Effects of a dorsal column lesion on somatosensory evoked potentials in primates

MISSING IMAGE

Material Information

Title:
Effects of a dorsal column lesion on somatosensory evoked potentials in primates
Physical Description:
vii, 171 leaves : ill. ; 29 cm.
Language:
English
Creator:
Makous, James Carl, 1963-
Publication Date:

Subjects

Subjects / Keywords:
Evoked Potentials, Somatosensory   ( mesh )
Spinal Cord -- physiology   ( mesh )
Cerebral Cortex -- physiology   ( mesh )
Spinal Cord Injuries -- physiopathology   ( mesh )
Neural Pathways   ( mesh )
Department of Neuroscience thesis Ph.D   ( mesh )
Dissertations, Academic -- College of Medicine -- Department of Neuroscience -- UF   ( mesh )
Genre:
bibliography   ( marcgt )
non-fiction   ( marcgt )

Notes

Thesis:
Thesis (Ph.D.)--University of Florida, 1993.
Bibliography:
Bibliography: leaves 164-170.
Statement of Responsibility:
by James Carl Makous.
General Note:
Typescript.
General Note:
Vita.

Record Information

Source Institution:
University of Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 002328727
oclc - 54535021
notis - ALT2371
System ID:
AA00011792:00001

Full Text











EFFECTS OF A DORSAL COLUMN LESION ON SOMATOSENSORY EVOKED
POTENTIALS IN PRIMATES











By

JAMES C. MAKOUS


A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN
PARTIAL FULFILLMENT OF THE REQUIREMENTS
FOR THE DEGREE OF DOCTOR OF PHILOSOPHY


UNIVERSITY OF FLORIDA


1993















ACKNOWLEDGMENTS


I would like to thank my advisor, Charles Vierck, for

taking me into his laboratory at such a late stage in my

graduate education and for his patience and support of my

sometimes unorthodox ideas and experiments. I also thank

the members of my ever-changing committee, David Green,

Bruce Hunter, Richard Johnson, Christiana Leonard, Louis

Ritz and Barry Whitsel, for their critical evaluation and

support of my research. I thank William Luttge for his many

hours spent in making this a better department, John

Middlebrooks for his support during the early stages of my

training and the Center for Neurobiological Sciences for

financial support.

I would also like to thank Jean Kaufman for teaching me

to handle and care for monkeys; Anwarul Azam for keeping the

computers on-line; Carol Martin-Elkins for teaching me basic

histology. I thank Laura Kasper, James Murphy, Anita Puente

and Karl Vierck for technical expertise. I also thank

Robert Friedman, Diana Glendinning and Douglas Swanson for

their support, advice and camaraderie during all the stages

of my education. I am especially indebted to Robert

Friedman for his generosity in sharing his time, knowledge,















laboratory space and equipment all of which were in short

supply (save for his knowledge). I also thank Babbette

Botchin and Dan Thiel for expert veterinary care.

Last, I would like to thank my family; my father whose

advice both personal and professional has never steered me

wrong; and my wife, Elizabeth, who has been an endless

source of love, support and inspiration during these years

of training.


iii















TABLE OF CONTENTS


Pg

ACKNOWLEDGEMENTS........................................

ABSTRACT.................................................

CHAPTERS


age

ii

V


1 GENERAL INTRODUCTION......................... 1

Behavior.......................................... 2
Anatomy and Physiology.......................... 3
Plasticity.................................... 10
Evoked Potentials ............................. 12
Overview of Dissertation..................... 13

2 RECORDINGS FROM THE WHITE MATTER TRACTS
OF THE SPINAL CORD ............................ 15

Methods......................................... 16
Results ....................................... 22
Discussion.................... ...................* 38

3 RECORDINGS FROM THE CEREBRAL CORTEX............ 42

Methods....................................... 44
Results......................................... 54
Discussion.................... ................ 112

4 PHYSIOLOGICAL CHANGES DURING RECOVERY FROM
A DORSAL COLUMN LESION....................... 119

Methods.............. ......................... 120
Results....................................... 123
Discussion............ .................. ...... 150

5 GENERAL DISCUSSION............................. 155

Neural Mechanisms.............................. 155
Conclusions............................. ....... 163

REFERENCES .............................................. 164

BIOGRAPHICAL SKETCH ...................................... 171

iv















Abstract of Dissertation Presented to the Graduate School of
the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy



EFFECTS OF A DORSAL COLUMN LESION ON SOMATOSENSORY EVOKED
POTENTIALS IN PRIMATES


By

James Carl Makous


May 1993



Chairperson: Charles J. Vierck, Jr.
Major Department: Neuroscience


A dorsal column (DC) lesion has a significant lasting

effect on behavioral tasks that require temporal processing

of tactile information (i.e. frequency and duration

discrimination). These experiments describe physiological

correlates of the behavioral deficits in temporal

discrimination observed in primates following DC lesions.

In experiment 1, compound action potentials were

recorded from the major ascending white matter tracts of the

cord. These experiments determined the extent to which

different sensory pathways of the spinal cord responded to

high frequency stimulation. At 10 pulses per second (pps)















pathways in the lateral spinal columns were suppressed to

70-80% of their control values, whereas the DCs were not

suppressed at the same frequency.

In experiment 2, epidural evoked potentials were

recorded from implanted electrodes before and after a DC

lesion. In response to mechanical stimulation, the cortical

evoked potential at 10 pps could not be distinguished from

the background activity (noise) following the DC lesion.

The response to electrocutaneous stimulation showed a

frequency-dependent suppression at 10 pps, after the lesion,

that approximated a 20% reduction in amplitude.

In experiment 3, the cortical evoked potential was

monitored over weeks following the lesion. Under some

conditions, there was a significant increase in the

amplitude of a late (90 ms) peak following the DC lesion.

This increase was correlated with behavioral recovery on a

grasping task. There was no significant recovery for the

early peaks (20 and 50 ms).

In summary, a DC lesion caused an overall decrease in

amplitude of the cortical evoked potential. There was a

significantly greater frequency-dependent suppression of

responses to electrocutaneous stimulation at 10 pps

following the DC lesion. The reduction could be explained in















part by the extent to which the spared spinal pathways

followed 10 pps. There also was an increase in the

amplitude of the 90 ms peak during the weeks following the

lesion that was correlated with a behavioral recovery in

grasping, suggesting a cortically mediated phenomenon.


vii















CHAPTER 1

GENERAL INTRODUCTION



Since the early 1900s investigators have attempted to

use clinical findings in patients with compromised dorsal

columns (DC) to infer the role of this pathway in

perception. These observations suggested that a lesion of

the DCs affects the ability of a patient to detect and

discriminate among a variety of tactile stimuli (see Nathan

et al., 1986). However, for most of these patients, the

lesion involved other spinal pathways as well. In other

instances, most or all of the deficits appeared to recover

over time. Nevertheless, there are deficits that are common

to most accounts of DC interruption, and these are related

to processes that required temporal processing of

somesthetic stimuli. For example, tactile direction

sensitivity or the ability to identify letters written on

the hand (graphesthesia) can be affected by such a lesion.

Graphesthesia requires the patient to integrate the stimulus

over a few seconds. It also has been reported that a

repeated tapping stimulus or an object placed in the

patient's hand will fade over time. However, there are

differing accounts with respect to absolute threshold shifts









2

in response to light tactile stimuli and with respect to

recovery from these deficits.



Behavior

Given the ambiguities in the clinical findings,

investigators have attempted to quantify sensory capacities

in animal models, where a lesion can be isolated to an

individual pathway and confirmed histologically. Vierck

(1977), using Von Frey hair stimulation, showed that there

was no threshold shift on the sole of the foot following a

mid-thoracic DC lesion. For a variety of other tasks,

considerable recovery of function has been observed

following a DC lesion. DeVito and colleagues (1964) found

that monkeys had only a transient deficit in weight

discrimination, and Vierck (1966) found no permanent effect

on limb position sense. Levitt and Schwartzman (1966) showed

a transient deficit in two-point discrimination;

discrimination of size and location of tactile stimuli also

recover (Vierck, 1973; Vierck et al., 1983, 1988). For many

years there was no firm evidence of any permanent sensory

deficit following interruption of the DCs. In fact, Wall

(1970, p. 518) was so bold as to say that the DCs are only

'involved in controlling the analysis of messages arriving

over the other somatosensory pathways.'

More recently, tasks that require some degree of

temporal resolution have shown long lasting sensory deficits









3

in monkeys. Vierck (1974) showed that tactile direction

discrimination was hampered for up to 1 year following a DC

lesion, whereas simple detection of movement across the skin

showed only a transient deficit. Later, Vierck et al.

(1985) showed that monkeys trained to discriminate a 10 Hz

tactile stimulus from a 14 Hz stimulus could not

discriminate 10 Hz from 35 Hz following interruption of the

DC pathway, even after extensive retraining (1 yr). Animals

trained on the same task with lesions of the dorsal-lateral

(DL) and/or antero-lateral (AL) columns showed no such

lasting deficit. In a more recent study (Vierck et al.,

1990), normal monkeys could discriminate between stimulus

trains of 3 pulses (at 10 Hz; 200 ms total duration) from

trains of 5 to 7 pulses (400-600 ms duration); however,

following interruption of the DC pathway, the animals could

not discriminate trains of 3 pulses from trains of as many

as 35 pulses. Thus, it appears that tasks requiring

temporal processing of tactile input, at least in a range of

frequencies from 10-35 Hz, depend uniquely upon input from

the DC.



Anatomy and Physiology

When the DCs are compromised, somatosensory information

reaching the cortex must be carried by the remaining

pathways. The only other spinal pathways that are thought

to contribute significantly to perception are long









4

somatosensory pathways (long tracts) in the dorsal-lateral

(DL), and antero-lateral (AL) columns (Vierck, 1984). Long

tracts in the DL column ascend the spinal cord ipsilateral

to the primary afferent fibers (Rustioni et al., 1979),

whereas the AL tract fibers cross the midline near the root

entry level and ascend the spinal cord in the

antero(ventral)-lateral quadrant of the spinal cord,

contralateral to the primary afferent fibers (Trevino and

Carstens, 1975).

The majority of the fibers constituting the DL and AL

pathways have their cell bodies in the dorsal horn, where

there are a number of modulatory influences. For example,

Jimenez et al. (1987) have shown that there are descending

fibers originating in the brainstem reticular formation and

red nucleus that have presynaptic inputs onto cutaneous

primary afferents entering the dorsal horn. Furthermore,

cutaneous information is carried up the cord to the

reticular formation (Salibi et al., 1980) and the red

nucleus (Berkley, 1986) via the DC nuclei. Therefore, a DC

lesion could result in a change in the regulation of

information entering the dorsal horn, owing to a loss of DC

connections with the reticular formation and red nucleus.

There is also evidence that primary afferents synapsing

in the dorsal horn are directly inhibited by cutaneous

stimulation (Wall, 1958; Janig, 1968). For example,

following cutaneous stimulation, there is an inhibition of









5

primary afferents in the dorsal horn lasting greater that

100 ms; however, primary afferents in the DCs show an

inhibition lasting only 25-35 ms (Wall, 1958). Therefore,

since normal neuronal excitability in the dorsal horn can

have a refractory period of more than 100 ms, it is possible

that changes in perception following interruption of the DC

results from a restriction of temporal processing of

cutaneous information in pathways emanating from the spinal

grey matter.

If pathways other than the DCs are to code for touch,

it is important that the pathways arising from the dorsal

horn carry information about non-noxious, natural cutaneous

stimuli as opposed to information strictly from muscle

afferents. Indeed, a number of studies have shown that the

AL and DL tract fibers respond to natural stimulation of the

skin (Willis et al.,1974; 1975; Surmeier et al., 1988;

Downie et al., 1988) and that they contribute to cortical

responses (Andersson et al., 1975). However, few studies

have looked at frequency coding in the AL tract; the stimuli

used in these studies usually involved a static displacement

of the skin or hair.

The DL tract has been studied much less extensively in

the primate than in the cat, and the two species appear to

differ substantially in the size of the DL pathway

(Nijensohn and Kerr, 1975). Thus, the information carried

in this pathway in primates is not well characterized.









6

However, it has been demonstrated in acute monkey

preparations that an SI cortical evoked potential is smaller

in amplitude but still intact (Andersson et al., 1972)

following interruption of the ipsilateral (to the stimulus)

DCs and the contralateral AL tract, sparing only the

ipsilateral DL tract. This result was obtained in response

to light mechanical stimulation of the skin. Although it is

still not clear the extent to which the DL and AL pathways

carry temporal information, it is apparent that both

pathways carry information that is cutaneous in nature.

The DC, DL and AL tracts synapse in the thalamus before

tactile information is transmitted to the cortex, where it

contributes to normal perception. All three long tracts

have converging inputs onto the ventro-basal complex of the

thalamus (Berkley, 1980; 1983), providing sites for

interaction among the various pathways. Indeed, a number of

studies have demonstrated the existence of excitatory and

inhibitory interaction in the ventro-basal complex of the

thalamus (Anderson et al., 1964; Janig et al., 1979; Salt,

1989). These interactions can last from 50 to 500 ms,

depending upon the cells, and are thought to shape the

spatio-temporal profile of responses that are sent to the

cortex. Thus, removal of one of the inputs to this region

(e.g., a DC lesion) could lead to an imbalance between the

remaining excitatory and inhibitory inputs. This could

change neuronal integration (filtering) in the thalamus,









7

which, in turn, could result in a change in perception.

A number of studies have attempted to ascertain the

role of the various spinal cord pathways in generating a

cortical response through evoked potential measurements. In

general the results have been inconclusive, owing to the

variety of stimuli used, the lack of repeated measures

within individual animals, and the different kinds of

anesthesia.

Gardner and Morin (1953; 1957) showed in pentobarbital

anesthetized monkeys that neither chronic nor acute DC

section altered the latency or threshold of somatosensory

cortical evoked potentials. In this study, they used

electrical stimulation directly on dissected cutaneous and

deep nerves. It was not clear how they determined threshold

or the amplitude of stimulation. It is also possible that

direct stimulation of the nerve recruited predominantly the

large diameter fibers originating in muscle (York, 1985).

Eidelberg and Woodbury (1972) also found no change in

the amplitude or latency of the cortical evoked potential

following chronic (>4 weeks) or acute DC or dorsal quadrant

lesions in monkeys. They stimulated the radial nerve

transcutaneously at 1-3 mA under chloralose anesthesia.

Unfortunately, these results were compared to unoperated

controls rather than preoperative controls from the same

animal.

On the other hand, Andersson et al. (1972) recorded









8

evoked potentials from acute Nembutal anesthetized monkeys

and found that there was an increased latency of 4-5 ms and

a decrease in amplitude of about 25% following DC section.

There was no change in the location of maximum activity from

intradermal electrocutaneous stimulation of the foot or the

hand. In the conditions when a mechanical stimulus was

used, the animal had already received a lesion of either the

ventral quadrant or dorsal quadrant before recordings were

made. This made it difficult to ascertain the contribution

of the DCs to the mechanical response.

A more recent study by Cusick et al.(1979) in monkeys

showed, under ketamine anesthesia, that acute DC section

almost completely eliminated the cortical evoked response to

exposed sciatic nerve stimulation. Following a 4-7 week

period after the lesion, there was a small increase in the

amplitude to about 25% of the preoperative control

amplitude. On the other hand, following a lesion of the DL

and AL pathways, sparing the DCs, there was no appreciable

attenuation of the evoked potential. Thus, the conclusion

from these studies is that the DCs provide sufficient input

to support a cortical evoked potential, but intact DCs are

not necessary for the presence of an evoked potential.

A few investigators have also looked at unit responses

following interruption of the DC pathway or dorsal

hemisection. Andersson et al. (1975) found, in lightly

anesthetized pentobarbitall) monkeys, that acute dorsal









9

hemisection of the cord decreased the number of units per

cortical penetration that could be driven by peripheral

stimulation. There was also a high number of units that

were spontaneously active and were driven unreliably. The

latencies of units did not appear to be longer than control

animals, but units were more often found on the hairy than

on the glabrous skin. Although most of the units tended to

be rapidly adapting, they typically could not follow

repetitive stimulation beyond 4 Hz.

These results are in general agreement with the results

of Dreyer et al. (1974), though they differ in the details.

Dreyer et al. found, in awake monkeys with chronic (21-72

days) DC lesions, that fewer numbers of units were driven by

peripheral stimulation, and the responses were less

reliable. In contrast to Andersson et al. (1975), they

found a high proportion of neurons that responded to deep

stimulation rather than light stimulation. There was a

cortical region, corresponding primarily to the center of

the hand and foot representations in areas 3b and 1, that

was nearly devoid of responses following DC lesion, whereas

the surrounding cortical areas (3a and 2) appeared more

normal in the nature and proportion of responses.

Until now the afferents in the fasciculus gracilis (FG)

from the foot have been presented as if they were

distributed in a pattern similar to that of the fasciculus

cuneatus (FC) from the hand. However, Whitsel et al. (1972)









10

found, in recordings from fibers at cervical levels, that

fibers in the FC innervated both cutaneous and deep

structures, whereas the fibers in the FG were predominantly

cutaneous. This is in contrast to recordings made more

caudally, where the fiber types were similar for both the FC

and the FG. This suggests that there is sorting of fibers

from the FG as it ascends the spinal cord, resulting in a

projection of proprioceptive afferents to the dorsal horn

and then the DL tract (Dreyer et al., 1974). Therefore, it

is possible that a lesion of the FG would produce behavioral

effects that differ from those produced by a lesion of the

FC.


Plasticity

A number of recent findings suggest there is some

degree of plasticity in the adult nervous system (see Wall,

1988 for review). For example, the results of Merzenich et

al. (1984) have shown in recordings from anesthetized

monkeys that the cortical representation of the fingers

adjacent to an amputated finger included the region of

cortex formerly dominated by input from the missing finger.

These changes in the borders of the cortical representation

were on the order of 1-2 mm and appear to represent an

unmasking of inputs that are normally suppressed by the

anesthesia. More recently, Pons et al. (1991) found that

long term (> 12 yrs) dorsal rhizotomy (C2-T4; C = cervical;









11

T = thoracic; L = lumbar) produced large (> 1 cm) changes in

the area and location of cortex represented by the face

relative to control animals. That is, the face

representation invaded the area normally occupied by the

hand. It is not clear where the actual reorganization took

place; however, given the large shift in the cortical map, a

cortical mechanism is unlikely. A more likely locus could

be the spinal cord or thalamus, where the representations of

adjacent body parts are in closer proximity.

Rausell et al. (1992) also looked at the thalamus of

the monkeys used in the Pons et al. (1991) study. They

found a down regulation of gamma-aminobutyric-acid (GABA)

receptors in the deafferented thalamus relative to controls.

They also found changes in calcium binding proteins that are

specific for either lemniscal or nociceptive pathways.

Thus, it appears that there were long term changes in the

thalamus that accompanied the reorganization of the map at

the cortical level.

Changes in spinal cord have also been observed

following deafferentation. For example, Goldberger and

Murray (1982) observed sprouting in the cord following

dorsal rhizotomy in cats. They found an increased

projection to nucleus gracilis and to the spinal laminae

that normally received projections from the spared root.

However, there was no sprouting into areas not normally

receiving projections from the spared root. There also was









12

a tendency for areas that are topographically organized to

not receive any sprouting fibers.

Thus, it appears that neural plasticity does exist in

the adult nervous system, and it is possible that changes

following injury to the brain or spinal cord could be taking

place at many possible points along the neuraxis. It is

likely, then, that a behavioral recovery following a lesion

could reflect changes in the neural connectivity in the

spared pathways.



Evoked Potentials

Evoked potentials are the summated neural response from

a group of neurons depolarizing and/or hyperpolarizing in

synchrony. The amplitude of the evoked potential is

proportional to stimulus amplitude under some stimulus

conditions (Beck and Rosner, 1968; Franzen and Offenloch,

1969; Kulics, 1982); however, this issue has been disputed

(Uttal and Cook, 1964; Schwartz, 1976). Evoked potentials

are often used as a diagnostic tool for the integrity of the

afferent pathways (Pillay, 1984). The first positive peak

in the somatosensory cortical evoked potential is thought to

arise from the depolarization of layer four (a current

sink), whereas the current source arises from the dendrites

at the surface of the cortex. The second peak (negative)

arises from a local spread of activity into immediately

adjacent cortex. The source is layer four and the sink is









13

the cortical surface where axons are contacting dendrites.

The second positive peak is derived from a more global

spread of information to association cortices. In this

condition, the SI cortex serves as a source and the other

cortical areas serve as sinks. This late positive potential

can be recorded over a large area of cortex (Arezzo et al.,

1981; Kulics and Cauller, 1986; Emerson and Pedley, 1990;

Allison et al., 1991).



Overview of Dissertation

The following chapters describe a series of experiments

that were designed to look at the effects of a DC lesion on

somatosensory evoked responses to repetitive stimulation.

These experiments describe physiological accompaniments of

the temporal discrimination deficits observed following such

a lesion. The second chapter looks at the extent to which

compound action potentials recorded in the three long spinal

tracts follow high frequency stimulation. It was expected

that such an experiment would provide evidence concerning

the origin of any frequency-dependent suppression that might

be obtained at the cortical level in the absence of the DC

pathway. The third chapter describes changes in the evoked

cortical response and its frequency-dependent suppression

following the DC lesion. The fourth chapter follows the

evoked potential as it changes over months following the

lesion. This experiment was designed to correlate changes









14

in the cortical evoked potential with behavioral changes

observed in animals recovering from DC lesions. The last

chapter discusses some possible mechanisms for the changes

in the cortical evoked potential that accompany DC

interruption.














CHAPTER 2

RECORDINGS FROM THE WHITE MATTER TRACTS OF THE SPINAL CORD



Following a DC lesion, the remaining pathways that

contribute to perception are the DL and AL tracts (Vierck,

1984). Where there is a deficit in an animal's behavior

following such a lesion, there must also be a deficiency in

the capacity of the spared pathways to carry the neural

information necessary for the animal to perform the task.

For example, interruption of the DC pathway leads to

deficits in frequency discrimination at frequencies above 10

pps and in duration discrimination for trains of pulses at

10 pps (Vierck et al., 1985; 1990). Thus, it is likely that

the spared pathways can not reliably carry these kinds of

information at frequencies of 10 pps or greater.

There is reason to believe that the spared pathways

emanating from the dorsal horn grey matter are subject to

modulatory influences. For example, cutaneous stimuli

suppress the responses to stimuli less than 100 ms later in

the dorsal horn (Wall, 1958). In addition, Jimenez et al.

(1987) have demonstrated presynaptic influences originating

in the brainstem and terminating on primary afferents

entering the dorsal horn. Thus, either one of these









16

mechanisms could regulate the extent to which spared

pathways contribute to perception.

It was this line of reasoning that led to the first set

of experiments. If the pathways emanating from the dorsal

horn were dramatically suppressed at 10 pps, then the dorsal

horn could be a neural locus of the behavioral deficit

observed following DC interruption. Therefore, experiments

were undertaken to record from the white matter tracts of

the cord in response to different stimulus frequencies, and

the extent to which each pathway was suppressed by

increasing frequencies was determined.



Methods



Animals

The subjects were three female macaques (1 Macaca

arctoides and 2 Macaca nemestrina) that weighed between 4.8

and 11.4 kg. They were housed in an outdoor communal

enclosure to maximize the animals' health and psychological

well being. The animals were maintained on a diet of Purina

monkey chow, fruit and water. Prior to the acute

experiments, these animals were used in two procedures: a

behavioral study in which they received a cortical lesion of

the arcuate gyrus, and an anatomical study in which they

received an injection of HPR into the lumbrical muscles of

each hand, one week prior to the acute experiment. Two









17

months prior to the acute experiment, one animal (monkey E)

had a surgical interruption of the dorsal column on the

right side between Til and T12 (see Histology). This

procedure was conducted under sterile surgical conditions by

Dr. C. J. Vierck, Jr. (see Chapter 3 Methods for details).

Preparation. Each animal was premedicated with

intramuscular (i.m.) ketamine HCL (100 mg) and atropine

sulfate (.54 mg), given intravenous (i.v.) fluids lactatedd

Ringers) and then intubated. Isoflurane anesthesia was

passed to the animal and maintained at a level of 1.5-2.5%

throughout the remainder of the experiment. Each animal's

core temperature was monitored with an esophageal

thermometer and was maintained between 37 and 38 degrees C

with a heating pad (K-pad) and warmed paraffin bags. The

back of the animal was then shaved and a rostral-caudal

incision was made over the low thoracic vertebrae. The

muscle and fascia were scraped away from the dorsum of 4 or

5 vertebrae, and 2 or 3 of the dorsal laminae were removed

with rongeurs. The underlying fat and connective tissue

were dissected away, exposing the dura mater. A rostral-

caudal slit was made in the dura, and it was reflected to

expose the pia. The cord was then bathed in warm (37

degrees C) mineral oil. The remaining two spinous processes

at the rostral and caudal extent of the exposure were then

clamped and suspended by a stereotaxic frame.

Stimulation and recording. Each foot was stimulated









18

electrocutaneously at the tibial nerve, posterior to the

medial malleolus at a current sufficient to elicit a small

twitch of the first toe (ca. 5 mA). The duration of each

pulse was 0.2 ms, and the frequency of stimulation was

varied systematically from 1.5 pulses per second (pps) to 50

pps. Stimulating electrodes consisted of silver/lead solder

molded into a small bead 4 mm in diameter. The anode and

cathode were separated by 2 cm (cathode proximal),

lubricated with sergi-lube and held in place with adhesive

pads and tape. The animal was grounded with a ring

electrode placed around each calf proximal to the

stimulating electrode.

Compound action potentials were recorded from the white

matter tracts of the spinal cord by lowering tungsten

electrodes into the cord with a X-Y-Z micro-positioner.

Tungsten electrodes were made by immersing a tungsten rod

(0.007 in. diameter) into a bath of saturated sodium nitrite

and passing an AC voltage (ca. 6 volts) between the tungsten

and a carbon rod, also in the bath. The tungsten was then

slowly removed from the bath, and, in this way, a blunt

point was etched onto the end of the electrode. The

sharpened end of the tungsten was then passed through the

back end of a glass micro-pipette. The tip of the pipette

was broken to allow about 1 cm of exposed tungsten. The

tungsten was held in place by dental wax, and the entire

exposed tip of the electrode was dipped in Insl-x









19

(insulation). The tip of the electrode was then briefly

touched against a hot soldering iron. This usually yielded

impedances in the range of 10 100 kohms (at 1000 Hz),

depending on the duration of contact with the hot iron.

Signals recorded in the spinal cord in response to

electrocutaneous stimulation were differentially amplified

(Bak, 10,000x) in reference to a muscle flap along the

margin of the surgical opening. The signals were digitized

(11 us, Vetter), and recorded on magnetic tape (VCR) for

off-line analysis. The signal was also filtered (3-300 Hz

pass-band) and averaged on-line (286-computer; RC

electronics software) to monitor the potential during each

recording session.

Procedure. The experimental procedure consisted of the

following. The electrode was lowered into each of the major

tracks of the cord after visualizing the dorsal root entry

zone and the midline under a dissecting microscope. For

dorsal column penetrations the electrode was positioned

sufficiently near the midline to evoke a large response to

ipsilateral stimulation of the foot without evoking a

response from the contralateral foot. Dorsal-lateral

penetrations were made slightly lateral to the dorsal root

entry zone and were shallow enough to record little or no

response from stimulation of the contralateral foot.

Antero-lateral penetrations were made slightly more lateral

and deeper than the DL penetrations, to maximize a response










to contralateral stimulation. These locations often

produced an ipsilateral response that was less than 40% the

amplitude of the contralateral response.

After an electrode was placed in one of the tracts by

one of the above functional definitions, a frequency series

was conducted. In the case of the DC and DL tracts the

ipsilateral foot was stimulated, whereas the contralateral

foot was stimulated for the AL recordings. These three

stimulus conditions will be referred to as stimulation of

the experimental foot, whereas stimulation of the non-

experimental foot will be referred to as stimulation of the

control foot. Each frequency series consisted of 100-150

sweeps by the averaging software over a sweep duration of 75

ms. Thus, for the higher frequencies, the animal received

more stimuli (i.e. at 1.5 Hz there was 1 stimulus/sweep and

at 50 Hz there were 3 stimuli/sweep). The stimulus sequence

was typically as follows: experimental foot (at 1.5, 2.5,

5.0, 7.5, 10, 25, 50 and then 1.5 Hz), control foot (1.5

Hz), and experimental foot (1.5 Hz).



Data analysis

The data were analyzed off-line. "Raw" evoked

potential records were read off magnetic tape, band-pass

filtered (3-300 Hz), digitized (300 us bin width) and

averaged by commercially available software (RC

electronics). The 75 ms long averages were converted into











ASCII files and saved on disk. ASCII files were imported

into a spreadsheet, where latencies and amplitudes could be

computed. Latency was computed from the peak of the

stimulus artifact to the peak of each distinct waveform.

The peak amplitude of the potential was computed by taking

the difference between the amplitude 6 ms after the stimulus

artifact and the amplitude at the mean latency of the peak.

The response amplitudes at each frequency were expressed as

a percentage of the amplitude of the response to stimulation

of the experimental foot at 1.5 Hz (referred to as the

standard amplitude).



Histology

At the end of each experiment the animals were given an

overdose of pentobarbital (195 mg i.v.). They were then

perfused through the heart, first with warm (37 degrees C)

phosphate buffer (7.4 pH) and heparin sulfate (1 units/L),

then with fixative (2.5% glutaraldehyde and 2.0%

paraformaldehyde). This was followed by a mixture of

fixative and 10% glycerol and then rinsed with 10% glycerol

in phosphate buffer. The spinal cord was removed from C2

down to LI; the dorsal roots were counted, and the cord was

blocked 2 levels rostral and caudal to the recording

exposure. In the animal with the DC lesion, the cord was

blocked rostral and caudal to the lesion as well. The cord

was cryo-protected in a 10% formalin and 30% sucrose (or









22

glycerol) solution for a minimum of 1 week. Transverse

sections were cut on a freezing microtome at 50 um, mounted

on slides and let dry for a minimum of 1 day. They were

then stained (cresyl violet) and cover-slipped (Permount).

The extent of the DC lesion was drawn, and the electrode

tracks were reconstructed.



Results



Histology

The locations of the recording electrodes were verified

histologically in all recordings. Each of the recording

sites for each animal is shown in Figure 2-1. The DC

recordings were typically near the midline and in the

ventral half of the DCs. In one animal (monkey K), the

recording site bordered the grey matter. The DL recordings

were usually confined to the DL quadrant but occasionally

bordered the ventral quadrant of the cord. The AL

recordings were usually ventral to, but occasionally

bordered, the DL quadrant. During the AL recordings it was

often difficult to eliminate the response from the control

foot, especially at 11 ms.

The DC lesion in monkey E was between Til and T12 and

is shown in Figure 2-2. There was complete interruption of

the right DC at this level. There was very little

involvement of the left DC or the grey matter and no

encroachment into the right DL.











Spinal Recordings

Table 2-1 shows the peak latencies obtained for each

animal in each tract, averaged across all frequencies. The

'n' column corresponds to the number of frequency series

that were conducted in that tract for that animal. Monkey E

received a right DC lesion more than 2 months prior to the

recordings. Note that the AL response had two peaks that

differed in latency by 3 ms in two of the three animals.

There was a tendency for potentials evoked in the post-

synaptic pathways to have a slightly longer latency (0.3 4

ms) than the DC pathway.

Representative averaged responses from each tract are

shown in Figures 2-3, 2-4 and 2-5. These were derived from

one animal (Monkey D) on three separate penetrations. The

response from the DC (Figure 2-3) had a latency of 9.6 ms

and showed little suppression with high frequency

stimulation. There was almost no DC response from the

control foot. The response from the DL tract (Figure 2-4)

had a latency around 10.1 ms, but showed a more pronounced

suppression with increasing frequencies. At 50 pps the

potential was little more than 50% of the standard response

to 1.5 pps. The response from the control foot was small.

The AL tract had the most complex response (Figure 2-5).

The pll response was suppressed to nearly 50% at 10 pps,

whereas the p14 response was not substantially suppressed

until 25 pps. However, the p14 was completely abolished at









Figure 2-1. Electrode recording locations from the
spinal cord white matter of three monkeys. All
recording sites are drawn on a standard mid-thoracic
(T8) section. Black ovals represent approximate
recording locations, while dashed lines connect
recordings sites within one electrode penetration.
Often, two frequency series were recorded at the
electrode location. Abbreviations as in text.





25






D L B ,o D L
64DL DL
DL

e#&I
AL


Monkey K


DC
4 DC DL
D LD
L\ ) I DL \
AL 0

AL


Monkey E


DC


DLDL
DOD
DLAL o A D
D AL A'
AAL
Moe D AL6

Monkey D

















0) C
.0 0 0
-r4 M

1 -4

o 0

o






-4
4 *) 0





0o


-H 0
o r.
0 0
M 4-)



0 4J














-4M 4-I (1)


*4*


1^4































w
>
Q>) CMl
C I-
0


-J1






















0







.r- E
t5 "







0 m




- D
I-i








E t-
>Co
a)


0**

O E-



L., C8
Ec 0
CU)
0-


a

00 M





c a)







mu- U)


C6


0 0
L6 1-


EI-t


cIq


0D co
5 );


co


- Jn
0 E,


00


00


co -o
0;05 -a0C
V- l IIv- v


Ci


00


QE oio


c
0
E


110
0


0) -W
>-'
(UCe









Figure 2-3. Compound action potentials recorded from
the DC of monkey D. (TOP) The responses to different
frequencies of ipsilateral stimulation of the tibial
nerve are shown and are expressed as a percentage of
the response to 1.5 pps. (BOTTOM) The response to
stimulation of the contralateral (control) foot at 1.5
pps expressed as percentage of the ipsilateral response
at 1.5 pps.

















Spinal cord recordings
dorsal column


pps


50 pps


10 15


control
foot


- -


1.5 pps


100



80



60



40


20-


- .


I I I I I I I I I I I I Il li
5 10 15 20
time (ms)


-* "^









Figure 2-4. Compound action potential recordings from
the DL tract of monkey D. (TOP) Responses to different
frequencies of ipsilateral stimulation of the tibial
nerve are shown and are expressed as a percentage of
the response to 1.5 pps. (BOTTOM) The response to
stimulation of the contralateral foot at 1.5 pps.


















Spinal cord recordings
dorsal-lateral


1.5 pps

5 pps-


pps


25 pps


50 pps.


10 15


/" \ /


/, -- j


100



80



60



40


-20


control
foot


I


5 10 15 20
time (ms)









Figure 2-5. Compound action potentials recorded from
the AL tract of monkey D. (TOP) Responses to different
frequencies of contralateral stimulation of the tibial
nerve are shown and are expressed as a percentage of
the response to 1.5 pps. (BOTTOM) The response to
stimulation of the ipsilateral foot at 1.5 pps.




















Spinal cord recordings
antero-lateral


p1 1


1.5 pps


pps


50 pps


", 10 pps


10 15


control
- foot


-1 a .


10 15 20
time (ms)


100


0
CL
4-0



C.
0
0


20-


! I

















r-4 4) W 0\P r (U
(0 Lt 0 in --4 r.
. (0-4 4J I4
naM n C
2~H 000
w w 0 0 0
. 1)W 0
4J r '
PE- 0 4-J 4-)

0 W a 0


C '0 E-4
O W 0G

+1 0
*"" + 14 *





0 04




" ~4 I W & .
0W 0 0

o0 0
"3 W CM 0
r (- H w1



0 0 0 -H

0 W 4 M M
rq 41 3 ,


0 w r. 3O
aO ) p to



04> 0 0r. r



0 o0 P 0 4j







Pi4 U 0 4J1 4J 4J> U













,- L






-O00


OCCx


/
/


NW I
/
/ 1


I --x


/
/


Q 1
Q_





-co
-to
-CM (I)


0
0

(sdd


I


g'I "lOJ1,uoo


%) pnl!;dLuv


I I


I









37

50 pps, whereas the pll was not. Moreover, the p14 appeared

to shift its latency at 25 pps, which suggests the existence

of yet a third peak that was unmasked at 25 pps. Both

animals with a p14 response exhibited this phenomenon. The

pll response from the control foot was 20% of the standard

response to stimulation of the experimental foot, which was

not uncommon for this electrode location.

Averaged (across animals) frequency-response functions

for each of the tracts are shown in Figure 2-6. The most

obvious result was that responses from the DL and AL columns

did not follow the 50 pps stimulus as well as the potentials

within the DC pathway. However, at 10 pps the late response

from the AL tract (pl4) was not significantly different from

the DC response. In fact, there was even a slight (but not

significant) augmentation of the p14 response at 2.5 pps.

Beyond 10 pps, however, there was a precipitous drop off in

the response at p14 to higher frequencies. The AL response

at pll showed a similar trend to the response from the DL

tract at the low frequencies, but beyond 10 pps the two

functions diverged. The DL response was never suppressed to

more than 70% of control at the highest frequencies, whereas

the pll response from the AL tract at 50 pps was down to 20%

of the standard amplitude.



Summary

In response to electrocutaneous stimulation of the









38

tibial nerve, the DC pathway showed a significantly (p<0.05)

smaller suppression than any other subdivision of the spinal

cord, at stimulus frequencies of greater than 10 pps. At 10

pps or less, the three tracts were suppressed by nearly

comparable amounts. At 10 pps, suppression of the p14

response from the AL tract was not significantly different

than the DC suppression. On the other hand, the DL and the

AL (pll) responses were significantly (p<0.05) more

attenuated at 10 pps than the DC response.



Discussion



The principal result of this study was the observation

that ascending pathways in the lateral columns, which

emanated from the dorsal horn, are not profoundly suppressed

by repetition rates up to 10 pps. Therefore, behavioral

deficits in detection of 10 pps stimulation after

interruption of the dorsal columns are unlikely to be

attributed to a suppression of input to supraspinal targets

of pathways in the DL and AL tracts. At 10 pps, where the

tracts respond at 70% of the standard response or greater,

it is possible that any one of the pathways could contribute

significantly to the perception of a 10 pps stimulus.



Latency

The relative latencies of responses within the three









39

tracts was as expected. The DC pathway had the shortest

latency to peak (10.3 ms), whereas the DL and AL had longer

latencies (10.7 and 10.8 ms, respectively). Since these

latter two pathways comprise post-synaptic populations of

fibers and do not include primary afferents (in contrast to

the DC), it follows that longer latency responses were

obtained from the lateral columns. The AL and DL responses

should be nearly 1 ms later than responses from the DC,

given the mandatory synapse (i.e. delay of 1 ms) in the

dorsal horn. It is possible that a combination of factors

contributed the 0.5 ms difference actually observed. For

one, the DC response from monkey E was about 1.5 ms longer

than the DC response from the other two animals. Since the

DCs also have a large number of post-synaptic cells arising

from the dorsal horn (Rustioni et al., 1979), it is possible

that this recording was made from a division of the DC with

a preponderance of post-synaptic fibers. Since there was a

DC lesion in monkey E on the contralateral side, responses

arising from the control foot were not a determining factor

for this electrode placement, as it could have been in the

other recordings. Another factor that could have

contributed to the small difference between the DC and AL

latencies was the large bin width of 0.3 ms.

In addition to peaks at 11 and 14 ms, there was a peak

at 15 ms that appeared only at 25 pps in the AL response,

but it was abolished at 50 pps. It is possible that this









40

peak was derived from a third group of fibers in the AL,

given its different latency and unique frequency

sensitivity.

One animal (monkey K) showed no appreciable pll

response in the AL tract. It is not clear why this

occurred, since the recording electrode in this animal did

not appear to be in a substantially different location than

the other two animals. It is possible that the high

electrode impedance (100 kohms) used in this animal drew

from a smaller population of fibers than the electrodes used

in the other two animals (15 and 65 kohms).



Frequency Response

The DC response showed less suppression than the other

two tracts at all frequencies above 10 pps. This result was

not unexpected since the DL and AL tracts do not contain

primary afferent fibers. Because each synaptic delay tends

to decrease the reliability of high frequency transmission,

the DC should be better equipped to carry high frequency

information up the cord. Conversely, one could argue that

the AL and DL tracts do carry high frequency information,

but it is lost in the averaged response, owing to a variable

latency of individual responses. If this were the case, one

would expect to see not only a decrease in the amplitude of

the response at high frequencies, but also a broadening of

the peak at high frequencies. This was not observed in

either the DL or the AL tracts.











Summary

These experiments demonstrated, through compound action

potentials recordings from large populations of spinal

fibers, that the extent of suppression for the different

pathways at 10 pps is similar. In the most extreme case,

the AL pll response was 70% of standard amplitude in

contrast to no suppression for the DC response (100% of

standard). Thus, it is likely that the pathways spared of a

DC lesion could reliably carry 10 pps stimulus up the cord.

However, at 50 pps there were large differences in the

frequency-dependent suppression of different pathways.

If any of the long spinal tracts were to be compromised

in some way (i.e. by a lesion), it is possible that, at the

cortex, there would be a decreased response amplitude,

because some percentage of the input to cortex was lost. It

is also possible, given the small difference (ca. 20%) in

slope between the DC and other pathways, that there would be

a concomitant change in slope of the frequency-response

function at the cortex following a DC lesion. If these were

indeed the outcomes of cortical recordings, then the animal

in a behavioral task might require a larger stimulus for

detection and could reveal an alteration in threshold with

stimulation frequency.














CHAPTER 3

RECORDINGS FROM THE CEREBRAL CORTEX



Following a DC lesion, the information arriving at the

cortex is limited by the extent to which the spared pathways

(DL and AL) carry somatosensory information and by the

responsivity of partially deafferented central relays of the

spared input. Since there are behavioral deficits in

discrimination of repetitive stimulation at frequencies of

10 pps and above, it should be instructive to look at

cortical responses to 10 pps before and after a DC lesion.

There are several possible outcomes of these recordings:

1) It is possible that, following a DC lesion, the

cortical evoked potential is diminished in amplitude.

Although this has been the subject of some dispute (see

Chapter 1), it is likely that a removal of one of the long

tracts would result in the diminution of the overall amount

of input to cortex. This result would suggest that an

elevation of detection thresholds could be seen following

the lesion. However, this has not been found with

application of von Frey hairs to the sole of the foot

(Vierck, 1977) or with brush stimuli to the lateral calf

(Vierck, 1974). Thus, it appears that input from infrequent









43

stimulation (e.g., von Frey hairs or brush stimuli) gains

access to the cortex more effectively than input which has

been preceded by recent stimulation (e.g., train of pulses).

2) It is possible that, following a DC lesion, there

is a change in the extent to which the entire cortical

evoked potential is suppressed by increasing frequency.

This result would suggest that information transmitted to

the cortex over the spared pathways is compromised by the

capacity for these alternate pathways (or partially

deafferented relays) to follow a 10 pps stimulus. Such a

frequency-dependent suppression of early and late peaks

would suggest that the locus of the effect is sub-cortical:

either at the dorsal horn or thalamus.

3) It is possible that, following the lesion,

different peaks in the cortical evoked potential would be

suppressed by different amounts, either in absolute

amplitude or in frequency-dependent suppression. For

example, it is possible that the first peak of the evoked

potential, which is associated with the depolarization of

layer four, would show no change in amplitude or frequency-

dependent suppression. If only the later peaks, associated

with the spread of information that arose from layer four,

showed an effect of the lesion, this would suggest that the

pathways spared by the lesion adequately carry the

information to the cortex, but that there are changes at the

cortical level following deafferentation that affect the









44

animal's ability to perform the task.

The above reasoning led to the experiments below. The

idea was to determine changes in the cortical evoked

potential following a DC lesion. In this way, the neural

mechanisms underlying the behavioral deficit might be

ascertained.



Methods



Chronic Recordings



Subjects and Training. The subjects were three female

Macaca arctoides (stumptailed macaques) that weighed between

8.6 and 11.3 kg. These animals were not used in the acute

spinal cord experiments. The animals were housed in an

outdoor communal enclosure to maximize their health and

psychological well being. They were tested five or six days

a week. Each day the animals were led into the laboratory

on a leash and seated comfortably in a plexiglas primate

chair. They were gradually trained to tolerate gentle

restraint of their wrists and ankles by leather straps. In

the chair, apple sauce was dispensed through a tube in

response to the animal pressing a bar. After 10-15 min in

the chair, each animal was returned to its home "run" for

the normal diet of Purina monkey chow, fruit and water (ad

libitum).









45

Surgery. Following chair training, epidural recording

electrodes were implanted over primary somatosensory cortex

(SI). This was done under sterile conditions at the

University of Florida animal surgery facility. The animals

were premedicated with ketamine HCL (100 mg) and atropine

sulfate (0.54 mg) before intubation; then Halothane

anesthesia was administered and maintained at 1.5-2.0%.

During the implantation, the skull was exposed, the

bone was cleared of muscle and fascia, and 10 surgical

grade, stainless steel bone screws were implanted. Five

screws were placed over each SI foot representation, in 3-4

mm increments along (3 mm lateral to) the midline (see

Figure 3-1). The stereotaxic coordinates were chosen so as

to cover and flank the foot representation on each side.

Two or four reference electrodes were also implanted over

each frontal lobe. A teflon coated platinum-iridium wire

(Medwire) was passed from the head of each screw to a

miniature D-25 connector that was encased in dental acrylic,

along with the implanted screws. The skin flaps were closed

to cover the skull not covered by acrylic. The animals were

given antibiotics (Dualpen, i.m.) for 1 week following the

surgery and Torbutrol (0.5mg/6hrs) for pain.

Another surgery was performed after normal (pre-lesion)

evoked potential recordings were collected from each animal

(typically for 2 months post-implant). During this surgery,

a unilateral fasciculus gracilis lesion was made by Dr. C.J.









46

Vierck, Jr. (sponsor) at a mid-thoracic level. A

laminectomy was performed on 2 mid-thoracic vertebrae; the

dura was pulled aside, and the dorsal cord was exposed. The

cord was visualized with a microscope, and the dorsal white

matter was interrupted with precision forceps between the

midline and the root entry zone on one side. The incision

was closed, and each animal was given antibiotics (Dualpen,

i.m.) for 1 week following the surgery.



Recording. The animals were seated in a primate chair

that provided loose restraint of the arms, legs and abdomen.

The male D-25 connector on the head was mated with a female

D-25. The female connector was connected, via individually

shielded cables, to a manual switch box located on the wall

adjacent to the animal. The switch box allowed a choice

between any reference electrode, located over frontal

cortex, and any "active" electrode over SI cortex. The two

outputs from the switch box were differentially amplified

(Bak; 10,000x), digitized (Vetter), and recorded on magnetic

tape (VCR) for off-line analysis. The signal was also

averaged on-line (286-computer; RC electronics software) to

monitor the evoked potential during each recording session.

Stimulation. While in the primate chair, the sole of

each foot was stimulated either electrocutaneously or

mechanically. An S88 stimulator (Grass) generated the

stimuli. The stimulus frequencies typically ranged from 1.5









47

pps to 10 pps for both electrocutaneous and mechanical

stimuli. The number of presentations of each stimulus

condition varied from 64-320, depending upon the stimulus

frequency.

For the electrocutaneous stimuli, the output of the S88

passed through a stimulus isolation unit, a constant current

unit, a current monitor, and then out to the animal. Each

animal received stimulation of the foot via two wells in

adhesive foam pads that were filled with either Sergi-lube

or electrode paste. The amplitude of stimulation (4 mA) was

chosen so as to elicit a recognizable evoked potential

without causing any discomfort to the animals or a

perceptible muscle twitch. The duration was short (0.2 ms)

to minimize the stimulus artifact. Two animals received

electrocutaneous stimulation of the ankle with ring

electrodes made of aluminum foil. Each 3/4 in. strip of

foil was coated on one side with sergi-lube, wrapped around

the ankle or the calf, and held in place with adhesive tape.

Stimulus amplitude was 12 mA, which caused a slight twitch

of the foot during about half of the recording sessions.

Mechanical stimulation was provided by a vibrator (Ling

102) that was driven by a locally designed and constructed

driver that took the voltage output of the S88 (a 10 ms

pulse) and matched it to the current requirements of the

vibrator. The vibrator was mounted on a track that could be

vertically advanced by turning a screw with a 1 mm pitch. A









48

9 mm diameter probe, mounted on the end of the vibrator,

passed up through a linear voltage displacement transducer

(LVDT; Schaevitz) and then through a 1.9 cm hole drilled in

a plexiglass foot rest. The monkey's foot was held in place

by a polyform cast, such that the thenar eminence of the

foot was directly over the 1.9 cm hole. The probe was

advanced to make a 1-2 mm indentation in the skin. The

stimulus amplitude was calibrated to 0.4 mm, monitored

(LVDT) throughout each session on an oscilloscope, and saved

on magnetic tape.



Data analysis. All data were analyzed off-line. The

"raw" evoked potential records were read off tape, low-pass

filtered (100 Hz), and digitized (1000 Hz, 1 ms bin width)

by commercially available software and an A-D board (RC

Electronics). The digitized records were converted into

ASCII files for further analysis by software written in 'C'

(by the author).

The averaged evoked potentials were constructed first

by discarding any traces that fell outside of a user

specified amplitude window (typically 300 uV). This was

intended to reject evoked potentials that contained movement

artifact. The remaining traces from an individual session

were then averaged, and the standard deviation (SD) was

computed for each time point (bin) in the average. Averaged

data from individual days were combined into averages across









49

blocks of days with similar stimulus conditions, and

comparisons between pre-lesion and post-lesion data were

based upon these blocked averages. To compare the evoked

potentials recorded under different conditions, the peak-to-

peak amplitudes of p20-to-n50 and n50-to-p90 were computed

(see Figure 3-3). This was done to avoid problems presented

by a fluctuating baseline, especially at different stimulus

frequencies and after the DC lesion, when baseline

fluctuations approached the amplitude of the evoked

potential.

Estimates of baseline levels of noise were made by

calculating the mean amplitude and variance of the

recordings obtained during the 250 ms before each stimulus

onset at 1.5 pps. This was done by taking a random sample

of 1 ms bins. Each sample was taken from a 10 ms window,

once every 20 ms. This procedure yielded results that did

not differ significantly (p<0.05) from sampling once every

50 ms (at 50 ms the autocorrelation function was zero).

However, this technique provided a better estimate of the

noise by increasing the number of samples.

The statistical analyses were all done by programs

written in 'C' by the author. A t-test was used to identify

significantly different slopes and y-intercepts (Mendenhall

et al., 1986; pg. 464) for functions relating response

amplitude to stimulus frequency. This was done first by

fitting a linear (least-squares-fit) function to the data.









50

Then, the mean and variances of the estimates of the slope

and y-intercept were used to compute the test statistic 't.'

This was a conservative test.



Acute Recordings



Animals. Acute recordings from SI cortex were obtained

from three female stumptailed macaques, two of which were

used in the awake experiments above, and three female rhesus

monkeys (Macaca mulatta). The rhesus monkeys were

previously used in behavioral experiments in vision, and had

received horseradish peroxidase injections in small muscles

of the hand and forearm one week prior to the acute

experiment.



Preparation. Each animal was premedicated with

atropine (1.0 cc), and then put into a 72 liter plastic

container that had been filled with isoflurane (about 4%).

The animal was removed once it was flaccid (about 5 mins).

It was then intubated, given isoflurane anesthesia and

intravenous fluids consisting of: lactated Ringer's, sodium

bicarbonate (1 meq/kg/6hrs), mannitol (1 g/kg), and a

corticosteroid (dexamethasone, 2 mg/kg/8hrs). Isoflurane

anesthesia was maintained at a level between 1.5-2.5%

throughout all surgical manipulations. Each animal's core

temperature was monitored with an esophageal thermometer and









51

was maintained between 37 and 38 degrees C with a heating

pad (K-pad) and warmed paraffin bags. After placement in a

stereotaxic head holding device, the head was shaved, and an

incision was made along the midline of the head. The bone

was scraped clean of tissue, and a single 1.4 cm hole was

made unilaterally over SI cortex with a trephine. Rongeurs

were used to extend the hole to the midline. The dura was

dissected away with iridectomy scissors and precision

forceps to expose the underlying pia. The cortex was then

bathed in artificial cerebral spinal fluid (CSF; Merlis,

1940).



Recording. A recording chamber was mounted over the

opening with stainless steel screws and dental acrylic. It

was filled with artificial CSF and closed-off with a device

designed to hold a recording electrode and microdrive. This

setup formed a tight seal, but it allowed the electrode to

be positioned anywhere within the chamber, while keeping

track of the x-y coordinates to within 1/4 mm. The head was

then supported by a screw imbedded in the dental acrylic on

the skull, and the ear and eye bars were removed.

Electroencephalogram (EEG) electrodes were attached to two

screws that had been threaded into the skull over the

frontal and occipital cortices. The EEG was then

differentially amplified (10,000x), filtered (low-pass 40

Hz) and displayed continuously on an oscilloscope. Once the








52

surgery was complete, and the animal's head was suspended by

the screw in the acrylic cap, the anesthetic was changed to

nitrous oxide (66%) and isoflurane 0.5-1.0% (Tigwell and

Sauter, 1992). The animal was then given a dose of

Pancuronium (0.03 mg/kg/3hrs i.v.) and put on an artificial

respirator (Harvard; 20 breaths/min; 10 ml/kg body mass).

The plane of anesthesia was monitored by the EEG, and the

expired CO2 was maintained between 35 and 40 mm Hg.

Glass and Insl-x coated tungsten electrodes (impedance

= 50-200 kohms at 1000 Hz; see chapter 2 for electrode

construction) were lowered into the cortex with the

microdrive (Narishige). The electrode depth was chosen to

maximize the early negative potential (ca. 20 ms). In this

way, the recordings were always near to, but deeper than,

layer IV. Evoked potentials were differentially amplified

(10,000x), digitized (Vetter) and saved on magnetic tape

(VCR). The potentials were also filtered (3-300 Hz) and

averaged on-line (1 ms bin width; RC Electronics A-D board

and software). The averages were saved on disk in ASCII

format.



Stimulation. Electrocutaneous stimuli were delivered

to the thenar eminence and ankle of each foot. These

stimuli were similar to those described above in the Chronic

Recordings. The amplitude of the thenar stimulus was

typically 5-6 mA, whereas the ankle stimulus was 12 mA.









53

Stimulus frequencies were 1.5, 2.5, 5, 7.5, and 10 pps.



Histology. At the end of each experiment, the animal

was given an overdose of sodium pentobarbital (195 mg i.v.)

and perfused through the heart. The stumptails were

perfused with saline and 1 ml/L of heparin sulfate, and then

with 10% formalin. The rhesus monkeys were perfused first

with warmed (36 degrees C) phosphate buffer (7.4 pH) and

heparin sulfate, and then with fixative (2% paraformaldehyde

and 2.5% glutaraldehyde). This was followed by a mixture of

fixative and 10% glycerol and rinsed with 10% glycerol in

phosphate buffer. The brain and spinal cord were removed,

and the cortex was photographed. The spinal cord and cortex

were blocked and cryo-protected in 30% sucrose (or glycerol)

and 10% formalin for a minimum of 1 week. Frozen sections

were cut at 50 um, and stained with cresyl violet. In

animals with DC lesions, the extent of the lesion in the

cord was reconstructed with camera lucida drawings.



Data Analysis. Averages were compiled on-line from

100-150 stimuli, converted to ASCII files and saved on disk

for later analysis. The data were analyzed with spreadsheet

software where peak values of n20 were computed relative to

a baseline value. The baseline was taken as the maximum

amplitude between about 10 and 15 ms after the stimulus.

For each recording location, evoked potential amplitudes









54

were expressed as percent of the response to 1.5 pps.



Results

Chronic Recordings



Histology. The locations of the recording electrodes

for the three animals are shown in Figure 3-1 (arrows). The

active electrodes were either directly over or bordered the

SI cortex in all three animals. In monkey C, the active

electrodes (indicated by arrows) were located over the

posterior half of the precentral gyrus (in the motor strip).

The electrode used to record from the DC deafferented cortex

(on the left) was actually over the midline in monkey C.

However, the potentials from the two sides of the animal

were similar in shape and latency, and these potentials were

similar to those obtained in the other animals. Therefore,

it is likely that recordings from all three animals arose

from similar sources in the cortex.

The dorsal column lesions in all three animals were

complete and are shown in Figure 3-2. Monkey F received two

consecutive lesions separated by four months. The first DC

lesion (right side) was at spinal cord level T7, included

some of the grey matter in the dorsal horn and crossed the

midline in the most dorsal part of the left DC. The second

DC lesion in monkey F was between level T3 and T4 on the

left, and it included gray matter in the dorsal horn at the









55

midline. This lesion crossed the midline to include the

dorsal aspect of the right DC, and it crossed the dorsal

root entry zone to include a small portion of the DL tract.

The DC lesion in monkey S was on the right, between T8 and

T9; it crossed the midline to include the ventro-medial

portion of the left DC, and it included dorsal grey matter

up to and including the central canal. The DC lesion in

monkey C was the largest of the three animals. It was

between T7 and T8 on the right, and it included much of the

left DC, sparing a dorsal-lateral portion. In addition,

part of the right DL tract was included in this lesion.



Mechanical stimulation. A normal epidural evoked

potential recorded over SI cortex (see Figure 3-1 for

location) in response to a mechanical displacement of the

skin (0.4 mm, 10 ms, 1.5 pps) is shown in Figure 3-3. The

solid line denotes the average response (monkey S),

accumulated from 210 stimulus presentations over 5 recording

sessions (days), and the dashed lines represent plus and

minus one standard error (SE) at each time point along that

average (1 ms bin width). The first major positive peak

occurred around 20 ms after the onset of the stimulus (see

also Table 3-la). The next two major peaks in the response

were around 50 ms and 90 ms. Although these peaks did not

always occur at exactly these latencies under all stimulus

conditions in all animals, p20, n50 and p90 will be used to









56

refer to these major peaks for ease of nomenclature.

The effects of different stimulus frequencies on the

evoked potential from normal monkeys are shown in Figure 3-

4. Stimulus frequencies ranged from 1.5 to 10 pps and are

listed to the right of each trace. The control frequency

(1.5 pps) was repeated at the end of each series to monitor

any possible changes in the response over the recording

session. In this Figure, the stimuli were averaged in

sweeps of 500 ms. At 10 pps the evoked potentials from

successive stimuli interact, since each response lasted

longer than the inter-stimulus-interval (100 ms). Following

interruption of the DC, the evoked potential in response to

mechanical stimulation was attenuated considerably, even at

1.5 pps (Figure 3-5, monkey S). At 10 pps the potential was

nearly buried in the noise. Note that the amplification

scale is the same in both Figures 3-4 and 3-5. After DC

lesions, a large attenuation of responses to all frequencies

was observed in all three animals when stimulated

mechanically.

This result is shown quantitatively for the three

animals in Figure 3-6 by expressing the amplitudes of the

p20-n50 complex (top) and p90-n50 complex (bottom) in terms

of average percent of the pre-lesion amplitude at 1.5 pps.

The pre-lesion (solid line, open triangles) frequency-

response function does not drop below 50% at 10 pps for

either complex. On the other hand, the post-lesion response









57

(long dashes, open squares) to 1.5 pps is nearly 20% of the

pre-lesion response to 1.5 pps, and the post-lesion response

to 10 pps is just above the noise (solid line = average

noise, short dashes = 95% confidence interval). For both

the p20-n50 and p90-n50 complexes there was a statistically

significant effect of the lesion on response amplitude (y-

intercept) at the p<0.01 level. Each individual animal also

showed a significant amplitude shift (see Tables 3-2 and 3-

3).

It was also of interest to determine if there was a

change in the slope of the frequency-response function with

the lesion. That is, did the lesion cause a greater

frequency-dependent suppression at 10 pps relative to 1.5

pps? This is shown quantitatively in Figure 3-7 by

expressing the amplitude of both pre- and post-lesion

complexes in terms of percent amplitude of the each response

to 1.5 pps. Neither the p20-n50 nor the p90-n50 complexes

showed a significant change in slope, as a result of the

lesion, in response to mechanical stimulation. This was

also true for individual animals (Tables 3-2 and 3-3);

however, the post-lesion potentials were limited by a floor

effect, starting with considerable attenuation of responses

to the low frequency (Figure 3-6).

Not only were responses to a mechanical stimulus

significantly suppressed at stimulus frequencies between 1.5

and 10 pps, but a similar result was obtained from one









58

animal (monkey C), with stimulation at 0.1 pps in both the

pre-lesion and post-lesion conditions. The post-lesion p20-

n50 and p90-n50 complexes were suppressed to 23.4% and 18.4%

of pre-lesion amplitudes, respectively. These results

suggests that, over a 100 fold range of stimulus

frequencies, interruption of the DCs suppressed responses to

mechanical stimulation by at least 40% at all frequencies,

and by greater than 70% at low frequencies.



Electrocutaneous Stimulation. A normal evoked

potential, recorded over SI cortex (monkey S) in response to

electrocutaneous stimulation of the foot (0.2 ms, 4 mA, 1.5

pps), is shown in Figure 3-8. This potential was recorded

from the same electrode location as the potential shown in

Figure 3-3; however, it was recorded on alternate days. The

average (solid line) and SE (dashed lines) of the response

at each time point are shown for 303 stimulus presentations

collected over five separate days. The latencies of the

different peaks were approximately the same as for the

mechanical stimulation (Table 3-1b). However, in this

figure a stimulus artifact can be seen at 0 ms.

Responses to different frequencies of electrocutaneous

stimulation are shown in Figure 3-9 for the normal cortex

and Figure 3-10 for the DC deafferented cortex. In general,

the shape of the normal evoked potential did not change

substantially in response to different stimulus frequencies,









59

except for 10 pps, where successive responses interacted.

There was also a shift in the n50 minimum to an earlier

latency at frequencies between 2.5 pps and 5 pps. The

effect of a DC lesion on the responses to electrocutaneous

stimulation was to decrease the overall amplitude, though

not to the extent seen for mechanical stimulation (shown in

Figures 3-3 and 3-4). This improvement in the post-lesion

signal-to-noise ratio provided an opportunity to examine

frequency-dependent suppression in the absence of a floor

effect, in contrast to the response from the mechanical

stimulus.

Figure 3-11 shows the effect of a DC lesion on averaged

frequency-response amplitudes, for both the p20-n50 (top)

and p90-n50 (bottom) complexes. The p20-n50 complex was

reduced in amplitude by 60% or more at all frequencies post-

lesion relative to the pre-lesion amplitude at 1.5 pps

(p<0.01). This decrease in amplitude was also significant

(p<0.01) for the p20-n50 complex in each of the three

animals. The y-intercept was not significantly altered for

the p90-n50 complex; however, since the average function was

not linear, this was not a valid test. When the difference

between the pre-lesion mean amplitude (across frequency) and

post-lesion mean amplitude of the p90-n50 complex was tested

for significance (t-test), there was a reliable difference

between means (p<0.05). One animal (monkey S) that had a

relatively linear function showed a significant change in









60

the y-intercepts (p<0.01) of the functions relating

amplitude of the p90-n50 complex to frequency (see also

Tables 3-2 and 3-3).

The DC lesion led to an increased frequency-dependent

suppression of the p20-n50 response to electrocutaneous

stimulation of the foot (Figure 3-12, top). This effect was

not observed in the p90-n50 complex (Figure 3-12, bottom).

The slope of the frequency response function was

significantly (p<0.01) steeper in the post-lesion cortex

than the pre-lesion cortex for p20-n50. There was about a

20% greater reduction in amplitude at 10 pps following the

lesion, relative to the pre-lesion suppression by 10 pps.

This frequency-dependent suppression was true for two of the

three individual lesions (p<0.05; see Table 3-2).

Thus, the p20-n50 complex of the evoked responses to

electrocutaneous stimulation was significantly decreased in

amplitude by a DC lesion, and frequency-dependent

suppression was increased. This was also true for the p90-

n50 complex in amplitude but not for frequency-dependent

suppression.



Ankle stimulation. A normal evoked potential, recorded

over SI cortex (monkey C) in response to electrocutaneous

stimulation of the ankle (0.2 ms, 12 mA, 1.5 pps), is shown

in Figure 3-13. This potential was recorded from the

electrode location shown in Figure 3-1. The average (solid









61

line) and SE (dashed lines) of the response at each time

point are shown for 320 stimulus presentations collected

over five separate days. The latencies of the different

peaks were approximately the same as for mechanical

stimulation and electrocutaneous stimulation to the foot

(Table 3-ic). In this Figure, the stimulus artifact is a

small negative deflection at 0 ms.

Averaged responses to different frequencies of ankle

stimulation are shown in Figure 3-14 for normal cortex and

Figure 3-15 for the DC deafferented cortex. In the normal

cortex, p20 remained relatively constant with changes is

stimulus frequency, whereas p90 suppressed with increasing

frequency. Following the DC lesion, there was about a 40%

reduction in the amplitude of the p20-n50 complex at all

frequencies (Figure 3-16). After the lesion, the p90-n50

complex showed about a 50% reduction in amplitude at 1.5 pps

but only a 10-15% reduction in amplitude at 10 pps. Both

the p20-n50 and p90-n50 complexes showed a significant

(p<0.01) decrease in amplitude following the lesion. The

individual animals also showed this effect (see Tables 3-2

and 3-3).

The frequency-dependent suppression following the DC

lesion was similar to the pre-lesion condition. There was a

20-50% reduction at 10 pps relative to 1.5 pps for both

complexes (Figure 3-17). One of the animals (monkey S)

showed a significant difference in slope for the p20-n50









62

complex between the normal and the DC-deafferented cortex

(p<0.05), but this was not true for the average of the

animals. It should be noted that, in monkey S, the

comparison between the sides was made after the DC lesion;

consequently, there was a partial DC lesion on the control

side of this animal.

The evoked potential recorded in response to

electrocutaneous stimulation of the ankle showed that

interruption of the DC led to a change in the amplitude of

both the p20-n50 and p90-n50 complexes. However, there was

no significant change in the slope of either frequency-

response function following the DC lesion, in contrast to

the results from electrocutaneous stimulation of the thenar

eminence of the foot.



Summary of Chronic Recordings. The awake recordings

showed that under all stimulus conditions there was a

statistically significant change in the amplitude of the

evoked potential following interruption of the DC. When

electrocutaneous stimulation to the sole of the foot was

used, there was a significant change in the frequency-

dependent suppression of the early response after a DC

lesion. This result was not observed under conditions of

mechanical stimulation of the foot or electrical stimulation

of the ankle.


















Q) '0II


+lJ 4. 0
U4J

0 0 II


U) 0 *

4- *J 0 :3
U *- "

u V En
r-4 >0 r1
0)
~$4
0 0 I



U 0 -(0




04 0)0
00C00




0p O








4j 4J *r-
(0 4O
0 C) 0 I
O *X H













n 1O u
( 0 O-


0 4 U C4 O





.9-I C ) 0 ON
-4 0 -I *p
FX4 9i ) O4 P



















































ia.


0



















U *O 44
a- 0) 0




S40 )











-ic






10
U) 0










.M 0
0










p u









I 0

-4




-4 Q
)U



--I 0 )



O U-H U


































LL



0


U


C,,
Qj-.













Table 3-1. Evoked potential latencies. Each block shows the peak latencies for each
component of the evoked potential in response to electrocutaneous, mechanical or ankle
stimulation before and after a DC lesion.

Monkey Stimulus Pre-les Pre-les Pre-les Pst-les Pst-les Pst-les
/Foot p20 (ms) n50 (ms) p90 (ms) p20 (ms) n50 (ms) p90 (ms)
F/R elec 21.0 48.0 96.0 23.0 48.0 93.0
F/L elec 21.0 48.0 95.0 23.0 45.0 87.0
S/R elec 19.5 55.0 83.0 23.0 56.0 94.0
Average 20.5 50.3 91.3 23.0 49.7 91.3
Std. Dev. 0.87 4.04 7.23 0.00 5.69 3.79
C/R mech 23.0 58.5 85.5 25.0 46.0 92.0
F/L mech 23.0 55.0 93.0 23.0 53.0 93.0
S/R mech 22.0 47.0 89.5 26.0 42.0 90.0
Average 22.7 53.5 89.3 24.7 47.0 91.7
Std. Dev. 0.58 5.89 3.75 1.53 5.57 1.53
S/L* ankle 17.0 33.0 88.0 -
S/R ankle 19.0 55.0 77.0
C/R ankle 19.0 55.0 84.0 19.0 47.0 87.0
Average 18.0 44.0 86.0 12.7 34.0 54.7
Std. Dev. 1.41 15.56 2.83 10.97 29.72 47.61

* Note: No pre-lesion evoked potentials were recorded from monkey S in response to ankle
stimulation. All recordings were made (after a right DC lesion) on the control and deafferented
sides.
















oo


UO
, 0
d -


dd- i-
>Jddd


U) LO
Co9o
0 0
c; o


d dd d
.~000 0
OOO* 0 0




0., w m w ) u
o0 ccc c





Em
(I D + + +
E




0 a: -j cc cc



0>
D Co

0
E cu
E A


tE
D -;
(U (c


I I


0)

0)















o
0
a)












5-
C



0

Co





eD
0O





0o









0
(D
a
0
C
0









0
E
0 .






0.
O -
0.



4-











Z.


I I + +


+ ++ ,

































m
cIa


00
oo
dd


I I


c~
-d


(- t- C--
0 0


( D + + + *




,000 0
.* o o 5
> oddd d




0 c c c




(DI + + +
E


LL- LL


I--
Ee

0 '

0a)
!e

c
(D T




aB o
e
Coc
CO
0Q.0
-CL

00C
z .C
> 0
a)
SC
0.0
(D
0)0)

0)0

0 (0

00)0


0 1


* I ++








Figure 3-3. Averaged pre-lesion evoked potential,
recorded over SI cortex in response to mechanical
stimulation of the foot (monkey S). Solid line
represents the average of responses to 210 stimulus
presentations, accumulated over 5 days. Dashed lines
are + 1 SE.
















stimulation


p20


p90


n50


-100 -50 0 50 100
time (ms)


150 200 250


Q)i


-0
<-1


Mechanical
























O
M (0



0 0L
o o m .
o 0*



41.
o- o




- Og
S04




tn 04
o H cz






SU-4 0
01>




rj 0)4-4







4) r -i U) 0

z4 In ) 40
pk(A- 0 NM






73










S a Ca

L.n 0n 0 L)


tO
0




--0
--
0 0



E.__
CO


oE










-0



pI I I I I I O


















Vn *
S0-) U

() l >1
%4 ) u4
4-4 >J o




o>
*H 0 N

0 e (4 0


WA0



"-4

.,1 4 ,.







rl ,4 U
V 0





0*O
0 0 -
4 En 0A u
o
*JC 0
S 0 (a









40) 0
I0 44 U
> 0








in U4





0 co

a 0-0


O t*H 0
j(n r-






75










U) cn U ) .n
O_ Q- Do Q- o o_
0- a 0a

LO U 0) I 0 l-
1 1 1







0 0













Qi)
C0
O O
. -0















- -






-PI oE
CPn !IduV








Figure 3-6. Combined averages of 3 monkeys' evoked
potentials in response to mechanical stimulation,
before and after a DC lesion. Amplitudes are expressed
as percent of the pre-lesion response to 1.5 pps.
Error bars are + 1 SE. (TOP) The amplitude of the
p20-n50 complex. (BOTTOM) The amplitude of the p90-
n50 complex. Solid line along each x-axis represents
the mean amplitude of the noise, and the short dashes
represent the 95% confidence interval of that noise.









P20-N50
Mechanical stimulation


A pre-lesion
0 post-lesion


----- ----- -----: ^ -- --- -

2 4 6 8 10
Stimulus frequency (pps)


P90-N50
Mechanical stimulation


A pre-lesion
a post-lesion


10
10


100-


100-


i I I .
---------- -------- --


2 4 6 8
Stimulus frequency (pps)









Figure 3-7. Combined averages of 3 monkeys' evoked
potentials in response to mechanical stimulation,
before and after a DC lesion. Amplitudes are expressed
as percent of either the pre-lesion or post-lesion
response to 1.5 pps. (TOP) The amplitude of the p20-
n50 complex. (BOTTOM) The amplitude of the p90-n50
complex. Error bars are + 1 SE.









P20-N50
Mechanical stimulation


-1


A pre-lesion
o post-lesion


4 6
ulus frequency


P90-N50
Mechanical stimulation


A pre-lesion
0 post-lesion


' I


2
Stimulus


4 6
frequency


100-







60-
" 60-

0
U
40-

Q) -

S20-
E


2S
Stim


(pps)


100-


8
(pps)


i i


I I M









Figure 3-8. The averaged pre-lesion evoked potential
recorded over SI cortex in response to electrocutaneous
stimulation of the foot (monkey S). Solid line
represents average of responses to 303 stimulus
presentations, accumulated over 5 days. Dashed lines
are + 1 SE.





81











Electrocutaneous stimulation
40-
p20 p90




20 stimulus i
artifact '


10~

0-i
' '10



-20-

-20-



-30- \Q
n50

-40- 1 1 ,1


50 100
time (ms)


A

-


150 200 250





















S0 U

44 > 4J

. H



0 4- 4 0
0



00 4

a)



.0 0









t o
rC *o








0 I0





0 0










0 0




.ri 4J 44 M
rZ4 til 0 Z4






83











CL 0") 0- 0-

O U 0 L0 0 Lo
l_ I n '-- f

0-






0-0
4-j
o l)


D (n


-00




0 -0



O0

0






9pni-dldwv



















C ) 0n
Q) ta Q) >4
Q) m ti 1o
4- 0) M-l
(-4 > 4> 0

o>O
-4 0 N


0o .
4- -- 4
044


00



134) 0
104


- 4-4






(/ 0- U

> 0

*U)4



*r-I 02




0



0C 0 to

C- 4 )



*0' 0
-I 4 C 4-4 M
44 M 0 QL4




















o uO
inr-,.


LO


9pnl!ldwV


c
0






O Q
0
0
0-4-



n


C)
0





o0--




oE

0.-o
CN-









Figure 3-11. Combined averages of 2 monkeys' evoked
potentials in response to electrocutaneous stimulation
before and after a DC lesion (3 lesions). Amplitudes
are expressed as percent of the pre-lesion response to
1.5 pps. Error bars are + 1 SE. (TOP) The amplitude of
the p20-n50 complex. (BOTTOM) The amplitude of the
p90-n50 complex. Solid line along each x-axis
represents the mean amplitude of the noise, and the
short dashes represent the 95% confidence interval of
that noise.










P20-N50
electrocutaneous stimulation
A pre-lesion
7T post-lesion


0-



- ---- --.



0- I I

0 2 4 6 8 10
Stimulus frequency (pps)


P90-N50
electrocutaneous stimulation
A pre-lesion
0 post-lesion


\ ---


100-


100-


80



60-



40-



20-



0-


2 4 6 8
Stimulus frequency (pps)


I








Figure 3-12. Combined averages of 2 monkeys' evoked
potentials in response to electrocutaneous stimulation
before and after a DC lesion (3 lesions). Amplitudes
are expressed as percent of either the pre-lesion or
post-lesion response to 1.5 pps. Error bars are + 1 SE.
(TOP) The amplitude of the p20-n50 complex. (BOTTOM)
The amplitude of the p90-n50 complex. Lines drawn
through data points are least-squares-fit.










P20-N50


mulation
A pre-lesion
3 post-lesion


N
N N


I 6I
2 4 6
Stimulus frequency


8
(pps)


P90-N50
electrocutaneous stimula


10





tion
A pre-lesion
a post-lesion


2 4S 6 8
Stimulus frequency (pps)


100-

0-
in
g_ 80-
uin


100-


O-
S80

LO


-0 60-
0

U
u






E


10


i









Figure 3-13. The averaged pre-lesion evoked potential
recorded over SI cortex in response to electrocutaneous
stimulation of the ankle (monkey C). Solid line
represents average of responses to 320 stimulus
presentations, accumulated over 5 days. Dashed lines
are + 1 SE.

















Ankle stimulation


p90


p20


n50


-30 -
-100


1
-50


I I I
0 50 100
time (ms)


1 I I
150 200 250


30-



20-



, 10-



-7 0-
-2-
E _
-Q


-10-



-20



















)C O -

44 ) >4 0

C) 0



0 4






.-4
0o0



.4 44

U)
,- 04-1



0 0 .



0










I 0p0
0 4-4


gw


I 4
44 tn 4J




M'-4 0
-H 4 O 4-4 P
r- U 5^






93










() I n in cn
C0 CL a O.
Q. a- ID CL -
U) LO 0 10 Ln
,- 1 0 r-,i -











C


.o


c-










I( I





9pnl!ldwV