Vestibulospinal monitoring in experimental spinal

5 downloads 0 Views 1MB Size Report
Vestibulospinal tract function was monitored in experimental contusion of the spinal cord in cats, and ..... corticospinal or rubrospinal tracts were involved,.
J Neurosurg 52:64-72, 1980

Vestibulospinal monitoring in experimental spinal trauma WISE YOUNG, M.D., PH.D., JOHN TOMASULA, PH.D., VINCENT DECRESCITO, PH.D., EUGENE S. FLAMM, M.D., AND JOSEPH RANSOHOFF, M.D.

Department of Neurosurgery, New York University Medical Center, New York, New York Vestibulospinal tract function was monitored in experimental contusion of the spinal cord in cats, and compared with somatosensory cortical evoked potentials. Both white and gray matter portions of the vestibular and somatosensory pathways were evaluated in cord injuries at T-7 and L-4. Severe contusions of 20 gm-20 cm force impact resulted in a rapid (less than 1 second) abolition of thoracic white matter conductivity, but a somewhat slower (4 to 5 minutes) loss of lumbar gray matter responses. A paradoxical transient recovery of white matter conductivity occurred 1 to 2 hours after injury, despite eventual progression to central hemorrhagic necrosis at the contusion site. In contrast, mild contusions (20 gm-10 cm force impact) produced only a temporary loss of neuronal activity: white matter for 1 to 2 hours, and gray matter for 30 to 40 minutes. In general, vestibular and somatosensory potentials showed similar sensitivity to contusion, although the former tended to recover earlier. We conclude that contusion injury causes two types of neuronal dysfunction in spinal cord: 1) a low-threshold concussion-related loss of activity lasting 30 to 120 minutes; and 2) a higher threshold necrotic process, requiring 1 to 2 hours to develop, which apparently spreads from gray to white matter.

KEY WORDS 9 gray matter

S

9

vestibulospinal tract 9 somatosensory evoked potentials 9 neuronal dysfunction 9 spinal cord injury

OMATOSENSORY evoked potentials (SEP) have been used extensively to monitor spinal cord function in experimental spinal injury, e ~,11,~,27,29al The significance of SEP changes in spinal cord injury has been questioned in several respects. First, although the continued absence of SEP for more than 3 hours after injury is a grave prognostic sign and portends permanent paraplegia in 90% of animals, e'sa7 transient losses of SEP for 1 to 2 hours are often reversible. Second, SEP may return as long as 24 hours after severe spinal wounding in monkeys without any correlation with sensory or motor recovery? 1 Third, SEP are believed to be primarily conducted through the dorsal columns 3,~,1~a8,29and the dorsolateral fasciculus?,* However, Wall, et al., 2e,Sa's4 and others 32 have graphically shown that isolated lesions of the dorsal cord seldom result in significant clinical sensory loss. Thus, the correlation of SEP loss with clinical dysfunction may be coincidental to damage of more clinically significant spinal tracts. It was, therefore, of importance to examine the function of another spinal pathway, the vestibulospinal tract (VST), and compare it with SEP in 64

9 white matter

reversible and irreversible lesions of the spinal cord. As an experimental monitoring technique, using the VST presents several distinct advantages: 1) The tract is easily and selectively activated by stimulation of the vestibular nerve; 2'1s-1e,22's52) The responses are carried in a well defined pathway situated in the ventral funiculus) '2s a part of the spinal cord not evaluated by SEP; and 3) Signal averaging is not required to record vestibulospinal responses in the spinal cord or in muscles. Consequently, the vestibulospinal responses can be tested within milliseconds of a spinal injury, whereas SEP usually require at least 1 minute of signal averaging. In this paper, we will describe the vestibulospinal evoked potential (VEP) and SEP responses before and after two types of injury to the spinal cord. The first injury, from dropping a 20-gin weight 20 cm (20 gm20 cm) onto the exposed spinal dura, results in permanent paraplegia in 90% of cats. 8,s,27The second, from a 20-gin weight dropped 10 cm (20 gm-10 cm), produces irreversible paraplegia in only 10% of the cats. 6,8 The experiments are divided into two groups. In one, the VEP and SEP were recorded before and after injury J. Neurosurg. / Volume 52 / January, 1980

V e s t i b u l o s p i n a l m o n i t o r i n g in e x p e r i m e n t a l spinal t r a u m a directed to the T-7 segment. This paradigm was meant to examine white matter conductivity in the thoracic cord. In the other, the VEP responses were recorded intraspinally from the ventral gray matter, and the peripherally evoked responses from the dorsal gray matter of the cord segment at L-4. The L-4 segment was then contused with 20 gm-20 cm and 20 gm-10 cm forces to ascertain the effect of injury on the evoked gray matter activity. Table 1 summarizes these paradigms. The purpose of the experiments was to compare the effect of impact injury on these two pathways in both gray and white matter.

TABLE 1

Summary of experiments* Experiment

No. of Cats

Site

I II III IV

3 3 3 3

T-7 T-7 L-4 L-4

Weight & Distance 20 20 20 20

gm-20 gm-10 gm-20 gm-10

cm cm cm cm

Monitoring Paradigm V& S V& S V, S, V, V, S, xT, S

*Monitoring was carried out before and after impaction. V = vestibulospinal evoked responses; S = somatosensory evoked potentials; x7 = vestibulospinal evoked field potentials; = peripherally evoked segmental field potentials.

Materials and Methods

For the first group of experiments, six cats were anesthetized with 25 mg/kg doses of pentobarbital injected intraperitoneally. They were then mounted on an apparatus for spinal procedures. Electrodes were introduced epidurally over the left somatosensory cortex. The right sciatic nerve was stimulated at 2/sec and twice twitch threshold, and 256 somatosensory responses evoked by the sciatic nerve stimulation were averaged on a Nicolet single-channel averager after being amplified • 20,000 and filtered (bandpass 150 to 3000 Hz). For the VEP, a hole was drilled into the right vestibule to expose the eighth nerve. Bipolar silver electrodes were placed in close proximity to the nerve and cemented into place. Single 0.1 msec pulses, twice threshold (10 to 20 V) were used for typical vestibulospinal muscular activation. Constant voltage stimulus isolation units were used. The responses were recorded with bipolar needle electrodes inserted into the ipsilateral triceps, rectus abdominis, quadriceps, and occasionally the contralateral external oblique muscle. A laminectomy was then performed over the T-7 vertebral segment, widely exposing the spinal cord for impaction. In the second group of experiments, six cats were initially anesthetized with 35 mg of methohexital sodium, a short-acting barbiturate, given intramuscularly. In each animal the cord was transected superior to the colliculi by surgical removal of the cerebral cortices and thalamus during carotid artery occlusion. The cavity was then packed with cotton and Gelfoam. Respiration in these animals was spontaneous and care was taken that no prolonged period of hypotension occurred during the decerebration. All the animals had blood pressure monitored with a femoral arterial line and intravenous replacement of fluids during the experiments. Blood gases were checked regularly to ensure that pO2 was greater than 100 mm Hg, pCO2 was approximately 20 to 30 mm Hg, and pH was between 7.3 and 7.4. These animals typically showed good spinal cord responses for more than 8 hours after decerebration. The vestibulospinal evoked triceps response was checked throughout the experiment. Since the triceps are innervated by spinal segments proximal to the T-7 and L-4 injury sites, its J. Neurosurg. / Volume 52 / January, 1980

presence suggests unimpaired cervical cord and brainstem function. A lumbar laminectomy was then carried out at L2--4. The intraspinal recordings were made at the L-4 cord segment, identified by peripherally evoked segmental field potentials. Glass micropipettes, filled with 2M NaCI solution and beveled to 5 MegOhm resistance, were used forextracellular field potential monitoring. The dura was incised to expose a small area of the lumbar cord for electrode insertion. For depth profiles, the electrodes were advanced with a micromanipulator to a depth of 2 to 3 mm, and then withdrawn in 100-~t steps during vestibular and sciatic nerve stimulation. The intracellular recordings were carried out using electrodes filled with 2M KCI solution and beveled to 20 MegOhm resistance. The spinal cords were contused with a 20-gm weight dropped on a small bronze striker plate that spanned the width of the thoracic cord. A polyethylene tube was used to direct the weight and to standardize the distances. Care was taken to make the laminectomies sufficiently wide so that the spinal cord sustained the full brunt of the impact. All the 20 gm-20 cm impacts to the T-7 area fulfilled the following criteria: 1) The animals had a massive burst of neuronal and muscle activity on impact; 2) There was a transient 30 to 40 mm Hg increase in arterial blood pressure, followed by a prolonged 50 to 60 mm Hg drop of pressure from a normal mean of 120 mm Hg; 3~3) On visual inspection, there was clear evidence of disrupted arachnoid vessels and subdural hemorrhage at the impact site; and 4) All the spinal cords were examined after the experiments for gross evidence of central hemorrhagic necrosis. The 20 gm-10 cm impacts at T-7 produced similar, but less prominent, manifestations of injury. Results

We will first summarize the features of normal VEP responses. Unilateral single-shock stimulation of the vestibular nerve, both in the intact and decerebrated cat, produced a consistent muscle activation pattern. Figure 1 illustrates some of the responses. The ster65

W. Young, et al.

Fro. 1. Axial musculature activated by vestibular stimulation (v). The responses of the following muscles, ipsi- and contralateral to the stimulus, are shown: sternomastoid (sm), trapezius (tpz), supraspinatus (ss), latissimus dorsi (ld), and external oblique (co). The spinal segments represented are indicated in parentheses. Bipolar needle electrodes were used to record the muscle responses evoked by single shocks delivered to the vestibular nerve. Note that the sternomastoid and external oblique activation occurs on the contralateral side (c) and the remainder on the ipsilateral side (i). ra = rectus abdominis; t = triceps.

Fro. 2. Comparison of vestibular responses in the ipsilateral triceps (top trace), and rectus abdominis (bottom trace). A: The longer latency of the rectus abdominis activation is shown. B: A short interstimulus period of 5 msec was used; the second response can be seen to fail. C: A double pulse was used to activate the vestibular nerve; the muscular response in both muscles followed appropriately.

66

nomastoid (C2-3) muscle was contralaterally activated with a latency of 2 msec. The ipsilateral trapezius (C3-4), supraspinatus (C5-6), latissimus dorsi (C6-8), and rectus abdominis (T8-12) were excited within 3 to 5 msec. The external oblique activation tended to occur on the contralateral side. Note that nearly every segment of the spinal cord from C-2 to C-8 and T-8 to T-12 is represented. In Fig. 2, the responses from the ipsilateral triceps and rectus abdominis are compared. Observe the short latency of the triceps response and the slightly longer latency of the rectus abdominis activation. When double-pulse stimuli were applied to the vestibular system, both responses can be seen to follow up to interstimulus intervals of 10 msec. At interstimulus intervals of 5 msec, the second response failed. These latencies and stimulus-following rates are consistent with a rapidly conducting disynaptic pathway. In the decerebrate animals, the vestibular responses were usually enhanced. In addition, it was possible to demonstrate the influence of head position on the patterns of vestibular activation. For example, in Fig. 3, the responses of ipsilateral triceps and biceps are shown. When the head was turned toward the stimulated side, the triceps response was enhanced and the biceps showed little or no early response. The biceps, however, did have a prominent late response which can be shown to arise from stretch receptor excitation of the biceps by the triceps contraction. When the head was turned away from the stimulated side,

J. Neurosurg. / Volume 52 / January, 1980

V e s t i b u l o s p i n a l m o n i t o r i n g in e x p e r i m e n t a l s p i n a l t r a u m a

Fio. 3. Influence of head position on vestibular evoked responses of the upper limb. The muscular activity of the ipsilateral triceps (t) and biceps (b) is shown. When the head was turned toward the left, left vestibular nerve activation resulted in enhancement of the triceps and suppression of the biceps. When the head was turned away from the left, there was enhancement of the biceps and suppression of the triceps.

there was a suppression of the triceps and the appearance of an early biceps activation. This pattern represents the influence of tonic neck reflexes on vestibulospinal evoked activity. The effect of vestibular stimulation on the lumbar cord was more subtle. In Fig. 4, the responses of the spinal cord to peripheral stimulation were recorded bipolarly from the dorsal cord surface and the ventral root at L-7. In Fig. 4A and B, activation of the anterior tibialis nerve, a flexor system, produced repetitive waves in the spinal cord and ventral roots. Conditioning the anterior tibialis stimulation with vestibular stimulation had no effect on the response. In contrast, vestibular stimulation influenced the responses evoked by gastrocnemius nerve stimulation, an extensor system. The responses to extensor stimulation, shown in Fig. 4C to E, consisted of an early reflex followed by a late discharge. A single shock to the vestibular nerve 20 msec preceding the gastrocnemius stimulation resulted in a suppression of both the early reflex and the late discharge. However, when a train of stimuli was applied to the vestibular nerve, the early response was enhanced, while the late discharge was suppressed. Intraspinal recordings of the gastrocnemius monosynaptic reflex revealed a similar depressive and enhancement effect by vestibular stimulation (Fig. 5). J. Neurosurg. / Volume 52 / January, 1980

FIG. 4. Vestibulospinal interaction with lower-limb reflexes. A: Anterior tibialis (a) stimulation results in a bursting discharge in both the cord dorsum (C) and ventral root (R) of L-7. B: Conditioning anterior tibialis stimulation (a) with vestibular stimulation (v) has no effect. C: Stimulation of the gastrocnemius (g) evokes an early monosynaptic response and a late discharge. D: Single vestibular shocks, preceded by 10 msec, suppressed both the early response and late discharge. E: A train of vestibular shocks (20/sec) apparently enhanced the early response while suppressing the late discharge. Note the increase in amplitude of the vestibular evoked spinal potential.

Figure 6 shows the field potentials recorded with a microelectrode inserted into the spinal cord approximately 500 u from the midline. The vestibular evoked field potentials were positive close to the dorsal surface, reversing to a negativity at 2 mm deep in the ventral horn. The latency of the response is consistent with vestibulospinal activation. In order to confirm that the negativity at 2 mm represents neuronal activation, intracellular recordings were made at these depths. An example of such a recording is shown on the bottom right trace. The majority of units responding to the vestibular stimulation were situated in the ventromedial gray matter and were not motoneurons, as tested by the absence of antidromic activation. The peripherally evoked field potentials, in contrast, were maximally negative in the dorsal horn, reversing to a 67

W. Y o u n g , e t al.

FIG. 5. Effect of vestibular stimulation on the monosynaptic reflex. The L-4 monosynaptic response to gastrocnemius nerve stimulation is shown. A: The effect of a train of vestibular stimulation. M = monosynaptic response; P = polysynaptic response. B: Single shocks to the vestibular nerve can be seen to suppress the late polysynaptic component of the response. C: The normal response.

positivity in the ventral horn. This distribution of field potentials is consistent with the monosynaptic activation of dorsal horn neurons by primary afferents. Both 20 gm-20 cm and 20 gm-10 cm impactions of the T-7 cord resulted in rapid loss of SEP and VEP, the latter within a second of impact. The recovery patterns, however, were quite different for the two types of injuries. Table 2 shows the pattern of loss and recovery of SEP and VEP. In Group I cats (with T-7 20 gm-20 cm injuries), there was a transient recovery of VEP and SEP at about 60 and 120 minutes, respectively; however, both disappeared by the 3rd hour for the remainder of the experiment. Figure 7 illustrates the sequence of VEP loss in one typical experiment. In Group II cats (with T-7 20 gm-10 cm injuries), both SEP and VEP recovered without the secondary deterioration seen with the more severe impaction. Figure 8 shows an example of this. Note that the triceps response to vestibular stimulation was not abolished by the T-7 injury; in fact, some facilitation of the response often occurred. To speak of the transient recurrence of both VEP and SEP in the 20 gm-20 cm injuries as normal recoveries would be misleading. The waveforms of the 68

FIG. 6. Field potential profiles evoked by vestibular nerve (v) and sciatic nerve (s) stimulation at L-4. The left column represents the vestibular responses. The right column represents the segmental reflex responses. The depths of the recordings are indicated in the middle. The bottom right trace is the intracellularly recorded response of a typical neuron at 2-mm depth to vestibular nerve stimulation; five superimposed traces were photographed.

responses were usually abnormal. For example, the transiently recurring VEP was often smaller in amplitude and tended to fail frequently. The SEP, as a rule, was smaller in amplitude and had increased latencies. Figure 9 illustrates one particular case. Note how the first positive wave of the SEP was diminished in size and shifted by 3 to 5 msec. The pattern of gray matter dysfunction, as studied with field potentials in the L-4 segment, was quite different from the T-7 white matter disturbances. As shown in Table 2, neither the vestibular response ('v') nor the peripherally evoked segmental response (~) were immediately abolished. Furthermore, in Group III cats (with 20 gm-20 cm impact), no consistent pattern of recovery was seen. In Group IV cats (with 20 gm-10 cm injuries), both the "v" and ~ recovered at an earlier time than did "r and ~ in Group II cats (with T-7 20 gm-10 cm impacts). On several occasions, seizure-like activity was recorded with the microelectrodes in the first few minutes after the impact.

Discussion Stimulation of the vestibular nerve in the cat results in distinct patterns of muscular activation (Figs. 1, 2, J. Neurosurg. / Volume 52 / January, 1980

V e s t i b u l o s p i n a l m o n i t o r i n g in e x p e r i m e n t a l spinal t r a u m a TABLE 2

Change in spinal cord parameters before and after spinal cord injury* Minutes Postinjury

Experiment

Cat No.

Control

I(20gm-20 em at T-7)

3 2

SV SV

1

SV

3 2 1

SV SV SV

----

III (20 gin-20 cm atE-4)

3 2 1

$9 $9 $9

-V $9 5V

-9 iV

IV (20 gin-10 cm at L-4)

3 2

i9 ~;9

i9 S9 iV

S9 sV

II(20gm-10 cm at T-7)

1

iV

1

2

5

10

Hours Postinjury 30

---

-V

-V

-V

--

-9 SV S9

1

2

3

-V -V

SV SV

-V

SV

-V

-V -V SV

SV SV SV

SV SV SV

--

-v

-v

-V S9 g9

S9 i9 S9

59 SV S9

4

5

SV SV SV

-V SV SV

lv i9 S9

i9 i9 i9

*V = vestibulospinal evoked responses; S = somatosensory evoked potentials; "v" = vestibulospinal evoked field potentials; and g = peripherally evoked segmental field potentials. - indicates the absence of a recognizable response. Control value was recorded prior to impact.

FIG. 7. Vestibulospinal trace inactivation during 20 gm20 cm (400 gm-cm) impacts to T-7. Both the rectus abdominis (A.R.) and the external oblique (E.O.) responses were abolished. Note the secondary deterioration of the vestibulospinal responses by the 2nd hour.

FIG. 8. Vestibulospinal tract inactivation during 20 gm10 cm (200 gm-cm) impacts to T-7. Left: The triceps responses were not abolished by the impact. Right: The quadriceps responses were lost within a second, but recovered rapidly.

and 3). What assurance is there that the pathway is indeed the vestibulospinal tract (VST)? First, the response had a remarkably short latency of 2 to 4 msec. This rapid conduction time virtually rules out any participation by the cerebellum or cerebral cortex in the response. It is consistent with a rapidly conducting disynaptic system, such as the VST. Second, the pattern of muscular activation corresponds with those reported in other laboratories, a4,~2involving primarily the trunk and proximal extensor limb muscles. If the corticospinal or rubrospinal tracts were involved, more distal limb and flexor musculature would be expected~ Finally, the enhancement of the vestibular nerve responses by decerebration further argues against corticospinal tract participation, and is consis-

tent with known facilitation of vestibulospinal reflexes by decerebration. 2~ However, it is well known that vestibular nerve stimulation excites the reticular formation and, consequently, the reticulospinal tract (RST)2 ,x4-~6 Gernandp4 had attributed some of the later components of the vestibular evoked radial nerve potentials to reticulospinal pathways. Llinas and Terzuolo24a5 have shown that reticulospinal fibers innervate many segments of the spinal cord, exerting largely an inhibitory influence on motor activity. This may, therefore, be the basis of the complex effects of vestibular nerve stimulation on the lower limb reflexes (see Figs. 4 and 5). It is likely that the responses evoked by vestibular nerve stimulation may be, in

J. Neurosurg. / Volume 52 / January, 1980

69

W. Young, e t al.

lh

I

3h

I I

S P

I

lOms

!

FIG. 9. Alterations in the somatosensory evoked potentials (SEP) waveform during transient recovery after impact injury. The SEP obtained from one experiment with 20 gm20 cm impactions to T-7 are shown. There was an immediate loss of the SEP but a subsequent recurrence of a delayed, but definitely recognizable, SEP response. By 2 to 3 hours, this response disappeared.

part, mediated and influenced by reticulospinal activation. Nevertheless, this is not necessarily a disadvantage for vestibulospinal monitoring since both the RST and VST are anatomically closely related. The effect of contusion on VEP was not remarkably different from SEP. Both the 20 gm-20 cm and 20 gm10 cm impacts result in rapid abolishment of the VEP and SEP (see Table 2). The VEP afforded an opportunity to examine the immediate effects of impact, since it could be tested within milliseconds of injury. The VEP was abolished almost as soon as it could be tested. Moreover, VEP allowed an assessment of segmental excitability at different levels of the cord. An apparent increase in excitability occurred shortly after impact in the cord proximal to injury, that is, the triceps response to vestibular stimulation (see Fig. 8). In general, however, the VEP tended to recover earlier and showed more resistance to injury than did SEP. This may be due to the more ventral location of the VST in relationship to the dorsal impact trauma. Perhaps the most interesting finding was the occurrence of a transient recovery of both VEP and SEP in the 20 gm-20 cm impacts, despite eventual progression to central hemorrhagic necrosis. This recovery period of 30 to 60 minutes invariably occurred between 1 and 2 hours after impact. In contrast, the 20 70

gm-10 cm impacts resulted in complete recovery after an initial conduction block. This pattern of functional loss suggests that two separate but possibly interrelated mechanisms of injury may be involved: the first, a concussion-induced conduction block lasting 1 to 2 hours; the second, the triggering of a secondary necrotic process that occurred only in severe injuries and required 2 to 3 hours to develop. It is perhaps not coincidental that the time course of central hemorrhagic necrosis in spinal cord injury 1,~ and its encroachment on white matter ~~ parallels the timing of the secondary loss. The absence of the secondary loss in the 20 gm-10 cm impacts implies that the threshold levels of the two types of dysfunction are different. Although great care was taken to implement the 20 gm-20 cm impact injuries properly, in the manner documented in our laboratory, 8,:8 to produce gross central hemorrhagic necrosis, the possibility remains that in this relatively small sample set the data may have been biased by insufficient injury to the cord. Several arguments can be made against this possibility. First, the 20 gin-20 cm impacts resulted in eventual loss of neuronal function, whereas the 20 gm10 cm impacts did not. Second, the animals were all examined after the experiments to verify the presence or absence of symmetrical central hemorrhagic necrosis. Third, the secondary loss of both SEP and VEP at 2 to 3 hours cannot be attributed to general deterioration of the animal since the triceps response to vestibular stimulation remained intact throughout the experiments. The transient recovery of white matter conductivity in the 20 gin-20 cm injuries was unexpected. Despite many studies of SEP changes during spinal trauma, 6-1~'~7,29 such recovery patterns have not been emphasized. On closer perusal of the literature, however, some examples of transient recoveries were found. For example, D'Angelo 7 observed transient SEP and H-reflex recovery with approximately the same time course in cervical cord trauma. Ducker, et al., 9,~2 found that SEP sometimes persisted after impact injury, albeit with bizarre waveforms. We suspect that the occurrence of the transient recovery is critically dependent on the degree of injury since it requires that the initial concussion-produced conduction block must recover within an hour before the secondary necrotic processes can take over. From the experiments involving injury to L-4 gray matter, a different pattern of neuronal dysfunction emerged. Neither the peripherally evoked dorsal horn potentials (S) nor the vestibular ventral horn potentials ('v') were immediately abolished by 20 gm-10 cm or 20 gm-20 cm impactions. Both the V and S continued to be present for 4 to 5 minutes after impact, in contrast to the immediate compromise of SEP and VEP in the T-7 injuries. With the exception of a brief and isolated appearance of a ~ response in one animal, no transient recovery of neuronal responses J. Neurosurg. / Volume 52 / January, 1980

Vestibulospinal monitoring in experimental spinal trauma was observed. As in the T-7 impactions, the dorsal cord responses in the L-4 injuries were abolished earlier than the ventral cord responses. Histological studies of experimental spinal cord injuries in this a8 and other laboratories 1,9,~~176 have clearly shown that spinal cord necrosis invariably begins in the gray matter and spreads outwards into the surrounding white matter. Shortly after the impact, the morphology of the cord appears remarkably normal with only a few areas of extravasation of blood. Within 30 minutes, scattered evidence of dead and dying neurons appears in the gray matter. By 1 hour, there is usually frank necrosis of gray matter. Progression of edema into the white matter occurs at 1 to 2 hours. The data obtained in this study roughly correspond to this sequence of events. Gray matter field potentials are lost within 5 minutes of impact. White matter conductivity is blocked immediately, but recovers after 30 to 60 minutes, only to succumb to some secondary process that requires 1 to 2 hours to occur. The mechanisms of these processes are unclear. This study raised some additional questions. First, the delayed loss of V and $, compared to the immediate effect of T-7 impact on SEP and VEP, is puzzling. The presynaptic fibers which activate the field potential responses presumably lie within the impact area, and yet their function is not as rapidly compromised by impact as in the thoracic cord. Second, what is the relationship of the sequence of neuronal dysfunction with blood flow changes? We have obtained evidence that the secondary loss of both VEP and SEP occur simultaneously with a decrease in blood flow in the contused tissue? 6 Third, is the white matter dysfunction a necessary concomitant of the gray matter necrosis? The need for more experimentation defining the nature of the neuronal response to injury is obvious. References 1. Allen AR: Remarks on the histopathological changes in the spinal cord due to impact. An experimental study. J Nerv Ment Dis 41:141-147, 1914 2. Andersson S, Gernandt BE: Ventral root discharge in response to vestibular and proprioceptive stimulation. J Neuraphysiol 19:524-543, 1956 3. Andersson SA, Norrsell K, Norrsell U: Spinal pathways projecting to the cerebral first somatosensory area in the monkey. J Physial 225:589-597, 1972 4. Bennett MH, McCallum JE, Stasiak TS: The effect of dorsal spinal cord infarction and the somatosensory evoked response in cats. Prac Sac Neurosci 696:1077, 1975 5. Brodal A, Pompeiano O, Walberg F: The Vestibular Nuclei and Their Connections; Anatomy and Functional Correlations. Edinburgh/London: Oliver and Boyd, 1962, 193 pp 6. Campbell JB, DeCrescito V, Tomasula J J, et al: Bioelectric prediction of permanent post-traumatic J. Neurosurg. / Volume 52 / January, 1980

7. 8.

9.

10. 11.

12.

13. 14.

15. 16. 17. 18. 19.

20. 21.

22. 23. 24.

paraplegia. Presented at the Annual Meeting of American Association of Neurological Surgeons, Houston, Texas, April, 1971 D'Angelo CM: The H-reflex in experimental spinal cord trauma. J Neurasurg 39:209-213, 1973 Donaghy RMP, Numoto M: Prognostic significance of sensory evoked potential in spinal cord injury. Proceedings af the 17th Spinal Card Injury Conference, 1969, pp 251-257 Ducker TB: Experimental injury of the spinal cord, in Vinken P J, Bruyn GW (eds): Handbaak of Clinical Neuralagy. New York: American Elsevier, 1976, Vol 25, pp 9-26 Ducker TB, Kindt GW, Kempe LG: Pathological findings in acute experimental spinal cord trauma. J Neurasurg 35:700-708, 1971 Ducker TB, Salcman M, Lucas JT, et al: Experimental spinal cord trauma, II: Blood flow, tissue oxygen, evoked potentials in both paretic and plegic monkeys. Surg Neural 10:64-70, 1978 Ducker TB, Salcman M, Perot PL Jr, et al: Experimental spinal cord trauma, I: Correlation of blood flow, tissue oxygen, and neurologic status in the dog. Surg Neural 10:60-63, 1978 Erulkar SD, Sprague JM, Whitsel BL, et al: Organization of the vestibular projection to the spinal cord of the cat. J Neuraphysial 29:626-664, 1966 Gernandt BE: Vestibulo-spinal mechanisms, in Kornhuber HH (ed): Handbook of Sensory Physiology, Volume VI/I. The Vestibular System, Part 1: Basic Mechanisms. Berlin/Heidelberg/New York: SpringerVerlag, 1974, pp 541-564 Gernandt BE, Gilman S: Descending vestibular activity and its modulation by proprioceptive cerebellar and reticular influences. Exp Neural 1:274-304, 1959 Gernandt BE, Iranyi M, Livingston RB: Vestibular influences on spinal mechanisms. Exp Neural 1:248-273, 1959 Giblin DR: Somatosensory evoked potentials in healthy subjects and in patients with lesions of the nervous system. Ann NY Acad Sci 112:93-142, 1964 Goodkin R, Campbell JB: Sequential pathological changes in spinal cord injury: a preliminary report. Surg Forum 20:430-432, 1969 Goodman JH, Bingham WG Jr, Hunt WE: Ultrastructural blood-brain barrier alterations and edema formation in acute spinal cord trauma. Neurasurg 44:418--424, 1976 Green BA, Wagner FC Jr: Evolution of edema in the acutely injured spinal cord: a fluorescence microscopic study. Surg Neural 1:98-101, 1973 Griffiths IR, Miller R: Vascular permeability to protein and vasogenic oedema in experimental concussive injuries to the canine spinal cord. J Neural Sci 22: 291-304, 1974 Grillner S, Hongo T: Vestibulospinal effects on motoneurones and interneurones in the lumbrosacral cord. Prag Brain Res 37:243-262, 1972 Halliday AM, Wakefield GS: Cerebral evoked potentials in patients with dissociated sensory loss. J Neural Neurasurg Psychiatry 26:211-219, 1963 Llinas R, Terzuolo CA: Mechanisms of supraspinal actions upon spinal cord activities. Reticular inhibitory mechanisms on alpha-extensor motoneurons. J Neuraphysial 27:579-591, 1964 71

W. Young, et al. 25. Llinas R, Terzuolo CA: Mechanisms of supraspinal actions upon spinal cord activities. Reticular inhibitory mechanisms upon flexor motoneurons. J Neurophysiol 28:413--422, 1965 26. Noordenbos W, Wall PD: Diverse sensory functions with an almost totally divided spinal cord. A case of spinal cord transection with preservation of one anterolateral quadrant. Pain 2:185-195, 1976 27. Numoto M, Flanagan ME, Wallman E, et al: Prognostic significance of the somatosensory evoked potential and H reflex in spinal cord injury. Proceedings of the 18th Spinal Cord Injury Conference, 1971, pp 227-230 28. Nyberg-Hansen R: Functional organization of descending supraspinal fibre systems to the spinal cord. Anatomical and physiological correlations. Ergeb Anat Entwicklungsgesch 39:3-48, 1966 29. Perot PL Jr: The clinical use of somatosensory evoked potentials in spinal cord injury. Clin Neurosurg 20:367-381, 1973 30. Rawe SE, Perot PL Jr: Pressor response resulting from experimental contusion injury to the spinal cord. J Neurosurg 50:58-63, 1979

72

31. Singer JM, Russell GV, Coe JE: Changes in evoked potentials after experimental cervical spinal cord injury in the monkey. Exp Neurol 29:449--461, 1970 32. Vierck CJ Jr: Comparison of forelimb and hindlimb motor deficits following dorsal column section in monkeys. Brain Res 146:279-294, 1978 33. Wall PD: The sensory and motor role of impulses travelling in the dorsal columns towards cerebral cortex. Brain 93:505-524, 1970 34. Wall PD, Dubner R: Somatosensory pathways. Annu Rev Physiol 34:315-336, 1972 35. Wilson V J: Physiological pathways through the vestibular nuclei. Int Rev Neurobiol 15:27-81, 1972 36. Young W, Tomasula J, DeCrescito V, et al: Blood flow changes and neuronal dysfunction in spinal cord injury. (In preparation)

Address reprint requests to: Wise Young, M.D., Ph.D., Department of Neurosurgery, New York University Medical Center, 550 First Avenue, New York, New York 10016.

J. Neurosurg. / Volume 52 / January, 1980