Acta Neurol. Scandinav. 53, 3-20, 1976

Department of Neurology, Haydarpasha Military Hospital, Istanbul and Division of Neurology, Medical School, Aegean University, Izmir.

STUDIES ON THE HUMAN EVOKED ELECTROSPINOGRAM

I . The Origin of the Segmental Evoked Potentials

CUMHUR ERTEKIN

ABSTRACT

Evoked potentials from the human spinal cord were studied in 39 normal volunteers. Intrathecal recordings from lower cervical and lower thoracic intervertebral levels were made after the supramaximal stimulation of the median, ulnar and posterior tibia1 nerves, respec- tively. I t was shown that the segmental cord potentials varied in shape and size according to the spatial relationship between the position of the electrode tip and the spinal cord and roots within the vertebral canal. Three main types of segmental evoked responses were obtained. One of thcm was recorded behind the cord dorsum and around the midline, and was composed of fast, sharp early, and slow, late com- ponents. This was called a CD potential and its first component was related to the activity of the ascending dorsal funiculus fibres. The second evoked response was the DR potential, and this triphasic com- pound action potential of very high amplitude and longer duration had no remarkable late component. It was recorded when the tip of the intrathecal electrode was lateral to the midline within the vertebral canal, and was then related mostly to activity of the spinal roots. Another kind of potential was called PH potential. I t had a very small triphasic spike and two later components with prominent negativities being higher than the first spike. This potential might be related to the electrode tip position facing, and close to, the posterior horn of the spinal gray matter. Late components of the segmental evoked potentials were related to the pre- and post synaptic activity of the horizontally oriented fibres within the segmental gray matter of the posterior horn.

Because of its firmly closed anatomical position, i t is difficult to approach the human spinal cord for recording of its electrical activity, although a few attempts have been made to record both spontaneous and evoked potentials directly from the cord.

Before the human application, the classical studies on the spinal cord

1’

4

by the pioneers of the 1930’s and 1940’s were all carried out with gross electrodes placed either on the surface of the cord or on the dorsal and ventral roots in cats. With the recording electrodes placed directly on the surface of the spinal cord, Gasser & Graham, in 1933, demonstrated that when an afferent spinal root is stimulated by a single electrical shock, the response recordable from the surface of the cord is a spike followed by slower, irregular potentials. The response was also record- able beyond the segment stimulated. The spike in this case was the action potential of the long axons of the dorsal root neurons which run up in the dorsal columns of the spinal cord. This spike was followed by a slow, negative potential and then by a long-lasting, positive com- ponent. These slow potentials were named by Gasser d? Graham, “the Negative and Positive Intermediary Cord Potentials”, respectively. After Gasser & Graham, many publications appeared in the literature dealing especially with the origin of these slow potentials.

Magladcry e t al. (1951) were able to record evoked potentials from the lumbosacral enlargement using intrathecal electrodes within the subarachnoid space in normal, adult human subjects. They evoked the potentials by stimulation of the posterior tibia1 nerve. However, the potential risks involved in this method must have prevented its wide- spread use, and the subject has been neglected until recent years. Recently Ertekin ( 1973) has shown that the intrathecal recording technique from the cervical and lumbosacral spinal cord of intact human subjects could be used with relative ease and safety. Although Shirnoji et al. (1972) and Cracco (1973) have described a simple method of recording the spinal evoked electrogram with electrodes either introduced into the epidural space in man or superficial electrodes placed on the skin. However, the methods of these authors, based on computer averaging technique, do not seem to allow further neurophysiological analysis in detail ; and these methods are unlikely to be used for clinical purposes on account of the very small size of the evoked potentials obtained. Furthermore, there is considerable distance between the spinal cord and the recording sites in these techniques, and so the great disadvantage of volume conduction could not easily be prevented. The physiological interpretation of the evoked potentials would thus remain within the uncertain and very restricted limits of the methods.

In the present study, an intrathecal recording technique has been used in 39 normal, young volunteers, and has shown that the origin of the spinal cord potentials evoked by adequate peripheral nerve stimulation varied with and depended on the position of the electrode tip within the vertebral canal. The results obtained in the present

5

study and the subsequent one ( E r f e k i n 1975) seem to offer opportunity for the clinical application to human spinal cord diseases.

MATERIAL AND METHODS

Thirty-nine young, adult men were investigated. They were volunteer soldiers and had no neurological complaints o r findings. The ethical principles of the Declaration of Helsinki concerning human experimentation were followed throughout the study. The mean age of the subjects studied was 21.6 years, and ranged between 20-35 years of age. The technique for recording the spinal cord evoked potentials was basically the same as was previously described (Ertekin 1973). The investiga- tions were all carried out with the subject in the lateral recumbent position. A large area around the puncture site on the skin was cleaned with antiseptic. For the insertion to the lower thoracic intervertebral levels, the subject was made to lie with his knees drawn up to his chin in order to stretch the lumbar spine as much as possible. For the lower cervical insertion, the neck was anteflexed toward the chest.

The recording electrodes used were sterilized, stainless steel needles, coated with Teflon@, except a t the tips; the long active electrode had a bared tip of 3 mm, and the tip diameter was somewhat less than 1 mm; the reference electrode had a bared tip of 5 mm. The active electrode was introduced into the subarachnoid space through the midpoint of an arbitrary line connecting the spinous processes of the vertebral levels investigated. The reference electrode was inserted subcutaneously, lateral to the intrathecal electrode at the same level. Recording electrodes were first connected to the output of the stimulator (DISA-Ministim 0, 14E 10) . During insertion of the intrathecal electrode, rectangular electrical pulses, with a duration of 0.1 msec and an intensity of 10-20 V at l/sec, were delivered. As soon as the subarachnoid space was entered, motor responses in the form of contractions of muscles in a segmental manner were observed in the limbs, according to the level of intrathecal electrode (such as upper limbs in the cervical region and lower limbs in the lower thoracic vertebral levels). A t this stage the subjects said they experienced a radicular type of sensation. Then the position of the electrode was gently readjusted until the lowest threshold for eliciting motor responses in the limb could be achieved. After the placement of the tip of the intrathecal electrode as close to the spinal cord as possibly by gentle manipulation, the electrodes were then connected to the input transformer (ratio: 1/30, 10-500 Hz) of the electro- myograph (DISA 14 A 30) . In order to obtain recordings from the lower thoracic vertebral levels, the posterior tibia1 nerve at the popliteal fossa (and for recordings from the lower cervical levels, the median and/or ulnar nerves at the elbow) was stimulated. During supramaximal stimulation of the peripheral nerves, the intra- thccal electrode was also adjusted and gently pushcd forwards, by subtle manipulations, to achieve the closest positions of the electrode to the spinal cord. In spite of this, the intrathecal electrode could not be directed onto the midline in some cases. In these cases, after the recording in this lateral site, a second attempt was made to replace the electrode on the midline and this position was the site for a second recording. Bipolar surface electrodes were used for peripheral nerve stimulation. Rectangular electrical pulses were used 0.1 msec long at 500 V and frequencies generally lower than 1 per sec. In 39 cases, potentials from 71 inter- vertebral (I.V.) levels were recorded. Among these, the simultaneous recordings from both lower cervical and thoracic levels were made in 29 cases. In a few

6

cases, three recording levels were used. The number of different I.V. levels recorded in these cases was as follows: C 5-6, 14; C 6-7, 22; Th I-C 7 , l ; Th 2-3, 1; Th 9-10,2; Th 10-11,9; Th 11-12,16; and Th 12-L 1, 6 recordings.

In some of the subjects, the electromyographic activities were simultaneously recorded from the back and neck muscles at the same level as the intrathecal electrodes by using a concentric needle electrode. A large, long ground electrode was connected circumferentially in between the lower cervical and thoracic regions. Another superficial ground electrode was also placed on the same extremity to which the stimulation was applied.

The sweep speed used for recording the evoked potentials on film was often 1.0 and 2.5 msec/mm; but 0.5 and 5.0 msec/mm was also used in some cases. A s well as single sweeps, 10-20 superimposed sweeps were photographed. The distance between the stimulating cathode and the skin insertion point of the intrathecal recording electrode was carefully measured with a tape-measure, paying attention to the usual anatomical course of the nerve stimulated.

During investigation, the general condition and the neurological status of each case were continuously controlled, and all cases were kept in neurological wards after the investigation. They were followed up and visited every day in the ward for a t least 20 days. After this time, the subjects were discharged from the hospital. The majority of them came for later controIs a t our request, and some of them were followed up for as long as 4 or 5 months at certain intervals. None of the cases showed any neurological or other permanent complaints or disturbances, and all were as normal as before the investigation. However, for a few days after the investigations, some of the subjects complained of neck and back pains, and a smaller number of cases had headaches for a few days because of “spinal fluid leakage” due to voluntary, early ambulation. These complaints disappeared within a few days. A few cases, especially those with cervical level recordings, showed body temperature elevation, 6-8 hours after the investigation. The temperature rose to between 38”-40” C for 6 6 h, and then fell to a normal level, either spontaneously or after administration of aspirin. However, a mild o r moderate hyperpyrexia continued for, on average, 2-3 days (range 1-10 days). No neurological ab- normalities were encountered in these cases either. Thus, it seems that the procedure presented in this paper did not give rise to any serious and permanent disturbances.

RESULTS

In all the subjects, clear-cut evoked potentials were invariably obtained from either lower thoracic I.V. or lower cervical I.V. level recordings after the supramaximal stimulation of either posterior tibia1 nerve or median/ulnar nerves respectively. The segmental evoked potentials appeared to have a different shape and size, probably owing to the spatial relationships between the tip of the intrathecal electrode and the spinal cord and roots. It was possible to differentiate three major types of potential. All three types are illustrated in Figures 1 and 2. The statisticaI results of the parameters of the potentials are given in Tables 1 and 2.

The segmental evoked potentials, illustrated in Figures 1 C and 2 C, were encountered most frequently. For the sake of brevity, they will be

7

Figure 1. Segmental euokeil potentials f r o m the tumbosarral spinal cord recorded after the supramaximal stimulation of the posterior tibia1 neroe at the popliteal fossa. Upper and lower rows are, respectivelgl, at faster and slower sweep speeds. .411 responses shown are 20 superimposed potentials, and a downward deflection of

the trace is positive in this and in all subsequent figures. Calibration: 15 msec and 0.2 mV. A j DR type o f potential; R ) PH type o f potential; C ) CD t y p e o f potential.

called the CD potentials. A CD potential consisted of two components: an early and a late wave. The earlier component was faster, more synchroneous, and often triphasic in shape. I t often started with a small, positive dip; this was followed by a negative deflection with a very fast rising phase; and then another positive deflection appeared. The amplitude of the descending limb of the early component was the most prominent. The late component started immediately after the positive peak of the first component, became negative with a slower rising phase, and either ended at the isoelectric line, or another positiv- ity appeared, as shown in the Figures. A second negative, slow wave of smaller amplitude could very seldom be recorded. The durations of the late components were longer and their contours less sharp than those of earlier components. The earlier sharp component was stable and constant in a given case if the stimulation and recording conditions were kept stable; while the late component was somewhat less stable in character and varied slightly from stimulus to stimulus. The amplitudes of both components varied from case to case, and the

8

higher amplitLide could he seen in either the first o r the second component; or else both of them had amplitudes of almost equal size. The mean amplitude of the late component in lower thoracic record- ings was found to be significantly higher than that of the earlier component ( P < 0.01 ) .

The segmental potential illustrated in Figures 1 A and 2 A was shown to have a diferent character from the CD potentials. This potential will be called the DK potential. The main feature of these potentials was that, besides the very large size of the earlier component, the late component, when obtained, was of proportionally very low amplitude or, in some recordings, was absent. A I)R potential first 1)egaii with a profound, positive deflection ; this was followed by a positive-negative phase with a very high amplitude; and then it devcloped n downward deflection. This first triphasic component of the DK potential was larger in size and higher in amplitude than those of the CD potential in both the cervical and lower thoracic I.V. levels ( P < 0.001 ) .

The segmental evoked potentials illustrated in Figures 1 R and 2 B

9

lable 1 . Results of segmental evoked potentials recorded from the lower thoracic intervertebral levels'.

Latencies of potential components (msec)

Onset of first slow component slow component

End of first Type of First positive dip Negative peak potential or peak of first of first fast

fast component component

CD n = 12 11.1 t 0.2 0.7 'f 14.0 t 0.3 1.0 18.1 k 0.3 1.0 36.0 t 0.5 1.7 10-12 (range) 11-16 15-20 33-39

DR n = 14 14.2 f 0.3 1.1 17.9 f 0.4 1.5 26.8 zk 0.9 3.3 48.7 f 2.7 9.4 12-16 14-21 21-35 35-65

PH n = 7 12.2 f 0.4 1.0 13.2 f 0.2 0.5 14.5 -+ 0.2 0.5 26.2 t 0.8 2.0 11-14 12-1 4 14-1 5 25-31

Duration of potential components (msec)

First slow component First fast component

CD n = 12 7.0 t 0.2 0.7 17.9 f 0.5 1.7 5-8 16-20

DR n = 14 15.1 t 0.9 3.3 9-23

PH n = 7 3.7 f 0.2 0.5 3-5

22.2 I+ 1.8 6.3 14-32

11.8 f 0.8 2.0 10-17

Amplitudes of potential components (microvolts)

First fast component First slow component

Ascending limb Descending limb Ascending limb

CD n = 12 21.8 t 2.0 7.1 39.5 * 5.3 18.5 40.0 k 5.9 19.9 12-32 12-72 16-80

DR n = 14 80.5 f 12.0 33.7 55.5 t 8.7 22.4 21.9 k 2.4 9.1 28-192 24-130 10-40

PH n = 7 16.0 f 4.5 11.9 9.7 f 1.4 3.6 25.1 k 4.9 12.7 6-40 6-1 6 12-48

Obtained by the supramaximal stimulation of the posterior tibia1 nerve at the popliteal fossa. n = number of potentials recorded. Mean f SEM Range

f Standard deviation

Table 2. Results o f segmental evoked potentials recorded from the lower cervical intervertebral levels’.

Latencies of potential components (msec)

End of the first Type of First positive dip Negative peak Onset of first Potential? o r peak of first of first fast

fast component component slow component slow component

CD n = 23 6.5 f 0.1 0.5 8.3 f 0.2 0.9 12.6 f 0.4 1.9 29.2 f 0.8 3.8 5-8 6-10 9-1 6 23-40

DR n = 9 7.7 f 0.3 0.9 10.2 f 0.3 0.9 19.7 f 1.2 3.6 43.2 f 4.1 9.0 6-9 9-1 2 15-25 31)-55

Durations of potential components (msec)

First slow component First fast component

CD n = 23 6.1 f 0.3 1.4 16.3 f 0.5 2.3 3-9 13-20

DR n = 9 13.7 k 1.2 3.6 9-19

23.4 * 2.9 6.4 13-30

~~~ ~

Amplitudes of potential components (microvolts)

First fast component First slow component Ascending limb Descending limb Ascending limb

CD n = 23 36.2 f 4.2 20.4 46.8 zk 2.5 25.3 35.3 & 4.3 20.3 12-88 16-104 10-80

DR n = 9 230.2 ?c 43.4 130.2 143.5 f 26.6 80.0 52.2 f 12.7 35.5 56-440 28-260 12-120

* Obtained by the supramaximal stimulation of the median or ulnar nerves at the level of the elbow. n = number of potentials recorded. Mean f SEM Standard deviation Range

included in the Table. f PH potentials were recorded on three occasions and therefore they could not be

were the least encountered potentials in the whole study. They will be called the PI1 potentials. The first and earlier component of a PH potential was smaller in amplitude and shorter in duration than those of CD and DR potentials. This smaller triphasic wave always started with a positivity, and the second positivity often occurred on the

Y 3

Fig

ure

3. E

voke

d po

tent

ials

fro

m t

he l

ower

cer

vica

l sp

inal

cor

d ob

tain

ed b

g su

pram

axim

al s

tim

ulat

ion

of

anal

ogou

s ne

rves

of

uppe

r an

d lo

wer

lim

bs i

n th

ree

diff

eren

t no

rmal

sub

ject

s (A

, B,

C).

The

num

bers

abo

ve t

he t

race

s de

note

the

evo

ked

pote

ntia

ls e

lici

ted

bg

the

stim

ulat

ion

of

ipsi

late

ral

(I)

and

cont

rala

tera

l (2

) m

edia

n ne

rves

at

the

elbo

w;

and

the

stim

ulat

ion

of t

he i

psil

ater

al

(3)

and

cont

rala

tera

l (4

) po

ster

ior

tibi

a1 n

erve

s at

the

pop

lite

al f

ossa

. T

he d

iagr

am

on

the

left

sho

ws

the

poss

ible

si

te

of

the

tip

of

the

reco

rdin

g in

trat

heca

l el

ectr

ode

in r

elat

ion

to t

he s

pina

l co

rd,

deno

ted

bg

A,

B an

d C

. C

alib

rati

on:

20

mse

c an

d 0.

1 m

V

(Sam

e am

plif

icat

ions

are

use

d in

col

umns

3 a

nd 4

).

12

rising of the late component. The late component, unlike the earlier one, was always considerably higher in amplitude and longer in duration. As can be observed from the Figures, this slow wave was frequently followed by a second positive-negative wave with smaller amp1 i t ude .

The weak and brief intrathecal electrical stimulation delivered f roni the tip of the recording electrode, where this position accounted for the segmental CD potential, evoked a sensory and motor clinical phe- nomenon propagated to either both arms or both legs, according to the spinal segments stimulated. The threshold of the sensorial clinical phenomenon was often lower than that of motor one. On the other hancl, the intrathecal stimulation at the “DR position” of the intra- thecal electrode caused unilateral clinical phenomena in a segmental manner; and this was observcd on the same limb in which stimulation of its major nerve evokecl a DR potential. A transient but brisk and phasic pain, radiating to one extremity in the manner of radicular distribution, was sometimes reported during introduction of the intrathecal electrode. 111 the case of the “PH position”, the clinical phenomena produced by intrathecal stimulation were in the form of variable segmental responses propagated either unilaterally or bi- laterally.

In all cases, the features of the evoked potentials obtained by the stimulations of the homologous nerves from four limbs were systern- atically compared, and Figure 3 illustrates such a comparison in dif- ferent subjects. DR and PH potentials were found to be unilateral segmental responses, while the CD potentials were always recorded bilaterally. In other ords, when the intrathecal stimulation delivered from the tip of the recording electrode produced a unilateral clinical segmental response, the DR and PH potentials were evoked by stimula- tion of the same limb. And the contralateral homologous nerve stimulation in a corresponding site caused only a small triphasic response with a different shape. h’evertheless, there was a difference between the DR and PH potentials from two points. Onc point was that the size difference between the potentials produced by the stiniulation of opposite homologous nerves was greater in DR than the PH potentials. The second point was a difference in their cervical intrathecal recordings ; when the cervical segmental potential was of D R type, there was n o recordable response from the stimulation of either posterior tibia1 nerves. On the other hand, the stimulation of the posterior tibia1 nerve ipsilateral to the a rm nerve producing the PH potential evoked a triphasic action potential with a clear shape and size and with a delayed latency. W e called this response the “TRACTUS

13

Figure 4. The ef fect o f the position o f the intrathecal electrode f o r the evoked cord potentials recorded f r o m the cervical 6-7 I.V. level in a normal subject. ( A ) denotes the lateral position of the intrathecal electrode and ( B ) the position near the midline. Stimulation of the median nerves at the elbow (median) and the posterior tibial nerves at the popliteal fossa (Post. Tib.) both on the right ( R ) and the le f t (L ) sides. Calibration: 25 msec and 80 FV (Same amplifications are used in lower

three traces in A and upper three traces in B columns).

POTENTIAL” (see Figures 3 A and 3 B) . CD potentials were obtained bilaterally in almo’st the same shapes and comparable sizes following peripheral stimuli from both homologous segmental nerves seperately. And in the case of cervical intrathecal recording, it was easily possible to evoke “TRACTUS POTENTIALS” on stimulation of the posterior tibial nerve in either leg (see Figure 3 C ) . Same spatial relationship, as mentioned above, has been proved by different electrode positions

14

Figure 5 . The effect o f the position of intrathecal electrode f o r the cord euoked potentials recorded f r o m the ceruical 5-6 I . V . level in another subject. Same arrangement of illustration was used us in Figure 4 . Calibration: 15 msec and 0.1 mV (Same amplifications are used in all except the uppermost trace in A column).

being used within the same intrathecal recording level in some subjects. Figures 4 and 5 illustrate the cases in whom such comparison was made. As can be seen from the Figures, only the ipsilateral peripheral afferent volley elicited from the median nerves could be recorded as DK potentials. In this position, a very small triphasic response was evoked by contralateral median nerve stimulation, and there was no recordable response from either posterior tibia1 nerve when stimulated (Figures 4 A, 5 A ) . When the position of the intrathecal electrode was changed to the midline in these cases, both segmental peripheral

15

inputs elicited by homologous median nerves were recorded as CD potentials, and their sizes were nearly equal (Figures 4 B and 5 B, upper two potentials). After stimulation of the posterior tibial nerves from both legs, it was possible to record Tractus potentials. In the majority of subjects, Tractus potentials could only be elicited uni- laterally. In these cases, a Tractus response could only be recorded by stimulation of the leg ipsilateral to the arm in which median nerve stimulation evoked the segmental CD response with higher amplitude of the first component. The contralateral posterior tibial nerve stimulation did not show Tractus potential in these cases (Figure 4 R ) , third and fourth traces). Nevertheless, in five cases, the stimulation of both posterior tibial nerves caused the appearance of Tractus potentials bilaterally (Figure 5 B, third and fourth traces).

The findings obtained from such an investigation support the conclusion that the segmental CD and DR potentials are different in origin. In this respect, there was further evidence that both potentials had significant differences in their shape and size, as already men- tioned. Both the amplitude and duration of the first component of the CD potential were found to be smaller than those of the DR potential ( P < 0.001). The same results could be obtained from different record- ing sites at the same level in the same subjects.

When the tip of the recording electrode came closer to the back of the midline of the spinal cord, the late component of the segmental evoked potentials became more prominent, as in the case of CD potential; when the tip of electrode was directed lateral from the midline, the late component became less significant and its amplitude diminished, as in the case of DR potential. The amplitude of the late component of the CD potential was often found to be almost the same in size as the homologous segmental afferent input on either side.

DISCUSSION

The segmental spinal evoked potentials obtained from the intrathecal electrode by the supramaximal stimulation of mixed peripheral nerves seem to be strictly determined by the location of the tip of the intra- thecal electrode, together with the anatomical condition of the spinal roots and spinal cord in the vertebral canal. When the tip of the recording electrode was near the midline (i.e. behind the cord dorsum), the segmental input from the periphery showed itself as a CD potential. Therefore, the origin of the first component of the CD potential was accepted as being the cord dorsum. Several reasons can be given for this conclusion. First of all, this is the only anatomical

16

position that explains how the bilateral clinical phenomena radiated to both extremities when the brief and very weak electrical shocks were delivered from the tip of the intrathecal electrode. In accordance with this, symmetrical and homologous afferent inputs, given to both extremities, gave rise to evoked potentials which were comparable and more or less of the same sizes and shapes. Furthermore, in lower cervical recording around the midline, some patients felt transient paraesthesiae-like sensation on one leg during the introduction of the cervical electrode ; and the afferent volley elicited from this leg always caused the compound action potentials obtained from the cervical intrathecal electrode (Er tek in 1975) . This indicates that the cervical recording electrode is anatomically close to the posterior median sulcus and behind the cord dorsum, because in the above-mentioned conditions it was the only way that the activities of the ipsilateral and posteriorly situated long afferent tracti could be recorded. To support this point, the Tractus potentials could be obtained by stimulation of the leg nerve on both sides in five cases. In fact, these cases indicate that the tip of the cervical recording electrode is very close to or at the midline, completely behind the cord dorsuin; and that there are almost equal distances between the tip of the electrode and both posterior halves of the spinal cord. The relationship between the amplitude of the first component of the cervical CD potentials and the size of the Tractus potentials is also in accordance with the same spatial relation- ship between the electrode and the posteriorly situated ipsilateral afferent tractus ; i.e. the posterior funiculus or, more specifically, “Fasciculus gracilis”.

Segmental DR potentials, on the other hand, indicate that the tip of the electrode is situated laterally within the vertebral canal. In other words, segmental DR potentials are related to the activities of the spinal roots. A very weak intrathecal stimulation in this position has always given clinical phenomena in the ipsilateral limb where the nerve volley coming from this extremity has always been large in size; while the afferent volley from the opposite homologous nerve recorded was a very small triphasic potential. These facts show that the asymmetrical and lateral position of the tip of the electrode greatly influenced the recordable reflections of the segmental inputs within the volume conduction medium. Hence the volley passing from the ipsi- lateral spinal roots was recorded as very large compound action potentials; while the contralateral volley coming to the spinal cord was a very small one, because of the longer distance between the recording site and the source of the electrical generation point. Further evidence to support the conclusions that the DR potential is related to

17

the volleys of the spinal roots, and that the tip of the electrode position is near to them but fa r from the cord dorsum, is the fact that, with the DR recording electrode positioned at the cervical level, it was never possible to record the afferent ascending volley coming from the leg nerves.

Segmental CD potentials are very similar to the cord dorsum potentials obtained from animal studies (Gasser & Graham 1933 and others), and this was also so in intrathecal recordings in human studies (Magladery e t al. 1951, Ertekin 1973). In animal studies, recording from the cord dorsum of the spinal cord with the stimuli given either to dorsal roots or to the peripheral nerves showed that the first potential is a triphasic spike, which is universally accepted as the activity of the intramedullary nerve fibers situated within the cord dorsum. And the positive peak of the initial spike response shows the arrival of the afferent volley in the cord (Gasser & Graham 1933, Campbell 1945, Lloyd & McZntyre 1949, Bernhard 1953, Gelfan & Tarlov 1955, Austin & McCouch 1955). This initial spike is followed by slow negative and even slower positive -. The late, slow potentials were thought to be related to the physiological events within the spinal segments investigated. Although the various kinds of the afferent inputs may alter the size and shape of the late, slow waves (Bernhard 1953), the negative intermediary cord potentials contain two components, named N-1 and N-2. The general opinion is that they represent the synaptic and/or postsynaptic activities created by the afferent segmental input within the gray matter of the posterior horn (Gasser & Graham 1933, Bernhard 1953). The latest and slowest positive potential or P-wave is usually considered as the electrographic counterpart of the negative dorsal root potentials, which show the primary afferent depolarization, and therefore the presynaptic in- hibition (Barron h Matthews 1938, Lloyd & Mclntyre 1949, Eccles 1964, Wall 1964). Although the initial spike and the negative inter- mediary slow potentials in mammalians are comparable to the human CD potentials, a very distinct P-wave similar to the animal studies could not be recorded in our study. Some later positivity could be encountered in a few records, but their duration and amplitude did not fit the P-wave.

If the electrode position was changed and recorded in two different sites at the same level in the same subject, the latency of the positive dip of CD potential was always shorter than the latency of the deep positive peak of the DR potential in the same segmental spinal level. This observation also indicates that the CD potential is not the volume conduction of the DR potential, and that the first component of the CD

2 ACTA NEUROL. SCAND. 53,1

18

potential is mainly composed of the activity of the myelinated fast central cord dorsum fibers situated longitudinally; while the segmental DR potential is the mixture of the recordable total activity passing through the spinal roots. Thus the DR potential is composed of a very dense volley of both large and medium size afferent nerve fibers of the dorsal roots, and probably of an antidromic volley from the anterior roots. This fact is also supported by the size and shape characteristics of the DR potentials.

The segmental PH potential also had different features, and there- fore i t appeared that somewhat different neural mechanisms are involved in its production. It can firstly be proposed that the PH potential is the recording obtained at the site in between the electrode positions for CD and DR potentials. This potential was always obtained by the unilateral segmental input, and the potentials recorded by the homologous contralateral nerve stimulation had not the same shape and size as that elicited by ipsilateral stimulation. Thus, according to the volume conduction rule, the position of the tip for the P H potential should be the intermediary position in between the DR and CD positions. This type of potential has been reported in cats by different authors (Gasser & Graham 1933, Bernhard 1953, Austin & McCouch 1955, Gelfan & Tarlou 1956). Our suggestion as to the origin of PH potentials is that they could well be related to the electrode tip position facing the posterior horn in a slightly lower spinal segment than the segment where the maximal afferent volley was coming into the spinal cord investigated. According to Gasser & Graham (1933), the height of the intramedullary spike falls rapidly to a small fraction of its initial value below the activated root, and i t is difficult to follow the wave further down to 6 to 7 mm; the evoked responses similar to PH potentials were clearly shown just in the beginning or vicinity of the posterior horn recordings in the tracking methods in frogs, cats and monkeys (Campbell 1945, Fernander de Molina & Gray 1957).

The late, slow components of all segmental evoked potentials of any type are related to the segmental spinal pre- and/or postsynaptic activities of the gray matter of the spinal cord. If the proportional amplitude difference between the first and late components is taken into consideration, one would observe that the amplitude of the slow component will become greater as the tip of the electrode is moved closer to the spinal gray matter. Thus the difference is most remark- able in the DR potential, while the size of the late component is most remarkable in the PH potential.

19

ACKNOWLEDGEMENTS

I am grateful to Dr. Fikret Turan for his help throughout the study. I also thank Mrs Lynne Kayan and Dr. Altan Kayan for reading the manuscript and for their invaluable criticisms.

REFERENCES

Austin, G. M. & G. P. McCouch (1955) : Presynaptic component of intermediary cord potentials. J. Neurophysiol. 28, 441-451.

Barron, D. H. & B. H. C. Matthews (1938) : The interpretation of potential changes in the spinal cord. J. Physiol. (Lond.) 92, 276-321.

Bernhard, C. G. (1953) : The spinal cord potentials in leads from the cord dorsum in relation to peripheral source of afferent stimulation. Acta physiol. scand. 29, Suppl. 106, 1-29.

Campbell, B. (1945): The distribution of potential fields within the spinal cord. Anat. Rec. 91, 77-88.

Cracco, R. Q. (1973) : Spinal evoked response: peripheral nerve stimulation in man. Electroenceph. clin. Neurophysiol. 35, 379-386.

Declaration of Helsinki (1964) : Recommendations guiding doctors in clinical research. Wld. med. J. 11, 281.

Eccles, J. C. (1964): Presynaptic inhibition in the spinal cord. Progress in Brain Research, Vol. 12: Physiology of Spinal Neurons, ed. J. C. Eccles & J. P. Schade, pp. 65-91. Elsevier, Amsterdam/London/New York.

Ertekin, C. (1973) : Human evoked electrospinogram. New Developments in Electro- myography and Clinical Neurophysiology, Vol. 2, ed. J. E. Desmedt, pp. 344- 351. Karger, Basel.

Ertekin, C. (1975): Studies on the human evoked electrospinogram. 11. The con- duction velocity along the dorsal funiculus. Acta neurol. scand. In press.

Fernandez de Molina, A. & J. A. B. Gray (1957) : Activity in the dorsal spinal gray matter after stimulation of cutaneous nerves. J. Physiol. (Lond.) 137, 126-140.

Gasser, H. S. & H. T. Graham (1933): Potentials produced in the spinal cord by stimulation of dorsal roots. Amer. J. Physiol. 189, 303-320.

Gelfan, S. & Tarlov, I. M. (1955): Differential vulnerability of spinal cord structures to anoxia. J. Neurophysiol. 18, 170-188.

Gelfan, S. & I. &I. Tarlov (1956): Physiology of spinal cord, nerve root and peripheral nerve compression. Amer. J. Physiol. 185, 217-229.

Lloyd, D. P. C. & A. K. McIntyre (1949) : On the origins of dorsal root potentials. J. gen. Physiol. 32, 409-443.

Magladery, J. W., W. E. Porter, A. M. Park & R. D. Teasdall (1951): Electro- physiological studies of nerve and reflex activity in normal man. IV. The two neurone reflex and identification of certain action potentials from spinal roots and cord. Bull. Johns Hopkins Hosp. 88, 499-519.

Shimoji, K., K. Tatsuhiko, H. Higashi, T. Morioka & E. 0. Henschel (1972) : Evoked spinal electrograms recorded from epidural space in man, J. appl. Physiol. 33, 468-471.

Wall, P. D. (1964) : Presynaptic control of impulses a t the first central synapse in the cutaneous pathway. Progress in Brain Research Vol. 1 2 : Physiology of

2'

Spinal Neurons, ed. J. E. Eccles & J. P. Schade, pp. 92-115. Elsevier, Amster- dam/London/New York.

Received April 11, 1975 Cumhur Ertekin, M.D. Division of Neurology Medical School Hospital Aegean University, Bornova Izmir, Turkey