Abstract
A train of electrical stimuli properly applied to the posterior structures of the human lumbar cord in chronically spinal cord injured (SCI) subjects is capable to build a certain organisation of spinal reflex pathways which depends on the frequency of the stimulus train, and is reflected in a particular shape of the EMG responses elicited in the lower limb muscles. A closer look to the features of these responses reveals that the neuronal organisation is slowly established following a well-defined, fixed order of configurations of the lumbar interneuronal network.
1.
Introduction
It
has been shown [2] that sustained (tonic) electrical stimulation of posterior
structures of the human lumbar cord, completely or partially isolated from the
brain by SCI, can induce various rhythmical and non-rhythmical patterns of
motor activity in paraplegic subjects (Fig. 1), while the characteristic
responses of the lower limb muscles observed in the electromyographic (EMG)
recordings appear to be the result of posterior root stimulation (posterior
root muscle reflex responses, PRMRR). For different frequencies of the
stimulus train, the EMG recordings reveal PRMRR of different shapes, presumably
reflecting distinct configurations of the interneuronal network within
the lumbar grey matter which is assumed to be activated by the externally
applied stimulation. Using a fixed combination of stimulus parameters,
particularly a constant frequency of the stimulus train, the shape of
successive responses may first change significantly for a certain period of
time from the onset of the stimulation until some kind of “equilibrium” has set
it. This observation led us to the hypothesis that, for each stimulus
frequency, the network of spinal interneurones tends to a temporally stable
organisation, depending on the frequency actually applied, by passing through
different configurations following a “fixed recruitment order” of interneuronal
pathways.
2.
Subjects and Methods
We explored the above formulated hypothesis while recording brain motor control assessment (BMCA) [4] patterns of spinal cord motor responses in five people with chronic, discomplete [1], accidential SCI to epidural stimulation of the posterior structures of the upper lumbar cord. The procedure for the placement of the epidural electrode, and the technique of spinal cord stimulation (SCS) have been described on several occasions, see, e. g., [3]. For SCS the four contacts of the stimulating electrode (0 – 3, 0 being the most rostral, 3 the most caudal one) were used in pairs (cathode, anode) in “bi-polar” stimulation mode with a stimulus rate of 2.1 to 50 Hz, and an amplitude of 1 – 10 V.
The
effects of SCS were verified by poly-EMG recording. EMG activity was recorded
with the subjects in the supine position and the surface electrodes placed
bilaterally over the quadriceps, adductors, hamstring, tibialis anterior,
triceps surae, abdominal and paraspinal muscles. For further analysis, and for
presentation purposes, the obtained EMG stripcharts were converted to
stimulus-triggered raster representations.
3.
Results
Fig.
2 shows the typical shape of a PRMRR observed for the quadriceps muscle group
immediately after the onset of stimulation (“early response”) using a stimulus
frequency of 5 Hz (broken line), as well as an average of 10 successive PRMRR
elicited during a later phase when the shape of the responses was (already)
temporally stable (full line). The diagram also shows the (averaged) latency
times of the responses, and the latencies of their potential peaks. When
successive responses are compared, there may be some “jitter” in the latency
times, but in the very most cases this jitter was within the accuracy of our
EMG recordings, i. e. the standard deviation (SD; calculated over at least 10
responses) of the latencies was smaller or equal 0.5 msec (according to the
sampling rate of our recordings of 2000 per second). When the latency time of a
PRMRR had a SD ³ 0.5 ms,
this was frequently “compensated” until the immediately following potential
peak, the latency of which had a SD £ 0.5 ms again. We also wondered whether the latency times as
well as their jitter depended on stimulus strength and frequency. They showed
no significant differences when the stimulus strength was changed, but the
latency times (of corresponding peaks) seem to decrease slightly with
increasing stimulus frequency. However, responses obtained at a constant
frequency are clearly time-correlated.
In
Fig. 3, 18 successive PRMRR illustrate the transition from the early response
shown in Fig. 2 to the later one which indicates some kind of “equilibrium
state”. In fact, there is a gradual change from one state to the other,
rather than “turning off” one and “turning on” the other. Fig. 4 gives a second
example for the smooth transition from one state, configuration, of the
“interneuronal system” to another. The diagram shows 50 successive PRMRR
elicited at a constant frequency of 31 Hz, and obtained in the quadriceps
muscles some seconds after the onset of the stimulation when the tonic muscle
activity output started to convert to a phasic, rhythmical one. The two
extremes observed (compare first and last response in this raster) may be
referred to as “the early”, short-latency, and “the late”, long-latency
component, respectively. When the “equilibrium state” has fully set in, the
early component may be completely suppressed resulting in a prolonged latency
time of the PRMRR during bursting activity.
The
shapes of the PRMRR shown in Figs. 2 and 4 (late component) were revealed as
typical for posterior root stimulation of the quadriceps muscle group as they
appear repeatedly in the EMG recordings regardless of the frequency
actually used for the stimulation (Fig. 5). In other words, for every
stimulus frequency the responses observed immediately after stimulation
onset had the same shape, and in the following the very same phases
(configurations) were passed through until the responses remained
temporally stable.
The
results obtained for other muscle groups (e. g. hamstring and triceps surae)
are analogous to those presented above for quadriceps.
4.
Summary and Conclusions
The results presented above suggest that a train of electrical pulses properly applied to the posterior structures of the human lumbar cord can be regarded as a conditioning stimulus capable to establish a certain organisation of spinal reflex pathways which depends on the frequency actually used. In parallel, each single stimulus within the pulse train may serve as a test stimulus in order to assess this organisation as well as the transitions between different configurations of the lumbar interneuronal network.
The
fact that PRMRR obtained at a constant frequency are time-locked suggests that
some neuronal organisation has been set up. Our results strongly support the
hypothesis that the transition from one organisation to the other follows a
strict, well-defined order of configurations involving more and more centrally
situated interneurones. Fig. 6 may be an expressive graphic representation of
how these transitions are accomplished. Simply speaking, the higher the
stimulus frequency, the “deeper” the “central” or “core” interneurones that are
finally recruited.
We do believe that the technique described above and used for this study will prove to be a powerful tool in the assessment of the organisation of spinal cord circuitry.
References
[1] M. R. Dimitrijevic. Motor control in chronic
spinal cord injury patients. Scand J Rehab Med, Suppl 30, pp 53–62,
1994.
[2] M. R.
Dimitrijevic, Y. Gerasimenko, M. M. Pinter. Evidence of a Spinal Central Pattern Generator in Humans. In: O. Kein,
R. M. Harris-Warrick, L. M. Jordan, H. Hultborn, N. Kudo (eds). Neuronal Mechanisms for Generating
Locomotor Activity. Annals of the New York Academy of Sciences, Vol. 860, pp
360–376, 1998.
[3]
M. Murg, H. Binder, M. R. Dimitrijevic. Epidural electrical stimulation of
posterior structures of the human lumbar spinal cord: 1. muscle twitches – a
functional method to define the site of stimulation. Spinal Cord 38, pp
394–402, 2000.
[4]
A. M.
Sherwood, W. B. McKay, M. R. Dimitrijevic. Motor Control after spinal cord
injury: assessment using surface EMG. Muscle & Nerve 19, pp
966–979, 1996.