~, the S.D. No, number of spindle pairs analyzed.
Spread of spindle waves in the cortex Although highly coherent spindles were found within definite and relatively small cortical zones corresponding to the main sensory receiving areas, some of this activity could be due to spread from neighbouring regions by intracortical or corticocortical connections. Therefore, experiments were performed in which the corticocortical connections between two recording sites in the cortex were cut (Fig. 4). Cortical spindles were recorded in SI from two loci 2 mm apart (A). Cross-correlation of the spindles resulted in correlation factors between -+-0.7 and +0.4, the mean value being +0.5 (B). A vertical incision into the cortex was made between the recording electrodes with the electrodes left in situ. Care was taken not to displace the electrodes. The lesion, verified histologically, extended through all cortical layers and between 1 and 2 mm into the underlying white matter. Cortical spindle activity returned immediately after the lesion in both recording sites. Contrary to what might be expected, cross-correlation of the spindles (C) revealed a significant increase in spindle wave coherences. The experiment, repeated several times, both in SI and in the lateral and suprasylvian gyri, always gave the same result. Start intervals of cortical spindles When recorded from closely adjacent cortical areas the majority of spindles started simultaneously. With increasing interelectrode distance the amount of strictly coincident spindles decreased and the amount of local spindle activity, defined as isolated appearance of spindles, increased correspondingly. In order to get an impression of the size of the cortical area with strictly coincident spindle activity, spindles were recorded continuously for 11 rain from three different sites in SI (Fig. 5). The incidence of spindle activity in this period was 7/min and the duration of each spindle varied between 1 and 2 sec. The intervals between the start of spindles developing at the different recording sites were measured. The criterion for onset of a spindle was that the first spindle wave to be accepted should have an amplitude of at least twice the random baseline variation and be followed by a regular sequence of at least three
18
IOO
D
C
B
A
I00'
U/M IO~ INTERM
%, FZ W
~ O,I
s
[]
~ 0.4.
s
[]
~2~4s
85 INTERV.
50 ¸
50
START INTERVAL
U/TR
J
n~ w 0-
START INTERVAL
START INTERVAL
Fig. 5. The degree of spindle coincidence in SI. A: recording situation. The electrodes U and M are 3 m m apart. B: intervals between onset of spindles in these two recording sites. C: the interelectrode distance has been increased to 6 ram. D : intervals between onset of spindles recorded from the cortical loci shown in C. For details see text.
spindle waves. The records were arbitrarily divided in periods of 4 sec, and the time difference between the spindle onsets within this period was measured with an accuracy of 100 msec. Thus, assuming that one spindle from each recording electrode occurs within the 4 sec, spindles from either recording site could appear first in one of the 40 bins, thus giving 40 possible start intervals. I f a consecutive series of spindles from two cortical recording sites were mutually unrelated, there would be a chance of 1/40 that one pair would show one particular start interval. Applying this method to the spindle records of the experiment illustrated in Fig. 5, it appeared that 18 % of the spindles recorded 3 mm apart started simultaneously (within 4- 100 msec), 48 % started within a time lag or lead of 400 msec while the amount of local spindles (interval between 2 and 4 sec) was 19 %. In Fig. 5C the electrodes were 6 mm apart, one recording locus being in the forelimb area and the other in the hindlimb area of SI. In this recording situation 6 % of the spindles started simultaneously (within -~: 100 smec), 25 % started within 4- 400 msec. Forty-one per cent were local spindles.
Effect of barbiturate on spindle coherence In lightly anaesthetized preparations the EEG was complex, consisting of well defined periods of typical spindle activity interspersed between a mixture of fast and slow background activity. Small amounts of barbiturate reduced the background activity temporarily, leaving the spindle waves more conspicuous. Consequently, cross-correlation of cortical spindles recorded from adjacent areas revealed a small overall increase in the correlation values after barbiturate administration. When no additional barbiturate was given, these changes were temporary. This apparent increase in the cortical spindle wave coherence evidently was related to the mentioned 'smoothing effect' of barbiturate. Experimental smoothing either by frequency filters in the preamplifiers or by computer smoothing subroutines had the same effect,
19
LL
r
"'U
g
,
f
T
-e.
T
l
?
[
I
e
-?
I
,~05 J 0 {.)
5 rnin
5'
5'
q
I'
I'
I'
1'
5`
5`
I'
i'
N
2
2
2
2
2
2
2
2
2
4
BARBITURATE
(mg)
]Fig. 6. Effect of small amounts of barbiturate on the cortical spindle wave synchrony. The recording situation is illustrated in the inset. Spindles from these loci were recorded and cross-correlated after sinai amounts of sodium pentobarbital (2 rag) were administered intravenously. Ten spindle pairs were analyzed alter each administration; the average of their maximal correlation coefficients is plotted in the diagram. The vertical lines through the plots are the S.D. The abscissa of the diagram indicates the amount of barbiturate given.
thus supporting the proposed explanation. Apart from this initial effect, increasing doses of barbiturates did not increase the cross-correlation coefficients between records taken from adjacent cortical points. Art example of this is shown in Fig. 6. A series of doses of sodium pentobarbital, each of 0.7 mg/kg, was given to a lightly anaesthetized preparation at intervals of 5 rain. Spindles were recorded and crosscorrelated between each injection. Each circle in Fig. 6 represents the average of the maximal correlation factor of I0 cross-correlated spindles. The vertical bars represent ± 1 S.D. Except for a moderate reduction in spindle wave coherence following the second dose the cortical spindle wave coherence did not change by barbiturate administration. Cortical points which did not show any spindle wave coherence also failed to reveal such coherence after barbiturate administration. DISCUSSION
Localized topographical organization of cortical rhythmic activity The thalamo-cortical (TC) systems taking part in rhythmic 10/sec activity appear to be topographically organized. Thus, Andersson 7 and Andersson and Wolpow 1°, by transection of the dorsal columns, induced continuous slow wave activity in the denervated zone of the cortex with a sharp delineation to activity in the neighbouring non-denervated areas. Similar results were later obtained from the thalamus 9. Furthermore, Andersen et al. 6, Ganes 15 and Ganes and Andersen 16 recorded spontaneous barbiturate spindles from a focus in the ventro-basal part of the thalamus and its corresponding cortical projection area. Maximal thalamo-cortical spindle wave synchrony was found when the recording electrodes were functionally on-line. The wave synchrony decreased rapidly when the cortical electrode was displaced from the on-line position. Although the TC spindle wave synchrony was not reduced uniformly in the
20 different directions from the on-line locus, the reduction was always paralleled by a similar reduction of the responses evoked in the cortex by contralateral cuta~lecms stimulation. Therefore, the extent of the TC synchronous areas seems determined by the projection from the thalamic recording locus to the cortex. The thalamic ceil group apparently serves as pacemaker for the cortical rhythmic activity. Based on these results one would expect that full wave synchrony between spindles recorded lrom two different cortical points would also be limited to small areas ot' the cortex and further that the size and shape of these areas would be determined by the inpul f'rom the corresponding thalamic pacemaker. Mapping experiments within SI and SII supported this hypothesis. However, the absolute size of these cortical areas showing synchronous spindle waves evidently depends on the criterion for spindle wave synchrony. Andersen et al. 6 found a i'ull wave synchrony only when the spindles were recorded at a maximal interelectrode distance of 1--2 ram. The applied criterion was that the peak latency of the waves in both spindles corresponded to each other. Up to an interelectrode distance of 3-4 mm a considerable fraction of the waves in a pair of spindle records were still synchronous while the wave synchrony fell sharply with larger interelectrode distances. Similar results were obtained in the present study using the cross-correlation coefficient between two records as a quantitative index of their wave synchrony. High cross-correlation values indicate that each wave in one spindle corresponds to a similar wave in the other; full spindle wave synchrony thus implies a correlation coefficient close to i-I or - I . Such a high wave correlation was only found with interelectrode distances of less than 1-2 ram, the iso-correlation lines of + 0 . 9 thus circumscribed a small cortical area in the vicinity of the reference electrode. A certain wave synchrony was, however, also found between spindles recorded with an interelectrode distance up to 3 4 mm. In these cases the cross-correlation factors were significantly lower indicating that only a fraction of the spindle waves were synchronous. The lowest correlation factor used in the present study was 0.5. This value indicates that only 257~, of the digitized values of one spindle record are linearly related to the corresponding digitized values of the other record. The cortical area enclosed by the 0.5 iso-correlation line was still relatively small and corresponded fairly well to the 30 ?/o amplitude line of the evoked potential. Lack o f cortico-cortical influence on spindle synchrony In addition to the size of the cortical projection from a given thalamic pacemaker, the size of the cortical entities for rhythmic activity might conceivably depend upon cortico-cortical triggering of rhythmic spindle activity. If, however, active or passive spread is important for the distribution of cortical spindle over larger distances, one would have expected a more diffuse topographical organization of the thalamocortical spindle systems and thereby larger cortical areas to show synchronous spindle waves. Furthermore, cutting of association fibres between two cortical sites should have reduced the spindle wave synchrony between these sites, a feature not observed, neither in previous studies 6 nor in the present investigation. In conclusion, therefore, spindle wave synchrony in the cortex seems limited to relatively small areas, the size and shape of which are to a great extent defined by the
21 c o r t i c o - p e t a l c o n n e c t i o n s f r o m the r h y t h m i c a l l y active t h a l a m i c p a c e m a k e r s a n d n o t significantly by i n t r a c o r t i c a l spread. In a g r e e m e n t with this view a series o f small intravenous doses o f b a r b i t u r a t e h a d no significant effect o n any o f the p a r a m e t e r s for cortical spindle wave s y n c h r o n y used in the present study. Cortical areas with simultaneous spindle onset Lehtinen 22 r e p o r t e d t h a t b a r b i t u r a t e spindles in the m o t o r cortex o f the cat started s i m u l t a n e o u s l y in a relatively large area, a n d he suggested the term m a c r o u n i t s to describe such areas. In the present study a similar m a p p i n g o f spindle starts was m a d e in the sensory cortex. A l t h o u g h a certain c o u p l i n g existed between onset o f spindles r e c o r d e d f r o m areas 5-6 m m a p a r t , full coincidence o f all spindles was observed only with a n interelectrode distance o f less t h a n 2 ram. Since b a r b i t u r a t e anaesthesia tends to m a k e spindle activity in different cortical, as well as thalamic, areas m o r e independent3,z,6, 8, the c o u p l i n g o f the spindles will evidently vary a c c o r d i n g to the d e p t h o f anaesthesia. D e t e r m i n a t i o n o f so-called m a c r o u n i t s a n d c o m p a r i s o n o f their size, can, therefore, be misleading. In a d d i t i o n , as also m e n t i o n e d by Lehtinen 2z, s y n c h r o n o u s l y starting spindles in two cortical zones do not imply t h a t their waves are in synchrony. A similar onset does n o t necessarily require a c o m m o n t h a l a m i c pacem a k e r b u t m a y rather reflect the fact that spindle activity in one p a r t o f the t h a l a m u s c a n quickly induce spindle activity in another.
REFERENCES l ADRIAN, E. D., Double representation of the feet in the sensory cortex of the cat, J. Physiol. (Lond.), 98 (1940) 16-18P. 2 ADRIAN,E. D., Afferent discharges to the cerebral cortex from peripheral sense organs, J. Physiol. (Lond.), 100 (1941) 159-191. 3 ANDERSEN,P., AND ANDERSSON, S. A., Physiological Basis of the Alpha Rhythm, Appleton-CenturyCrofts, New York, 235 p. 4 ANDERSEN,P., ANDERSSON,S. A., AND LOMO, T., Thalamocortical relations during spontaneous barbiturate spindle potentials, J. Physiol. (Lond.), 186 (1966) 37-38P. 5 ANDERSEN, P., ANDERSSON, S. A., AND LOMO, T., Patterns of spontaneous rhythmic activity within
various thalamic nuclei, Nature (Lond.), 211 (1966) 881-889. 6 ANDERSEN, P., ANDERSSON,S. A., AND LOMO, T., Nature of thalamocortical relations during spontaneous barbiturate spindle activity, J. Physiol. (Lond.), 192 (1967) 283-307. 7 AND~RSSON,S. A., Projections of different spinal pathways to the second somatic sensory area in cat, Acta physiol, scand., 56, Suppl. 194 (1962) 1-74. 8 ANDERSSON,S. A., HOLMGREN, E., AND MANSON, J. R., Synchronization and desynchronization in
the thalamus of the unanaesthetized decorticate cat, Electroenceph. clhz. Neurophysiol., 31 (1971) 335-345. 9 ANDERSSON,S. A., HOLMGREN, E., AND MANSON, J. R., Localized thalamic rhythmicity induced by spinal and cortical lesions, Electroenceph. clin. Neurophysiol., 31 (1971) 347-356. 10 ANDERSSON,S. A., ANDWOLPOW,E. R., Localized slow wave activity in the somatosensory cortex of the cat, Acta physiol, scand., 61 (1964) 130--140. 11 BENJAMIN, R. M., AND WELKER, W. 1., Somatic receiving areas of cerebral cortex of squirrel
monkey, J. Neurophysiol., 20 (1957) 286-299. 12 CREUTZFELDT,O. D., WA'rANABE,S., AND Lux, H. D., Relations between EEG phenomena and potentials of single cortical cells. I. Evoked responses after thalamic and epicortical stimulaton, Electroenceph. clin. Neurophysiol., 20 (1966) 1-18. 13 CREtJTZFELDT,O. D., WATANABE,S., AND LUX, H. D., Relations between EEG phenomena and
22 potentials of single cortical cells. II. Spontaneous and convulsoid activity, Etectroenceph~ clh~. Neurophysiol., 20 (1966 ) t 9- 37. 14 DARIAN-SMI'rH, I., I~ISTER, J., MOK, H., AND YOKOTA, T., Somatic sensory cortical projectio~ areas excited by tactile stimulation of the cat, a triple representation, J. Ph~'shd. (Lond.). 182 (1966) 671 689. 15 GANES, T., Barbiturate spindle activity in the thalamic lateral ventroposterior nucleus and lhe second somatosensory area of the cortex, Brain Research, 98 (1975) 473-483. 16 GANES, T., AND ANDF,RSEN, P., Barbiturate spindle activity in functionally corresponding thalamic and cortical somatosensory areas in the cat, Brain Research, 98 (1975) 457-472. 17 GUILLERY. R. W., ADRIAN, H. G., WOOLSEY, C. N., AND ROSE, J. E., Activation of somatosensor3, areas I and 11 of cat's cerebral cortex by focal stimulation of the ventrobasal complex. In D. P. PURPURA AND M. D. YAHR (Eds.), The Thalamus, Columbia Univ. Press, New York, 1966, pp. 197-207. 18 HAND, P. J., AND MORRISON, A. R., Thalamocortical relationships in somatic sensory system as revealed by silver impregnation techniques, Brain Behav. Evol., 5 (1972) 273-302. 19 HAND, P. J., AND MORRISON, A. R., Thalamocortical projections from the ventrobasal complex to somatic sensory areas I and lI of the cat, Exp. Neurol., 26 (1970) 291-308. 20 HAWS, G., AND WOOLSEY, C. N., The pattern of organization within the primary tactile area of the cerebral cortex of the cat, Fed. Proc., 3 (1944) 18. 21 JONES, E. G., AND POWELL, T. P. S., The cortical projection of the ventroposterior nucleus of lhe thalamus in the cat, Brain Research, 13 (1969) 298-318. 22 LErITINEN, 1., Temporal distribution of barbiturate spindle activity and recruiting responses in tl~e neocortex and pyramidal tract of cat, Ann. Aead. Sci. fenn. As, 149 (1972) 1--40. 23 LIVINGSTON,A., AND PHILLIPS, C. G., Maps and thresholds for the sensorimotor cortex of the cat, Quart. J. exp. PhysioL, 42 (1957) 190-205. 24 SPENCER, A., AND BROOKHART, J. M., Electrical patterns of augmenting and recruiting waves in depths of sensory motor cortex of cat, J. Neurophysiol., 24 (1961) 26-50. 25 SPENCER, A., AND BROOKHART, J. M., A study of spontaneous spindle waves in sensory motor cortex of cat, J, Nearophyskd., 24 (1961) 50-65. 26 WOOLSEY, C. N., AND FAIRMAN, D., Contralateral, ipsilateral and bilateral representation of cutaneous receptors in somatic areas I and 11 of the cerebral cortex of pig, sheep and other mammals, Surgery, 19 (1946) 684-702. 27 WOOLSEY, C. N., AND WANG, G. H., Somatic sensory areas 1 and II of the cerebral cortex of the rabbit, Fed. Proc., 4 (1945) 79.
