Abstract
This study tested the idea that there are sets on neurons in
the spinal cord responsible for generating elementary synergies (movement
primitives) that could have implications for
.
Introduction
It has been suggested
that electrical stimulation within intermediate laminae of the spinal cord can
activate sets of neurons to produce synergistic movements termed “movement
primitives” [3]. In addition to electrical stimulation of muscles and nerves,
stimuli have also recently been applied to the spinal cord and ventral roots as
potential avenues for restoring functional movement [1, 2]. To date, very few studies have compared results
using different methods in the same subjects. In addition, many of the basic
biomechanical and neurophysiological properties of
movements produced by stimulation are poorly understood. This paper compares
these properties using electrical stimulation of muscles, nerves, ventral and
dorsal roots, and the spinal cord. The results of comparing all these methods
in the same preparation should be helpful to understand the concept of movement
primitives and to provide some preliminary data for clinical
Methods
Eight
adult cats were anesthetized either with halothane (n=5) or pentobarbital
(n=3). In 5 animals the following nerves were exposed in the right hindlimb and cuff or epineural
electrodes were implanted: tibial nerve (TB), common peroneal nerve (CP), hamstring branches of the sciatic nerve supplying
knee flexors (KF) and hip extensors (HE),
and femoral nerve branches supplying hip flexors (HF) and knee extensors
(KE).
Each animal was positioned in a conventional spinal frame. The
right paw was fixed in a foot holder. The foot holder connected to one edge of
a 50cm rod that had a joystick-like fulcrum 40cm from the foot holder end. The
foot holder was designed to be able to move perpendicularly to the stick to
allow all saggital movements. The other end of the
rod was connected to a spring allowing the cat’s right hindlimb
to move in a uniform stiffness field in the saggital
plane. The three-dimensional movements of the hip, knee and ankle joints and
the paw were monitored using motion tracking sensors (6D-ResearchTM,
Skill Technologies, Inc.,
In all cats, bipolar intramuscular EMG electrodes were placed
close to the motor points in seven right hindlimb
muscles (lateral gastrocnemius (LG), tibialis anterior (TA), vastus lateralis (VL), semimenbranosus
(SM), posterior biceps femoris (BF), sartorius (SA), and iliopsoas
(IP). The electrodes were insulated except for a 3-mm tip.
In 3
animals, the spinal cord was stimulated through 17-19 microwire
electrodes inserted into the right side of the cord. The electrodes were 30-mm diameter, stainless-steel wires (California Fine Wire,
The resting position of the limb was adjusted to
approximate a normal standing posture. The
following procedures were performed:
1. The muscles
were stimulated individually and the kinematics were
recorded. The stimulus consisted of a train of monophasic
cathodic impulses lasting 0.8 sec. Each pulse had a duration
of 300 ms and a frequency of 50 Hz. The
stimulation current needed to produce a threshold muscle movement was
determined for each muscle using visual observation. We activated each muscle
at stimulus intensities of 1.2, 1.5, 2.0, 2.5, 3.0 and 4.0 times threshold (T).
2. Nerves were stimulated individually through epineural
or cuff electrodes and the EMG and kinematics were recorded. For the
stimulation of nerves, we used stimulus levels of 1.2, 1.5, 2.0, 2.5, and 3.0
T. 3. Intraspinal electrodes were independently
stimulated at stimulus levels of 1.2, 1.5, 2.0, 2.5, and 3.0 T.
In 3 animals, the right dorsal and ventral roots from L4 to S1
were exposed. Dorsal roots were sequentially cut peripherally and hook
electrodes were placed on each root. The roots were stimulated at levels of
1.2, 1.5, 2.0 and 3.0 T. Each ventral root was also stimulated at stimulus
levels of 1.2, 1.5, 2.0 and 3.0 T.
Results
As an
example of the motion analysis, Fig. 1 shows the trajectory of the hip, knee, ankle and paw when
BF was stimulated with 3 T. The distance of the paw movement was 6.3 cm from
the rest to the extreme position. The direction was 143° with respect to the forward direction.
Fig. 1. Sagittal trajectory of the hindlimb during stimulation of BF muscle. The O
and * respectively represent the position of sensors near the hip, knee, ankle
joints and the paw at rest and extreme positions during stimulation. The curved
lines represent the real trajectory of the hindlimb
segments during stimulation. The straight line represents the total vector of the paw.
In
the same way, vectors induced by stimulation of muscles, nerves, ventral and dorsal roots and spinal
cord were analyzed and plotted in polar coordinates. Fig. 2a- e show typical results elicited from single animals.
Stimulation of either muscles or nerves evoked distinct, reproducible movements
to 6 separate directions. Intraspinal stimulation was
able to induce movements to all directions. Ventral root stimulation
produced only downward and backward movements. In contrast, dorsal root
stimulation produced only upward and forward movements. Interestingly, vectors produced by microstimulation
of intermediate spinal cord regions often changed abruptly depending on the
stimulus level, which was very different from muscle, nerve and root
stimulation (Fig. 3).
Fig. 2. Lines radiating from the center of the polar graphs
represent the vectors from rest to extreme position during stimulation for all
stimulus strengths.
The
threshold current for muscle stimulation was 1437 ± 434 mA,
whereas those for nerve, spinal cord, and root stimulation were less than 130 mA. Thus,
muscle stimulation required over an order of magnitude higher currents to
induce a minimal movement.
Fig. 3 Comparison of maximal directional changes produced by
varying the level of stimulus intensity. Maximal degree
differences by spinal cord stimulation (41 ± 37°)
were significantly larger than those obtained by the other four stimulation
locations (p<0.01).
We also compared recruitment curves for stimulation of muscles, nerves,
spinal cord, and roots in terms of amplitude and direction (Fig. 3). Using repeated measures ANOVA, the
following results were obtained: 1. The recruitment curves
elicited by intramuscular stimulation had the most gradual slope and those
elicited by nerve and root stimulation had the steepest slopes. 2. The recruitment curves elicited by intraspinal stimulation curve were intermediate. However,
the coefficient of variation of movements induced by muscle stimulation was significantly larger than
that of nerve stimulation-induced movements.
Discussion
When stimulation pulses were applied through intraspinal
electrodes targeting “movement primitive” regions in the lumbar enlargement,
movements were activated with low stimulus intensity. The generated movement
vectors demonstrated that single muscles or synergistic muscle groups can be
activated [4]. This may not be surprising because individual motoneuronal pools exist as a cluster along the lumbo-sacral segments. However, movement recruitment curves
obtained by stimulating intermediate gray matter regions were rather variable
depending on electrode location. Furthermore, the direction of movement vectors
could dramatically change depending on stimulus strength. Presumably, these
stimuli activated a large number of interneurons,
whose state may change, as well as neighboring motoneuron
pools. Therefore, generating reproducible and selective movements by targeting
“movement primitive” locations may be a challenging aspect for intraspinal stimulation.
Application of
One
difficulty commonly associated with direct neural stimulation is that of
topological selectivity of activation because nerves typically contain axons
that innervate multiple different muscle groups. However, this issue may be
circumvented since we have found that flexor and extensor muscles of the hip,
knee and ankle can be activated selectively through nerve branches. Since nerve
stimulation produced more than 80% of the limits of the passive range of motion
(not shown), this approach has the potential to reproduce whole, lower limb
movements. In addition, the stimulus current required to activate nerve axons
using epineural or cuff electrodes was much lower
than muscle stimulation. One disadvantage of nerve stimulation is the
relatively narrow range of stimulation intensity generating minimal to maximal
movements.
Since ventral roots are composed of mixed nerve groups,
selectivity is relatively poor compared to single nerve or muscle stimulation
and mainly extensor movements are produced. Therefore, stimulation of ventral
roots alone will not induce good functional movements such as walking. However,
stimulation of ventral roots can produce strong extensor activity and may be a
good candidate to restore standing or leg powered cycling for ergometer exercise in which extensor muscles are
predominantly used.
References
[1] Mushahwar,
V.K. and K.W. Horch, Proposed specifications for a
lumbar spinal cord electrode array for control of lower extremities in
paraplegia. IEEE Trans Rehab Eng, 1997. 15: 237-43.
[2] Rushton, D.N., et
al., Lumbar root stimulation for restoring leg function: results in paraplegia.
Artif
Organs, 1997. 21: 180-2.
[3] Giszter, S.F., et
al., Convergent force fields organized in the frog's spinal cord. J Neurosci, 1993. 13: 467-91.
[4] Mushahwar, et al., Spinal cord microstimulation generates functional limb movements in chronically implanted cats. Exp Neurol, 2000. 63: 422-9.
[5] Kralj, A. & Bajd, T.
(1989). Functional Electrical Stimulation, Standing and
Walking after Spinal Cord Injury.
Acknowledgments
This study was supported by CIHR and AHFMR.