Articles in PresS. J Neurophysiol (October 24, 2012). doi:10.1152/jn.00649.2012
1
Rhythmic activity of feline dorsal and ventral spinocerebellar tract neurons during
2
fictive motor actions
3
Brent Fedirchuk2, Katinka Stecina1, Kasper Kyhl Kristensen1, Mengliang Zhang1, Claire F.
4
Meehan1, David J. Bennett3, and Hans Hultborn1
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
Affiliations:
1. University of Copenhagen, Faculty of Health Sciences, Department of Neuroscience and
Pharmacology, Copenhagen, Denmark
2. University of Manitoba, Faculty of Medicine, Department of Physiology, Winnipeg,
Manitoba, Canada
3. University of Alberta, Center for Neuroscience, Edmonton, Alberta, Canada
Running Head: Spinocerebellar activity during fictive motor behaviors
Keywords: cerebellum, Clarke's column, locomotion, scratch, sensorimotor integration
CONTACT INFORMATION/ Correspondence to:
Katinka Stecina, Ph.D.
Copenhagen University
INF-Panum Institute
Blegdamsvej 3-33.3.
Copenhagen 2200 Denmark
e-mail: stecina@sund.ku.dk
Copyright © 2012 by the American Physiological Society.
1
25
26
Abstract
27
Neurons of the dorsal spinocerebellar tracts (DSCT) have been described to be
28
rhythmically active during walking on a treadmill in decerebrate cats, but this activity ceased
29
following de-afferentation of the hindlimb. This observation supported the hypothesis that
30
DSCT neurons primarily relay the activity of hindlimb afferents during locomotion, but lack
31
input from the spinal central pattern generator (CPG). The VSCT neurons, on the other hand,
32
were found to be active during actual locomotion (on a treadmill) even after de-afferentation,
33
as well as during fictive locomotion (without phasic afferent feedback).
34
In this study we compared the activity of DSCT and VSCT neurons during fictive
35
rhythmic motor behaviors. We used decerebrate cat preparations in which fictive motor tasks
36
can be evoked while the animal is paralyzed and there is no rhythmic sensory input from
37
hindlimb nerves. Spinocerebellar tract cells with cell bodies located in the lumbar segments
38
were identified by electrophysiological techniques and examined by extra- and intracellular
39
microelectrode recordings. During fictive locomotion 57/81 DSCT and 30/30 VSCT neurons
40
showed phasic, cycle-related activity. During fictive scratch 19/29 DSCT neurons showed
41
activity related to the scratch cycle. We provide evidence for the first time that locomotor and
42
scratch drive potentials are present not only in VSCT but also in the majority of DSCT
43
neurons. These results demonstrate that both spinocerebellar tracts receive input from the
44
CPG circuitry often sufficient to elicit firing in the absence of sensory input.
45
46
2
47
Introduction
48
The feline dorsal and ventral spinocerebellar tracts (DSCT and VSCT) are two main
49
ascending pathways which have been perceived to be serving different functional roles. They
50
are different entities because of their anatomical organization and they have also been
51
thought to convey different input to the cerebellum (Bosco and Poppele 2001; Lundberg
52
1971; Oscarsson 1965). The commonly accepted view is that the DSCT neurons primarily
53
relay sensory input from afferents and they are less influenced by the activity of other spinal
54
neurons than the VSCT cells (Arshavsky et al. 1986; Bosco and Poppele 2001). There is
55
strong evidence that VSCT cells monitor premotoneuronal activity in relation to motor
56
commands (Lundberg 1971; Mann 1973; Oscarsson 1965). Cortical input via disynaptic
57
pathways to DSCT cells (Hongo and Okada 1967; Hongo et al. 1967) and some unidentified
58
spinal input to DSCT cells has been recognized that may serve to maintain their tonic activity
59
in the absence of sensory input following section of the dorsal roots (Holmqvist et al. 1956).
60
Recently, cortical control of DSCT cells in neonatal mice has been recognized (Hantman and
61
Jessell 2010) and it was suggested that DSCT neurons may also be important components of
62
spinal circuits used for planning and evaluation of motor actions.
63
Early reports on the activity of spinocerebellar tract cells during locomotion described
64
DSCT cells to be rhythmically active during actual walking on a treadmill in decerebrate cats,
65
but de-afferentation of the hindlimb by dorsal root transection abolished the step-related
66
modulation (Arshavsky et al. 1972b). This observation supported the hypothesis that the
67
function of the DSCT neurons is to relay the activity of hindlimb sensory afferents during
68
locomotion while not receiving excitation from the spinal neurons comprising the central
69
pattern generator (CPG) for locomotion. VSCT neurons, on the other hand, were found to be
70
active not only during actual locomotion on a treadmill, (Arshavskii et al. 1972a) but also
3
71
following deafferentation (Arshavskii et al. 1972b), as well as during fictive locomotion
72
without rhythmic sensory feedback (Orsal et al. 1988). In addition, VSCT neurons have been
73
found to be phasically active during fictive scratch (Arshavskii et al., 1975; Arshavsky et al
74
1978), while DSCT neurons were reported to have no rhythmic modulation of firing during
75
fictive scratch (Arshavskii et al. 1975). Some years later, indirect evidence lead to the
76
hypothesis that spinal timing generators may also convey information to the cerebellum via
77
DSCT neurons (Perciavalle et al. 1995) but it is still unclear whether signals from the CPG
78
circuitry reach DSCT cells. Therefore the role of DSCT cells during rhythmic motor actions,
79
namely during fictive locomotion and fictive scratch was examined in this study.
80
Fictive locomotion and scratch refers to motor output monitored by
81
electroneurograms from hindlimb nerves in decerebrate animals that are paralyzed by
82
pharmacological blockade of the neuromuscular junctions. The fictive motor output closely
83
resembles that during real locomotion and scratch activity but there is no movement and
84
therefore no rhythmic sensory feedback. In this state the activity of the spinocerebellar tract
85
cells can be ascribed to inputs from the neuronal networks involved in the generation of the
86
motor activity. In this study, we examined the activity pattern of DSCT and VSCT neurons
87
during two different types of fictive motor activity by extra- and/or intracellular recordings of
88
identified tract cells. Fictive locomotion was induced by the electrical stimulation of the
89
mesencephalic locomotor region (MLR) and fictive scratch was induced by mechanical
90
stimulation of the skin covering the ears or the face in decerebrate cats following application
91
of curare and/or bicucculine at the dorsal root entry zone of the first and second cervical
92
segments. Preliminary results have been presented as abstracts (Fedirchuck et al. 1995;
93
Stecina et al. 2008).
4
94
95
Materials and Methods
96
Preparation
97
Experiments were performed during two series of studies. The first series consisted of
98
15 animals in which fictive locomotion was evoked, and the second series consisted of 7
99
animals in which fictive scratch was evoked as rhythmic motor activity. Thus a total of 22
100
adult cats of either sex weighing 3.0 – 4.6 kg were used. All surgical and experimental
101
procedures were conducted in accordance with EU regulations (Council Directive
102
86/609/EEC) and with National Institutes of Health guidelines for the care and use of
103
laboratory animals (National Institutes of Health publication no. 86-23, revised 1985). All
104
procedures were approved by the Danish Animal Experimentation Inspectorate. There were
105
4/7 animals in the second series from which data were also used for other studies (Stecina et
106
al. 2007).
107
In the first series of studies (experiments on fictive locomotion) anesthesia was first
108
induced by an intravenous injection of Saffan (1 ml/kg; alphaxalone 9 mg/ml + alphadolone 3
109
mg/ml; for n=15, 3.0 - 4.4 kg animals). In the second series (experiments on fictive scratch)
110
induction was attained by halothane (2-3% halothane, 70% N2O and 30% O2; for n=7, 3.2 –
111
4.6 kg animals). After a tracheotomy and intubation, anesthesia was maintained throughout
112
the surgery with Halothane (0.8-1.5%) delivered in an oxygenated mixture of nitrous oxide
113
(60% N2O, 40% O2). The blood pressure was monitored continuously via a carotid artery
114
catheter and cannulae were also placed in both forelimb brachial veins for administration of
115
drugs. Atropine (0.1 mg/kg, subcutaneous), dexamethasone (1.0 mg/kg, intravenous) or
116
solumedrol (2.5 mg/kg i.v.) and a glucose/bicarbonate buffer solution (10% dextrose and
5
117
1.7% NaHCO3) at a rate of 2.5 to 4.0 ml/hr, intravenously were routinely administered early
118
in the experiment.
119
In both series of experiments, the nerves innervating the following hindlimb muscles
120
were dissected on the left side: the multifunctional hamstring muscles posterior biceps and
121
semitendinosus (PBSt) that are often active during both flexion and extension of fictive
122
locomotion and during extension of fictive scratch, the semimembranosus and anterior biceps
123
(SmAB) active during extension, the medial and lateral gastrocnemious and soleus (GS),
124
plantaris (Pl) that are both active during extension, the tibialis anterior (TA) and extensor
125
digitorum longus (EDL), often TA and EDL together as deep peroneal (DP) both active
126
during flexion of fictive locomotion and during both flexion and extension of fictive scratch,
127
the and digit mover flexor digitorum and hallucis longus (FDHL) that is often active during
128
both phases of fictive locomotion and of fictive scratch. In addition, the posterior tibial (Tib)
129
nerve innervating ankle extensor muscles and carrying cutaneous input; the cutaneous sural
130
(Sur) nerve, and the cutaneous superficial peroneal (SP) nerve were also dissected. In the
131
second series of experiments, the nerve innervating the peroneus longus (PerL) muscle that is
132
an ankle flexor during fictive locomotion but it becomes synergist of ankle extensors during
133
fictive scratch was also dissected in some animals. In both series of experiments, the
134
sartorius (Sart), and quadriceps (Q) branches of the femoral nerve were dissected and they
135
were placed in implanted plastic cuff electrodes. The Sart (depending on the nerve branch
136
dissected) is active mostly during flexion of fictive locomotion and scratch but it can be also
137
active during both phases. On the right side the PBSt, SmAB and Sart nerves were dissected
138
in some of the experiments. All peripheral nerves were cut distally and the proximal stump
139
was freed from connective tissue.
6
140
In the first series of experiments both sciatic nerves and all the dissected branches
141
were sufficiently dissected to allow the hindlimbs to hang pendant thus avoiding extension of
142
the hip when constructing the hindlimb paraffin pool, which could prevent locomotor
143
activities while the dissected nerves were laid in a plastic tray, filled with mineral oil, where
144
they were placed on bipolar silver hook electrodes to be either stimulated or recorded.
145
In the second series of experiments the hindlimbs were fixed in an extended position
146
to allow for the construction of a paraffin pool using the skin covering the hindlimbs and the
147
dissected nerves were placed on bipolar silver hook electrodes. In this case all branches from
148
the femoral, obturator and sciatic nerves were sectioned and the tendons of muscles crossing
149
the hip joint (which were not denervated) were cut in order to prevent sensory feedback
150
signaling the hip extension. All the hindlimb nerves as well as the exposed spinal cord were
151
covered with mineral oil. The temperature of the animal´s core and the mineral oil pools were
152
maintained at physiological levels using a feedback heating system.
153
In both series of experiments, laminectomy of the L3-L6 vertebrae exposed the
154
lumbo-sacral segments, and in 2 animals the lower thoracic segments were also exposed by a
155
laminectomy of the 12th to 13th thoracic vertebrae. The first cervical vertebra was also
156
removed and in the second series the second cervical vertebrae was removed as well.
157
In both series of experiments, a craniotomy was performed and the animal was
158
mechanically decerebrated at a precollicular postmammillary level and all brain tissue rostral
159
to the transection was removed. At this time the anesthetic was discontinued and decreases in
160
blood pressure associated with the decerebration were countered by intravenous
161
administration of Oxypherol (an oxygen carrying volume expander, <10ml) and/or Gentran
162
(3000 mM dextran solution, <10ml). The animal was paralyzed with intravenous Pavulon
7
163
(pancuronium bromide; 0.2mg/kg, supplemented every 40-60 min) and ventilated to maintain
164
end tidal CO2 at 4-6%. The tentorium was removed to expose the brainstem and the
165
cerebellum for later electrical stimulation. When the blood pressures became less than 80
166
mmHg the drop was counteracted with intravenous administration of a volume expander (see
167
above) or noradrenaline as needed.
168
Evoking fictive motor behaviors
169
Fictive locomotion was elicited in the paralyzed preparation by electrical stimulation
170
of the mesencephalic locomotor region. Insulated monopolar steel electrodes were placed
171
bilaterally in the midbrain (Horsley-Clarke coordinates: P1 - 2; L3.5 - 4; H0 -1.5 and
172
electrical stimulation (30-200 μA, rectangular current pulses delivered at 15-20 Hz) elicited
173
fictive locomotor activity, which was recorded from the peripheral nerves. The location of the
174
electrodes was adjusted to obtain the lowest electrical threshold for locomotion and the most
175
stable locomotor pattern possible. Normally stimulation was unilateral, but occasionally
176
bilateral stimulation was required to improve locomotor activity.
177
Fictive scratch was evoked by topical application of D-tubocurarine solution 0.1 – 1%, and/or
178
bicucculine solution 0.1 – 1% onto the first and second cervical dorsal root entry zone on the
179
left side followed by mechanical stimulation of the skin of the ear on the left side.
180
Recording Techniques
181
Both extracellular and intracellular recordings were obtained using pulled-pipette
182
glass microelectrodes filled with 2 M potassium acetate (1.4 to 2.0 μ tips; 3 to 10 MΩ) and
183
amplified with an Axoclamp 2A microelectrode amplifier. Electroneurograms (ENGs) of
184
hindlimb nerves were also recorded, or alternatively, each nerve could be stimulated. A ball
185
electrode was placed on the dorsal surface of the spinal cord typically at the L6-L7 segment
8
186
to record the incoming afferent volley associated with peripheral nerve stimulation (0.1 ms
187
pulse) or descending volley evoked by supraspinal stimulation. The strength of peripheral
188
nerve stimulation was given in multiples of the threshold that is the stimulus strength
189
necessary to recruit the most excitable fibers (i.e. that produced an incoming volley recorded
190
on the surface of the spinal cord at lumbar levels).
191
Data capture and analyses
192
In the first series of experiments, the microelectrode recordings, incoming volley
193
recordings and rectified ENG recordings were digitized usually at a rate of 20 KHz, 5 KHz,
194
and 650 Hz, respectively, on a Concurrent/Masscomp 5400 series computer and a custom
195
made software from the Spinal Cord Research Center at the University of Manitoba
196
(Winnipeg, Canada). In the second series ENG signals were digitized at a rate of 10 kHz and
197
filtered (5 Hz to 1 kHz) while the microelectrode and spinal cord potential recordings were
198
digitized at a rate of 20 kHz by using CED 1401 and Spike 2 version 5.21 Software
199
(Cambridge Electronic Design, Cambridge, UK) and a personal computer with Pentium
200
processor.
201
Post-hoc analysis of the data (i.e. offline analysis) consisted of calculating the
202
averaged instantaneous firing frequencies of extracellular units or averaged membrane
203
potential change of intracellularly recorded cells based on normalized and averaged
204
locomotor cycles. Built-in spike sorting algorithms within the software identified specific
205
cells which were determined to fulfill the criteria of being spinocerebellar tract cells (see
206
paragraph below). The analysis of the locomotor cycles was based on flexor and extensor
207
bursts of identified muscle nerves (TA, EDL or Sart as markers for onset of the flexor phase
208
and GS, SmAB and sometimes PerL during scratch as markers for the extensor phase). Cycle
9
209
duration was determined as the time between consecutive onsets of flexor activity;
210
flexor/extensor phase duration was determined as the time period when the flexor/extensor
211
ENG activity exceeded a set threshold. This threshold was visually determined by inspecting
212
typically 1 – 3 min long activity with consecutive alternating flexor and extensor bursts.
213
Locomotor and scratch cycles were divided into 30 bins and normalized. The instantaneous
214
firing frequency (IFF) of each identified neuron was calculated in each of the bins.
215
Overlaying the averaged ENGs with the graph of the IFF was used to visually determine the
216
phase of activity of each neuron (i.e. when the IFF was maximal).
217
Criteria for identification of units
218
DSCT neurons are known to ascend in the ipsilateral spinal white matter and VSCT
219
neurons ascend in contralateral spinal white matter (Eccles et al. 1967; Lundberg and
220
Oscarsson 1961; Mann 1973). Therefore, spinocerebellar tract cells were identified by
221
differential electrical stimulation applied at either ipsilateral or contralateral sites with respect
222
to the intraspinal microelectrode as illustrated schematically in Fig. 1A-C. Stimulation on the
223
surface of the cord at the first cervical (C1 stim see in Fig. 1C) segment was applied to
224
identify ascending projections of a unit. Units were classified to be spinocerebellar tract
225
neurons if they could be antidromically activated either by electrical stimulation at the
226
surface of the cerebellum (surface stim in Fig. 1A) or by intra-cerebellar stimulation (intra-
227
CB stim in Fig. 1B). Spinocerebellar tract neurons projecting on the ipsilateral (ipsi) and
228
contralateral (contra) sides were compatible with DSCT and VSCT origin, respectively (see
229
review by Mann 1973).
230
231
Fig. 1 here
Surface stimulation was applied using insulated monopolar steel electrode placed on
10
232
the anterior cerebellar cortex with the anode placed in the neck muscles near the base of the
233
head, thus presumably activating deeper structures such as the peduncle. Surface stimulation
234
was used in all experiments of the first series (i.e. those with fictive locomotion).
235
Furthermore, we often verified that in addition to antidromic activation from the cerebellar
236
surface, VSCT neurons could be antidromically activated by a high strength stimulus pulse
237
(up to 300 μA, 1 ms pulse) from the contralateral MLR electrode since most VSCT fibers
238
enter the cerebellum through the superior cerebellar peduncle (Oscarsson 1965) and this
239
structure is close to the cuneiform nucleus and the area stimulated for evoking fictive
240
locomotion albeit some VSCT neurons are known to enter via the inferior cerebellar peduncle
241
as in e.g. (Kitamura and Yamada 1989).
242
Intra-CB stimulation was applied by using parylene-coated tungsten electrodes (0.1 –
243
0.3 MΩ, World Precision Instruments, Sarasota, FL, USA) initially inserted into the
244
cerebellum 1-2 mm dorsal and caudal to its junction with the inferior colliculi and about 2-3
245
mm lateral from the midline as shown in Fig. 1B. In Fig. 1D we illustrate cord dorsum
246
potentials recorded at the 6th-7th lumbar segment following intra-cerebellar stimulation
247
applied in one preparation. The most prominent descending volleys were seen at intra-CB
248
depth of 5-7 mm. Extracellular recordings in Fig. 1E illustrate the spike evoked by the intra-
249
CB stimulation at 6 mm depth. The collision of the antidromic spike with a spontaneous
250
spike is shown by the arrow. The area where the approximate location of our identification
251
points were in the cerebellum corresponds well with previous reports on the optimal
252
antidromic activation sites used for DSCT neurons (Edgley and Gallimore 1988).
253
254
Results
11
255
Not only the VSCT neurons, but the majority of the DSCT neurons was found to be
256
active during fictive locomotor activity in a cyclic, phase-related manner even though there
257
was no actual hindlimb movement or phasic afferent input. In addition, we also found that
258
about two-thirds of the DCST neurons show phase-related activity during fictive scratch.
259
Activity of DSCT neurons during fictive locomotion
260
Extracellular or intracellular recordings were obtained from 81 DSCT neurons. 69 of
261
these neurons were recorded in the L3 - L5 spinal segments and 12 neurons were recorded in
262
the L1 or L2 spinal segments. Figure 2 shows the extracellular recording of a DSCT neuron
263
during MLR-evoked fictive locomotion. The period shown in A starts just as the MLR
264
stimulus (100 μA, 20 Hz) was turned off and shows that the rhythmic activity is related to the
265
fictive motor pattern and not directly linked to the MLR stimulus. After normalizing and
266
averaging the fictive step cycles (n=12) using the onset of Sart ENG activity as the cycle
267
onset, the occurrence of action potentials during different times of the step cycle could be
268
plotted. The upper trace of Fig. 2B shows the averaged instantaneous firing frequency (IFF)
269
for the normalized and averaged step cycle in relation to the averaged ENG activity from the
270
flexor (Sart) and the extensor (SmAB) muscle nerves. Then the step cycle-related change in
271
instantaneous firing frequency was measured as shown in Fig. 2B for this cell. The maximal
272
firing frequency was 78 Hz. Similar analysis was done for each cell recorded extracellularly
273
during fictive locomotion.
274
Fig. 2 here
275
Intracellular recordings were obtained from 12 DSCT neurons during fictive
276
locomotion. Figure 3 illustrates oscillations of the membrane potential associated with the
277
fictive step cycle on the intracellular microelectrode recording shown in the top panel (A).
12
278
When such oscillations of postsynaptic potentials were recorded from motoneurons they were
279
called locomotor drive potentials or LDPs (Shefchyk and Jordan 1985). Spikes were absent in
280
this trial because the sodium channels had been inactivated by a prolonged depolarizing
281
current injection just before this recording period. The lower panels (B) show the averaged
282
membrane potential and flexor and extensor ENG activity during the normalized and
283
averaged step cycles (n=8). The activity of 9/12 cells was deemed to be modulated in phase
284
with the step cycle based on analysis of intracellular recordings; but not all of these cells were
285
recorded from an extracellular position and their firing patterns could not be analyzed.
286
However, all 9 cells exhibited some degree of LDPs ranging from 0.7 to 6.0 mV.
287
Fig 3 here
288
A total of 57/81 (70%) of DSCT cells in this study showed activity that was
289
modulated in relation to the fictive step cycle. There was no difference in the propensity for
290
DSCT neurons from different spinal segments to exhibit phasic modulation during fictive
291
locomotion, or to be active during a particular phase of the step cycle (see Table 1). Of the
292
extracellularly recorded cells (n=48), 8 fired exclusively during flexion as in Fig. 2; 20 fired
293
only in extension, and 20 fired throughout the fictive step cycle but at a higher frequency in
294
one phase compared to the other. There was no significant difference in the cycle related
295
changes of the instantaneous firing frequency between extension and flexion related DSCT
296
neurons (Mann-Whitney rank sum test p>0.05). Of the intracellularly recorded DSCT (n=9)
297
cells, 3 were depolarized (and/or had action potentials) during extension, 4 were excited
298
during flexion and 2 were excited during flexion as well as during part of extension. The
299
remaining 24/81 of DSCT cells tested had tonic activity without phasic modulation in relation
300
to fictive locomotion. Prior to the onset of fictive locomotion (i.e. MLR stimulation) 33
13
301
DSCT cells showed tonic background activity with a mean of 22 Hz IFF while the mean IFF
302
of the same cells during fictive locomotion was 20 Hz. The background firing frequency of
303
7/33 cells was comparable to the peak firing rates seen during fictive locomotion. In 2/33
304
cells the firing frequency during fictive locomotion was actually lower than the background
305
rate. In the remaining 24 cells the firing frequencies during fictive locomotion were higher
306
than the background rates.
307
308
309
Activity of VSCT neurons during fictive locomotion
Extracellular and intracellular recordings were obtained from 30 VSCT neurons
310
within the L2 to L5 spinal segments. Fig. 4 illustrates an extracellular recording of the
311
activity of a VSCT neuron during fictive locomotion. This unit started firing at the peak of
312
ipsilateral extension, i.e. before the onset of Sart ENG activity (see Fig. 4B). In the example
313
in Fig. 4, the cell had a 300 Hz change in instantaneous firing frequency during the averaged
314
(n=12) fictive step cycle (see Fig. 4B). Overall, 100% of the VSCT cells recorded showed
315
phasic activity with fictive locomotion. A total of 19 extracellularly recorded VSCT units
316
fired exclusively in flexion, 8 fired only during extension and 3 units fired throughout the
317
fictive step cycle but at a higher frequency in one phase (2 in flexion and 1 in extension).
318
There was no significant difference in the degree of the cycle related changes in
319
instantaneous firing frequency between extension and flexion-coupled VSCT neurons (Mann-
320
Whitney rank sum test p>0.05).
321
322
323
Fig. 4 here
Prior to the onset of fictive locomotion there were 14/30 VSCT cells with tonic
background activity. For 13 of these VSCT cells, the firing rates during fictive locomotion
14
324
were much greater than the background frequencies prior to locomotor activity. The
325
remaining one VSCT cell had comparable firing rates prior to and during fictive locomotion.
326
Locomotor drive potentials in VSCT cells are illustrated in Fig. 5A and during the
327
normalized and averaged steps (n=24) the peak-to-peak amplitude was 6.8 mV (Fig. 5B). All
328
of the intracellularly recorded VSCT cells (n=7) were depolarized during the flexion phase of
329
fictive locomotion. In 4/7 cells the action potential generation ceased spontaneously (i.e.
330
without hyperpolarization of the membrane potential by current injection) so the LDPs could
331
be averaged, and their amplitude was measured. For the other 3 cells, hyperpolarizing current
332
injection (1.0, 6.4 and 7.8 nA) was used to transiently suppress action potential production in
333
order to measure LDP amplitude. Each of the 7 VSCT neurons recorded intracellularly
334
displayed LDPs ranging from 1.4 to 9 mV peak-to-peak amplitude.
335
Fig. 5 here
336
337
338
Activity of DSCT neurons during fictive scratch
Figure 6A illustrates the extracellular recording of the activity of a DSCT cell with
339
action potentials during the extension phase of a 4 s long bout of fictive scratch activity. This
340
is the same cell as illustrated in Fig. 1C and D while identifying it with intra-cerebellar
341
stimulation. The ENG recordings were normalized and averaged (n=25) based on the onset of
342
an extensor (PerL) ENG activity as the start of the cycle and the averaged and normalized
343
instantaneous firing frequency is shown in Fig. 6B.
344
345
346
Fig. 6 here
Intracellular records from one DSCT cell during fictive scratch are illustrated in Fig.
7. Note that the depolarization occurs in phase with the activity of the ankle extensor Pl
15
347
nerve. This membrane potential modulation during fictive scratch is similar to those
348
described previously in hindlimb motoneurons (Perreault 2002). The averaged modulation of
349
the membrane potential during the scratch cycles (n=7) was based on the Pl ENG activity is
350
shown in Fig. 7B with the averaged ENGs of the Pl and TA nerves. The extracellularly
351
recorded IFF of this DSCT cell averaged and normalized during another bout of scratch
352
(n=25) is overlaid on the membrane potential and it shows that the modulation of the IFF was
353
44 Hz.
354
Fig. 7 here
355
Overall, extracellular recordings were made from 29 DSCT cells (16 from within the
356
L3-L4 spinal segments and 13 in L2), and 19/29 DSCT were phasically active during fictive
357
scratch. The phasic activity of 3 DSCT cells coincided with flexion, 6 cells were active
358
during extension and 10 cells fired throughout both phases of the scratch cycle but at a higher
359
frequency in one phase compared to the other. Prior to fictive scratch 12 DSCT cells had
360
tonic activity and their mean IFF was 79 Hz while during fictive scratch the mean IFF of the
361
same cells was 85 Hz. There were 4/12 DSCT cells which had lower IFF during fictive
362
scratch than prior to it. Intracellular records during fictive scratch without action potential
363
generation in DSCT cells were collected from 2 other cells in addition to the one illustrated in
364
Fig. 7B and the “scratch-drive” potential amplitudes were 0.7 and 3.8 mV.
365
366
367
Differences in firing frequency between VSCT and DSCT cells
Our data of the DSCT and the VSCT cell activity during fictive locomotion were
368
collected as interspersed recordings with similar robustness of locomotor network activity
369
while recording from one or the other cell types. We have not made a systematic comparison
16
370
of cellular activity from one locomotor or scratch bout to the next, but in 13 DSCT cells we
371
have recordings during two or more bouts of fictive scratch and none of these cells showed
372
changes in their firing pattern from one bout to another. There were no attempts made to
373
quantify the ENG recordings in relation to the firing frequency for either cell type.
374
It became apparent that although DSCT and VSCT neurons could both be phasically
375
active during fictive locomotion, there were differences in the degree of modulation of
376
activity exhibited by the two cell types. Fig. 8A is a summary graph showing the change in
377
instantaneous firing frequency during the fictive step cycles of all DSCT and VSCT cells.
378
Those DSCT cells that were not phasically modulated with fictive locomotion are shown with
379
a step cycle-related frequency change of "0" Hz. VSCT neurons tended to have larger
380
changes in instantaneous firing frequency than DSCT neurons (see also Figs. 2 and 4) and
381
this difference was statistically significant (Mann-Whitney rank sum test p<0.001; non-
382
modulated DSCT neurons were excluded from the test sample). The amplitudes of the LDPs
383
of the few intracellularly recorded VSCT neurons also tended to be larger than those recorded
384
in DSCT neurons. However, the small sample size in this study precludes the verification of
385
this difference by statistical means. The firing frequency modulation of the DSCT neurons
386
during fictive scratch is illustrated in Fig. 8B and it was found to be in the same range as that
387
during fictive locomotion.
388
Fig. 8 here
389
390
Relation of excitatory peripheral input and phase of rhythmic activity
391
We have categorized the DSCT cells examined in the present study based on the
392
synaptic input from peripheral afferents as it has been described in the reviews by Mann
17
393
(1973) and Oscarsson (1965). Table 1 shows the source of excitatory synaptic input to the
394
DSCT neurons tested in this study. There was no relation between the source of afferent
395
excitation to a DSCT neuron and whether or not the activity of the neuron was modulated
396
with the fictive step cycle. For those DSCT neurons with activity that was modulated with the
397
fictive locomotion, there was no relation between the source of synaptic excitation and the
398
cycle phase that the unit was active—even in the case when DSCT cells had monosynaptic
399
input from group I afferents. There was no difference in the degree of modulation between
400
the DSCT neurons with different patterns of inputs (one way ANOVA, p=2.97). As shown in
401
Table 1, DSCT cells with excitation from extensor group I afferents were not only active
402
during extension, but 7/18 were active during flexion.
403
Table 1 here
404
The DSCT neurons with phasic activity during fictive locomotion could belong to any
405
of the categories that we have defined (see Table 1). The DSCT cells examined during fictive
406
scratch were not included in the Table due to the small sample size. During fictive scratch 2/3
407
flexor-related DSCT cells had no discernible sensory input and the third cell had group II
408
muscle afferent input from Sart and Quad. There were 2/6 of the extensor-related cells that
409
had no discernible sensory input and 2 with input from Tib and FDHL (and 2 that were not
410
tested for sensory inputs). In, 6/10 excitation from multiple sources (Sart, Q, PBSt or Tib)
411
was evoked, 2/10 had no sensory input and 1/10 had only cutaneous input from Sural (1/10
412
cell was not tested for inputs).
413
The VSCT neurons were also divided into groups based on the pattern of synaptic
414
input that they receive. In our sample, 6 of the VSCT neurons received excitation from
415
extensor group I muscle afferents, 2 from flexor group I muscle afferents, 6 received
18
416
excitation from higher threshold group II muscle and cutaneous afferents, while 9 received
417
polysynaptic inhibition from various sources. Seven VSCT neurons had no discernible
418
sensory inputs. As with DSCT neurons, we detected no trend for a difference between the
419
degrees of modulation within the fictive step cycle for the different categories of the VSCT
420
neurons, but note our small sample size.
421
422
423
Discussion
The results presented in this paper show that not only VSCT but also DSCT cells
424
discharge phasically during fictive locomotion evoked by electrical stimulation of the MLR
425
in pre-collicular/post-mamillary decerebrated cat preparations. We also demonstrate that
426
many DSCT cells are phasically active during fictive scratch. Given the absence of phasic
427
sensory activity during these fictive motor outputs, the results imply that inputs from the CPG
428
are often sufficient to induce firing in these ascending tract neurons.
429
430
431
Comparison of DSCT and VSCT activity during fictive motor actions
The vigorous rhythmic activity of VSCT neurons during fictive locomotion reported
432
in this study supports previous findings (Arshavskii et al. 1972a; b) and the concept that
433
VSCT cells convey information about the activity of spinal interneurons as well as about
434
input from sensory afferents (Arshavsky et al. 1972a; Lundberg 1971). Previously DSCT
435
neurons in the upper lumbar segments have been described to have phasic activity during
436
over-ground locomotion, but also that the phasic activity was abolished by sectioning the
437
ipsilateral dorsal roots, although a tonic activity of around 9 Hz remained (Arshavskii et al.
438
1972c). During the fictive locomotion used here, no phasic sensory input is generated, thus
19
439
our observations on the phasic activity in the 70% of the recorded DSCT neurons are contrary
440
to previous conclusions. In addition, our results demonstrate that 66% (19/29) of the DSCT
441
cells are rhythmically active during fictive scratch. We did not record from VSCT cells
442
during fictive scratch since an extensive study by Arshavsky and colleagues (Arshavskii et al.
443
1975) has demonstrated that virtually all VSCT cells in thalamic (74 cells) and decapitate cats
444
(44 cells) were discharging rhythmically in relation to the scratch cycle. There are no
445
previous published results on DCST cells during fictive scratch, but in the discussion on the
446
activity of VSCT cells during actual and fictive scratching (Arshavsky et al. 1978) it is
447
mentioned that DSCT cells “were found to have no rhythmical modulation during fictious
448
scratching”. These observations by Arshavsky and colleagues (Arshavskii et al. 1972c;
449
Arshavsky et al. 1978) imply that the phasic activation of DSCT neurons during locomotion
450
and scratch was attributable solely to their activation by hindlimb proprioceptive systems and
451
are therefore seemingly contradictory to our present observations of rhythmically active
452
DSCT neurons during fictive locomotion and scratch.
453
In the single report on the lack of phasic DSCT activity post-deafferentation (during
454
locomotion) there were only 11 cells investigated so there may have been a sampling bias of
455
the DSCT neurons recorded. The differences between the preparations used i.e. thalamic or
456
decapitate (Arshavskii et al. 1972c; Arshavsky et al. 1978) vs. pre-collicular/post-mamillary
457
decerebration (present study) may also account for some of the apparent differences.
458
Alternatively, phasic activity could have been dependent on the rostro-caudal location of
459
DSCT cells therefore we extended our recordings to DSCT neurons located in the L1 to L5
460
spinal segments. We found that DSCT neurons from all segments could exhibit rhythmic
461
activity during fictive locomotion as well as during fictive scratch (but our sample size during
20
462
scratch is relatively small). There was no relation between the phasic activity during fictive
463
locomotion and the conduction velocity or the types of excitatory afferent input of a DSCT
464
neuron (see Table 1). There was no relation between the phasic activity during fictive scratch
465
and the types of excitatory afferent input of a DSCT neuron. It is noteworthy that DSCT
466
neurons active during the extension phase of fictive locomotion could be activated by group I
467
afferents from either extensors or flexors.
468
469
470
What is driving tonic and rhythmic activity of the DSCT cells?
Tonic background (or “resting” activity) activity of DSCT cells in unanaesthetized
471
decerebrate preparations without motor activity has been well documented (Arshavskii et al.
472
1972a; Holmqvist et al. 1956). In principle there could be at least three sources for this
473
activity; firstly, a tonic drive from sensory afferents, secondly a tonic excitatory input from
474
spinal interneurons and/or descending pathways, or, thirdly, a spontaneous activity
475
maintained by intrinsic properties in the DSCT neurons.
476
It is known that sensory afferents exert monosynaptic excitation of VSCT neurons to
477
a smaller extent than that of DSCT neurons (Arshavsky et al. 1986; Lundberg 1971;
478
Oscarsson 1965). This is also supported by the different distribution of vesicular
479
glutamatergic transporters (VGlut1 and 2) in the glutamatergic terminals on VSCT and DSCT
480
cells (Shrestha et al. 2012). Myelinated primary afferent terminals contain the VGlut1
481
(Alvarez et al. 2004; Todd et al. 2003; Varoqui et al. 2002), while spinal excitatory
482
interneurons and most descending tracts neurons on the other hand express the VGlut2
483
(Shrestha et al. 2012; Todd et al. 2003; Varoqui et al. 2002). In labeled VSCT cells it is seen
484
that VGlut2 terminals are dominating, while the opposite is true for DSCT cells in Clarke´s
21
485
column (Shrestha et al 2012). If the tonic activity of the DSCT neurons in our present
486
preparation is indeed maintained by tonic excitatory synaptic drive that would either be
487
maintained by activity in the sensory afferents in nerves that were not sectioned during the
488
preparation or originate from the relatively sparse innervations from segmental excitatory
489
interneurons. The most obvious explanation for rhythmic excitatory drive from the spinal
490
CPG circuit of locomotion and scratch would be the activation of spinocerebellar cells via
491
excitatory interneurons which could activate motoneurons in parallel. However, as the overall
492
evidence on the excitation of DSCT neurons belonging to Clarke´s column by spinal
493
excitatory interneurons is sparse (see review by Mann 1973 for electrophysiological data,
494
also see Krutki et al. 2011 and Shrestha et al. 2012 for primarily anatomical evidence) a
495
prominent role of inhibitory spinal interneurons should be considered in “sculpting” the tonic
496
background activity.
497
Populations of inhibitory interneurons that project to VSCT cells include reciprocal Ia
498
inhibitory interneurons (Lindstrom and Schomburg 1974), non-reciprocal “Ib” inhibitory
499
interneurons (Jankowska et al. 2010; Lundberg and Weight 1971) and group “Ib/group II”
500
inhibitory interneurons (Jankowska et al., 2010). No disynaptic inhibitory input from
501
reciprocal Ia inhibitory interneurons was found in DSCT cells (Hongo et al. 1983a;
502
Lindstrom and Takata 1977), but ”Ib” inhibitory interneurons (Hongo et al. 1983a; b) as well
503
as group Ib/II inhibitory interneurons (Jankowska and Puczynska 2008) contact DSCT cells.
504
The Ib inhibitory interneurons are known to be silenced (or inhibited) during fictive
505
locomotion (McCrea et al. 1995) and fictive scratch (Perreault et al. 1999), while the Ia
506
inhibitory (Geertsen et al. 2011) and sub-populations of the Ib/II inhibitory interneurons
507
(Shefchyk et al. 1990; Stecina 2006) are rhythmically active during both fictive locomotion
22
508
and scratch. Thus there is certainly the possibility of phasic inhibitory inputs to both DSCT
509
(from the phasically active Ib/II inhibitory interneurons) and VSCT cells (from the Ia
510
inhibitory interneurons) during these motor tasks.
511
Holmqvist et al. (1956) reported that the background firing following sectioning of all
512
the lumbar and sacral dorsal roots remains unchanged in DSCT cells. Later Arshavskii and
513
colleagues (1972c) described that the background firing was reduced on average from a firing
514
frequency of 12 Hz to 9 Hz following hindlimb deafferentation. The maintenance of some
515
tonic activity following de-afferentation would then suggest that there is a tonic excitation
516
from central sources. The VSCT cells receive direct excitation from several descending tracts
517
(see Arshavsky et al 1986) but no monosynaptic input from the medial longitudinal fasciculus
518
to DSCT cells have been found (Hammar et al. 2011; Krutki et al. 2011). Thus the longer
519
latency inputs could be mediated by the segmental excitatory interneurones even though there
520
may be only few interneurones involved in this loop. Inhibitory input from supraspinal
521
centers to both DSCT and VSCT cells is likely to be mediated by the same spinal
522
interneurons as those used for sensory-evoked inhibition (Baldissera et al. 1981; Hammar et
523
al. 2011).
524
There is no direct evidence for the sustained (or the rhythmic) activity of DSCT cells
525
to be caused by intrinsic cellular properties. However, the Clarke’s column neurons are
526
labeled strongly for the voltage gated CaV1.3 Ca++ channels (Zhang et al. 2008) which carry
527
a persistent inward current (PICs) in many neurons. In motoneurons 5-HT is facilitating the
528
PICs (Hounsgaard et al. 1988; Hounsgaard and Kiehn 1989), and it is therefore interesting
529
that the dorsal horn component of the DCST cells (Jankowska et al. 1995) as well as Clarke´s
530
column cells (Pearson et al. 2000) receive rather intense serotonergic innervations.
23
531
Serotonergic innervations of VSCT cells are indeed similar to that of DSCT cells and
532
hindlimb motoneurons (Hammar and Maxwell 2002), but there is no electrophysiological
533
studies on how serotonin affects intrinsic properties of VSCT neurons during fictive motor
534
activity. In addition, we would also like to bring up the possibility that VSCT and DSCT
535
neurons may undergo a state-dependent enhancement of their excitability during fictive
536
locomotion and scratch. Lumbar motoneurons have been shown to have their voltage
537
threshold for action potential production lowered (i.e. hyperpolarized) during fictive
538
locomotion (Krawitz et al. 2001) and fictive scratch (Power et al. 2010). In addition the post-
539
spike afterhyperpolarization is reduced in spinal motoneurons during both fictive locomotion
540
(Brownstone et al. 1992) and fictive scratch (Power et al. 2010). These changes facilitate
541
motoneuron recruitment and repetitive firing during these activities, and it is possible that
542
similar changes occur in spinocerebellar tract neurons and contribute to their rhythmic
543
activity during these motor outputs. Further investigation of the state-dependent regulation of
544
DSCT cell excitability (especially in relation to the seemingly sparse excitatory synaptic
545
input to Clarke´s column DSCT cells yet a phasic activation during rhythmic motor activity)
546
is required to address this possibility.
547
Finally, we would like to address whether the primary afferent depolarization (PAD)
548
and the subsequent antidromic discharges of the primary afferents evoked by the locomotor
549
and/or scratch networks (Bayev and Kostyuk 1982; Bayev and Kostyuk 1981; Bayev et al.
550
1978; Beloozerova and Rossignol 1999) could contribute significantly to the activity of
551
spinocerebellar tract cells during fictive locomotion and scratch. Firstly, the tonic discharge
552
of afferent fibers in the resting decerebrate state is actually reduced during activation of the
553
locomotor CPG –and even more reduced during fictive scratch when compared to resting i.e.
24
554
no motor activityin the same preparation (Cote and Gossard 2003). Secondly, the CPG
555
activity leads to a phase-related modulation of the dorsal root potentials but at the same time
556
it also leads to reduction of transmission in sensory afferent-evoked PAD pathways (Cote
557
and Gossard 2003). All in all, we would have expected a large difference in the modulation of
558
the firing frequency in DSCT cells during the two behaviors i.e. lower changes during fictive
559
scratch than during fictive locomotion. Our results, however, show overlapping frequency
560
modulation (see Fig. 8) therefore we find it unlikely that PAD evoked by the locomotor and
561
scratch CPG is the cause of the firing activity of DSCT cells during fictive motor actions.
562
563
Functional implications of rhythmic firing in the dorsal and ventral spinocerebellar
564
pathways during motor activity
565
An extensive discussion on the role of the spinocerebellar pathways with regard to
566
sensory feedback to the cerebellum during motor activity is outside the scope of this study.
567
Our present results emphasize the central -both spinal and descending- inputs to the DSCT
568
cells which have been known (see previous section) but have been underestimated because of
569
the emphasis on the strong sensory input to DSCT cells that has been prevalent since the
570
1960s.
571
Our data show that there is an underlying “CPG-driven” activity of DSCT cells during
572
rhythmic motor tasks. Even if the actual recruitment and final firing rate is more strongly
573
influenced by peripheral afferent activity, the underlying locomotor and scratch potentials
574
would provide a fluctuating baseline on which the afferent input is superimposed. The
575
convergence of sensory input with the baseline excitation or inhibition may serve as a gate to
576
allow the selective transmission of sensory input to the cerebellum.
25
577
Any hypothesis on the role of the information transmitted by the DSCT (in general,
578
and as compared to that by the VSCT) has to take into account the terminations and
579
interactions at cerebellar level, starting with the convergence at the granule cells e.g. (De
580
Zeeuw et al. 2011; Ekerot and Jorntell 2008) and continuing with the interactions and the
581
convergence from the olivocerebellar projections at Purkinje cell level and at the cerebellar
582
cortical layer (Valle et al. 2012). Several recent reviews focus on the possibility that specific
583
spatiotemporal firing patterns may be of particular significance for information processing in
584
the cerebellum (De Zeeuw et al. 2011; Perciavalle et al. 1995). In a long series of
585
publications Bosco, Poppele and colleagues emphasized the wide, but organized,
586
convergence onto the DSCT as reviewed by (Bosco and Poppele 2001). They argued that the
587
DCST could work as parallel distributed networks that relay information on limb
588
biomechanics and kinematics. In the discussion of the recent report on the convergence of
589
cortico-spinal excitation and inhibition and afferent input to DSCT cells Hantman and Jessell
590
(2010) is referring to internal models of the planned motor activity, and the corollary
591
discharge to distinguish exafference (sensory signals generated from external stimuli in the
592
environment) from re-afference (sensory signals resulting from an animal's own actions) due
593
to the planned movement. Thus they place the DSCT cells in a more central position than
594
previously viewed for motor planning and evaluation. Our present results add to previous
595
evidence demonstrating that the early idea of the DSCT pathway as primarily mediating
596
sensory afferent information cannot be maintained. Even though there are differences
597
between the convergence of sensory afferent input and central excitation versus inhibition for
598
DSCT and VSCT neurons, the overall similarities seem to be dominating during rhythmic
599
motor actions. While the hypothesis of VSCT cells being an input-output comparator stands
26
600
yet unchallenged, the role of DSCT cells with respect to sensory-motor integration must be
601
re-evaluated with an increased appreciation that transmission through the DSCT reflects the
602
convergence of activity in spinal motor generating circuitry and peripheral sensory afferents.
603
604
Acknowledgements
605
The authors wish to thank Lillian Grøndahl, Ingrid Kjær and Gilles Detillieux for their
606
excellent technical assistance, Monica Gorassini for participating in some experiments of the
607
first series of experiments, and E. Jankowska for constructive comments on the interpretation
608
of the results. B. Fedirchuk was a Medical Research Council of Canada Postdoctoral
609
Fellowship recipient, K. Stecina was supported by the EU FP6 MarieCurie Actions Intra-
610
European Fellowship and the Danish Agency for Science Technology and Innovation.
611
27
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
References
Alvarez FJ, Villalba RM, Zerda R, and Schneider SP. Vesicular glutamate
transporters in the spinal cord, with special reference to sensory primary afferent
synapses. J Comp Neurol 472: 257-280, 2004.
Arshavskii I, I., Berkinblit MB, Gel'fand IM, Orlovskii GN, and Fukson OI. Activity
of neurons of the ventral spino-cerebellar tract during locomotion. Biofizika 17: 883890(translated in Biophysics 817:926-935), 1972a.
Arshavskii I, I., Berkinblit MB, Gel'fand IM, Orlovskii GN, and Fukson OI. Activity
of ventral spino-cerebellar tract neurons during locomotion of cats with deafferented
hindlimbs. Biofizika 17: 1112-1118(translated in Biophysics 1120:1762-1764), 1972b.
Arshavskii I, I., Berkinblit MV, Gel'fand IM, Orlovskii GN, and Fukson OI. Activity
of neurons of the dorsal spinocerebellar tract during locomotion. Biofizika 17: 487494(translated in Biophysics 417:508-514), 1972c.
Arshavskii II, Gel´fand IM, Orlovskii GN, and Pavlova GA. Letter: Neuron activity
in the the ventral spinocerebellar tract during "fictious" scratching. Biofizika 20: 748749, 1975.
Arshavsky YI, Berkinblit MB, Fukson OI, Gelfand IM, and Orlovsky GN. Origin of
modulation in neurones of the ventral spinocerebellar tract during locomotion. Brain
Res 43: 276-279, 1972a.
Arshavsky YI, Berkinblit MB, Fukson OI, Gelfand IM, and Orlovsky GN.
Recordings of neurons of the dorsal spinocerebellar tract during evoked locomotion.
Brain Res 43: 272-275, 1972b.
Arshavsky YI, Gelfand IM, and Orlovskii GN. Cerebellum and Rhythmical
Movements. Berlin: Springer-Verlag, 1986.
Arshavsky YI, Gelfand IM, Orlovsky GN, and Pavlova GA. Messages conveyed
by spinocerebellar pathways during scratching in the cat. II. Activity of neurons of the
ventral spinocerebellar tract. Brain Res 151: 493-506, 1978.
Baldissera F, Hultborn H, and Illert M. Integration in spinal neuronal systems. In:
Handbook of Physiology - The Nervous System II, edited by Brookhart JM,
Mountcastle VB, Brooks VB, and Geiger SR. Bethesda, Maryland: American
Physiological Soc., 1981, p. 509-595.
Bayev KV, and Kostyuk PG. Polarization of primary afferent terminals of
lumbosacral cord elicited by the activity of spinal locomotor generator. Neuroscience
7: 1401-1409, 1982.
Bayev KV, and Kostyuk PG. Primary afferent depolarization evoked by the activity
of spinal scratching generator. Neuroscience 6: 1981.
Bayev KV, Panchin Yu V, and Skryma RN. Primary afferent depolarization during
fictive scratching in thalamic cats. Neurophysiology (Kiev) 10: 173-176, 1978.
Beloozerova I, and Rossignol S. Antidromic discharges in dorsal roots of
decerebrate cats. I. Studies at rest and during fictive locomotion. Brain Res 846: 87105, 1999.
Bosco G, and Poppele RE. Proprioception from a spinocerebellar perspective.
Physiol Rev 81: 539-568, 2001.
Brownstone R, Jordan LM, Kriellaars DJ, Noga BR, and Shefchyk SJ. On the
28
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
regulation of repetitive firing in lumbar motoneurones during fictive locomotion in the
cat. Exp Brain Res 90: 441-455, 1992.
Cote M-P, and Gossard J-P. Task-dependent presynaptic inhibition. J
Neuroscience 23: 1886-1893, 2003.
De Zeeuw CI, Hoebeek FE, Bosman LW, Schonewille M, Witter L, and
Koekkoek SK. Spatiotemporal firing patterns in the cerebellum. Nat Rev Neurosci
12: 327-344, 2011.
Eccles JC, Ito M, and Szentagothai J. The cerebellum as a neuronal machine.
Berlin: Springer, 1967.
Edgley SA, and Gallimore CM. The morphology and projections of dorsal horn
spinocerebellar tract neurones in the cat. J Physiol 397: 99-111, 1988.
Ekerot CF, and Jorntell H. Synaptic integration in cerebellar granule cells.
Cerebellum 7: 539-541, 2008.
Fedirchuck B, Hultborn H, D.J. B, and Gorassini M. Dorsal spinocerebellar tract
neurons can be influenced by the neural circuitry producing fictive locomotion in the
cat. In: Society for Neuroscience1995.
Geertsen SS, Stecina K, Meehan CF, Nielsen JB, and Hultborn H. Reciprocal Ia
inhibition contributes to motoneuronal hyperpolarisation during the inactive phase of
locomotion and scratching in the cat. J Physiol 589: 119-134, 2011.
Hammar I, Krutki P, Drzymala-Celichowska H, Nilsson E, and Jankowska E. A
trans-spinal loop between neurones in the reticular formation and in the cerebellum.
J Physiol 589: 653-665, 2011.
Hammar I, and Maxwell DJ. Serotoninergic and noradrenergic axons make
contacts with neurons of the ventral spinocerebellar tract in the cat. J Comp Neurol
443: 310-319, 2002.
Hantman AW, and Jessell TM. Clarke's column neurons as the focus of a
corticospinal corollary circuit. Nat Neurosci 13: 1233-1239, 2010.
Holmqvist B, Lundberg A, and Oscarsson O. Functional organization of the dorsal
spino-cerebellar tract in the cat. V. Further experiments on convergence of excitatory
and inhibitory actions. Acta Physiol Scand 38: 76-90, 1956.
Hongo T, Jankowska E, Ohno T, Sasaki S, Yamashita M, and Yoshida K.
Inhibition of dorsal spinocerebellar tract cells by interneurones in upper and lower
lumbar segments in the cat. J Physiol 342: 145-159, 1983a.
Hongo T, Jankowska E, Ohno T, Sasaki S, Yamashita M, and Yoshida K. The
same interneurones mediate inhibition of dorsal spinocerebellar tract cells and
lumbar motoneurones in the cat. J Physiol 342: 161-180, 1983b.
Hongo T, and Okada Y. Cortically evoked pre-, and postsynaptic inhibition of
impulse transmission to the dorsal spinocerebellar tract. Exp Brain Res 3: 163-177,
1967.
Hongo T, Okada Y, and Sato M. Corticofugal influences on transmission to the
dorsal spinocerebellar tract from hindlimb primary afferents. Exp Brain Res 3: 135149, 1967.
Hounsgaard J, Hultborn H, Jespersen B, and Kiehn O. Bistability of alphamotoneurones in the decerebrate cat and in the acute spinal cat after intravenous 5hydroxytryptophan. J Physiol 405: 345-367, 1988.
Hounsgaard J, and Kiehn O. Serotonin-induced bistability of turtle motoneurones
29
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
caused by a nifedipine-sensitive calcium plateau potential. J Physiol 414: 265-282,
1989.
Jankowska E, Krutki P, and Hammar I. Collateral actions of premotor interneurons
on ventral spinocerebellar tract neurons in the cat. J Neurophysiol 104: 1872-1883,
2010.
Jankowska E, Maxwell DJ, Dolk S, Krutki P, Belichenko PS, and Dahlstrom A.
Contacts between serotoninergic fibres and dorsal horn spinocerebellar tract
neurons in the cat and rat: a confocal microscopic study. Neuroscience 67: 477-487,
1995.
Jankowska E, and Puczynska A. Interneuronal activity in reflex pathways from
group II muscle afferents is monitored by dorsal spinocerebellar tract neurons in the
cat. J Neurosci 28: 3615-3622, 2008.
Kitamura T, and Yamada J. Spinocerebellar tract neurons with axons passing
through the inferior or superior cerebellar peduncles. A retrograde horseradish
peroxidase study in rats. Brain Behav Evol 34: 133-142, 1989.
Krawitz S, Fedirchuk B, Dai Y, Jordan L, and McCrea D. The voltage threshold for
action potential production in hindlimb motoneurons is lowered during fictive
locomotion in the cat. J Physiol 532: 271-281, 2001.
Krutki P, Jelen S, and Jankowska E. Do premotor interneurons act in parallel on
spinal motoneurons and on dorsal horn spinocerebellar and spinocervical tract
neurons in the cat? J Neurophysiol 2011.
Lindstrom S, and Schomburg ED. Group I inhibition in Ib excited ventral
spinocerebellar tract neurones. Acta Physiol Scand 90: 166-185, 1974.
Lindstrom S, and Takata M. Lack of recurrent depression from motor axon
collaterals of IaIPSPs in dorsal spinocerebeller tract neurones. Brain Res 129: 158161, 1977.
Lundberg A. Function of the ventral spinocerebellar tract. A new hypothesis. Exp
Brain Res 12: 317-330, 1971.
Lundberg A, and Oscarsson O. Three ascending spinal pathways in the dorsal part
of the lateral funiculus. Acta Physiol Scand 51: 1-16, 1961.
Lundberg A, and Weight F. Functional organization of connexions to the ventral
spinocerebellar tract. Exp Brain Res 12: 295-316, 1971.
Mann MD. Clarke's column and the dorsal spinocerebellar tract: a review. Brain
Behav Evol 7: 34-83, 1973.
McCrea DA, Shefchyk SJ, Stephens MJ, and Pearson KG. Disynaptic group I
excitation of synergist ankle extensor motoneurones during fictive locomotion in the
cat. J Physiol 487: 527-539, 1995.
Orsal D, Perret C, and Cabelguen JM. Comparison between ventral
spinocerebellar and rubrospinal activities during locomotion in the cat. Behav Brain
Res 28: 159-162, 1988.
Oscarsson O. Functional Organization of the Spino- and Cuneocerebellar Tracts.
Physiol Rev 45: 495-522, 1965.
Pearson JC, Sedivec MJ, Dewey DE, and Fyffe RE. Light microscopic
observations on the relationships between 5-hydroxytryptamine-immunoreactive
axons and dorsal spinocerebellar tract cells in Clarke's column in the cat. Exp Brain
Res 130: 320-327, 2000.
30
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
Perciavalle V, Bosco G, and Poppele R. Correlated activity in the spinocerebellum
is related to spinal timing generators. Brain Res 695: 293-297, 1995.
Perreault M-C. Motoneurons have different membrane resistance during fictive
scratching and weight support. J Neurosci 22: 8259-8265, 2002.
Perreault MC, Enriquez-Denton M, and Hultborn H. Proprioceptive control of
extensor activity during fictive scratching and weight support compared to fictive
locomotion. J Neurosci 19: 10966-10976, 1999.
Power KE, McCrea DA, and Fedirchuk B. Intraspinally mediated state-dependent
enhancement of motoneurone excitability during fictive scratch in the adult
decerebrate cat. J Physiol 588: 2839-2857, 2010.
Shefchyk S, and Jordan L. Excitatory and inhibitory postsynaptic potentials in
alpha-motoneurons produced during fictive locomotion by stimulation of the
mesencephalic locomotor region. J Neurophysiol 53: 1345-1355, 1985.
Shefchyk S, McCrea D, Kriellaars D, Fortier P, and Jordan L. Activity of
interneurons within the L4 spinal segment of the cat during brainstem-evoked fictive
locomotion. Exp Brain Res 80: 290-295, 1990.
Shrestha SS, Bannatyne BA, Jankowska E, Hammar I, Nilsson E, and Maxwell
DJ. Excitatory inputs to four types of spinocerebellar tract neurons in the cat and the
rat thoraco-lumbar spinal cord. J Physiol 590: 1737-1755, 2012.
Stecina K. Preferential suppression of transmission and candidate neurones
mediating reflex actions from muscle group II afferents during fictive motor activityElectronic Thesis and Dissertations Collections (public access URL). In: PhD thesis
in the Department of Physiology. Winnipeg: Manitoba, 2006, p. 167.
Stecina K, Zhang H, and Hultborn H. Spinal interneurones involved in mediating
extensor group I muscle afferent actions during fictive locomotion and scracth. In:
Society for Neuroscience 78-11/II10 2007.
Stecina K, Zhang M, Sukiyasian N, and Hultborn H. Repetitive firing of feline
dorsal spinocerebellar neurones during fictive motor activity. In: Federation of
European Neuroscience Societies 2008.
Todd AJ, Hughes DI, Polgar E, Nagy GG, Mackie M, Ottersen OP, and Maxwell
DJ. The expression of vesicular glutamate transporters VGLUT1 and VGLUT2 in
neurochemically defined axonal populations in the rat spinal cord with emphasis on
the dorsal horn. Eur J Neurosci 17: 13-27, 2003.
Valle MS, Eian J, Bosco G, and Poppele RE. The organization of cortical activity in
the anterior lobe of the cat cerebellum during hindlimb stepping. Exp Brain Res 216:
349-365, 2012.
Varoqui H, Schafer MK, Zhu H, Weihe E, and Erickson JD. Identification of the
differentiation-associated Na+/PI transporter as a novel vesicular glutamate
transporter expressed in a distinct set of glutamatergic synapses. J Neurosci 22:
142-155, 2002.
Zhang M, Moller M, Broman J, Sukiasyan N, Wienecke J, and Hultborn H.
Expression of calcium channel CaV1.3 in cat spinal cord: light and electron
microscopic immunohistochemical study. J Comp Neurol 507: 1109-1127, 2008.
31
796
Figure Captions
797
798
Figure 1. Schematic illustration of methods
799
800
A. For the identification of spinocerebellar neurons stimulation of the cerebellar surface
801
(surface stim; indicated by the open arrow) was used in the first series of experiments. The
802
mesencephalic locomotor region was stimulated (MLR stim) as shown by the double arrows
803
in order to evoke fictive locomotion.
804
B. For the identification of spinocerebellar neurons intra-cerebellar (intra-CB) stimulation
805
was used at approximate sites indicated by the filled arrow in the second series of
806
experiments.
807
C. Verification of antidromic activation of tract cells by the stimulation of the ipsilateral
808
(ipsi) or the contralateral (contra) dorsolateral funiculus at the level of the first cervical (C1)
809
vertebra was used prior to fictive motor activity (monitored by electroneurogram recordings,
810
ENGs, of hindlimb muscle nerves). Extra-, and/or intracellular recordings in the lumbar
811
segments (L1 – L5) were collected from identified spinocerebellar tract cells.
812
D. Cord dorsum potential (cdp) recordings after intra-cerebellar stimulation at identified
813
depths with reference to the cerebellar surface. Note that maximal volleys were evoked at
814
depths ranging between 5-7 mm.
815
E. Single microelectrode recordings extracellularly (e.c.) with antidromic DSCT activation
816
upon stimulation at 6.0 mm depth in the CB. The last trace on the bottom is the cdp recorded
817
at L6 with the dotted line indicating the arrival of the descending volley. Note the collision of
818
the antidromic spike when the cell was firing spontaneously (arrow).
819
32
820
821
Figure 2. Rhythmic activity of an extracellularly recorded DSCT neuron during fictive
822
locomotion
823
824
A. The extracellular microelectrode recording (upper trace) and the rectified and filtered
825
ENG recordings from a variety of hindlimb muscle nerves on the left (L) and right (R) side
826
(Sart, SmAB, PBSt and DP nerves).
827
B. The step-cycle based average of the normalized (30 bins per cycle) instantaneous firing
828
frequency (IFF) and ENGs of Sart and SmAB.
829
33
830
831
Figure 3. Intracellular recording from a DSCT neuron during fictive locomotion
832
833
A. Intracellular microelectrode recording (DSCT i.c.) and rectified and filtered ENG
834
recordings (same ENG abbreviations as in Fig. 2) during MLR evoked (100 μA, 20 Hz)
835
fictive locomotion. This is the same neuron that was recorded from an extracellular position
836
in Fig. 2. Locomotor related depolarizations in the membrane potential (gray boxes) are
837
apparent.
838
B. The step-cycle based average of the normalized (30 bin per cycle) membrane potential
839
(DSCT i.c., black line) that was 3.5 mV after action potential generating sodium channels had
840
been inactivated just prior to the intracellular recording period by injection of a depolarizing
841
current overlaid on the IFF (grey line) obtained from recording in an extracellular position
842
(same as in Fig. 2B). The last two lines show average and normalized (30 bin per cycle)
843
ENGs of Sart and SmAB.
844
34
845
846
Figure 4. Extracellular recording from a VSCT neuron during fictive locomotion
847
848
A. Extracellular microelectrode recording (VSCT e.c., top trace) and rectified and filtered
849
ENG recordings during MLR evoked (150 μA, 20 Hz) fictive locomotion. See ENG
850
abbreviations as in Fig. 2 and an additional ankle extensor, gastrocnemious, GS, is shown.
851
B. The step-cycle based average of the normalized (30 bin per cycle) IFF ENGs of Sart and
852
SmAB.
853
35
854
855
Figure 5. Intracellular recording from a VSCT neuron during fictive locomotion
856
857
A. Intracellular microelectrode (VSCT i.c.) recording of the activity of the same neuron as in
858
Fig. 4 and hindlimb ENG activity (ENG abbreviations as in Fig. 4). . Locomotor related
859
depolarizations in the membrane potential (gray boxes) are apparent..
860
B. The step-cycle based average of the normalized (30 bins per cycle) membrane potential
861
(VSCT i.c., black line) that was 7.1 mV after action potential generating sodium channels had
862
been inactivated just prior to the intracellular recording period by injection of a depolarizing
863
current overlaid on the IFF of this cell (grey line same as in Fig. 4).
864
36
865
866
Figure 6. Extracellular recording from a DSCT neuron during fictive scratch
867
868
A. The extracellular microelectrode recording ( DSCT e.c., top trace) the discriminated unit
869
after spike-sorting (second trace from top) and the rectified and filtered; ENG recordings
870
from a variety of hindlimb muscle nerves with abbreviations as in Fig. 2 and in addition,
871
peroneus longus (PerL) is illustrated.
872
B. The step-cycle based average of the normalized (30 bins per cycle) IFF (upper traceand
873
ENGs from Sart, PerL and GS.
874
37
875
876
Figure 7. Intracellular recording from a DSCT neuron during fictive scratch
877
878
A. Intracellular microelectrode recording (DSCT i.c.) and rectified and filtered ENG
879
recordings (of plantaris, Pl and TA) during fictive scratch. Scratch-cycle related
880
depolarizations in the membrane potential (gray boxes) are apparent.
881
B. The scratch-cycle based average of the normalized (30 bins per cycle) membrane potential
882
(DSCT i.c., black line) that was 5 mV after action potential generating sodium channels had
883
been inactivated just prior to the intracellular recording period by injection of depolarizing
884
current overlaid on the IFF of this cell (grey line) obtained from recording in an extracellular
885
position in another bout of fictive scratch.
886
38
887
888
Figure 8. Comparison of DSCT and VSCT firing frequency modulation during fictive
889
locomotion and scratch
890
891
Histograms (with 10 Hz binning) showing the number of cells and the change in their
892
instantaneous firing frequency (IFF) during fictive locomotion (A) and during fictive scratch
893
(B). The DSCT cells (filled bars) showed significantly lower modulation of the firing rates
894
then the VSCT cells (open bars). Note that similar firing frequencies were observed during
895
fictive locomotion and scratch.
896
39
897
898
Table 1. Summary of DSCT activity during fictive locomotion and excitatory input from
899
peripheral afferents
900
901
Excitatory input from extensor group I (Extensor gr I), flexor group I (Flexor gr I), cutaneous
902
and group II muscle afferents (Group II & Cutaneous), Group II muscle afferents only, or no
903
input from any of the tested nerves was used for grouping DSCT cells. The percent of cells in
904
each group per total DSCT tested is shown as “% of this category” with the percentage of
905
total summarized in the “Total” column. The number of cells with phase-related modulation
906
of firing frequency are shown in the first row. The number of cells with higher peak firing
907
frequency in the flexion (F) and the extension (E) phase is shown as labeled accordingly. The
908
spinal segments that the neurons were recorded from (L1-L5) are indicated for each group,
909
and summarized in the respective columns. The second row shows the segmental location of
910
those cells that had no step-cycle related modulation of firing frequency. The third row
911
summarized the overall distribution of cells in the groups based on excitatory input (sample
912
size and % of total).
913
A
MLR stim
C
surface
stim
curare/bicucculine
application
recording microelectrode
in lumbar segments
C1 stimulation
contra / ipsi
~
~
B
intra-CB
stim
hindlimb
ENGs
MLR
stim
~
E
D
intra-CB stim depth
cdp at L6
intra-CB stim depth = 6.0 mm
e.c.
4 mm
5 mm
6 mm
7 mm
cdp
10 mV
8 mm
2 ms
2 ms
A
DSCT e.c.
L Sart
L SmAB
L DP
R Sart
R PBSt
1s
B
80
Instantaneous
Firing
Frequency Hz
78 Hz
0
IFF at rest = 6.4 Hz
L Sart
L SmAB
2
averaged step cycle
A
4 mV
DSCT i.c.
-52 mV
L Sart
L SmAB
L DP
R Sart
R PBSt
1s
B
-51
DSCT i.c.
-52
3.5 mV
Em
(mV) -53
-54
-55
Em at rest =
- 52.8 mV
IFF From
Figure 2B
L Sart
L SmAB
2
averaged step cycle
A
VSCT e.c.
L Sart
L SmAB
L GS
R Sart
R SmAB
1s
B
300
Instantaneous
Firing
Frequency Hz 200
300 Hz
100
0
IFF at rest = 0 Hz
L Sart
L SmAB
2
averaged step cycle
A
4mV
VSCT i.c.
-52
L Sart
L SmAB
L GS
R Sart
R SmAB
1s
B
-44
-46
VSCT i.c.
mV
-48
-50
-52
7.1 mV
IFF from
Figure 3B
L Sart
L SmAB
2
averaged step cycle
Em at rest =
-53.7 mV
A
DSCT e.c.
GS
PerL
PBSt
Sart
B
Instantaneous 100
Firing
Frequency
Hz
1s
65 Hz
IFF at rest = 37 Hz
0
Sart
PerL
GS
2
averaged scratch cycle
A
DSCT i.c.
-75
mV
-80
Pl
TA
B
DSCT i.c.
-78
0.5 s
IFF
44 Hz
Em at rest
- 78.6 mV
mV -80
-82
Pl
TA
2
averaged scratch cycle
Fictive Locomotion
A
24
DSCT
VSCT
number of cells
10
8
6
4
2
0
0
number of cells
B
50
100
150
200
250
300
change in IFF (Hz)
Fictive Scratch
10
8
6
4
2
0
0
50
100
150
200
250
300
change in IFF (Hz)
Table 1
DSCT Neuron Activity
During Fictive Locomotion
Source of excitatory synaptic input
Group I from
Extensors
Modulated with Step Cycle; n
Active Phase of Step Cycle
Segmental Location
(n-segment)
7=F
18
11=E
Group I from
Other Nerves
6
2=F
4=E
Group II &
Cutaneous
5=F
9
4=E
Group II
Only
4=F
12
2-L1
1-L2
3-L3
3-L4
11-L4
2-L3
4-L4
2-L4
1-L5
2-L4
2-L5
8=E
1-L2
3-L3
4-L4
2-L3
2-L4
No excitatory
inputs
7=F
12
1-L1
5-L3
1-L4
Total
5=E
25=F
1-L2
1-L3
3-L4
3-L1
1-L2
12-L3
8-L4
1-L5
57
32=E
2-L2
4-L3
24-L4
2-L5
% of this category
32%
10%
16%
21%
21%
(70% of total)
Not Modulated; n
7
1
7
3
6
24
1-L3
1-L3
6-L4
1-L2
4-L2
2-L4
2-L4
6-L2
2-L3
16-L4
Segmental Location
(n-segment)
1-L2
% of this category
29%
4%
29%
13%
25%
(30% of total)
Total
% of total
25
31%
7
9%
16
20%
15
18%
18
22%
81
6-L4