Download Intersegmental synchronization of spontaneous activity of dorsal

Exp Brain Res (2003) 148:401–413
DOI 10.1007/s00221-002-1303-6
RESEARCH ARTICLE
E. Manjarrez · I. Jimnez · P. Rudomin
Intersegmental synchronization of spontaneous activity
of dorsal horn neurons in the cat spinal cord
Received: 30 May 2002 / Accepted: 2 October 2002 / Published online: 28 November 2002
Springer-Verlag 2002
Abstract Extracellular recordings of neuronal activity
made in the lumbosacral spinal segments of the anesthetized cat have disclosed the existence of a set of neurons
in Rexed’s laminae III–VI that discharged in a highly
synchronized manner during the occurrence of spontaneous negative cord dorsum potentials (nCDPs) and
responded to stimulation of low-threshold cutaneous
fibers (<1.5T) with mono- and polysynaptic latencies.
The cross-correlation between the spontaneous discharges
of pairs of synchronic neurons was highest when they
were close to each other, and decreased with increasing
longitudinal separation. Simultaneous recordings of
nCDPs from several segments in preparations with the
peripheral nerves intact have disclosed the existence of
synchronized spontaneous nCDPs in segments S1–L4.
These potentials lasted between 25 and 70 ms and were
usually larger in segments L7–L5, where they attained
amplitudes between 50 and 150 V. The transection of the
intact ipsilateral hindlimb cutaneous and muscle nerves,
or the section of the dorsal columns between the L5 and
L6, or between the L6 and L7 segments in preparations
with already transected nerves, had very small effects on
the intersegmental synchronization of the spontaneous
nCDPs and on the power spectra of the cord dorsum
potentials recorded in the lumbosacral enlargement. In
contrast, sectioning the ipsilateral dorsal horn and the
dorsolateral funiculus at these segmental levels strongly
decoupled the spontaneous nCDPs generated rostrally
from those generated caudally to the lesion and reduced
the magnitude of the power spectra throughout the whole
I. Jimnez · P. Rudomin ())
Department of Physiology, Biophysics and Neurosciences,
Centro de Investigacin y Estudios Avanzados del Instituto
Politcnico Nacional, Apartado Postal 14-740, Mxico, DF 07000,
Mxico
e-mail: [email protected]
Tel.: +1-525-7477099
Fax: +1-525-7477105
E. Manjarrez
Instituto de Fisiologa, Benemrita Universidad Autnoma de Puebla,
Mxico
frequency range. These results indicate that the lumbosacral intersegmental synchronization between the spontaneous nCDPs does not require sensory inputs and is
most likely mediated by intra- and intersegmental
connections. It is suggested that the occurrence of
spontaneous synchronized nCDPs is due to the activation
of tightly coupled arrays of neurons, each comprising one
or several spinal segments. This system of neurons could
be involved in the modulation of the information transmitted by cutaneous and muscle afferents to functionally
related, but rostrocaudally distributed spinal interneurons
and motoneurons, as well as in the selection of sensory
inputs during the execution of voluntary movements or
during locomotion.
Keywords Dorsal horn neurons · Neuronal
synchronization · Cutaneous afferents · Presynaptic
inhibition · Spinal cord
Introduction
In a previous study (Manjarrez et al. 2000) we analyzed
the origin of the spontaneous activity that is recorded
from the dorsum of the lumbosacral enlargement of the
spinal cord in the anesthetized cat. There were clear
indications that the low frequency components of these
potentials (3–20 Hz) were produced by a set of spontaneously active neurons located in the dorsal horn, some of
which responded with monosynaptic latencies to stimulation of low-threshold cutaneous afferents. During the
generation of these spontaneous cord dorsum potentials
(nCDPs), the monosynaptic intraspinal field potentials, as
well as the dorsal root potentials (DRPs), produced by
stimulation of low threshold cutaneous and by Ib muscle
afferents, were facilitated, while the intraspinal field
potentials and the DRPs produced by stimulation of Ia
afferents were depressed. In addition, the spontaneous
nCDPs and the fluctuations of monosynaptic reflexes
were reduced by conditioning stimulation of cutaneous
nerves, suggesting a continuous modulation of the motor
402
output by this set of dorsal horn neurons (see Rudomin
and Dutton 1969; Rudomin et al. 1974).
In the earlier study (Manjarrez et al. 2000), we
suggested that the ensemble of active dorsal horn neurons
that leads to the spontaneous nCDPs fires in a highly
synchronized manner and is longitudinally distributed
throughout several spinal segments. In 1979, Brown and
colleagues examined in the spinal cord of the cat the
correlation between the spontaneous firing of dorsal horn
neurons, and found a significant number of neuron pairs
with non-flat cross-correlograms, suggesting the existence
of a common input to these neurons. Subsequent studies
by Eblen-Zajjur and Sandkuhler (1997) in the rat have
confirmed the findings of Brown et al. (1979). In addition,
they indicated that the dorsal horn neurons with correlated
discharges showed the highest incidence of overlapping
receptive fields. However, in these studies, very little
information was presented on mechanisms and/or pathways contributing to the intersegmental synchronization
of the spontaneous activity of these neurons.
One question that emerged from the work of Manjarrez
et al. (2000) was the extent to which the arrays of neurons
involved in the generation of the spontaneous nCDPs
operated mainly within the L5–L7 spinal segments, or
whether synchronized spontaneous nCDPs were also
generated in other spinal regions. Another question was
the extent to which the generation of spontaneous nCDPs
was a consequence of the acute transection of muscle and
cutaneous nerves that was routinely performed during the
dissections. Both issues were taken up recently by
Manjarrez et al. (2002c). Preparations with intact peripheral nerves showed spontaneous nCDPs in segments C3–
C4 as well as in segments L5–L7. Yet, spectral analysis of
the cord dorsum potentials simultaneously recorded from
the lumbar and cervical regions showed very little
coherence between them, suggesting they were generated
by two independent populations of neurons.
The purpose of the present series of observations has
been to further characterize the population of dorsal horn
neurons located within the lumbosacral regions of the
spinal cord that discharges in synchrony with the
spontaneous nCDPs, to examine the segmental distribution of the involved networks and, by means of spinal
lesions, to disclose the intraspinal pathways mediating
their synchronized intersegmental firing. Some of these
observations have been published in abstract form
(Manjarrez et al. 1998).
Materials and methods
General procedures
Guidelines contained in NIH publication 85–23 (revised in 1985)
on the principles of laboratory animal care were followed
throughout. The experimental procedures were similar to those
followed previously (Manjarrez et al. 2000). Briefly, observations
were carried out in adult cats initially anesthetized with pentobarbitone (35 mg/kg weight, i.p.), supplemented during the dissection
and recording periods with additional doses of 10 mg/kg i.v. as
necessary to maintain an adequate level of anesthesia, which was
assessed by verifying that the pupils were constricted, that blood
pressure was stable (between 100 and 120 mmHg), and that it was
unaffected by noxious stimulation to the skin.
In all experiments the lumbosacral and low thoracic spinal
segments were exposed and the left L4–S1 ventral roots sectioned.
In 18 experiments the left posterior biceps and semitendinosus
(PBSt), gastrocnemius (GS), deep peroneus (DP), sural (SU) and
superficial peroneus (SP) nerves were dissected free, sectioned and
their central ends prepared for stimulation. In four other experiments these nerves were exposed and left intact. A pair of stainless
steel needles was inserted into the left foot pad for stimulation.
After the surgical procedures, the animals were paralyzed with
pancuronium chloride (0.1 mg/kg) and artificially ventilated. Tidal
air volume was adjusted to a 4% CO2 concentration in the expired
air. Exposed tissues were covered with mineral oil to prevent
desiccation and kept at constant temperature (37C) by means of
radiant heat.
Stimulation and recording
Evoked and spontaneous potentials were recorded from the dorsal
surface of the lumbosacral spinal cord along several segments. To
this end we used pairs of silver ball electrodes, one placed on the
surface of the spinal cord and the other on the paravertebral
muscles at the same segmental level. The preamplifier filters were
set to 0.3 Hz in the low range and 10 kHz in the high range. In all
figures, negativity in recordings of cord dorsum potentials is
upwards. Action potentials from dorsal horn neurons and field
potentials were recorded from the L6–L7 segments using glass
micropipettes filled with 1.2 M NaCl solution. The resistance of the
microelectrodes varied between 7 and 15 MW. In several experiments simultaneous recordings were made from pairs of dorsal
horn neurons in layers III–VI within the L6–L7 segments using two
separate micropipettes inserted at angles to each other. The
approximate distance between neurons was estimated from the
longitudinal separation of the tips of the recording micropipettes
measured when both were touching the cord surface. Clearly this is
an overestimate because the micropipettes converged at deeper
regions.
Sampling and averaging of potentials was triggered with pulses
generated by a window discriminator activated either by the
spontaneous nCDPs recorded from a given (reference) spinal
segment, or by stimuli applied to the cutaneous nerves or the
ipsilateral foot pad in preparations with intact nerves (single pulses
1.1–1.5 times the strength of the most excitable fibers, determined
relative to the threshold of the smallest afferent volley recorded
from the cord dorsum, T). The averages were made up of 32–64
successive potentials. All potentials and triggering pulses were
stored on a digital tape for further processing.
The interaction between the neuronal sets producing the
spontaneous nCDPs and neurons responding to stimulation of
cutaneous afferents (evoked responses) was examined using the
same protocol as Manjarrez et al. (2000). Briefly, every time a
spontaneous nCDP crossed a fixed voltage level, set with a window
discriminator, a TTL pulse was sent to the input channel of a
control device that triggered a sequence of no stimulus, a stimulus
at a delay after the nCDP, or a stimulus >500 ms from an nCDP
(control). Potentials for each condition were collected separately.
On the completion of one cycle of stimulation, the control device
prevented initiation of the next cycle until after 2 s.
Spectral analysis
Spectral analysis of 2-min recordings of cord dorsum potentials
obtained in the absence of stimulation of sensory nerves was done
in ten experiments, using the Fourier transform of the Chaos Data
Analyzer program of the American Institute of Physics Academic
Software. Spontaneous cord dorsum potentials were acquired with a
sampling rate of 500 Hz.
403
Coincidence histograms
The relationship between spontaneous neuronal firing and the
spontaneous nCDPs was established for each neuron by aligning
the raster displays of the spontaneous action potentials with the
peak of the averaged spontaneous nCDPs recorded close to the site
of micropipette insertion (see Fig. 1). Coincidence histograms were
obtained by adding the raster displays of all neurons (see Figs. 1, 2).
Histograms were constructed using 1-ms bins within a range of
200–250 ms.
Cross-correlation between action potentials generated
by pairs of neurons
Cross-correlation between the spontaneous activity of pairs of
neurons was calculated using the cross-correlation function (see
Fig. 3A, B). To this end we used recordings of the activity of pairs
of neurons taken during the occurrence of a spontaneous nCDP.
Correlation strength (F) was expressed as the ratio between the
peak correlation and the basal correlation (Fig. 3C; see EblenZajjur and Sandkuhler 1997).
Spinal lesions
In ten experiments we investigated the effects of acute spinal
lesions on the spontaneous nCDPs that were simultaneously
recorded from several segments in the lumbosacral enlargement.
Fig. 1A–F Synchronized spontaneous and evoked activity of
neurons in laminae III–VI. A Raster display of the spontaneous
activity of 17 neurons during the occurrence of spontaneous
nCDPs. Neuronal activity was recorded between 0.71 and 1.92 mm
below the cord surface. All responses were aligned relative to the
peak of the spontaneous nCDPs recorded at the site of electrode
penetration. The small boxes to the right indicate data obtained
from the same interneuron. B Histogram generated by adding the
raster displays illustrated in A. C Upper trace Extracellular
recording of the activity of a single neuron in lamina IV at
1.26 mm depth during the generation of a spontaneous nCDP; lower
The lesions were made under the dissecting microscope by teasing
the spinal tissue with two fine forceps between the L6 and L7 or
between the L5 and L6 segments.
Histology
At the end of the experiment the animals were sacrificed with a
lethal dose of pentobarbitone, perfused with 10% formalin, and the
spinal cord removed. After fixation and dehydration, the lumbosacral segments were placed in a solution of methyl salicylate for
clearing and cut transversally to obtain 100-m sections including
the lesions (see Wall and Werman 1976).
Results
Spontaneous activity of neurons
In six experiments, spontaneous activity was recorded
extracellularly from neurons located within the lumbar
segments L6 and L7, in regions corresponding to Rexed’s
laminae III–VI (0.7–2.0 mm depth). No tests were
performed for their antidromic activation from thoracic
segments to determine if these were interneurons or
ascending tract neurons (see “Discussion,” however). The
trace nCDP simultaneously recorded at the site of the electrode
penetration. Note the neuronal burst-like discharge and the
extracellular field potential appearing in synchrony with the nCDP.
D–F Responses of the same neurons produced by single pulses
(1.1–1.5T) applied to the SU or SP nerves. Same format as in A–
C. Raster displays in D were ordered according to response latency.
In all figures, recordings of cord dorsum potentials are with
negativity upward. Recordings of intraspinal potentials are with
positivity upward. Voltage calibrations in C and F are for cord
dorsum potentials
404
Fig. 2A–F Spontaneous activity and evoked responses of unsynchronized neurons. Same format as that of Fig. 1. Neuronal activity
was recorded between 0.75 and 2.0 mm depth. Neuronal recordings
in C and F were made in laminae IV at 1.31 mm depth. Voltage
calibrations in C and F are for cord dorsum potentials
activity of 17 neurons was increased during the occurrence of the spontaneous nCDPs recorded from L6 or L7
(synchronic neurons, see Fig. 1A–C). This is consistent
with the observations of Sandkuhler and Eblen-Zajjur
(1994) in the anesthetized rat, who also found neurons in
laminae III–VI with spontaneous burst-like activity. There
were, in addition, 11 neurons located in the same spinal
regions whose activity was not synchronized with the
spontaneous nCDPs (see Fig. 2A–C).
Figure 1A shows the raster displays of the spontaneous
discharges of 17 dorsal horn synchronic neurons that were
recorded in different experiments during the generation of
the spontaneous nCDPs. Recording of neuronal activity
was made between 0.7 and 1.92 mm below the cord dorsal
surface, 0.7–1.0 mm from the midline. The histogram in
Fig. 1B was constructed by adding the activity of all
neurons before and during the spontaneous nCDPs
recorded close to the site of electrode penetration. The
shape of the histogram is similar to that of the spontaneous nCDPs (bottom trace in Fig. 1C).
For comparison, Fig. 2A–C shows the raster displays
and histogram of the spontaneous activity of neurons
whose activity was not clearly synchronized with the
spontaneous nCDPs. These neurons were in the same
spinal region as the synchronic neurons (0.75–2.0 mm
depth). In contrast with the histogram of Fig. 1B, the
histogram displayed in Fig. 2B is rather flat, and shows no
resemblance to the spontaneous intraspinal field potentials recorded with the same electrode, or with the
spontaneous nCDPs (Fig. 2C).
Responses to stimulation
of low-threshold cutaneous afferents
To explain the facilitatory interactions observed between
the spontaneous nCDPs and the responses evoked by
stimulation of a cutaneous nerve, Manjarrez et al. (2000)
proposed that the dorsal horn neurons that were activated
during the spontaneous nCDPs also responded to stimulation of low-threshold skin afferents. To test this
proposal further, we have examined the responses of the
synchronic and non-synchronic neurons to stimulation of
low-threshold cutaneous afferents from the SU and SP
nerves.
Figure 1D shows the raster displays obtained from the
responses of 17 dorsal horn synchronic neurons to
stimulation of the SP or SU nerves with single pulses
1.1–1.5T. The neuronal activity was aligned relative to
the time of arrival of the afferent volley recorded from
L6. The raster displays of each neuron were ranked
according to their response latencies. It may be seen that
the histogram constructed with the whole set of neuronal
responses has a shape that resembles that of the evoked
intraespinal field and cord dorsum potentials (Fig. 1E, F).
As illustrated in Fig. 1D, most neurons responded with
onset latencies between 1 and 3 ms after the arrival of the
afferent volley. The shortest values (0.7–1.2 ms) suggest
monosynaptic activation and the longer values (1.5–3 ms),
oligosynaptic connections (see Brown 1981; Koerber and
Brown 1982; Willis and Coggeshall 1991). The activation
threshold of the synchronic neurons to stimulation of the
cutaneous afferents varied between 1.1 and 1.5T, and
had no clear relationship to response latency.
405
Fig. 3A–C Correlation between spontaneous firing of pairs of
synchronic neurons. A Cross-correlation histogram of spontaneous
activity of two synchronic neurons located within lamina IV at 1.26
and 1.25 mm depth, separated by 2 mm along the rostrocaudal axis.
B Cross-correlation histogram of two synchronic neurons located
between laminae IV and V at 1.45 and 1.36 mm depth and
longitudinally separated by 7.4 mm. C Strength of correlation (F)
of four pairs of neurons versus longitudinal separation between the
recording micropipettes. For further explanations see text
Figure 2D, E shows the raster display and histogram
obtained from the responses of 11 non-synchronized
neurons to stimulation of the SP or SU nerves (1.1–
1.5T). These neurons responded with latencies ranging
from 1.9 to 8.3 ms (Fig. 2D). There was no significant
difference in the activation thresholds of synchronic
(1.2€0.2T) and non-synchronic neurons (1.2€0.5T)
(P>0.05).
firing relative to the other neuron during the generation of
the spontaneous nCDPs was lower.
Altogether it was possible to have stationary recordings of the activity of four pairs of synchronic neurons. As
shown in Fig. 3C, the correlation strength (F) between the
spontaneous activity of these four pairs of synchronic
neurons during the spontaneous nCDPs decreased with
increasing separation between the tips of the two recording micropipettes, but was still detectable by 7.4 mm.
Although correlation strength is not a direct measure of
the coefficient of correlation, these data indicate that the
synchronization between neuronal activity at the time of
the spontaneous nCDPs decreases with increasing separation between neurons.
Cross-correlation between the activity of pairs of neurons
The results described in the preceding sections have
indicated that a substantial number of dorsal horn neurons
discharge in synchrony during the generation of the
spontaneous nCDPs in the same spinal segment. This
raised the question as to the extent to which longitudinal
distance between neurons affected the magnitude of their
cross-correlation. To this end we made simultaneous
recordings of the spontaneous activity of pairs of
synchronic neurons located in laminae III–VI within the
L7 and L6 segments by means of two micropipettes
whose tips were separated one from the other between 2
and 7.4 mm along the rostrocaudal axis in different
experiments.
Figure 3A shows the cross-correlation histogram of the
activity of two dorsal horn synchronic neurons located in
the more ventral part of lamina IV that were separated by
2 mm. This histogram had a distinct peak relative to the
basal values, indicating highly correlated activation
during the generation of the spontaneous nCDPs (see
Inbar et al. 1979). The calculated correlation strength (F)
for this pair of neurons was 5.2. For comparison, the
histogram in Fig. 3B shows the cross-correlation histogram of a pair of synchronic neurons located between
laminae IV and V, which were separated by 7.4 mm. This
pair of neurons had a lower correlation strength (2.4).
This means that the probability of having one neuron
Intersegmental synchronization
of the spontaneous nCDPs
Segmental patterns of spontaneous nCDPs
We have suggested previously (Manjarrez et al. 2000)
that the set of dorsal horn neurons that fire in synchrony
with the spontaneous nCDPs is distributed throughout
several spinal segments in a longitudinal array. The data
depicted in Fig. 3C suggest that this array may not extend
more than 10 mm rostrally and 10 mm caudally in the L6–
L7 segments. However, the small number of neurons
examined precludes a proper estimate of the longitudinal
extension of this array.
To gain more information on the longitudinal extension of the synchronized activity of the dorsal horn
neurons during the occurrence of the spontaneous nCDPs,
simultaneous recordings of cord dorsum potentials were
made from several spinal segments (usually from S1–L4).
Figure 4 shows samples of spontaneous nCDPs that were
recorded in the same preparation simultaneously from
four spinal segments (L4–L7). In some cases they were
larger in L4 and L5 than in L6 and L7 (open square in
406
Fig. 4A–F Synchronized cord
dorsum potentials in several
spinal segments. The figure
shows samples of nCDPs simultaneously recorded from
four different spinal segments
in a single experiment within a
10-min time period. Vertical
dotted lines show onset of the
earliest potential in each array
of synchronous discharges.
Symbols indicate different patterns of intersegmental nCDPs.
For further explanations see text
Fig. 4A), while in other cases they were larger in L6 and
L5 than in L4 or L7 (open square in Fig. 4B), or in L7 and
L6 than in L5 and L4 (filled diamond in Fig. 4F).
Synchronized spontaneous nCDPs could also appear in
three (filled circles in Fig. 4C, E) or in two segments
(filled squares in Fig. 4C, F), and also remain confined
within one segment only (either in L6, open diamonds in
Fig. 4A, B, or in L7, open circles in Fig. 4D, E).
Although we have not performed a statistical analysis
of the likelihood of occurrence of specific patterns of the
synchronized intersegmental nCDPs, some of them
appeared to be generated more often of what one may
expect from random variations (see Roerig and Feller
2000). To illustrate this point, in Fig. 5A we have
superimposed recordings obtained from the same experiment as that of Fig. 4, in which the spontaneous nCDPs
appeared simultaneously in four spinal segments (L4–L7).
Figure 5B shows instead recordings in which the spontaneous nCDPs appeared only in three segments (L4, L5
and L6), together with a small slow positive wave in L7.
The nCDPs displayed in Fig. 5C appeared mainly in
segments L4 and L5, while the potentials illustrated in
Fig. 5D appeared only in segment L7 and were timelocked to a small, slow positive wave in segments L4–L6.
Averaging procedures were used to disclose those
nCDPs that appear in synchrony in different spinal
segments. To this end, the nCDPs recorded at a given
location (reference nCDPs) were fed to a window
discriminator. A single pulse was generated whenever
the reference nCDPs exceeded a predetermined value,
usually set to include responses above 30% of maximal
amplitude. These pulses were used to trigger the averaging of the cord dorsum potentials in all segments.
Fig. 5A–D Intersegmental patterns of spontaneous nCDPs. Data
derived from the same set of records as that of Fig. 4. A–D
Superposed records of the spontaneous nCDPs that were simultaneously recorded from four different spinal segments. These
potentials were selected on the basis of their intersegmental
patterns. Traces were aligned according to the onset time of the
nCDPs recorded in one segment (L6 in A and B, L5 in C and L7 in
D). The time relation between the reference nCDPs and the
potentials recorded in the other segments was preserved in all cases.
Vertical dotted lines indicate estimated onset times of the reference
potentials. For further explanations see text
407
Fig. 6A–L The intersegmental distribution of synchronized spontaneous nCDPs in the lumbosacral enlargement is not affected by
peripheral nerve transection. A–C Superposed traces of averaged
spontaneous nCDPs simultaneously recorded in four different
segments before and after transection of cutaneous and muscle
nerves in the ipsilateral hindlimb. The triggering pulses for the
averaging were generated by the spontaneous nCDPs recorded in
segments L4, L5, L6 or L7–S1, as indicated.The vertical arrows
show the estimated onset of the L5 and L6 nCDPs. In A, the other
arrow points to the nCDP recorded after the nerve transection. D–F
Superposed averages of spontaneous nCDPs triggered by cord
dorsum potentials recorded in different segments and of spontaneous nCDPs followed by stimulation of the ipsilateral foot pad with
pulses 1.2T taken before peripheral nerve transection. G–I Cord
dorsum potentials produced by stimulation of the foot pad with
pulses of different strengths, as indicated. J–L Superposed power
spectra of cord dorsum potentials recorded from different segments
before and after peripheral nerve section. For further explanations
see text
Figure 6A–C shows the superposed averages of the
spontaneous nCDPs recorded from four different segments in a preparation with intact hindlimb nerves and
after some of these nerves were sectioned (see below).
The intersegmental distribution of the spontaneous
nCDPs varied according to which nCDPs were used as
reference for the averaging procedure. In the experiment
of Fig. 6, when the triggering pulses were generated by
the L5 or L6 nCDPs, large synchronized nCDPs were
recorded in segments L5 and L6 and small potentials were
recorded more caudally or rostrally (Fig. 6A, B). In
contrast, when the averages were obtained using as
reference the nCDPs recorded in the border between L7
and S1, the nCDPs remained confined within that region
(Fig. 6C). It thus seems that in this particular experiment
the neuronal arrays involved in the generation of the L4–
L6 spontaneous nCDPs were different from those involved in the generation of the L7–S1 nCDPs. In other
experiments the averages obtained using as reference the
spontaneous L6 or L7 nCDPs disclosed time-locked
nCDPs in segments L5 to S1 (Fig. 7A, B).
Time shifts between spontaneous nCDPs
Differences in onset latencies between averaged nCDPs
recorded from different segments were difficult to
estimate because of their slow rising phase. Yet, as
shown by the arrows in Fig. 6A, when triggering the
averages with the L5 nCDPs, the potentials generated in
L5 seemed to appear 4.6 ms before those recorded in L6.
In contrast, when the averages were obtained using as
reference the L6 nCDPs, the time shift between the L5
and L6 nCDPs was rather small (Fig. 6B).
Latency shifts between the nCDPs recorded in different segments were also observed without averaging. As
shown in Fig. 5A, the L5 and L6 nCDPs had a similar
latency and preceded the L4 and L7 nCDPs by 5.8 and
9.0 ms, respectively. However, when the nCDPs appeared
only in the L4–L6 segments, they started at about the
same time in the three segments (Fig. 5B). It is likely that
nCDPs with similar onset times were generated by a
common input. On the other hand, the time shifts between
spontaneous nCDPs recorded from different segments
could indicate that these potentials were generated in one
segment and propagated to other segments. Yet, they
could be also generated by a common input and the
408
Fig. 7A–J Sectioning the ipsilateral dorsolateral funiculus and the
ipsilateral dorsal horn reduces segmental synchronization of
spontaneous nCDPs. Data obtained from preparations with acute
transection of ipsilateral muscle and cutaneous hindlimb nerves. A–
E Effects on the averaged spontaneous nCDPs by a lesion
comprising the ipsilateral dorsolateral funiculus between segments
L6 and L7. Averages were obtained by triggering with pulses
generated by the spontaneous L6 or L7 nCDPs, as indicated. The
extent of the lesion is shown in C. F–J Effects produced by
sectioning the ipsilateral dorsal horn and columns. Histology shown
in H. For further explanations see text
latency shifts would result from differences in conduction
time to the different segments.
As reported by Manjarrez et al. (2000), the first
(monosynaptic) component of the cord dorsum responses
evoked by stimulation of cutaneous afferents was facilitated when preceded by spontaneous nCDPs. We have
found that the magnitude of the facilitation of the evoked
responses recorded in different segments varied according
to which nCDPs were used as reference for the averaging.
In the experiment of Fig. 6, facilitation of the monosynaptic component of the evoked responses recorded in L5
and L6 was larger when using as reference the L7–S1
potentials (Fig. 6F) than when using the L5 or the L6
potentials (Fig. 6D, E). In other experiments (not
illustrated) the evoked nCDPs recorded in these segments
were all facilitated when preceded by spontaneous
nCDPs, regardless of the segmental origin of the nCDPs
used as reference for the averaging.
Evoked CDPs
In addition to the spontaneous nCDPs, we also recorded
the cord dorsum potentials produced by stimulation of the
foot pad in preparations with intact hindlimb nerves, or by
stimulation of either the ipsilateral SU or SP nerve in
preparations with acutely transected nerves. In these
experiments, stimulus strength was adjusted between 1.1
and 1.3T to produce responses of about the same
amplitude as the spontaneous nCDPs. Figure 6G, H and I
shows the cord dorsum potentials produced by stimulation
of the ipsilateral foot pad with pulses 1.15, 1.2 and 1.25T
strength. These responses appeared in the four segments,
but were clearly larger in the L5 and L6 segments. A
similar rostrocaudal distribution was obtained following
stimulation of the SP or SU nerves, in confirmation of
previous studies (Bernhard 1953; Bernhard and Widen
1953; Koerber and Brown 1982). It should be noted in
Fig. 6 that the rostrocaudal distribution of the evoked
responses (Fig. 6G–I) was similar to that of the spontaneous nCDPs triggered by the L5 and L6 nCDPs
illustrated in Fig. 6A, B.
Effects of spinal cord lesions
on the intersegmental synchronization
of the spontaneous nCDPs
Most of the observations performed in this study were
made in preparations in which several muscle and
cutaneous nerves were acutely transected during the
dissection (see “Materials and methods”). This raised the
409
Fig. 8A–C Effects of spinal lesions on the frequency power
spectra recorded from different
spinal segments. Data obtained
from preparations with ipsilateral acute transection of muscle
and cutaneous hindlimb nerves.
Upper graphs in each panel
show the power versus frequency spectra recorded at various segmental levels, as
indicated. Power spectra are
expressed as a percentage of
maximum before the spinal lesion. The lower sets of graphs
show the areas of the power
spectra. Thick lines before, fine
lines after, a spinal lesion between segments L6 and L7. A
Effects produced by a lesion to
the ipsilateral dorsal columns.
For histology see open circle in
Fig. 9A. B Effects produced by
a lesion comprising the ipsilateral dorsal horn (see filled triangle in Fig. 9C). C Effects
produced by a dorsal hemisection (see square in Fig. 9E). For
further explanations see text
question as to the extent to which acute nerve transection
was responsible for the limited longitudinal distribution of
the spontaneous nCDPs within the lumbar enlargement. In
four experiments we recorded the spontaneous nCDPs
from preparations in which the left PBSt, GS, SU, SP and
DP nerves were dissected free, but left intact (see
“Materials and methods”). We found that transection of
these nerves had a relatively small effect on the amplitude
and segmental distribution of the synchronized nCDPs
(Fig. 6A–C).
In three preparations in which these nerves were cut
during the dissection, we found that sectioning the
ipsilateral dorsal columns between the L5 and L6
segments also had very small effects on the intersegmental synchronization of the spontaneous nCDPs. This lesion
was aimed to interrupt not only residual sensory inputs
from the periphery, but also the possible synchronization
due to transsegmental conduction of action potentials
generated by primary afferent depolarization in the
intraspinal terminals of the sensory fibers (see Rossignol
et al. 1998).
In contrast to the ineffectiveness of these lesions, we
found that transection of the ipsilateral dorsal horn (three
experiments) and/or the dorsolateral funiculus (three
experiments) markedly reduced the intersegmental synchronization of the spontaneous nCDPs. As shown in
Fig. 7A–E, after a lesion to the ipsilateral dorsolateral
funiculus, the spontaneous nCDPs recorded rostrally to
the lesion were largely (but not completely) uncoupled
from the potentials recorded caudally to the lesion. The
uncoupling became more pronounced when the lesion
also included the dorsal horn (Fig. 7F–J), and was almost
complete following a dorsal bilateral hemisection (not
illustrated; see Fig. 9E for histology).
410
Fig. 9A–F Summary of the changes in the segmental distribution
of frequency power spectra produced by spinal lesions. Data
obtained from preparations with ipsilateral transected hindlimb
nerves. A, C, E Segmental distribution of the normalized areas of
the frequency power spectra. Integration procedures as in Fig. 8. B,
D and F The same but after sectioning the ipsilateral dorsal
columns (B), the ipsilateral dorsal horn (D) and the dorsal spinal
half (F), between L6 and L7. Vertical dotted lines indicate the
segmental level of the spinal lesion. Data and histological
reconstruction of spinal lesions from individual experiments are
indicated by the same symbol
Power spectra
experiment of Fig. 6, the power spectrum of the potentials
recorded in the L5 and L6 segments was highest in the
low frequency range (between 2 and 4 Hz; see Fig. 6J–L).
In other experiments (Fig. 8A) the power spectrum of the
potentials recorded at L6 showed two peaks, one at 5 and
the other at 10 Hz, while at L5 and L7 the power spectra
showed one peak only (at about 7 Hz). In the experiment
of Fig. 8B the 5-Hz peak was of about the same
magnitude in segments L6–S1, while the 10-Hz component was largest in L7 and S1. Finally, in the experiment
of Fig. 8C, the 5- and 10-Hz components were largest in
L6. The factors determining the features of the power
spectra of the cord dorsum potentials recorded in different
preparations have not been elucidated.
In two preparations with intact hindlimb nerves we
observed that transection of the left PBSt, SU and SP
produced a small increase of the power spectra of the
spontaneous cord dorsum potentials in segment L6 and
practically no changes in the adjacent segments (Fig. 6J–
L)
To facilitate the analysis of the effects produced by the
spinal lesions, the power spectra were integrated. Clearly
this procedure precludes detection of changes in a specific
frequency range. Yet, it provides qualitative information
In a previous study made in anesthetized preparations
with acutely transected muscle and cutaneous nerves (see
Manjarrez et al. 2000), we reported that the cord dorsum
potentials recorded from L6 or L7 had dominant
frequencies between 3 and 20 Hz. These frequency
spectra appear to be basically the same as those recorded
in anesthetized preparations with intact hindlimb nerves
(see Manjarrez et al. 2002c).
Simultaneous recordings of cord dorsum potentials
from different spinal segments in anesthetized preparations with intact or acutely transected muscle and
cutaneous nerves have now indicated that within the
lumbosacral segments the low frequency components
have a restricted rostrocaudal distribution (see Figs. 6J–L,
8 and Manjarrez et al. 2002c). In two preparations the
region of maximal power was confined within the L6 and
S1 segments (Fig. 8A, B), and in a third preparation it was
located between the L5 and L7 segments (Figs. 6J–L, 8C,
9A, C, E).
Another feature that emerged from the analysis of
these power spectra was that the dominating frequencies
were not the same in all segments. For example, in the
411
of changes in the whole frequency range. Figure 8A
shows that a small lesion comprising the ipsilateral dorsal
column between segments L6 and L7 made in a
preparation with acutely transected ipsilateral hindlimb
nerves had very little effect on the frequency spectrum of
the spontaneous cord dorsum potentials, both rostrally and
caudally to the lesion (see also Fig. 9A, B).
In contrast, as illustrated in Figs. 8B and 9D,
sectioning the dorsolateral funiculus in the ipsilateral
side had more pronounced effects on the power spectrum,
particularly on the low frequency components of the
potentials recorded from L7 and S1, that is, caudally to
the lesion. A dorsal bilateral transection had even more
dramatic effects, both rostrally and caudally to the lesion
(Figs. 8C, 9F).
Discussion
Features of the synchronized spontaneous activity
of dorsal horn neurons
The present study has indicated that in Rexed’s laminae
III–VI of the lumbosacral spinal cord of the anesthetized
cat, there is a substantial number of neurons that fire in
synchrony during the generation of the spontaneous
nCDPs. These findings are in agreement with our
previous proposal based on the analysis of field potentials
(Manjarrez et al. 2000) that the neurons that discharge
during the generation of the spontaneous nCDPs are
located within the same region where low-threshold
cutaneous afferents terminate. We also found neurons in
the same spinal regions that did not fire in synchrony with
the spontaneous nCDPs. Since these neurons also responded to stimulation of cutaneous nerves, the absence
of correlated activity could be due to the small excitatory
drive, below firing threshold, received at the time of
generation of the spontaneous nCDPs (see O’Donovan
1999). However, since the available samples were rather
non-homogeneous, they might also represent another
functional type of neuron.
Our data also indicate that the population of neurons
with synchronized spontaneous activity is longitudinally
distributed along several lumbosacral spinal segments, in
a similar manner to the neurons responding to stimulation
of cutaneous nerves (see Fig. 6, and Bernhard 1953;
Bernhard and Widen 1953; Koerber and Brown 1982).
However, in agreement with the findings of Lidierth and
Wall (1998) in the rat, the spontaneous synchronous
discharge of these neurons appears not to depend on the
input supplied by primary afferents, because sectioning
several ipsilateral hindlimb muscle and cutaneous nerves
(see Fig. 6), as well as the dorsal columns between L6 and
L7, had very little effect on the segmental distribution of
the synchronized nCDPs.
It thus seems that the synchronization of the spontaneous activity of the laminae III–VI neurons emerges
from local interactions, as suggested by the finding that a
section of the ipsilateral dorsal horn and of the ipsilateral
dorsolateral funiculus largely uncouples the spontaneous
nCDPs generated rostrally from those generated caudally
to the lesion (Fig. 7; see also Kerkut and Bagust 1995;
Lidierth and Wall 1998). Yet, these may not be the only
pathways mediating the intersegmental synchronization
of the neuronal activity (see Kremer and Lev-Tov 1997).
In this regard, it should be noted that the transections
comprising the dorsal horn and the dorsolateral funiculus
also reduced the magnitude of the integrated frequency
power spectra of the spontaneous cord dorsum potentials
recorded both rostrally and caudally to the lesion (Figs. 8,
9). This could be due to the reduction of mutually
excitatory, self-potentiating interactions between adjacent
sets of neurons that increase the excitability of the whole
ensemble and synchronize their spontaneous activity (see
Lidierth and Wall 1996; O’Donovan 1999; Roerig and
Feller 2000).
The generation of a particular segmental combination
of synchronized nCDPs could be due to the activation of
specific sets of tightly coupled arrays of neurons. The
excitation of any neuron within a given array, either by
sensory or by spinal inputs, would activate, in a rather
stereotyped manner, the neurons belonging to that
particular ensemble (see Silberberg et al. 2002). This
explains the remarkable similarity between the segmental
distribution of the responses produced by stimulation of
the ipsilateral foot pad and the spontaneous nCDPs
recorded before as well as after the transection of muscle
and cutaneous nerves in the ipsilateral hindlimb (see
Fig. 6A, C, G–I).
In the absence of synchronized descending influences,
such as those generated during the execution of voluntary
movements (Baker et al. 1999), or during real or fictive
locomotion (Arshavsky et al. 1997; Grillner 1981), the
activation of the different arrays would be random, as
suggested by the data depicted in Fig. 4, which show that
the different intersegmental patterns of nCDPs may occur
more or less independently during the same recording
period. In the anesthetized cat, this activity appears to be
confined within the lumbosacral segments and has little
relation to the neuronal arrays involved in the generation
of the spontaneous nCDPs in the cervical segments
(Manjarrez et al. 2002c). Yet, these two arrays could
become synchronized by common inputs in other conditions, such as during locomotion (Kremer and Lev-Tov
1997; Pearson 2000).
Morphological substrates
for intersegmental synchronization of nCDPs
Although there are many studies on the different types of
neurons in the dorsal horn, information on their intra- and
intersegmental connectivity is rather limited (see Willis
and Coggeshall 1991 for review), and does not allow
disclosure of the neurons that are involved in the
intersegmental synchronization of the spontaneous
nCDPs.
412
Lidierth and Wall (1996) have concluded that the
isolated spinal cord of the anesthetized rat contains an
oscillatory mechanism that is synchronized by way of
intrinsic axons in the Lissauer tract and produces lowvoltage, low-frequency, DRPs on all lumbar dorsal roots
(see also Lidierth and Wall 1998). It is therefore possible
that the synchronization of the spontaneous activity of the
neurons in laminae III–VI observed in our experiments
during the spontaneous nCDPs is due, at least in part, to
the inputs they receive from neurons in lamina II.
However, using the Golgi method, Matsushita (1969)
found in the cat that cells of lamina II not only connect
with other cells in laminae II–VI, but also that cells in
laminae IV and V are connected with neighboring cells in
laminae III–VI. In addition, Schneider (1992) and
Schneider et al. (1995) have described in the spinal cord
of the hamster cells in laminae III–V with local axons and
mutual excitatory and inhibitory interconnections
(Schneider 2001), as well as neurons with deep axons
that usually bifurcate into dorsal and caudal daughter
branches up to 2.5 mm long, giving off collaterals ventral
to the cell body and dendrites. It is therefore possible that
the local axon interneurons are involved in the generation
of spontaneous nCDPs that remain confined within the
same segment, while the intersegmental synchronization
is mediated, at least in part, by the deep axon neurons.
Recent evidence supports the existence of mixed
synapses (chemical and electrical) on neurons throughout
the spinal cord of adult rats in Rexed’s laminae III–IX in
cervical, thoracic and lumbar segments (Rash et al. 1996).
If present in the spinal cord of the adult cat, it is not
unlikely that electrical synapses also contribute to the
synchronization of the spontaneous interneuronal discharges. In this case, gap junctions could provide
additional pathways for bidirectional excitatory communication between neurons (see Tresch and Kiehn 2000;
Roerig and Feller 2000).
One pending issue relates to the pathways and
mechanisms involved in the termination of the discharges
of neurons in layers III–VI during the spontaneous
nCDPs. Available evidence suggests that these neurons
have excitatory as well as inhibitory connections with
other neurons in the region (Schneider 2001). It is not
unlikely that these inhibitory inputs play an important role
in the termination of the episodes of synchronized
spontaneous activity of the neurons, as suggested by the
finding that stimulation of cutaneous nerves transiently
depresses the generation of spontaneous nCDPs (see
raster displays in Fig. 1D and Manjarrez et al. 2000), and
also by the finding that in some cases the spontaneous
nCDPs generated in one segment appear together with
positive waves in other segments (Fig. 5B, D), probably
because of activation of inhibitory neurons. Yet, depression of neuronal excitability by previous activity could
also play a relevant role in this process (see O’Donovan
1999).
Functional role of the intersegmental coupling
of spontaneous activity of dorsal horn neurons
Further understanding of the functional role of the
synchronization of dorsal horn neurons during the nCDPs
will require information on their axonal projections, that
is, if they have only local connections or if their axons
also contact neurons with ascending projections (see
below and Schneider 1992). In this context, it should be
mentioned that in a preliminary study Manjarrez et al.
(2002a) recorded a series of potentials in the contralateral
posterior sigmoid gyrus that appeared about 14 ms after
the L6 spontaneous nCDPs, and were abolished by
sectioning the dorsal columns and the ipsilateral dorsolateral funiculus at the thoracic level, suggesting involvement of the spinocervical tract (see Bryan et al. 1973;
Willis and Coggeshall 1991). During the occurrence of
spontaneous nCDPs in L6, the cortical potentials evoked
by electrical stimulation of cutaneous nerves or by
mechanical stimulation of the skin (Manjarrez et al.
2002b), as well as the segmental monosynaptic reflexes
(Manjarrez et al. 2000), were facilitated. Whether the
cortical and segmental actions associated with the generation of the spontaneous nCDPs were mediated by the
same neurons (e.g., ascending tract neurons with local
collaterals) or by separate sets of neurons, both receiving
a common input, remains an open question.
It is also important to determine the location and
extension of the receptive fields in response to natural
stimulation of the skin of the neurons that are synchronized and non-synchronized with the spontaneous nCDPs.
According to Brown et al. (1979), all pairs of dorsal horn
neurons in the lumbosacral cord of the cat with non-flat
cross-correlograms had overlapping receptive fields.
These results agree with those reported by Eblen-Zajjur
and Sandkuhler (1997) for the rat. It thus seems reasonable to assume that the neurons that fire in synchrony
during the spontaneous nCDPs presently analyzed could
have overlapping receptive fields. This would increase
their influence on spinal reflex pathways when naturally
activated.
The coupling between the spontaneous activity of
dorsal horn neurons could allow the constitution of a
distributed assembly, which encodes a stimulus by means
of correlated increases of neuronal discharge rates
(Aertsen et al. 1989). There is evidence suggesting that
synchronization plays a functional role in the coding of a
cutaneous stimulus. Long-lasting radiant skin heating
induced in the vast majority of neuron pairs an increase in
the strength of synchronization (Eblen-Zajjur and Sandkuhler 1997).
Acknowledgements We thank Profs. E. Jankowska and A. LevTov for their useful comments on the manuscript, D. Chvez, C.A.
Garca and S. Pia for their participation in some of the
experiments, and A. Rivera, E. Velzquez and P. Reyes for
technical assistance. This work was partly supported by NIH grant
NS 09196, CONACyT grants 41739, 3908N and by Sistema
Nacional de Investigadores, Mxico.
413
References
Aertsen AMHJ, Gerstein GL, Habib MK, Palm G (1989) Dynamics
of neuronal firing correlation: modulation of “effective connectivity”. J Neurophysiol 61:900–917
Arshavsky YI, Deliagina TG, Orlovsky GN (1997) Pattern
generation. Curr Opin Neurobiol 7:781–789
Baker SN, Kilner JM, Pinches EM, Lemon RN (1999) The role of
synchrony and oscillations in the motor output. Exp Brain Res
128:109–117
Bernhard CG (1953) The spinal cord potentials in leads from the
cord dorsum in relation to peripheral source of afferent
stimulation. Acta Physiol Scand 29:1–29
Bernhard CG, Widen L (1953) On the origin of the negative and
positive cord potentials evoked by stimulation of low threshold
cutaneous fibres. Acta Physiol Scand 29:42–54
Brown AG (1981) Organization in the spinal cord. Anatomy and
physiology of identified neurones. Springer, Berlin Heidelberg
New York
Brown PB, Koerber HR, Yezierski RP (1979) Cross-correlation
analysis of connectivities among cat lumbo-sacral dorsal horn
cells. J Neurophysiol 42:1199–1211
Bryan RN, Trevio DL, Coulter JD, Willis WD (1973) Location
and somatotopic organization of the cells of origin of the spinocervical tract. Exp Brain Res 17:177–189
Eblen-Zajjur AA, Sandkuhler J (1997) Synchronicity of nociceptive
and non-nociceptive adjacent neurones in the spinal dorsal horn
of the rat: stimulus-induced plasticity. Neuroscience 76:39–54
Grillner S (1981) Control of locomotion in bipeds, tetrapods and
fish. In: Brooks VB (ed) Handbook of physiology, sect. 1, vol.
2. The nervous system, motor control. American Physiology
Society, Bethesda, pp 1179–1236
Inbar G, Madrid J, Rudomin P (1979) The influence of the gamma
system on cross-correlated activity of Ia muscle spindles and its
relation to information transmission. Neurosci Lett 13:73–78
Kerkut GA, Bagust J (1995) The isolated mammalian spinal cord.
Prog Neurobiol 46:1–48
Koerber HR, Brown PB (1982) Somatotopic organization of
hindlimb cutaneous nerve projections to cat dorsal horn.
J Neurophysiol 48:481–489
Kremer E, Lev-Tov A (1997) Localization of the spinal network
associated with generation of hindlimb locomotion in the
neonatal rat and organization of its transverse coupling system.
J Neurophysiol 77:1155–1170
Lidierth M, Wall PD (1996) Synchronous inherent oscillations of
potentials within the rat lumbar spinal cord. Neurosci Lett
220:25–28
Lidierth M, Wall PD (1998) Dorsal horn cells connected to the
Lissauer tract and their relation to the dorsal root potential in
the rat. J Neurophysiol 80:667–679
Manjarrez E, Jimnez I, Snchez J, Rudomin P (1998) Lumbar
spontaneous cord dorsum potentials are originated by synchronized activity of dorsal horn neurons. Soc Neurosci Abstr
453.11
Manjarrez E, Rojas-Piloni JG, Jimnez I, Rudomin P (2000)
Modulation of synaptic transmission from segmental afferents
by spontaneous activity of dorsal horn neurones. J Physiol
529:445–460
Manjarrez E, Rojas-Piloni JG, Vazquez D, Flores A (2002a)
Cortical neuronal ensembles driven by dorsal horn spinal
neurones with spontaneous activity in the cat. Neurosci Lett
318:145–148
Manjarrez E, Rojas-Piloni JG, Martnez L, Vzquez D, Vlez D,
Mndez I, Flores A (2002b) Amplitude of somatosensory
cortical evoked potentials is correlated with spontaneous
activity of spinal neurones in the cat. Neurosci Lett 323:187–
190
Manjarrez E, Prez H, Rojas-Piloni JG, Vlez D, Martnez L,
Flores A (2002c) Absence of coherence between cervical and
lumbar spinal cord dorsal surface potentials in the anaesthetized
cat. Neurosci Lett 328:37–40
Matsushita M (1969) Some aspects of the interneuronal connections in cat’s spinal gray matter. J Comp Neurol 136:57–80
O’Donovan MJ (1999) The origin of spontaneous activity in
developing networks of the vertebrate nervous system. Curr
Opin Neurobiol 9:94–104
Pearson KG (2000) Neural adaptation in the generation of rhythmic
behavior. Ann Rev Physiol 62:723–753
Rash JE, Dillman RK, Bilhartz BL, Duffy HS, Whalen LR,
Yasumura T (1996) Mixed synapses discovered and mapped
throughout mammalian spinal cord. Proc Natl Acad Sci U S A
93:4235–4329
Roerig B, Feller MB (2000) Neurotransmitters and gap junctions in
developing neural circuits. Brain Res Rev 32:86–114
Rossignol S, Beloozerova I, Gossard J-P, Dubuc R (1998)
Presynaptic mechanisms during locomotion. In: Rudomin P,
Romo R, Mendell L (eds) Presynaptic inhibition and neural
control. Oxford University Press, New York, pp 385–397
Rudomin P, Dutton H (1969) Effects of muscle and cutaneous
afferent volleys on excitability fluctuations in Ia terminals.
J Neurophysiol 32:158–169
Rudomin P, Nfflez R, Madrid J, Burke RE (1974) Primary afferent
hyperpolarization and presynaptic facilitation of Ia afferent
terminals induced by large cutaneous afferents. J Neurophysiol
37:413–429
Sandkuhler J, Eblen-Zajjur AA (1994) Identification and characterization of rhythmic nociceptive and non-nociceptive spinal
dorsal horn neurons in the rat. Neuroscience 61:991–1006
Schneider SP (1992) Functional properties and axon terminations
of interneurons in laminae III–V of the mammalian spinal
dorsal horn in vitro. J Neurophysiol 68:1746–1759
Schneider SP (2001) Properties of synaptic potentials evoked at
connections between local axon interneurons in spinal dorsal
horn. Soc Neurosci Abstr 392.13
Schneider SP, Sandiford DR, Kavookjian AM, Johnson BD (1995)
Synaptic connectivity of local circuit neurons in laminae III and
IV of the hamster spinal cord. J Comp Neurol 335:380–391
Silberberg G, Gupta A, Markram H (2002) Stereotypy in neocortical microcircuits. Trends Neurosci 25:227–230
Tresch MC, Kiehn O (2000) Motor coordination without action
potentials in the mammalian spinal cord. Nature Neurosci
3:593–599
Wall PD, Werman R (1976) The physiology and anatomy of long
ranging afferent fibers in the spinal cord. J Physiol 255:321–
334
Willis WD, Coggeshall RE (1991) Sensory mechanisms of the
spinal cord. Plenum Press, New York
Willis WD, Weir MA, Skinner RD, Bryan RN (1973) Differential
distribution of spinal cord field potentials. Exp Brain Res
17:169–176