0006-Ch08.pdf

Transformation of

Descending Commands

into Muscle Activity by

Spinal Interneurons in

Behaving Primates

Steve I. Perlmutter and Yifat Prut

CONTENTS

8.1 Introduction

8.2 Methodology

8.3 Interneuronal Activity Related to the Parameters of Movement

8.4 Non-Parametric Features of Interneuronal Firing

8.5 Spatial Tuning of Spinal Interneurons

8.6 Activity during an Instructed Delay Period

8.7 Temporal Properties of Firing

8.8 Concluding Remarks

References

8.1 INTRODUCTION

Ever since Sherrington’s pioneering experiments led him to conclude that reﬂexes

were “the unit reaction in nervous integration,” 1 physiologists have conceived of

descending systems recruiting spinal circuitry to produce voluntary movements. This

insight has evolved to the present view that transmission through spinal pathways,

which execute the stereotyped, stimulus-driven motor behaviors referred to as

reﬂexes, is mutable, dynamically modulated, and under differential control by

supraspinal structures. 2 By using spinal pathways to specify motoneuron activity,

the job of the motor cortices is presumably simpliﬁed, and sensory information can

be employed rapidly and “on-the-ﬂy” to guide movements and correct errors.

Sherrington believed spinal reﬂexes were anatomical and functional elements

that could be isolated and that operate in a highly interactive and combinative manner.

He inferred that cortical systems also take advantage of this organization, and his

discussion of behavioral control by the cerebral hemispheres (“nervous organs of

control”) is very much in accord with the modern view:

The way in which we ourselves acquire a new skilled movement, the means by which

we get more precision and speed in the use of a tool, the handling of an instrument,

or marksmanship with a weapon, is by a process of learning in which nervous organs

of control modify the activities of reﬂex centres, themselves already perfected for

other though kindred actions. 1

The development of these ideas originated in the early appreciation of the

ﬂexibility of spinal reﬂexes, as discussed by Goldstein. 3 He described “reﬂex rever-

sals” and the dependence of the stretch reﬂex “on the position of the limb, on the

behavior of the rest of the organism, and on whether or not attention is paid to it.”

The neural substrates for this state-dependence were elucidated in the second half

of the 20th century by elegant, meticulous experiments that revealed the extensive

convergence of information from multiple afferent and descending sources onto indi-

vidual spinal interneurons. 4–7 More recent work has demonstrated that the activation of

particular cutaneous or proprioceptive receptors can produce different motor outputs

during different phases of real or ﬁctive movements, and that the excitability of moto-

neurons and interneurons can be modulated by monoamines and neuropeptides.

Our conceptual framework of spinal function during movement has progressed

signiﬁcantly as these studies have ascertained the functional organization of spinal

interneurons. 8,9 However, it has been difﬁcult to verify and elaborate these concepts

in action, during natural movements, because of the inaccessibility of spinal circuits

to direct study in awake, behaving animals. Preparations exhibiting ﬁctive behaviors

have enabled the direct investigation of the neural mechanisms of locomotion and

scratching, 10–12 but other types of movement have been studied only in anesthetized

or decerebrate animals, or with indirect measurements of neural activity in moving

animals or human subjects. We believe that information on the properties of spinal

interneurons during normal behaviors is needed to fully explicate the recruitment

and modulation of spinal pathways by descending systems. Accordingly, we have

been recording the activity of spinal neurons during voluntary wrist movements in

monkeys, using techniques developed in the laboratory of Dr. Eberhard Fetz. 13–15

In our initial studies, we have examined the activity of spinal interneurons during,

and in preparation for, normal hand movements, and have compared these results

to data from motor cortex, red nucleus, and peripheral afferents during a similar

task. 15–17 Our efforts to classify interneurons have focused on identifying their output

linkages to motoneurons with spike-triggered averaging of electromyographic activ-

ity (EMG), and we have characterized the distribution of connections from premotor

interneurons to synergistic and antagonistic muscles of the forearm. 18

We believe our experimental technique provides a new approach to studying

spinal control of movements. It enables the direct measurement of the interactions

between descending commands and proprioceptive signals that generate voluntary

muscle activity. This approach complements the extensive data from previous studies

on the wiring and ﬂexibility of spinal pathways. In addition, information on the

properties of spinal interneurons during normal behavior will help interpret data on

descending signals, and advance hypotheses about information processing in supraspi-

nal motor structures. For example, current perspectives on the computational organiza-

tion of the motor system envision cortical commands in terms of abstract features, such

as movement direction referenced to a head- or shoulder-centered coordinate frame. 19,20

The task of completing the transformation into joint- or muscle-centered representations,

patterns of muscle activities that execute the desired movement, is often ascribed to

spinal circuits. We will address this hypothesis in this chapter by summarizing our

results on the features of interneuronal activity that do, and do not, relate to movement

parameters, and on the temporal properties of ﬁring of spinal neurons.

8.2METHODOLOGY

To study the properties of individual spinal cord neurons during voluntary movements,

we have applied classical, chronic recording techniques developed in studies on the

brain during the last 30 to 40 years. These techniques allow us to measure the activity

of single spinal neurons in awake, trained macaque monkeys performing natural behav-

iors. 18 A laminectomy exposes the dura mater overlying three segments in the lower

cervical cord and bone screws are inserted into the lateral masses of the vertebrae

overlying these, and adjacent, segments. A stainless steel chamber with a removable

cap is cemented to the screws with dental acrylic, providing long-term access to the

cord for daily recording sessions. Monkeys exhibit a stiffened posture of the upper back

with this implant, which fuses several vertebrae together, but are otherwise undisturbed

and behave normally. While the monkeys work at trained hand and arm tasks, the head

and spinal column are immobilized by ﬁxing the skull and vertebral implants to the

apparatus. Glass-coated tungsten electrodes are introduced through the dura with a

hydraulic microdrive mounted to the chamber. To date, we have recorded spinal neurons

during single-joint wrist movements, a power grip, and small amplitude, free-form

reaching movements. We are able to maintain stable recordings of single, well-isolated

neurons for more than 30 min if the monkey refrains from making large postural shifts.

At the end of the recording session, the electrode is removed, the chamber is re-sealed,

and the monkey is returned to his home cage as in brain-recording studies.

Our electrodes encounter many spinal neurons with spontaneous and task-related

activity in the awake monkey. We have taken two approaches to classifying these

neurons. First, input and output connectivity can be characterized in the awake

animal using well-developed criteria from earlier studies. Motoneurons and premotor

interneurons produce characteristic post-spike effects in spike-triggered averages of

EMG 21 from arm muscles ( Fig. 8.1 ). Consequently, we simultaneously record the

EMG of multiple forearm muscles and neuron activity. Most interneurons with post-

spike effects in muscles are probably last-order cells, but some may inﬂuence

motoneurons through oligosynaptic pathways. 18 Inputs from peripheral afferents can

be identiﬁed by responses to electrical stimulation of peripheral nerves or muscles

and by responses to passive displacements of the wrist and mechanical stimulation

of the skin. These techniques provide tests for inputs from large-diameter muscle

and cutaneous afferents. However, in the awake animal inputs from groups I and II,

muscle afferents can be difﬁcult to distinguish, and tests for inputs from high-

threshold group III and IV ﬁbers are not feasible. We have not focused on identifying

FIGURE 8.1 Spike-triggered averages of EMG for a motoneuron (A) and a spinal premotor

interneuron (B). EMG of multiple forearm muscles was averaged for 15 ms before and 35 ms

after the occurrence of action potentials of well-isolated neurons. Top traces show average

waveform of the action potential (A) or the autocorrelation of ﬁring (B) for the recorded

neurons. Bottom traces show average unrectiﬁed (A, except APL rectiﬁed) or rectiﬁed

(B) EMG aligned on the time of spike occurrence for independently recorded forearm muscles

that were co-active with the neuron. Spike-triggered averages of the motoneuron show a motor

unit action potential in APL (A). Premotor interneuron produced a post-spike facilitation in

FCR, and possibly in PL (B). The short, dashed horizontal line on the spike-triggered average

of FCR shows the mean EMG level during the baseline, pre-spike period (identiﬁed by the

short, dashed vertical lines). Post-spike effects that were sustained increases or decreases in

average EMG above or below the baseline mean ± 2 standard deviations (long, dashed

horizontal lines) were considered signiﬁcant. Note that post-spike effects of cells classiﬁed

as motoneurons and premotor interneurons were noticeably different. Cells were classiﬁed

as motoneurons only if they also were recorded at an appropriate depth in the cord, had a

low peak ﬁring rate and unidirectional activity (see 8.3 , Interneuronal activity related to the

parameters of movement). Muscles are APL, abductor pollicis longus; EDC, extensor digi-

torum communis; ECR, extensor carpi radialis; ED-2,3, extensor digitorum 2,3; ED-4,5,

extensor digitorum 4,5; FCR, ﬂexor carpi radialis; PT, pronator teres; PL, palmaris longus;

FDS, ﬂexor digitorum superﬁcialis; ECU, extensor carpi ulnaris.

descending inputs to interneurons, but this is possible with electrical stimulation of

supraspinal structures, such as the pyramidal tract, motor or sensory cortex, red

nucleus, or reticulospinal centers.

Second, we foresee our studies developing a new classiﬁcation scheme based

on the response properties of interneurons during normal behavior. Our hope is that

this new classiﬁcation scheme will augment rather than parallel the well-established

organizational framework developed by previous studies, and that the comparison

of the two will provide new insights into interneuronal function.

Flexion

Extension

w2305

32 trials

33 trials

-2

Time (s)

FIGURE 8.2 C6 premotor interneuron with phasic-tonic activity for extension torques. The

monkey performed isometric, ramp-and-hold ﬂexion (left) and extension (right) torques start-

ing from rest. Torque proﬁles for all trials are superimposed (bottom). Rasters (middle)

indicate timing of action potentials (small vertical lines) and torque onset (diamonds) in

individual trials (rows). Histograms (top) show the average activity of the neuron. All traces

are aligned on torque onset (time = 0). Average ﬁring rate was 11 sp/s when the monkey

exerted no torque at the wrist (baseline). During extension, the average rate peaked at 71 sp/s

during dynamic torques and then decreased to a steady level of 43 sp/s during the static hold

period; this response was termed phasic-tonic activity. Firing rate decreased during ﬂexion

relative to baseline. The neuron produced a post-spike facilitation in the spike-triggered

average of rectiﬁed EMG of one forearm muscle, EDC.

This chapter will discuss primarily data on spinal interneurons recorded in the

C6-T1 segments as monkeys generated isometric or auxotonic (against an elastic

load)ﬂexion and extension torques about the wrist. Torque controlled the position

of a cursor on a CRT screen in front of the animal. The monkey was required to

move the cursor into target boxes displayed on the screen. The position of each box

speciﬁed a ﬁxed level of wrist torque in ﬂexion or extension. The monkey received

an applesauce reward when he produced a ramp-and-hold torque that met velocity

and amplitude criteria.

8.3INTERNEURONAL ACTIVITY RELATED

TO THE PARAMETERS OF MOVEMENT

The activity of many C6-T1 interneurons was related to components of the torque

generated during extension and ﬂexion of the wrist. 22 The responses of interneurons

with and without effects on muscle activity, identiﬁed by spike-triggered averages

of EMG, were similar. Figure 8.2 shows the activity of a premotor interneuron

recorded in the C6 segment as the monkey generated isometric torques. The neuron

produced a post-spike facilitation of activity in the extensor digitorum communis

Plik z chomika:

Inne pliki z tego folderu:

Inne foldery tego chomika: