- Email: [email protected]
Propagation of afterdischarges along the septo-temporal axis of the rat hippocampus: a quantitative analysis
Electroencephalography and clinical Neurophysiology , 1989, 73:172-178 Elsevier Scientific Publishers Ireland, Ltd.
172
EEG 02384
Short communication
Propagation of afterdischarges along the septo-temporal axis of the rat hippocampus: a quantitative analysis C. Nunes Filipe 1, j.p. Pijn, V.M. Fernandes de Lima 2 and F.H. Lopes da Silva Dept. of Experimental Zoology, Universiteit van Amsterdam, Biologisch Centrum Gebouw 1I, 1098 S M Amsterdam (The Netherlands) (Accepted for publication: 1 April 1989)
Summary Afterdischarges (ADs) were elicited in the hippocampus of the rat under halothane anesthesia. Records were made along the septo-temporal axis of the hippocampal formation (dorsal (DHF), splenial (SP) and ventral (VHF) regions). For comparison records were also made from the contralateral DHF. The propagation of ADs was quantified using linear and non-linear regression analysis. The values of non-linear association were in general larger than those of linear association. Values of association and time delays between pairs of EEG signals recorded from different regions were estimated as function of elapsed time during the ADs. The association measures for ADs elicited by stimulation of the DHF were relatively large between the ipsi- and contralateral DHF and between the DHF and SP of the same side in contrast to those found between DHF and VHF. The threshold to elicit ADs in the VHF was considerably higher than in the DHF. We conclude that the DHF and SP form a functional entity only loosely coupled to the VHF and that simple spatial continuity is not enough for the propagation of epileptiform ADs between different areas. The estimate of time delays indicated a high degree of synchrony during ADs between homotopical sites on ipsi- and contralateral DHF whereas the degree of synchrony between DHF and SP was smaller. Key words: Hippocampus; Epileptic seizures; Propagation; Septo-temporal axis; Non-linear regression
The mechanisms of propagation of seizure discharges are still little understood in spite of a relatively large number of investigations (for review see Gotman 1986). Even within the same brain structure it is not yet clear to what extent and by which mechanisms seizure activity may spread, by neural pathways through chemical synaptic transmission (Prince 1978) or by simple contiguity through non-synaptic mechanisms (Taylor
i Supported by a grant from Junta Nacional de Investiga~o Cientifica e Tecnol6gica (JNICT), Portugal. Permanent address: Departamento de Fisiologia, Faculdade de CiSncias M&iicas da Universidade Nova de Lisboa, Campo de Santana 130, 1198 Lisboa Codex, Portugal. 2 Supported by Grant No. 20.0944/85 CNPq, Brazil. Permanent address: Departamento de Fisiologia, Faculdade de Medicina de RibeirAo Preto, 14.100 Universidade de S. Paulo, S. Paulo, Brazil.
Correspondence to: F.H. Lopes da Silva, Dept. of Experimental Zoology, Universiteit van Amsterdam, Biologisch Centrum Gebouw II, Kruislaan 320, 1098 SM Amsterdam, (The Netherlands).
and Dudek 1984; Konnerth et al. 1986). In this type of study an important limitation has been the lack of appropriate methods of quantification of EEG seizure activity. In this respect the most recent investigations have made use either of linear coherence and phase methods of analysis (Gotman 1983, 1987; Lieb et al. 1987) or of methods derived from information theory (Mars and Lopes da Silva 1983). However, in these studies the main concern has been either the interhemispheric spread (Gotman 1983, 1987; Lieb et al. 1987) or the spread between related but distant anatomical structures (Mars and Lopes da Silva 1983). We asked the question whether a local form of seizure activity will propagate within a structure such as the hippocampal formation (HF) along its whole length. In this study we employed methods of signal analysis enabling the quantification of both linear and non-linear relationships between EEG signals. Linear relations were quantified using the correlation coefficient r 2. We calculated r 2 as a function of timeshift between the signals. This is the square of the crosscorrelation function. It is the counterpart in the time domain of the coherence function in the frequency domain (Gotman 1983, 1987; Lieb et al. 1987). An advantage of the cross-correlation function is that it can be estimated on the basis of smaller epochs than those that are needed to obtain reliable estimates of coherence; r E expresses the linear statistical dependence of one signal on another. Epileptiform signals re-
0013-4649/89/$03.50 © 1989 Elsevier Scientific Publishers Ireland, Ltd.
PROPAGATION OF AFTERDISCHARGES corded from different areas are not necessarily linearly related. This led us to develop a procedure to calculate also the non-linear relationship between signals. We did this by applying the method of the non-linear correlation ratio h 2 (Guilford and Fruchter 1981; Pijn et al. 1989a,b). This measure gives the general statistical dependence of one signal on another signal even if the constraint of linearity does not hold. When the signals are linearly related h 2 is, in principle, equal to r 2. An alternative way to estimate relations between EEG signals has been introduced (Mars and Lopes da Silva 1983), namely the computation of the average amount of mutual information (AAMI). We preferred here to use the non-linear correlation method by means of which h 2 is calculated, instead of AAMI, because the former can be directly interpreted in terms of general statistical dependence between signals and it can give information about the type of relationship, whereas AAMI is a more complex measure of relationship. Besides the interest of investigating the general phenomenon of seizure propagation in the HF, this investigation may help to clarify how the different parts of the hippocampus are interrelated: on the one hand, there exists evidence that the different parts of the HF are related by means of an intrinsic system of longitudinal fibers extending along its length (Lorente de N6 1934; Hjorth-Simonsen 1973; Laurberg 1979; Bartesaghi et al. 1983) and that the epileptiform seizures elicited either in the dorsal or ventral part of the HF may spread to the opposite pole (Elul 1964); on the other hand, anatomical and electrophysiological data show that there are considerable differences between the functional anatomies of the dorsal (DHF), splenial (SP) and ventral hippocampal (VHF) formations (Siegel and Tassoni 1971; Ruth et al. 1982; Witter and Groenewegen 1984; Lopes da Silva et al. 1985; Van Groen and Lopes da Silva 1985).
Methods
Under halothane anesthesia 4 adult male Wistar rats (250-350 g) were implanted with electrodes symmetrically in the right and left D H F and in SP and VHF on the right side. After induction of anesthesia, the animals were tracheostomized, artificially ventilated and placed in a stereotaxic frame. The expired CO 2 concentration was monitored continuously by a capnograph and the body temperature was controlled. Seven bundles of stainless steel electrodes (diameter 100 ktm, insulated except at the cut ends), 4 for recording (one for each DHF, one for SP and one for the VHF) and 3 for stimulation (one for each D H F and one for the VHF) were implanted according to stereotaxic coordinates (Pellegrino et al. 1979). In different animals the positions of the bundles along the hippocampal axis were as follows (in relation to the bregma) at DHF: - 1 . 8 / - 2 . 4 mm, at SP: - 2 . 6 / - 4 . 4 ram, at VHF: -2.4/-3.6 mm. Their laminar positions were adjusted according to the recorded evoked field potentials (EPs) (Leung 1978, 1979a,b; Wadman et al. 1983). The reference was a screw in the frontal bone. EP measurements were obtained using monopolar derivations against this common frontal reference.
173 However, all records of afterdischarges (ADs) that were analyzed were obtained using bipolar derivations between electrodes of the same bundle from which EPs of opposite polarity had been previously recorded against the common reference. Bipolar recording of ADs was necessary to emphasize local activity and avoid contamination of the association measures with volume conducted potentials generated at a distance from the recording sites. The positions of the electrodes aimed at both DHFs and SP were adjusted according to the recorded potentials evoked by stimulation in the stratum radiatum (Schaffer collaterals) of the DHF. These electrodes were considered to be in good position when typical polarity-reversed responses were obtained (Fig. 1C). The position of the electrodes aimed at the V H F was also adjusted in order to record polarity-reversed EPs by local stimulation (Fig. 1D). The positions of the electrodes were later histologically verified; an example is shown in Fig. lB. In order to evoke seizure discharges, trains (4-5 sec) of pulses of 0.2 msec duration at 10, 25 or 50 Hz (tetanic stimulation), with an intensity 2-3 times the threshold of the EPs (usually 500-800 ~A) were applied. The signals were recorded on paper (16-channel Siemens-Elema Mingograph EEG machine, 0.5-700 Hz bandpass filtering), digitized at 1.4 kHz and stored on digital tape ( P D P l l / 3 4 computer). Signal analytical methods were used to study the propagation of epileptiform activity. We assumed that the epileptiform activity could be delayed and transformed in the course of spreading, the transformation being linear a n d / o r non-linear. The analysis was aimed at estimating the statistical relationship between a pair of EEG signals and the timeshift for which the amount of relationship was maximal, using linear and non-linear analytical methods (Pijn et al. 1989a.b). Two types of measure were used: the squared correlation coefficient (r 2) and the non-linear correlation ratio (h2). To apply the latter, a linear relationship between the signals need not be assumed (Guilford and Fruchter 1981; Pijn et al. 1989a,b). These methods were applied to short segments of the EEG signals (epochs of about 1 sec) selected in such a way that a burst of paroxysmal discharges was always contained in one epoch (stationarity was assumed within an epoch). The association values (r 2 and h 2) were calculated for each epoch. This was done by considering pairs of samples of the x and y signals. The regression curve giving the relation between the signals was estimated as follows: a scatterplot was made of the amplitude values of the samples. When the signals are linearly related the regression curve is given by a straight line that is calculated by applying least-squares linear regression; r 2 expresses the relative variance of the points around the line. When a straight line is a poor approximation of the regression curve (non-linearly related signals) a procedure to estimate the curve is to split the x amplitude range in bins (we used 10 equal sized bins for segments of about 1 sec). Each bin is characterized by a point with coordinates p and q. For p we took the midpoint of the x range of the bin and for q the average of the y amplitude values that correspond to the bin. The points (p, q) of successive bins were connected by straight lines. In this way a piecewise linear approximation of the
174
C. N U N E S FILIPE ET AL.
M
56
(~
87
~.~~'~'~
3 4
M
2_/---~
ZY--~
3_/---~ ~
3J--~.
4_J--~ ~'
5
~
8_F--L___
~'
4J---~
5_F--L 6J--~ 8J---L_.
2
'V'-
I
1
P R O P A G A T I O N OF A F T E R D I S C H A R G E S (non-linear) regression curve was obtained. Finally h 2 expresses the relative variance of the points around the non-linear regression curve. To compute r 2 and h 2 as functions of timeshift, one signal was shifted in steps in relation to the other and r 2 and h 2 were calculated for each step. The step was equal to the sample interval (1/512 sec) and the range of shifts used was from - 2 0 to + 20 msec. The maximal relationship was taken as the association value for the considered epoch and the corresponding timeshift as an estimator of the time delay between the 2 signals. Before doing these calculations the offset component was removed from each epoch of the EEG signal. The values of r 2 or h 2 were expressed as percentages (identical signals corresponding to values of 100%) and plotted as a function of the time of the AD.
Results
Evoked potentials In order to have a frame of reference for the relative positions of the recording and stimulating sites in relation to the structure of the H F field potentials evoked by stimulation either of the dorsal or the ventral stratum radiatum were recorded, as shown in Fig. I. Stimulation of the Schaffer collaterals in D H F never evoked clear responses in the V H F (Fig. 1C, channels 7 and 8). Stimulation of the V H F elicited EPs with similar morphology and polarity as the described above when recorded from the same region (Fig. 1D). However, stimulation of the V H F never evoked EPs in the D H F . Even when repetitive stimulation of the D H F was carried out (up to 5 sec) with frequencies ranging from 0.5 to 50 Hz, no clear EPs were recorded in the VHF. The reverse was also done: repetitive stimulation of the VHF, under the same conditions as indicated above, did not elicit measurable EPs in DHF, whereas in SP only small amplitude EPs similar to those found with paired pulse stimulation were observed. Afterdischarges Following local tetanic stimulation of the DHF, A D s were recorded both in the D H F s (ipsilateral and contralateral) and
175 in the SP. The A D s recorded m the SP showed burst morphology, time sequence and duration similar to those recorded in the D H F (Fig. 2). However, A D s recorded in the V H F following local tetanization of the D H F had always small amplitudes without clear spikes. The tetanic stimulation of the V H F never induced A D s in other hippocampal regions, even for a wide range of tetani, with duration up to 10 sec, frequency between 10 and 50 Hz and pulse intensity up to 1 mA. Only once a short A D (about 2 sec long) was recorded locally after 10 sec 25 Hz tetanic stimulation of the VHF. The degree of association between pairs of EEG signals was computed for 6 ADs, recorded from 3 animals. In all cases, the h 2 values were significantly larger than the r 2 values for the same epochs for the D H F - D H F and DHF-SP associations ( P < 0.001, Wilcoxon matched-pairs signed-ranks test). The difference between the 2 measures were specially marked when the bursts consisted of complex wave forms with a large number of spikes. This indicates that the h 2 is more powerful than the r 2 for determining the relatedness between complex EEG signals. For the same reason we preferred this measure. The h 2 values are plotted in Fig. 3A for the first 7 epochs of all 6 ADs. The EEG signals recorded from ipsi- and contralateral D H F s and from D H F and ipsilateral SP showed degrees of association during A D s of about 60%. In contrast, the values of association found between the EEG seizure activities of D H F and V H F were invariably low. The statistical comparison of the association values between signals recorded from D H F versus SP and D H F versus V H F showed that the former were significantly larger ( P < 0.00l, Wilcoxon matched-pairs signed-ranks test). Since the individual length of the analyzed A D s was variable (7-20 sec), we do not present in Fig. 3 the h 2 values for the late bursts. However, statistical analysis showed a significant difference ( P < 0.05) in the sense of a decrease in the h 2 of the terminal in relation to the initial epoch for the D H F - S P association (the mean value of the 1st epoch was 30.1%_+9.9 larger than the mean value of the terminal one). For the interhemispheric D H F association, the same tendency was not found. Time delays between bursts recorded from both D H F s and between D H F and SP were estimated using the h 2 method. Time delays could not be
Fig. 1. Positions of the electrodes and averaged field potentials evoked by stimulation in the stratum radiatum (Schaffer collaterals) in the dorsal hippocampal formation (DHF), the splenium (SP) and ventral hippocampal formation (VHF). A: positions of the recording electrodes are shown on schematic frontal brain sections, adapted from Pellegrino et al. (1979). The distances from the planes of the sections to the bregma (mm) are written at the top of each section. Electrode positions and corresponding record numbers are indicated: 1 and 2 = left D H F , 3 and 4 = right DHF, 5 and 6 = right SP and 7 and 8 = right VHF. R = right; L = left; M = midline. B: schematic camera lucida drawings of transverse sections of the hippocampal formation (HF) showing the positions of the recording electrodes. C: position of one stimulating electrode. The records of EPs were taken against a c o m m o n frontal reference. They represent averages of 40 sweeps. Note that stimulation of the Schaffer collaterals in the D H F evoked typical polarity-reversed responses that were similar in symmetrical regions of both D H F s (ipsi- and contralateral) (channels 1, 2, 3 and 4) and in SP (channels 5 and 6). N o clear responses were recorded in the V H F (channels 7 and 8). The arrows indicate the artefacts of the stimuli. Paired pulse stimulation was used. The records start with squared calibration pulses of 1 mV amplitude and 10 msec duration, positive upgoing. D: stimulation of the V H F evoked polarity reversed evoked potentials (EPs) in this same region (channels 7 and 8). Stimulation of the V H F did not evoke EPs in the D H F (channels 1, 2, 3 and 4). In the SP only a long latency small amplitude positive response was recorded, just below the stratum pyramidale.
176 L DHF
v u
-.
F J,,w
'
"
I
']
~
'
!
~
~
!
:
'
R DHF
,
'
, d
['
' i ,
R SP
R
VHF
~
~
~
-
~
-
-
~
-
--~-~,-~
5 mV[ ls Fig. 2. Bipolar records of an afterdischarge (AD) elicited by a 50 Hz, 4 sec long local tetanic stimulation in the right DHF. The arrow indicates the artefact of the last pulse of the tetanus. The ADs recorded in both D H F (ipsi- and contralateral) and hippocampal SP showed similar burst morphology, time sequence and duration. No clear AD was recorded in the VHF. (Abbreviations as in Fig. 1.)
A
B 100 -
20
90" 80 -
--~ DHFipsi-DHFcontralat,
DHFipsi-DHFcontralat. delays
18 16
,...,
70-
~
60:
.~
so-
"~4o:
~ lo ~ s
N 20"• ~1o-
,~~ a
14
0 0
,
~
,
,
,
,
'
l
2
3
4
5
6
7
12
s4 2 0
.
,.
.
, .. ,
~ rIl' u1 "
•
-16-14-v~-10-8 -6 .4 -2
u
2
•
I , I
4
,u
6
•
e p o c h sequence
100-
•" ~
90-
u
-u,
u,u
8 10121,~ 16
Delay (ms) DHF-SP delays
20
DHF - SP
18
DHF - VHF
80 -
16 14
--- 7 0 .
12
"-" 6 0 -
50" 4o: 30:
10 t_ a~
-~ 2 0
E
10' "<
4-
Z
0 0
u
i
i
n
1
2
3
4
I
I
u
5 6 7 epoch s e q u e n c e
0
I
I
I
~
-16-14-12-10-8-6-4-2
n
I
u
I
I
u[~]
0 2 4 6 8 10121416 Delay (ms)
Fig. 3. Association values (h 2) and time delays estimated on 6 a fterdischarges (ADs). A: averaged h z association values for the first 7 epochs. Note that the degree of association during the A D s between ipsi- and contralateral D H F (top) and between D H F and SP (bottom, closed squares) was large in comparison with that between the D H F and V H F (bottom, open squares). B: histograms of the distribution of time delays estimated by the h 2 method on the 6 analyzed ADs, between the ipsi- and contralateral D H F s (top) and between the D H F and SP (bottom). 82% of the delays estimated between ipsi- and contralateral D H F s were smaller than 1 msec. For the DHF-SP relation a more widespread distribution was found.
PROPAGATION OF AFTERDISCHARGES estimated as regards the relationship between D H F or SP and VHF because the corresponding association values were too low. The distributions of the time delays estimated for the interhemispheric D H F relation and for the DHF-SP relation are shown in Fig. 3B. The distribution of time delays for the interheimspheric association between DHFs showed a narrow peak around zero delay (82% of the delays < 1 msec) but we did not find any difference between delays estimated in the initial and terminal parts of the AD. For the DHF-SP relation a smaller degree of synchrony (40% of the delays < 1 msec) and a wider range of values was found.
Discussion The results show that the association measures (both h 2 and r 2) between ADs elicited by stimulation of the D H F Schaffer collaterals and bipolarly recorded are relatively large for the ipsi- and contralateral D H F and for the D H F and SP of the same side, in contrast to those found between D H F and VHF. These results are in agreement with anatomical data: commissural connections in the rat arise in the hilus of the dentate gyrus and on CA3 field; the latter, stimulated in the present experiments, are very dense in stratum oriens and radiatum (Van Groen and Wyss 1988). However, as regards longitudinal connections it was shown recently that the number of longitudinal or septo-temporal associational fibers which arise in the D H F and SP markedly decreases as one proceeds in the ventral direction and, conversely, the number of fibers with their origin in the V H F diminishes in the SP direction (for review see Witter 1986). The values of the association measures were not constant in the course of ADs. However, only for the DHF-SP association a significant difference was found between the initial and the terminal epochs, the latter values being smaller than the former. Using the h 2 method to estimate the delays between channels, a high degree of synchrony during ADs between ipsi- and contralateral DHFs was found (82% of the delays estimated were < 1 msec). Between D H F and S P a wider range of values was found. This seems to indicate that although D H F and SP are highly coupled during the ADs, the degree of synchrony between these areas is not as large as between homotopical places in ipsi- and contralateral DHF. It appears that the ipsiand contralateral D H F act as strongly coupled oscillators. In a complementary series of experiments we found that sectioning a specific part of the ventral hippocampal commissure makes this coupling disappear (Fernandes de Lima et al. submitted). We found, like others in cat (Elul 1964), that the threshold to elicit ADs in the V H F was considerably higher than in the DHF. The fact that no clear ADs were ever recorded in the V H F following tetanization of the D H F might be interpreted as due to this high V H F threshold. However, this is not likely to be the only explanation since EPs could also not be recorded from the VHF following stimulation of the DHF. Therefore, we may conclude that, besides differences in threshold, the fact that the D H F and V H F do not appear to be anatomically related by a direct pathway precludes ADs to
177 spread between the D H F and the VHF, in contrast with the propagation from ipsi- to contralateral D H F and from D H F to SP. This finding allows drawing 2 general conclusions: (i) the D H F and SP form a functional entity only loosely coupled to the VHF; (ii) an important condition for the propagation of epileptiform ADs between different areas is the existence of direct and dense connections between those areas, whereas simple spatial continuity is not enough.
References Bartesaghi, R., Gessi, T. and Sperti, L. Interlamellar transfer of impulses in the hippocampal formation. Exp. Neurol., 1983, 82: 550-567. Elul, R. Regional differences in the hippocampus of the cat. I. Specific discharge patterns of the dorsal and ventral hippocampus and their role in the generalized seizures. Electroenceph, clin. Neurophysiol., 1964, 16: 470-488. Fernandes de Lima, V.M., Pijn, J.P., Nunes Filipe, C. and Lopes da Silva, F.H. The role of hippocampal commissures in the interhemispheric transfer of epileptiform afterdischarges in the rat: a study using linear and non-linear regression analysis. Submitted. Gotman, J. Measurement of small time differences between EEG channels: method and application to epileptic seizure propagation. Electroenceph. clin. Neurophysiol., 1983, 56: 501-514. Gotman, J. Computer analysis of the EEG in epilepsy. In: F.H. Lopes da Silva, W. Storm van Leeuwen and A. Rrmond (Eds.), Handbook of Electroencephalography and Clinical Neurophysiology. Rev. Ser. Vol. 2. Clinical Applications of Computer Analysis of EEG and Other Neurophysiological Signals. Elsevier, Amsterdam, 1986: 171-204. Gotman, J. Interhemispheric interactions in seizures of focal onset: data from human intracranial recordings. Electroenceph. clin. Neurophysiol., 1987, 67: 120-133. Guilford, J.P. and Frnchter, B. Fundamental Statistics in Psychology and Education, 6th edn. McGraw-Hill, Singapore, 1981: 296-304. Hjorth-Simonsen, A. Some intrinsic connections of the hippocampus in the rat: an experimental analysis. J. Comp. Neurol., 1973, 147: 145-162. Konnerth, A., Heinemann, U. and Yaari, Y. Nonsynaptic epileptogenesis in the mammalian hippocampus in vitro. I. Development of seizure-like activity in low extracellular calcium. J. Neurophysiol., 1986, 56: 409-423. Laurberg, S. Commissural and intrinsic connections of the rat hippocampus. J. Comp. Neurol., 1979, 184: 685-708. Leung, L.S. Hippocampal CA1 region - - demonstration of antidromic dendritic spike and dendritic inhibition. Brain Res., 1978, 158: 219-222. Leung, L.S. Orthodromic activation of hippocampal CA1 region of the rat. Brain Res., 1979a, 176: 49-67. Leung, L.S. Potentials evoked by alvear tract in hippocampal CA1 region of rats. I. Topographical projection, component analysis, and correlation with unit activities. J. Neurophysiol., 1979b, 42: 1557-1570.
178 Lieb, J.P., Hoque, K,, Skomer, C.E. and Song, X.-W. Interhemispheric propagation of human mesial temporal lobe seizures: a coherence/phase analysis. Electroenceph. clin. Neurophysiol., 1987, 67: 101-119. Lopes da Silva, F.H., Groenewegen, H.J., Holsheimer, J., Room, P., Witter, M.P., Groen, Th. van and Wadman, W.J. The hippocampus as a set of partially overlapping segments with a topographically organized system of inputs and outputs: the entorhinal cortex as a sensory gate, the medial septum as a gain-setting system and the ventral striatum as a motor interface. In: G. Buzsaki and C.H. Vanderwolf (Eds.), Electrical Activity of the Archicortex. Akademiai Kiad6, Budapest, 1985: 83-106. Lorente de N6, R. Studies on the structure of the cerebral cortex. II. Continuation of the study of the amrnonic system. Psychol. Neurol., 1934, 46: 113-177. Mars, N.J.I. and Lopes da Silva, F.H. Propagation of seizure activity in kindled dogs. Electroenceph. clin. Neurophysiol., 1983, 56: 194-209. Pellegrino, L.J., Pellegrino, A.S. and Cushman, A.J. A stereotaxic Atlas of the Rat Brain. Plenum Press, New York, 1979:122 pp. Pijn, J.P.M., Vijn, P.C.M., Lopes da Silva, F.H. and De Lima, V.M.F. Evolution of interactions between brain structures during an epileptic seizure in the kindled rat. In: Advances in Epileptology. Raven Press, New York, 1989a: in press. Pijn, J.P.M., Vijn, P.C.M., Lopes da Silva, F.H., Van Erode Boas, W. and Blanes, W. Localization of epileptogenic foci using a new signal analytical approach. Neurophysiol. clin., 1989b, in press. Prince, D.A. Neurophysiology of epilepsy. Annu. Rev. Neurosci., 1978, 1: 395-415.
C. NUNES FILIPE ET AL. Ruth, E.R., Collier, T.J. and Routtenberg, A. Topography between the entorhinal cortex and the dentate septotemporal axis in rats. I. Medial and intermediate entorhinal projecting cells, J. Comp. Neurol., 1982, 209: 69-78. Siegel, A. and Tassoni, J.P. Differential efferent projections from the ventral and dorsal hippocampus of the cat. Brain Behav. Evol., 1971, 4: 185-200. Taylor, C.P. and Dudek, F.E. Synchronization without active chemical synapses during hippocampal afterdischarges. J. Neurophysiol., 1984, 52: 143-155. Van Groen, Th. and Lopes da Silva, F.H. Septotemporal distribution of entorhinal projections to the hippocampus in the cat: electrophysiological evidence. J. Comp. Neurol., 1985, 238: 1-9. Van Groen, Th. and Wyss, M.J. Species differences in hippocampal commissural connections: studies in rat, guinea pig, rabbit, and cat. J. Comp. Neurol., 1988, 267: 322-334. Wadman, W.J., Lopes da Silva, F.H. and Leung, L.S. Two types of interictal transients of reversed polarity in rat hippocampus during kindling. Electroenceph. clin. Neurophysiol., 1983, 55: 314-319. Witter, M.P. and Groenewegen, H.J. Laminar origin and septotemporal distribution of entorhinal and perirhinal projections to the hippocampus in the cat. J. Comp. Neurol., 1984, 224: 371-385. Witter, M.P. A survey of the anatomy of the hippocampal formation, with emphasis on the septotemporal organization of its intrinsic and extrinsic connections. In: R. Schwarcz and A. Yehezkel-Ari (Eds,), Excitatory Amino Acids and Epilepsy. Plenum, New York, 1986: 67-82.
172
EEG 02384
Short communication
Propagation of afterdischarges along the septo-temporal axis of the rat hippocampus: a quantitative analysis C. Nunes Filipe 1, j.p. Pijn, V.M. Fernandes de Lima 2 and F.H. Lopes da Silva Dept. of Experimental Zoology, Universiteit van Amsterdam, Biologisch Centrum Gebouw 1I, 1098 S M Amsterdam (The Netherlands) (Accepted for publication: 1 April 1989)
Summary Afterdischarges (ADs) were elicited in the hippocampus of the rat under halothane anesthesia. Records were made along the septo-temporal axis of the hippocampal formation (dorsal (DHF), splenial (SP) and ventral (VHF) regions). For comparison records were also made from the contralateral DHF. The propagation of ADs was quantified using linear and non-linear regression analysis. The values of non-linear association were in general larger than those of linear association. Values of association and time delays between pairs of EEG signals recorded from different regions were estimated as function of elapsed time during the ADs. The association measures for ADs elicited by stimulation of the DHF were relatively large between the ipsi- and contralateral DHF and between the DHF and SP of the same side in contrast to those found between DHF and VHF. The threshold to elicit ADs in the VHF was considerably higher than in the DHF. We conclude that the DHF and SP form a functional entity only loosely coupled to the VHF and that simple spatial continuity is not enough for the propagation of epileptiform ADs between different areas. The estimate of time delays indicated a high degree of synchrony during ADs between homotopical sites on ipsi- and contralateral DHF whereas the degree of synchrony between DHF and SP was smaller. Key words: Hippocampus; Epileptic seizures; Propagation; Septo-temporal axis; Non-linear regression
The mechanisms of propagation of seizure discharges are still little understood in spite of a relatively large number of investigations (for review see Gotman 1986). Even within the same brain structure it is not yet clear to what extent and by which mechanisms seizure activity may spread, by neural pathways through chemical synaptic transmission (Prince 1978) or by simple contiguity through non-synaptic mechanisms (Taylor
i Supported by a grant from Junta Nacional de Investiga~o Cientifica e Tecnol6gica (JNICT), Portugal. Permanent address: Departamento de Fisiologia, Faculdade de CiSncias M&iicas da Universidade Nova de Lisboa, Campo de Santana 130, 1198 Lisboa Codex, Portugal. 2 Supported by Grant No. 20.0944/85 CNPq, Brazil. Permanent address: Departamento de Fisiologia, Faculdade de Medicina de RibeirAo Preto, 14.100 Universidade de S. Paulo, S. Paulo, Brazil.
Correspondence to: F.H. Lopes da Silva, Dept. of Experimental Zoology, Universiteit van Amsterdam, Biologisch Centrum Gebouw II, Kruislaan 320, 1098 SM Amsterdam, (The Netherlands).
and Dudek 1984; Konnerth et al. 1986). In this type of study an important limitation has been the lack of appropriate methods of quantification of EEG seizure activity. In this respect the most recent investigations have made use either of linear coherence and phase methods of analysis (Gotman 1983, 1987; Lieb et al. 1987) or of methods derived from information theory (Mars and Lopes da Silva 1983). However, in these studies the main concern has been either the interhemispheric spread (Gotman 1983, 1987; Lieb et al. 1987) or the spread between related but distant anatomical structures (Mars and Lopes da Silva 1983). We asked the question whether a local form of seizure activity will propagate within a structure such as the hippocampal formation (HF) along its whole length. In this study we employed methods of signal analysis enabling the quantification of both linear and non-linear relationships between EEG signals. Linear relations were quantified using the correlation coefficient r 2. We calculated r 2 as a function of timeshift between the signals. This is the square of the crosscorrelation function. It is the counterpart in the time domain of the coherence function in the frequency domain (Gotman 1983, 1987; Lieb et al. 1987). An advantage of the cross-correlation function is that it can be estimated on the basis of smaller epochs than those that are needed to obtain reliable estimates of coherence; r E expresses the linear statistical dependence of one signal on another. Epileptiform signals re-
0013-4649/89/$03.50 © 1989 Elsevier Scientific Publishers Ireland, Ltd.
PROPAGATION OF AFTERDISCHARGES corded from different areas are not necessarily linearly related. This led us to develop a procedure to calculate also the non-linear relationship between signals. We did this by applying the method of the non-linear correlation ratio h 2 (Guilford and Fruchter 1981; Pijn et al. 1989a,b). This measure gives the general statistical dependence of one signal on another signal even if the constraint of linearity does not hold. When the signals are linearly related h 2 is, in principle, equal to r 2. An alternative way to estimate relations between EEG signals has been introduced (Mars and Lopes da Silva 1983), namely the computation of the average amount of mutual information (AAMI). We preferred here to use the non-linear correlation method by means of which h 2 is calculated, instead of AAMI, because the former can be directly interpreted in terms of general statistical dependence between signals and it can give information about the type of relationship, whereas AAMI is a more complex measure of relationship. Besides the interest of investigating the general phenomenon of seizure propagation in the HF, this investigation may help to clarify how the different parts of the hippocampus are interrelated: on the one hand, there exists evidence that the different parts of the HF are related by means of an intrinsic system of longitudinal fibers extending along its length (Lorente de N6 1934; Hjorth-Simonsen 1973; Laurberg 1979; Bartesaghi et al. 1983) and that the epileptiform seizures elicited either in the dorsal or ventral part of the HF may spread to the opposite pole (Elul 1964); on the other hand, anatomical and electrophysiological data show that there are considerable differences between the functional anatomies of the dorsal (DHF), splenial (SP) and ventral hippocampal (VHF) formations (Siegel and Tassoni 1971; Ruth et al. 1982; Witter and Groenewegen 1984; Lopes da Silva et al. 1985; Van Groen and Lopes da Silva 1985).
Methods
Under halothane anesthesia 4 adult male Wistar rats (250-350 g) were implanted with electrodes symmetrically in the right and left D H F and in SP and VHF on the right side. After induction of anesthesia, the animals were tracheostomized, artificially ventilated and placed in a stereotaxic frame. The expired CO 2 concentration was monitored continuously by a capnograph and the body temperature was controlled. Seven bundles of stainless steel electrodes (diameter 100 ktm, insulated except at the cut ends), 4 for recording (one for each DHF, one for SP and one for the VHF) and 3 for stimulation (one for each D H F and one for the VHF) were implanted according to stereotaxic coordinates (Pellegrino et al. 1979). In different animals the positions of the bundles along the hippocampal axis were as follows (in relation to the bregma) at DHF: - 1 . 8 / - 2 . 4 mm, at SP: - 2 . 6 / - 4 . 4 ram, at VHF: -2.4/-3.6 mm. Their laminar positions were adjusted according to the recorded evoked field potentials (EPs) (Leung 1978, 1979a,b; Wadman et al. 1983). The reference was a screw in the frontal bone. EP measurements were obtained using monopolar derivations against this common frontal reference.
173 However, all records of afterdischarges (ADs) that were analyzed were obtained using bipolar derivations between electrodes of the same bundle from which EPs of opposite polarity had been previously recorded against the common reference. Bipolar recording of ADs was necessary to emphasize local activity and avoid contamination of the association measures with volume conducted potentials generated at a distance from the recording sites. The positions of the electrodes aimed at both DHFs and SP were adjusted according to the recorded potentials evoked by stimulation in the stratum radiatum (Schaffer collaterals) of the DHF. These electrodes were considered to be in good position when typical polarity-reversed responses were obtained (Fig. 1C). The position of the electrodes aimed at the V H F was also adjusted in order to record polarity-reversed EPs by local stimulation (Fig. 1D). The positions of the electrodes were later histologically verified; an example is shown in Fig. lB. In order to evoke seizure discharges, trains (4-5 sec) of pulses of 0.2 msec duration at 10, 25 or 50 Hz (tetanic stimulation), with an intensity 2-3 times the threshold of the EPs (usually 500-800 ~A) were applied. The signals were recorded on paper (16-channel Siemens-Elema Mingograph EEG machine, 0.5-700 Hz bandpass filtering), digitized at 1.4 kHz and stored on digital tape ( P D P l l / 3 4 computer). Signal analytical methods were used to study the propagation of epileptiform activity. We assumed that the epileptiform activity could be delayed and transformed in the course of spreading, the transformation being linear a n d / o r non-linear. The analysis was aimed at estimating the statistical relationship between a pair of EEG signals and the timeshift for which the amount of relationship was maximal, using linear and non-linear analytical methods (Pijn et al. 1989a.b). Two types of measure were used: the squared correlation coefficient (r 2) and the non-linear correlation ratio (h2). To apply the latter, a linear relationship between the signals need not be assumed (Guilford and Fruchter 1981; Pijn et al. 1989a,b). These methods were applied to short segments of the EEG signals (epochs of about 1 sec) selected in such a way that a burst of paroxysmal discharges was always contained in one epoch (stationarity was assumed within an epoch). The association values (r 2 and h 2) were calculated for each epoch. This was done by considering pairs of samples of the x and y signals. The regression curve giving the relation between the signals was estimated as follows: a scatterplot was made of the amplitude values of the samples. When the signals are linearly related the regression curve is given by a straight line that is calculated by applying least-squares linear regression; r 2 expresses the relative variance of the points around the line. When a straight line is a poor approximation of the regression curve (non-linearly related signals) a procedure to estimate the curve is to split the x amplitude range in bins (we used 10 equal sized bins for segments of about 1 sec). Each bin is characterized by a point with coordinates p and q. For p we took the midpoint of the x range of the bin and for q the average of the y amplitude values that correspond to the bin. The points (p, q) of successive bins were connected by straight lines. In this way a piecewise linear approximation of the
174
C. N U N E S FILIPE ET AL.
M
56
(~
87
~.~~'~'~
3 4
M
2_/---~
ZY--~
3_/---~ ~
3J--~.
4_J--~ ~'
5
~
8_F--L___
~'
4J---~
5_F--L 6J--~ 8J---L_.
2
'V'-
I
1
P R O P A G A T I O N OF A F T E R D I S C H A R G E S (non-linear) regression curve was obtained. Finally h 2 expresses the relative variance of the points around the non-linear regression curve. To compute r 2 and h 2 as functions of timeshift, one signal was shifted in steps in relation to the other and r 2 and h 2 were calculated for each step. The step was equal to the sample interval (1/512 sec) and the range of shifts used was from - 2 0 to + 20 msec. The maximal relationship was taken as the association value for the considered epoch and the corresponding timeshift as an estimator of the time delay between the 2 signals. Before doing these calculations the offset component was removed from each epoch of the EEG signal. The values of r 2 or h 2 were expressed as percentages (identical signals corresponding to values of 100%) and plotted as a function of the time of the AD.
Results
Evoked potentials In order to have a frame of reference for the relative positions of the recording and stimulating sites in relation to the structure of the H F field potentials evoked by stimulation either of the dorsal or the ventral stratum radiatum were recorded, as shown in Fig. I. Stimulation of the Schaffer collaterals in D H F never evoked clear responses in the V H F (Fig. 1C, channels 7 and 8). Stimulation of the V H F elicited EPs with similar morphology and polarity as the described above when recorded from the same region (Fig. 1D). However, stimulation of the V H F never evoked EPs in the D H F . Even when repetitive stimulation of the D H F was carried out (up to 5 sec) with frequencies ranging from 0.5 to 50 Hz, no clear EPs were recorded in the VHF. The reverse was also done: repetitive stimulation of the VHF, under the same conditions as indicated above, did not elicit measurable EPs in DHF, whereas in SP only small amplitude EPs similar to those found with paired pulse stimulation were observed. Afterdischarges Following local tetanic stimulation of the DHF, A D s were recorded both in the D H F s (ipsilateral and contralateral) and
175 in the SP. The A D s recorded m the SP showed burst morphology, time sequence and duration similar to those recorded in the D H F (Fig. 2). However, A D s recorded in the V H F following local tetanization of the D H F had always small amplitudes without clear spikes. The tetanic stimulation of the V H F never induced A D s in other hippocampal regions, even for a wide range of tetani, with duration up to 10 sec, frequency between 10 and 50 Hz and pulse intensity up to 1 mA. Only once a short A D (about 2 sec long) was recorded locally after 10 sec 25 Hz tetanic stimulation of the VHF. The degree of association between pairs of EEG signals was computed for 6 ADs, recorded from 3 animals. In all cases, the h 2 values were significantly larger than the r 2 values for the same epochs for the D H F - D H F and DHF-SP associations ( P < 0.001, Wilcoxon matched-pairs signed-ranks test). The difference between the 2 measures were specially marked when the bursts consisted of complex wave forms with a large number of spikes. This indicates that the h 2 is more powerful than the r 2 for determining the relatedness between complex EEG signals. For the same reason we preferred this measure. The h 2 values are plotted in Fig. 3A for the first 7 epochs of all 6 ADs. The EEG signals recorded from ipsi- and contralateral D H F s and from D H F and ipsilateral SP showed degrees of association during A D s of about 60%. In contrast, the values of association found between the EEG seizure activities of D H F and V H F were invariably low. The statistical comparison of the association values between signals recorded from D H F versus SP and D H F versus V H F showed that the former were significantly larger ( P < 0.00l, Wilcoxon matched-pairs signed-ranks test). Since the individual length of the analyzed A D s was variable (7-20 sec), we do not present in Fig. 3 the h 2 values for the late bursts. However, statistical analysis showed a significant difference ( P < 0.05) in the sense of a decrease in the h 2 of the terminal in relation to the initial epoch for the D H F - S P association (the mean value of the 1st epoch was 30.1%_+9.9 larger than the mean value of the terminal one). For the interhemispheric D H F association, the same tendency was not found. Time delays between bursts recorded from both D H F s and between D H F and SP were estimated using the h 2 method. Time delays could not be
Fig. 1. Positions of the electrodes and averaged field potentials evoked by stimulation in the stratum radiatum (Schaffer collaterals) in the dorsal hippocampal formation (DHF), the splenium (SP) and ventral hippocampal formation (VHF). A: positions of the recording electrodes are shown on schematic frontal brain sections, adapted from Pellegrino et al. (1979). The distances from the planes of the sections to the bregma (mm) are written at the top of each section. Electrode positions and corresponding record numbers are indicated: 1 and 2 = left D H F , 3 and 4 = right DHF, 5 and 6 = right SP and 7 and 8 = right VHF. R = right; L = left; M = midline. B: schematic camera lucida drawings of transverse sections of the hippocampal formation (HF) showing the positions of the recording electrodes. C: position of one stimulating electrode. The records of EPs were taken against a c o m m o n frontal reference. They represent averages of 40 sweeps. Note that stimulation of the Schaffer collaterals in the D H F evoked typical polarity-reversed responses that were similar in symmetrical regions of both D H F s (ipsi- and contralateral) (channels 1, 2, 3 and 4) and in SP (channels 5 and 6). N o clear responses were recorded in the V H F (channels 7 and 8). The arrows indicate the artefacts of the stimuli. Paired pulse stimulation was used. The records start with squared calibration pulses of 1 mV amplitude and 10 msec duration, positive upgoing. D: stimulation of the V H F evoked polarity reversed evoked potentials (EPs) in this same region (channels 7 and 8). Stimulation of the V H F did not evoke EPs in the D H F (channels 1, 2, 3 and 4). In the SP only a long latency small amplitude positive response was recorded, just below the stratum pyramidale.
176 L DHF
v u
-.
F J,,w
'
"
I
']
~
'
!
~
~
!
:
'
R DHF
,
'
, d
['
' i ,
R SP
R
VHF
~
~
~
-
~
-
-
~
-
--~-~,-~
5 mV[ ls Fig. 2. Bipolar records of an afterdischarge (AD) elicited by a 50 Hz, 4 sec long local tetanic stimulation in the right DHF. The arrow indicates the artefact of the last pulse of the tetanus. The ADs recorded in both D H F (ipsi- and contralateral) and hippocampal SP showed similar burst morphology, time sequence and duration. No clear AD was recorded in the VHF. (Abbreviations as in Fig. 1.)
A
B 100 -
20
90" 80 -
--~ DHFipsi-DHFcontralat,
DHFipsi-DHFcontralat. delays
18 16
,...,
70-
~
60:
.~
so-
"~4o:
~ lo ~ s
N 20"• ~1o-
,~~ a
14
0 0
,
~
,
,
,
,
'
l
2
3
4
5
6
7
12
s4 2 0
.
,.
.
, .. ,
~ rIl' u1 "
•
-16-14-v~-10-8 -6 .4 -2
u
2
•
I , I
4
,u
6
•
e p o c h sequence
100-
•" ~
90-
u
-u,
u,u
8 10121,~ 16
Delay (ms) DHF-SP delays
20
DHF - SP
18
DHF - VHF
80 -
16 14
--- 7 0 .
12
"-" 6 0 -
50" 4o: 30:
10 t_ a~
-~ 2 0
E
10' "<
4-
Z
0 0
u
i
i
n
1
2
3
4
I
I
u
5 6 7 epoch s e q u e n c e
0
I
I
I
~
-16-14-12-10-8-6-4-2
n
I
u
I
I
u[~]
0 2 4 6 8 10121416 Delay (ms)
Fig. 3. Association values (h 2) and time delays estimated on 6 a fterdischarges (ADs). A: averaged h z association values for the first 7 epochs. Note that the degree of association during the A D s between ipsi- and contralateral D H F (top) and between D H F and SP (bottom, closed squares) was large in comparison with that between the D H F and V H F (bottom, open squares). B: histograms of the distribution of time delays estimated by the h 2 method on the 6 analyzed ADs, between the ipsi- and contralateral D H F s (top) and between the D H F and SP (bottom). 82% of the delays estimated between ipsi- and contralateral D H F s were smaller than 1 msec. For the DHF-SP relation a more widespread distribution was found.
PROPAGATION OF AFTERDISCHARGES estimated as regards the relationship between D H F or SP and VHF because the corresponding association values were too low. The distributions of the time delays estimated for the interhemispheric D H F relation and for the DHF-SP relation are shown in Fig. 3B. The distribution of time delays for the interheimspheric association between DHFs showed a narrow peak around zero delay (82% of the delays < 1 msec) but we did not find any difference between delays estimated in the initial and terminal parts of the AD. For the DHF-SP relation a smaller degree of synchrony (40% of the delays < 1 msec) and a wider range of values was found.
Discussion The results show that the association measures (both h 2 and r 2) between ADs elicited by stimulation of the D H F Schaffer collaterals and bipolarly recorded are relatively large for the ipsi- and contralateral D H F and for the D H F and SP of the same side, in contrast to those found between D H F and VHF. These results are in agreement with anatomical data: commissural connections in the rat arise in the hilus of the dentate gyrus and on CA3 field; the latter, stimulated in the present experiments, are very dense in stratum oriens and radiatum (Van Groen and Wyss 1988). However, as regards longitudinal connections it was shown recently that the number of longitudinal or septo-temporal associational fibers which arise in the D H F and SP markedly decreases as one proceeds in the ventral direction and, conversely, the number of fibers with their origin in the V H F diminishes in the SP direction (for review see Witter 1986). The values of the association measures were not constant in the course of ADs. However, only for the DHF-SP association a significant difference was found between the initial and the terminal epochs, the latter values being smaller than the former. Using the h 2 method to estimate the delays between channels, a high degree of synchrony during ADs between ipsi- and contralateral DHFs was found (82% of the delays estimated were < 1 msec). Between D H F and S P a wider range of values was found. This seems to indicate that although D H F and SP are highly coupled during the ADs, the degree of synchrony between these areas is not as large as between homotopical places in ipsi- and contralateral DHF. It appears that the ipsiand contralateral D H F act as strongly coupled oscillators. In a complementary series of experiments we found that sectioning a specific part of the ventral hippocampal commissure makes this coupling disappear (Fernandes de Lima et al. submitted). We found, like others in cat (Elul 1964), that the threshold to elicit ADs in the V H F was considerably higher than in the DHF. The fact that no clear ADs were ever recorded in the V H F following tetanization of the D H F might be interpreted as due to this high V H F threshold. However, this is not likely to be the only explanation since EPs could also not be recorded from the VHF following stimulation of the DHF. Therefore, we may conclude that, besides differences in threshold, the fact that the D H F and V H F do not appear to be anatomically related by a direct pathway precludes ADs to
177 spread between the D H F and the VHF, in contrast with the propagation from ipsi- to contralateral D H F and from D H F to SP. This finding allows drawing 2 general conclusions: (i) the D H F and SP form a functional entity only loosely coupled to the VHF; (ii) an important condition for the propagation of epileptiform ADs between different areas is the existence of direct and dense connections between those areas, whereas simple spatial continuity is not enough.
References Bartesaghi, R., Gessi, T. and Sperti, L. Interlamellar transfer of impulses in the hippocampal formation. Exp. Neurol., 1983, 82: 550-567. Elul, R. Regional differences in the hippocampus of the cat. I. Specific discharge patterns of the dorsal and ventral hippocampus and their role in the generalized seizures. Electroenceph, clin. Neurophysiol., 1964, 16: 470-488. Fernandes de Lima, V.M., Pijn, J.P., Nunes Filipe, C. and Lopes da Silva, F.H. The role of hippocampal commissures in the interhemispheric transfer of epileptiform afterdischarges in the rat: a study using linear and non-linear regression analysis. Submitted. Gotman, J. Measurement of small time differences between EEG channels: method and application to epileptic seizure propagation. Electroenceph. clin. Neurophysiol., 1983, 56: 501-514. Gotman, J. Computer analysis of the EEG in epilepsy. In: F.H. Lopes da Silva, W. Storm van Leeuwen and A. Rrmond (Eds.), Handbook of Electroencephalography and Clinical Neurophysiology. Rev. Ser. Vol. 2. Clinical Applications of Computer Analysis of EEG and Other Neurophysiological Signals. Elsevier, Amsterdam, 1986: 171-204. Gotman, J. Interhemispheric interactions in seizures of focal onset: data from human intracranial recordings. Electroenceph. clin. Neurophysiol., 1987, 67: 120-133. Guilford, J.P. and Frnchter, B. Fundamental Statistics in Psychology and Education, 6th edn. McGraw-Hill, Singapore, 1981: 296-304. Hjorth-Simonsen, A. Some intrinsic connections of the hippocampus in the rat: an experimental analysis. J. Comp. Neurol., 1973, 147: 145-162. Konnerth, A., Heinemann, U. and Yaari, Y. Nonsynaptic epileptogenesis in the mammalian hippocampus in vitro. I. Development of seizure-like activity in low extracellular calcium. J. Neurophysiol., 1986, 56: 409-423. Laurberg, S. Commissural and intrinsic connections of the rat hippocampus. J. Comp. Neurol., 1979, 184: 685-708. Leung, L.S. Hippocampal CA1 region - - demonstration of antidromic dendritic spike and dendritic inhibition. Brain Res., 1978, 158: 219-222. Leung, L.S. Orthodromic activation of hippocampal CA1 region of the rat. Brain Res., 1979a, 176: 49-67. Leung, L.S. Potentials evoked by alvear tract in hippocampal CA1 region of rats. I. Topographical projection, component analysis, and correlation with unit activities. J. Neurophysiol., 1979b, 42: 1557-1570.
178 Lieb, J.P., Hoque, K,, Skomer, C.E. and Song, X.-W. Interhemispheric propagation of human mesial temporal lobe seizures: a coherence/phase analysis. Electroenceph. clin. Neurophysiol., 1987, 67: 101-119. Lopes da Silva, F.H., Groenewegen, H.J., Holsheimer, J., Room, P., Witter, M.P., Groen, Th. van and Wadman, W.J. The hippocampus as a set of partially overlapping segments with a topographically organized system of inputs and outputs: the entorhinal cortex as a sensory gate, the medial septum as a gain-setting system and the ventral striatum as a motor interface. In: G. Buzsaki and C.H. Vanderwolf (Eds.), Electrical Activity of the Archicortex. Akademiai Kiad6, Budapest, 1985: 83-106. Lorente de N6, R. Studies on the structure of the cerebral cortex. II. Continuation of the study of the amrnonic system. Psychol. Neurol., 1934, 46: 113-177. Mars, N.J.I. and Lopes da Silva, F.H. Propagation of seizure activity in kindled dogs. Electroenceph. clin. Neurophysiol., 1983, 56: 194-209. Pellegrino, L.J., Pellegrino, A.S. and Cushman, A.J. A stereotaxic Atlas of the Rat Brain. Plenum Press, New York, 1979:122 pp. Pijn, J.P.M., Vijn, P.C.M., Lopes da Silva, F.H. and De Lima, V.M.F. Evolution of interactions between brain structures during an epileptic seizure in the kindled rat. In: Advances in Epileptology. Raven Press, New York, 1989a: in press. Pijn, J.P.M., Vijn, P.C.M., Lopes da Silva, F.H., Van Erode Boas, W. and Blanes, W. Localization of epileptogenic foci using a new signal analytical approach. Neurophysiol. clin., 1989b, in press. Prince, D.A. Neurophysiology of epilepsy. Annu. Rev. Neurosci., 1978, 1: 395-415.
C. NUNES FILIPE ET AL. Ruth, E.R., Collier, T.J. and Routtenberg, A. Topography between the entorhinal cortex and the dentate septotemporal axis in rats. I. Medial and intermediate entorhinal projecting cells, J. Comp. Neurol., 1982, 209: 69-78. Siegel, A. and Tassoni, J.P. Differential efferent projections from the ventral and dorsal hippocampus of the cat. Brain Behav. Evol., 1971, 4: 185-200. Taylor, C.P. and Dudek, F.E. Synchronization without active chemical synapses during hippocampal afterdischarges. J. Neurophysiol., 1984, 52: 143-155. Van Groen, Th. and Lopes da Silva, F.H. Septotemporal distribution of entorhinal projections to the hippocampus in the cat: electrophysiological evidence. J. Comp. Neurol., 1985, 238: 1-9. Van Groen, Th. and Wyss, M.J. Species differences in hippocampal commissural connections: studies in rat, guinea pig, rabbit, and cat. J. Comp. Neurol., 1988, 267: 322-334. Wadman, W.J., Lopes da Silva, F.H. and Leung, L.S. Two types of interictal transients of reversed polarity in rat hippocampus during kindling. Electroenceph. clin. Neurophysiol., 1983, 55: 314-319. Witter, M.P. and Groenewegen, H.J. Laminar origin and septotemporal distribution of entorhinal and perirhinal projections to the hippocampus in the cat. J. Comp. Neurol., 1984, 224: 371-385. Witter, M.P. A survey of the anatomy of the hippocampal formation, with emphasis on the septotemporal organization of its intrinsic and extrinsic connections. In: R. Schwarcz and A. Yehezkel-Ari (Eds,), Excitatory Amino Acids and Epilepsy. Plenum, New York, 1986: 67-82.