| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
1Unité de Physiologie et Biomécanique de la Locomotion, Département D'Éducation Physique et de Réadaptation, Université Catholique de Louvain, Louvain-la-Neuve, Belgium; and 2Department of Physiology, Université de Montréal, Montreal, Quebec, Canada
Submitted 31 March 2006; accepted in final form 8 July 2006
 |
ABSTRACT |
|---|
|
|
|
INTRODUCTION |
|---|
|
) and some neurons have axons that cross the midline at the cervical or lumbar levels and innervate the gray matter of both sides (Matsuyama et al. 1988, 1997, 1999). In addition, reticulospinal axons also terminate on commissural interneurons so that an action on contralateral activity may also be mediated by this pathway (Jankowska et al. 2003; Matsuyama et al. 2004). As such, there is a firm anatomical substrate demonstrating that reticulospinal neurons can produce widespread effects throughout the neuraxis, including both ipsi- and contralateral to the cell body.
Functional studies provide evidence that this anatomical substrate is used to facilitate the coordination of interlimb activity and to produce the complex patterns of muscular activity that are used to provide postural support in response to voluntary movements or unpredictable perturbations. In some of the earliest studies of the PMRF, Sprague and Chambers (1954) expanded on Magoun's original thesis of a structure producing global excitation or inhibition (Magoun 1944; Magoun and Rhines 1946) to demonstrate that lower-intensity stimulation of the PMRF in the decerebrate cat produced coordinated patterns of flexion and extension in the four limbs with flexion predominating in ipsilateral limbs and extension in contralateral limbs. More recent studies in awake animals have further amplified our understanding of the complex nature of the reticulospinal pathways. Microstimulation studies in awake cats (Drew and Rossignol 1990a) confirmed Sprague and Chamber's (1954) original findings of a predominant ipsilateral flexion and contralateral extension bias (together with ipsilateral head turning), although recordings of electromyographic (EMG) activity also showed that the stimulation was equally effective in activating ipsilateral extensors and contralateral flexors (Drew and Rossignol 1990b). Moreover, a recent study has shown strong bilateral projections to shoulder muscles in the primate (Davidson and Buford 2004). Together, these studies confirm that the PMRF is capable of influencing muscle activity at all levels of the neuraxis, both ipsi- and contralateral to the stimulation site.
Although stimulation to the PMRF produced a single characteristic pattern of activity when the cat was at rest, stimulation during walking transformed this pattern of activity into a more functional one that produced appropriate phase-dependent activation of muscles (Drew 1991; Drew and Rossignol 1984; Orlovsky 1972; Perreault et al. 1994). In the intact cat (Drew 1991), stimulation during the swing phase of the ipsilateral forelimb increased responses in ipsilateral flexors and, simultaneously, in contralateral extensors. Stimulation during contralateral swing facilitated activity in the contralateral flexors and produced facilitation or suppression of ipsilateral extensors. Responses were also observed in hindlimb flexor and extensor muscles and were, likewise, phase dependent. We have argued that this pattern of activity provides a flexible substrate that would appropriately facilitate the integration of postural responses into the locomotor pattern.
In support of this proposition, recent studies (Drew et al. 2004; Prentice and Drew 2001) showed that a large proportion of cells, including identified reticulospinal neurons (RSNs), exhibited multiple increases in activity as each limb in turn stepped over the obstacle. We suggested that the discharge activity of these cells provided information concerning the time and magnitude of the postural activity that occurred in the supporting limbs as any one limb stepped over the obstacle. However, we suggested that they could not specify the details of the postural pattern which would depend on the excitability of the spinal circuits and would be contingent on activity in other descending pathways. In this hypothesis, a substantial aspect of the final determination of the postural pattern is thus dependent on the rhythmical changes in the excitability of the spinal interneurons onto which these fibers impinge. Discharge activity occurring during ipsilateral swing would facilitate ipsilateral flexor muscles and contralateral extensor muscles, whereas activity occurring during ipsilateral stance would facilitate ipsilateral extensors and contralateral flexors (see Fig. 3 in Drew et al. 2004 and Fig. 16 in this manuscript).
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
). The question we now address is how the discharge activity of these cells is modified when the cat makes a reach with the right limb, contralateral to the recording site. In this situation, in which there is no base rhythmical activity in the spinal networks (as during locomotion), does the reticulospinal system continue to produce a similar pattern of activity during reaches of both the ipsilateral and contralateral limbs, or does it, instead, produce a pattern of activity specific to the requirements of each reach? For example, reaches of the left limb require increased activity in the extensors of the right forelimb and the left hindlimb while movements of the right limb require increased extensor activity in the left forelimb and the right hindlimb, i.e., reciprocal patterns of activity (see Figs. 2 and 3). The results favor a contribution to postural control similar to that observed during locomotion.
Preliminary results from these experiments have been published in abstract form (Schepens and Drew 2000, 2001, 2003b)
|
|
METHODS |
|---|
|
These experiments were performed on the same two cats (RS22 and RS23) used in our previous publications (Schepens and Drew 2003a, 2004). All surgical and experimental procedures are described in detail in those two papers.
In brief, cats were trained to stand quietly on four force platforms and to reach to a tube to obtain a food reward (Fig. 1). A shutter over the end of the tube, and controlled by computer, denied access outside the confines of the task. Cats were trained to reach with both the left and right forelimbs. After training, the cats were prepared for surgery under general anesthesia (2–3% isoflurane with oxygen) and in sterile conditions. A rectangular base-plate (internal measurements: 10 * 8 mm) was implanted over the left cerebellum to provide access to the brain stem reticular formation, and pairs of Teflon-insulated, stainless-steel wires were sewn, bilaterally, into the muscle bellies of selected fore- and hindlimb muscles as well as some axial muscles. The muscles used in these studies and details of their activation patterns can be found in Schepens and Drew (2004). Three microwires (50 µm diam) were inserted into the spinal cord at L2 to provide a means of antidromically identifying reticulospinal cells projecting to the ipsilateral, lumbar spinal cord. After surgery the cats were administered analgesics during a period of 48 h (buprenorphrine 5 µg/kg) and antibiotics during a period of 
10 days (penicillin G 40,000 IU). All surgical procedures were approved by the institutional ethics committee and followed National guidelines.
After recovery from the surgery, an electrode was introduced into the PMRF and the discharge activity of single neurons was recorded. All isolated single neurons were recorded, regardless of whether they could be antidromically identified from the microwire electrodes inserted at L2 (RSNs) or not (unidentified neurons). After isolation of a single neuron, cell discharge activity was recorded during a period of treadmill locomotion, and the cat was then transferred to the reaching apparatus. All isolated cells were recorded during both tasks, when possible, regardless of the pattern of discharge activity on the treadmill. Discharge activity was recorded during a minimum of five reaches from each forelimb (generally, 5 reaches with the left forelimb, 10 reaches with the right forelimb, and then 5 more with the left forelimb). Supplementary reaches, if present, were made in blocks of five with each limb. The cat was then transferred back to the treadmill and another cell isolated.
Data analysis
All data were recorded on-line. Forces exerted against the platforms (vertical, V; mediolateral, ML; and anteroposterior, AP) and EMG data were recorded at 1 kHz while the untreated unit trace was digitized at 100 kHz for off-line discrimination using custom routines.
Analysis routines were identical to those used in a previous publication (Schepens and Drew 2004). In brief, cells were classified as showing significant increases or decreases in discharge if they exceeded ±2 SDs of the average control activity, calculated from a 500-ms period preceding the onset of the cue signal (Fig. 1). Cells showing increased activity were further classified as phasic, phasic/tonic, or tonic based on a comparison of the discharge activity in the control period, during the dynamic part of the reach [from the onset of cleidobrachialis (ClB) activity to entry of the paw in the tube], and during the static period (between 1,000 and 1,500 ms after ClB onset). Phasic cells were defined as those cells showing a significant increase during the dynamic period but in which discharge activity during the subsequent static period was <150% of that during the control period. Cells with a tonic component had a discharge rate during the static period that was >150% of the control period. Cells with a phasic/tonic discharge additionally had a discharge during the dynamic period that was >125% of that during the static period (see Schepens and Drew 2004). When classifying cell activity during the period before the onset of the GO signal (the pretrigger period), cell discharge was likewise considered to be significantly modified if it exceeded ±2 SD of the control activity.
The onset of activity in the cell discharge and in selected muscles and force traces was measured from the computer traces during individual trials and was used to determine the temporal relationships of cell activity to different behavioral events. Linear regressions between either the latency of the onset of cell discharge or the lead time (latency of ClB onset - latency of cell onset) as a function of the time of onset of the ClB were used to determine if cell discharge was better related to movement onset or to the GO signal. Cells better related to the movement show a slope of 1.0 when cell discharge is plotted as a function of ClB onset; cells better related to the GO signal show a slope of 1.0 when the lead time is plotted against the ClB onset (Chapman et al. 1986; Schepens and Drew 2003a, 2004; Vicario et al. 1983). Linear regressions were also calculated between different epochs of the cell discharge and the onset and offset of activity in different muscles and force traces.
Quantitative analysis of the relationship between cell discharge activity and the magnitude of the EMG or force activity was calculated from temporal slices. For each individual trial, both the instantaneous frequency of the cell activity and the amplitude of the EMG and force traces were integrated over consecutive 50-ms bins. A trial of 8 s duration would, therefore, yield 160 values for the cell discharge and for each of the 64 traces of analog activity. These temporal slices were repeated for each trial included in the analysis, and the linear regression analysis was performed between the cell trace and each of the analog channels. Linear regression analysis was also performed using only the changes in cell and analog activity during the period from the GO signal until the time that the paw entered the tube. As this period includes both the initial, preparatory, anticipatory postural adjustments (pAPA) and the dynamic period from the onset of the ClB until the time that the paw enters the tube, we refer to this as the pAPA + dynamic period. In this case temporal slices of 20 ms were used. Multiple regressions were compiled from the same data using the manual stepwise function in Systat (v9.0). In this case, traces were added one at a time in an order specified by the user (see RESULTS).
For all of the recorded cells, we performed spike triggered averaging (STA) (Fetz and Cheney 1979, 1980) using the computer-rectified traces from all recorded EMGs. Averages were computed using all action potentials in the entire trial (whole trial), action potentials occurring from the period after the GO signal until the end of the trial (posttrigger), and those occurring in the period prior to the GO signal (control). Averages were only retained if >1,000 action potentials were available for any one period. Responses were considered to be significant if they exceeded ±2 SD of the control activity (50 ms before the trigger action potential) for a minimum of 3 ms. The latency of the responses was determined at the point where the averaged trace crossed the 2 SD line used to determine significance.
To avoid confusion, we always refer to muscles in RESULTS with respect to the side of the body on which they were recorded, either left or right. In the DISCUSSION, we use ipsilateral with respect to the site of the recording site, in the left PMRF and not with respect to the limb performing the reaching task.
|
|
RESULTS |
|---|
|
Cells (127, including 56 RSNs) were recorded during a minimum of five reaches with both the left (ipsilateral to the recording site) and the right limb. These cells form a subset of the dataset (142 cells) used in a previous publication (Schepens and Drew 2004). The location of this subset (primarily within the nucleus reticularis gigantocellularis) (see Fig. 3, Schepens and Drew 2004) and the conduction velocity of the axons of these cells [97.7 ± 17.0 (SD) m.s–1, n = 54 (the latency of 2 neurons was not measured)] was almost identical to that of the full dataset (98.1 ± 16.7 m.s-1). Because our previous analysis (Schepens and Drew 2004) showed no major differences in the discharge characteristics of the RSNs and the unidentified cells, they are treated together in this manuscript. Sixty seven (67/127) neurons (including 36 RSNs) showed increased activity prior to the onset of the ipsilateral ClB (iClB) muscle during the left reach; these cells discharging in advance of the prime flexor muscles responsible for the reach will form the major focus of this manuscript.
General task characteristics
Reaching movements of the forelimb are characterized by stereotypical changes in ground reaction forces (GRFs) and EMG activity in all four limbs (Schepens and Drew 2003a). Figure 2 illustrates some of the more prominent and pertinent changes. The initial changes during reach of the left forelimb (ipsilateral to the recording site), are observed as an increase in activity in the left lateral head of triceps (lTriL) and in FV under the left limb (lFLV) together with a decrease in activity in the right TriL and in FV under the right limb (Fig. 2, left; see also Fig. 3). These anticipatory postural adjustments preceding the movement itself are thought to serve to displace the center of mass into the triangle created by the three supporting limbs (Ioffé et al. 1982; Schepens and Drew 2003a). These pAPAs are followed by a large increase in the level of activity of shoulder protractors (e.g., ClB) and elbow flexor muscles that begin shortly before the limb is lifted from the ground and continue throughout the reach, together with a large increase in FV under the right supporting forelimb and a decrease in FV in the reaching limb. There is also increased activity in the left vastus lateralis (VL) and in FV under the left hindlimb and decreases in activity in the right VL and in FV under the right hindlimb. These postural adjustments accompanying the movement (aAPAs) are also anticipatory in nature in that the latency of activation is coincident with the onset of the activity in the prime movers, such as the ClB (Alstermark and Wessberg 1985; Schepens and Drew 2003a). During movements of the right forelimb, the changes in GRF are the exact reciprocal of those observed during movement of the left limb (Fig. 2, right).
The activity patterns in the major extensor muscles of the forelimbs were also reciprocal during left and right reaches. This is illustrated in Fig. 2 but is clearer in Fig. 3A. During left reach (thicker line) the lTriL showed an initial, phasic, increase in activity that occurred just prior to the pAPA and that was time locked to the GO signal and a second, brief, burst of activity that was time locked to the onset of movement (see Schepens and Drew 2003a for a detailed examination of the temporal relationships of different muscles to the major events occurring during the reach). This second period of activity terminated at approximately the time that the limb reached the target and the paw was inserted into the tube. In contrast, during a right reach (thinner lines), this muscle showed an initial decrease in activity followed by a sustained increase as the weight of the animal was transferred to the left side. This pattern of activity was observed in the other extensors of the forelimbs that we recorded, namely the palmaris longus (PaL) and the supraspinatus (SSp) (not illustrated). The forelimb flexor muscles became strongly active subsequent to the pAPA and just before the limb was lifted from the platform. In the brachialis (Br) muscle, which is a pure flexor of the elbow, there was a sharp peak of activity that was followed by a lower level of sustained activity throughout the reach (Fig. 3B). During movements of the right limb, there was a very low level of activity that was sustained throughout the movement. A similar pattern of activity was observed in the shoulder retractor, the teres major (TrM) as well as in the wrist dorsiflexor, the extensor digitorum communis (EDC). The shoulder protractors, the ClB and the cleidotrapezius (ClT), also exhibited a similar pattern of activity with the exception that in these muscles there was one large burst that was maintained throughout the movement (see e.g., Fig. 2). The onset of activity in all of these flexor muscles was tightly linked to the onset of the movement (Schepens and Drew 2003a). The activity patterns in the shoulder muscles were more heterogeneous, but all of them showed reciprocal changes in activity during the pAPA. As for the triceps muscle, the acromiotrapezius (AcT; Fig. 3C) and the spinodeltoideus (SpD; Fig. 3D) muscles showed an increase in activity during the pAPA during the left reach and a reciprocal decrease in activity during the right reach. During the reaching movement, the AcT also showed a similar pattern to the TriL although the level of activity during the left reach was maintained at a relatively higher level. This was even more the case in the SpD in which there was as high, or higher, a level of activity during the right reach as during the left reach. In other words, while this muscle showed a reciprocal pattern of activity during the pAPA, it showed a similar pattern during the latter part of the reach. A similar pattern of activity was also seen in the acromiodeltoideus.
As for the forelimb extensor muscles, the major hindlimb extensor muscles, such as the gastrocnemius (GL; Fig. 3E), also showed clear reciprocal patterns of activity. During the left reach, the level of EMG activity increased and during the right reach it decreased. A similar pattern was seen in the VL (see Fig. 2) and in the gluteus medius (GlM). This pattern of activity reflects the diagonal pattern of support that is observed as the weight of the cat is transferred onto the contralateral forelimb and the hindlimb diagonal to the supporting forelimb. Activity patterns in the hindlimb flexor muscles were more variable although they were generally reciprocal to the activity patterns observed in the hindlimb extensors. As illustrated in Fig. 3F, the semitendinosus (St) on the left side was generally inactive during the left reach but increased during a right reach when the left hindlimb was unloaded. The sartorius (Srt) showed a similar pattern of activity, although during the left reach, the tonic activity present in this muscle during standing was frequently inhibited.
Last, in the axial muscles, the averaged pattern of activity was frequently similar throughout the reach for both left and right reaches. In the biventer cervicus (BvC), there was an initial short-latency decrease in activity and then a prolonged increase in activity. Both of these periods of activity were temporally related to the GO signal. Similar patterns of activity were observed in the splenius (Spl) and the complexus (Com). The longissimus dorsi (LoD) muscle generally showed greater activity during the right, contralateral, reach. It was also quite noticeable that the activity patterns of these axial muscles were very variable on a trial-by-trial basis (and even from one average to another) and in many trials lacked any evidence of phasic activity at all. Such was not the case for the limb and shoulder muscles which showed a much more consistent pattern of activity.
General cell characteristics
Despite the clear reciprocal nature of the vertical forces and the EMG activity of many of the major extensor and flexor muscles, most of the reticular neurons that we recorded showed broadly similar patterns of activity during reaches of both the left (ipsilateral to the recording site) and the right (contralateral) forelimb. Indeed, all except 4/67 neurons showed increases in activity during both the left and right reaches. Figure 2 shows one example of a tonically discharging neuron and Fig. 4 shows four other examples, one of which discharged in a tonic pattern (R22T32A), one in a phasic/tonic pattern (R22T30C) and two phasically. In each of these examples, the cells discharged with the same pattern of activity during the right reach and either with similar or elevated discharge frequencies.
Of the 20 cells that discharged phasically (see METHODS) (see also Schepens and Drew 2004) during the reach of the left, ipsilateral, forelimb, 15/20 also discharged phasically during the reach of the right, contralateral forelimb (see examples in Fig. 4), 3 further cells showed slightly elevated increases of activity in the static phase of the movement causing their classification to change to phasic/tonic and only 2 cells showed a reciprocal decrease during the contralateral forelimb reach (Table 1). Among those cells classified as discharging in a phasic/tonic manner during the left reach there was slightly more change. Eighteen (18/37) discharged in a similar manner during the left and right reach, whereas a further 10 showed a reduction in the ratio between the phasic and tonic components that resulted in a reclassification as tonic cells. Five other cells showed a reduction in the tonic component so that they were now classified as discharging in a purely phasic manner during the right reach. Among the cells classified as discharging in a tonic manner during the left reach, 4/10 showed an identical pattern of discharge during the right reach and 6/10 were reclassified as discharging in a phasic/tonic manner. Despite these changes, it should be emphasized that, overall, 38/47 (81%) of those cells discharging with a tonic component (phasic/tonic and tonic cells) during the left reach continued to discharge with a tonic component during the right reach. The changes in classification during the left and right reach support our previous statement (Schepens and Drew 2004) that the cell discharge patterns form a continuum, from those cells discharging in a purely phasic fashion to those discharging in a purely tonic pattern, but with the majority of cells showing both a phasic and a tonic component.
|
Temporal relationships of discharge activity during reach
As we previously emphasized (Schepens and Drew 2004), during a reach with the left limb, the onset of the discharge activity of neurons in the PMRF may be time locked either to the GO signal or to the onset of the prime flexor muscles, the Br and the ClB. We have proposed that cells the discharge of which is time locked to the GO signal contribute to the initiation of the pAPA. Cells with a discharge activity that is time -locked to the onset of the flexor muscles are suggested to contribute to the initiation of the movement and the accompanying postural adjustments (aAPAs). In many cells, we saw evidence of both relationships, and most cells also continued to discharge throughout the dynamic phase of the movement with many continuing to discharge throughout the entire movement, until the reaching limb was replaced on the support surface.
Some of these characteristics changed during reaching with the right (contralateral) limb, whereas others remained the same. In the following sections we therefore address separately the temporal relationships of the different features of the cell discharge, both to the onset of the movement, as well as to later aspects, such as the termination of the dynamic phase of the reach. Moreover, because our initial analyses showed that most of the important features of the discharge activity during right as compared with left reaches were observed in cells of all three types, phasic, phasic/tonic, and tonic, we will present results from these cells together, emphasizing differences where apparent.
Discharge activity related to the GO signal
Initial changes in discharge activity that were time locked to the GO signal could occur as either a decrease in activity (observed only in cells with a tonic discharge prior to the GO signal) or as an increase in activity. Eleven neurons exhibited a decrease in activity during the left reach that was time-locked to THE go signal. During the right reach, all 11 of these cells showed a reciprocal pattern of activity, in this initial period of activity, in that they now exhibited an increased period of activity, equally time locked to the GO signal.
An example of one such cell is illustrated in Fig. 6A. During the left reach, the neuron showed a clear decrease in the level of activity that occurred at a fixed time after the GO signal. This decrease in activity was clearly time locked to the GO signal as illustrated by the graph of Fig. 6B (left), which illustrates the linear relationship between lead time and the onset of activity in the lClB (
). During the reach of the right limb, contralateral to the recording site, the initial change in the discharge of the cell was a short-latency increase in the activity (Fig. 6A, right). As illustrated in the graph of Fig. 6B (right), there was a significant, linear relationship between the lead time of the cell discharge and the onset of activity in the rClB. Two other examples of neurons showing similar reciprocal changes in the early period of the discharge activity are illustrated in Fig. 6C.
The relationship between lead time and the onset of activity in the ClB for all trials from the nine cells in which the early discharge could be accurately measured during both left and right reaches is plotted in Fig. 6D. As for the example of Fig. 6, A and B, the population data show a clear relationship to the GO signal for both the left and right reach. The value of the intercepts for the population data were 57 ms for the left limb and 50 ms for the right limb.
During left reaches, we previously reported (Schepens and Drew 2004) that, after the initial decrease in activity, some of these cells (8/11) showed subsequent increases in activity that were also time locked to the GO signal. Such is the case, for example, for the increase in activity following the initial decrease in activity during the left reach for the cell illustrated in Fig. 6A. We referred to these as secondary increases. This distinction is not pertinent during the reach with the right limb because the initial change was a short-latency increase in activity. Any secondary increase, if present, is, therefore not easily detectable in the individual trials. This is the case for the cell in Fig. 6A, although the averaged activity in the PEH suggests that such a secondary increase might occur.
In contrast to the cell illustrated in Fig. 6, the example neuron illustrated in Fig. 7 showed a nonreciprocal pattern of activation during the earliest part of the discharge. During the left reach, Fig. 7, A and B, left, there was a short-latency increase in cell discharge activity that was time locked to the GO signal. During reaches made with the right, contralateral, limb, this cell discharged in an identical manner (Fig. 7, A and B, right). This relationship was similar in all 10 cells for which latencies could be measured during both left and right reaches (Fig. 7, C and D).
A similar pattern of nonreciprocal activity was also observed in many of the neurons discharging phasically. Figure 8, A and B, illustrates one example in which the discharge was time locked to the GO signal during the left and right reach in the same manner as for the more tonically discharging cells illustrated in Fig. 7. In the illustrated example, as in many of the other phasically discharging cells, the neuron showed a short-latency increase in discharge that preceded, and that was maximal during, the pAPA (gray bar). Increases in activity that were significantly related to the GO signal during both left and right reach could be measured in four other cells (see e.g., Fig. 8C), and the relationship of this small population to the GO signal was homogenous (Fig. 8D) and similar to that observed for the single cell illustrated in Fig. 8B.
Discharge activity related to the movement onset
During the left reach, the initial increase in activity in 16 neurons with a phasic/tonic or tonic discharge, and in 3 neurons with a phasic discharge, was significantly related to the onset of the lClB onset, and therefore the movement (Schepens and Drew 2003a). During the right reach, only 1/19 of these cells showed a similar, significant, relationship between the initial increase in activity and lClB. Indeed, in 11/19 cells, the initial increase in activity during the right reach was significantly related to the GO signal. Although this change in temporal linkage might suggest a change in function, it should be emphasized that the comparison is complicated by the reciprocal nature of the initial change in cell discharge of the type illustrated in Fig. 6. Although a movement-related increase can be readily determined in a cell that shows an initial decrease in cell activity (the situation during the left reach), it can only with great difficulty be detected after an initial GO-related increase (the situation during the right reach). In other words, the time of onset of the movement-related activity during the right reach could be masked by the preceding GO-related activity.
Discharge activity related to other events during the movement
Many of the phasic/tonic cells showed a pronounced phasic phase of activity during the right reach that, as illustrated in Fig. 9B, was frequently followed by a clear depression of the discharge. This decrease in the discharge was clearly better related to the movement as it was only evident when the cell discharge was synchronized to the onset of the rClB (compare Figs. 6A, right, with 9B). Synchronizing the discharge on the lClB also revealed a similar but smaller depression during the reach made with the left limb (Fig. 9A). Linear regression analysis confirmed the relationship between the end of the period of cell discharge and several different movement-related events. For example, during both the left (Fig. 9C) and right (D) reach, the end of the phasic period of activity was significantly related to the end of the period of activity in the major muscles active during the reach, e.g., the Br (not illustrated) and the TrM. This decrease in activity also correlated with the transient decrease in force seen in the supporting limb (FV unloading) at this same time. In addition, there was also a significant relationship with the second period of phasic activity in the AcT, the end of the phasic period of activity in the back muscle, LoD, as well as with the period of activity of several other muscles (not illustrated). This underlines the close relationship in the postural changes in different groups of muscles with quite different anatomical functions.
Similar relationships between the end of the period of phasic cell discharge and the termination of the phasic period of activity in different muscle groups were found in many cells. These relationships are shown in Fig. 9, E and F, for all of those cells with a tonic component (solid lines) in which the relationship between the end of the phasic discharge and the end of the period of activity in the Br (or the TrM; see legend) was significant (P < 0.05). The slope of most of these regressions approached 1.0 and most had an intercept that was close to 0; this was more evident for the relationships during right reach than during the left reach. Similar relationships were found for 5/6 of the cells discharging with a phasic pattern of activity during the left and right reach and for which the offset of cell discharge could be measured in individual trials (dotted lines). The homogeneity of these responses is illustrated by the scatterplots in Fig. 9, E and F, which show the data from individual trials in 23 cells.
Pretrigger activity
Cell discharge frequency in some cells increased prior to the onset of the GO signal (Schepens and Drew 2004). This pretrigger activity was related to the changes in the level of the EMG activity that occurred following the appearance of the cue. The traces of weight distribution during the reach show that, during both left and right reaches, there was a shift of the weight forward over the forelimbs (see Figs. 10, A and B) and that this shift in weight generally started following the cue tone (open circles) but prior to the GO signal. The mediolateral traces on the other hand showed a reciprocal change in activity with the weight being shifted over the right limbs during the left reach and over the left limb during the right reach. Again, there was sometimes a similar change in weight distribution after the cue onset in the pretrigger period as seen in Fig. 10, A and B.
In the example illustrated in Fig. 10, A and B, the cell showed a qualitatively similar and significant (see METHODS) increase in activity before the GO signal during both the left and right reach. Concomitantly, there was an increase in the activity of the lAcT during both reaches and a forward shift of the weight distribution in both cases. There was also symmetrical increases in activity in several other muscles, such as the rAcT, the lTrM and the rTrM (not illustrated). As indicated in the preceding paragraph, the changes in the mediolateral weight distribution were reciprocal. Linear regression analyses between cell activity and each of the 64 EMG and force traces that were measured showed the highest coefficient of determination for the lAcT muscle during both the left and right reaches (Fig. 10, C and D). These relationships were substantially higher than those made using the change in the AP weight distribution of the cat.
Altogether, 11/67 cells showed significant increases in activity prior to the GO signal during both the left and the right reaches. In 8/11 cells, the best correlations with the cell discharge frequency during the pretrigger period were found with one of the muscles acting around the shoulder, either the AcT or the TrM. In addition, there were also slightly less high correlations with the activity in the Srt in 5/11 cells. Those cells showing strong linear relationships were also those that showed the largest changes in activity during the pretrigger period as a percentage of the control activity. In some cases, increases in activity in the shoulder muscles in the pretrigger period were as clear as those illustrated in Fig. 10. In 9/11 of these cells, the initial change in discharge activity following the GO signal was a short-latency increase in discharge activity during both the left and right reach. As such most of these cells would be classified as discharging in a nonreciprocal fashion, similar to those in Fig. 7, although in only two of these cells was it possible to measure the latency of the cell discharge.
Another 10/67 cells showed increased activity in the pretrigger period during the left reach but no significant change during the right reach. In 6/10 of these cells, there was no indication of any change in the weight distribution in the AP direction during the right reach. As such, the lack of any change during the right reach may simply reflect the fact that the cat made adjustments to its posture in the pretrigger period only during the left reach. In the other 4/10 cells, there were clear changes in posture during the pretrigger period during the right reach. All four of these cells showed an initial reciprocal pattern of discharge activity after the GO signal, similar to the example in Fig. 6.
A further 4/67 cells showed decreased activity during the pretrigger period during both the left and right reach. No consistent relationships were observed between cells discharge activity and EMG or force activity in these cells, although there was a tendency for decreased muscle tone in the BvC during the pretrigger period.
Quantitative relationships
We determined detailed relationships between cell discharge frequency and the magnitude of the EMG activity for that subset of neurons with a tonic component for which we were able to measure temporal relationships (27/47) as well as for 12/15 neurons that discharged phasically during both the left and right reach.
Because of the close relationships between changes in activity in different muscles, high coefficients of determination (R2) were found for multiple muscles during the left reach, although the correlation with the change in the rTriL was consistently high for muscles with a static component in their discharge (see Fig. 13 in Schepens and Drew 2004). During the right reach, this relationship was lost, and the correlations were much reduced. This is illustrated in Fig. 11A (2nd from left) by plotting R2 for the population of cells showing a tonic component during the left and right reach. All of the lines joining the coefficients of determination for the left and right reach are negative illustrating the greater, positive, relationship during the left reach. The opposite pattern was seen for the lTriL, i.e., correlations were negative during the left reach and positive during the right reach (Fig. 11A, left). Similar relationships during left and right reach were observed for the AcT and the SSp (not illustrated), and a reciprocal relationship was observed in the VL (i.e., positive responses in the lVL during left reach and in the rVL during the right reach, not illustrated). Relationships in the forelimb and hindlimb flexors as well as in axial muscles were weaker and more variable. From the point of view of the change on overall body posture, there was a positive relationship between cell discharge and the shift of the weight to one or the other side during the left and right reaches, respectively; in contrast, there was only a weak relationship to the forward shift of the body during the reach (Fig. 11A).
Figure 11B that these relationships were maintained when considering only the level of activity during the time from the GO signal until the time that the paw entered the tube (pAPA + dynamic), although the relationships were frequently weaker. In addition, the nonreciprocal cells (blue lines) showed a poor relationship with the rTriL during the left reach. This is primarily explained by the fact that the rTril shows an initial decrease in activity, followed by an increase, whereas the cell shows only an increase (Fig. 7). The population of cells with a purely phasic pattern of discharge during the left and right movements (Fig. 11C) showed some major differences in comparison with the population with a tonic component. In particular, the relationship between cell discharge frequency and the level of activity in the rTriL, and that with the lateral shift in body weight, was the reverse of that observed in the phasic/tonic and tonic cells.
Not only did cells show high correlation coefficients with multiple muscles, but there was a relationship between the strength of the relationship with different muscles. This is illustrated in Fig. 11D for three pairs of muscles as well as for the measures of the center of vertical pressure. When a cell showed a strong positive relationship with the rTriL, it normally showed a strong negative relationship with the lTriL (Fig. 11D, left). Similarly a strong positive relationship with the rTriL was associated with a strong positive relationship in the diagonally located lVL, and there was also a positive relationship between the left and right ClB. There was no relationship between the strength of the correlation with the forward and lateral shift of weight, suggesting independent control of these two behavioral events. Good relationships were also observed between lBvC and rBvC but not between the lBvC and the rTriL (not illustrated).
Because of the widespread anatomical branching of many reticulospinal axons (see Introduction), we considered the possibility that better relationships between cell discharge and EMG activity might be obtained by examining multiple muscles. To test this, we performed a multiple regression in which we added muscles in a stepwise fashion based on the results from the linear regression analyses and on considerations of the goal of the behavior. During the left reach, the rTriL was the first muscle entered as it was most frequently among the muscles showing the best relationship to cell activity in the linear regression analysis (Fig. 11A). The lClB in the left limb (the one performing the reach) was added subsequently, followed by the rAcT because it was active during both the dynamic and static phases and also produced high correlation coefficients in the linear regression analysis for many cells (not illustrated). Subsequently, we added the complementary muscles in the left forelimb, the Srt and the VL, as representative muscles for the hindlimbs, and finally we added representative axial muscles. For the right reach, the muscles were added in the reverse order.
Figure 12A shows the results of this analysis for all of those cells showing a reciprocal pattern of activity (red lines) in the period immediately after the GO signal (see e.g., Fig. 6) as well as for those showing a nonreciprocal pattern of activity (blue lines) similar to that illustrated in Fig. 7. In general, during the left reach, the reciprocal cells showed an increase in the value of R2 when the lClB and the rAcT were added to the regression. Subsequently, however, the addition of further muscles produced very little further change in the value of R2. During the right reach (Fig. 12D), two differences were observed. First, the overall value of the regression was lower during the reach of the contralateral limb than during the left reach, that is, the regression explained less of the variance in this condition. Second, although there was an increase in the value of R2 as the other forelimb muscles were added, as for the left reach, there was an additional increase in several cells as the left (contralateral) Srt was added to the equation.
The overall change in the value of R2 is summarized for the reciprocal cells in the population averages of Fig. 12B. The black triangles indicate the cumulative value of R2 as each muscle in turn was added to the equation showing that, on average, 
60% of the variance in the behavior during the left reach could be explained by combining the activity in the first 4 muscles. During the right reach, the same population of cells explained <40% of the variance for the equivalent four muscles and only attained 40% after adding representative muscles from all four limbs as well as axial muscles. The gray circles in the plots of Fig. 12B indicate the correlations obtained when using only the data from the dynamic period of the reach, whereas the green squares indicate the correlations obtained during the static period of the reach. These data show that the cells with a reciprocal pattern of activity are much better correlated during the dynamic part of the reach than during the static portion when activity in the muscles is primarily confined to the extensors of the supporting limbs.
The pattern of correlations obtained from the nonreciprocal cells (blue lines) was somewhat different from that observed in the neurons showing a reciprocal pattern of activity. In particular, these cells showed a much more variable pattern of correlation than did the reciprocal neurons for both the left and right reaches, suggesting that they form a less homogenous population. As a result, the cumulative values of R2 observed in the population average of the activity during the left reach (Fig. 12C) are substantially lower from those observed for the reciprocal population. Moreover, Fig. 12, A and C, shows that the strength of the correlations is relatively similar for both the left and the right reaches. Interestingly, for the left reach, the strength of the correlations during the static period of the behavior is almost equal to that observed during the dynamic period and the overall behavior. This is in part because of an increased strength of correlation during the static period and in part to a decreased correlation during the dynamic period.
Other cells with a tonic discharge showed results similar to those for the reciprocal group with higher values of R2 for left reaches than for right reaches (not illustrated).
Causal relationships
Spike-triggered averaging (STA) was used to examine causality in all 67 cells forming the major database. For this analysis, we used all action potential from all trials from the whole trial. Only cells for which we recorded >1,000 spikes were included in the final analysis. Overall, this yielded results for a total of 50 cells during the left, ipsilateral, reach and for 49 cells during the right, contralateral, reach. In both cats, EMG electrodes were implanted into 24 muscles, although some electrodes were lost over time.
Postspike facilitation (PSF) or depression (PSD) was observed in a total of 16 muscles from 21/50 cells during the left reach and in 44 muscles from 17/49 cells during the right reach. In most muscles, 49/60 (82%), PSD was observed. An example of the most typical responses that were observed is illustrated in Fig. 13 for the cell illustrated in Figs. 6A and 9. In this cell, STA using a total of 11,537 action potentials recorded during the left reach showed no signs of PSF or PSD in any of the 24 muscles recorded during this experiment. In contrast, STA during the right reach, using 8,924 action potentials showed significant PSD in four proximal muscles, the lAcT, lSSp, lTrM, and the lTriL (Fig. 13B); there was additionally PSD in the lPaL (not illustrated). In other words, these responses were observed in the extensors of the supporting limb, ipsilateral to the recording site but contralateral to the reaching limb. These responses were equally seen from averages made using action potentials only occurring prior to the onset of the GO signal (control, Fig. 13C) and from those occurring only after the GO signal (Fig. 13D). Similar responses were, of course, observed in the control period preceding the movements of the left limb. The validity of the responses is shown by Fig. 13E, which shows the similarity in the PSDs obtained from compiling STAs from successive groups of eight trials during the right, contralateral, reach. The fact that STA was observed in the control period, prior to the GO signal, but was observed only during movements with the right, contralateral, limb suggests that there must be some modulation, or gating, of the synaptic efficacy during the left reach.
Almost identical responses were obtained from seven other cells; two examples are shown in Fig. 14A. In both of these cells, STA using action potentials from the complete trial showed a clear postspike depression of activity in the lAcT and the lSSp during the right reach but no significant effects during the left reach. There was also depression of the activity in the lTrM, lTriL, lPaL, and lGlM in the cell illustrated in Fig. 14A and of the lTrM of the cell illustrated in Fig. 14B (not illustrated). As for the cell illustrated in Fig. 13, there was clear evidence of postspike depression in the control period, before the GO signal, for both the left and the right reach despite the small number of action potentials available to construct the average. Similar modulation of the responses were seen in the other 5/7 cells showing this pattern of response. All of the cells displaying strong PSD in the left forelimb muscles during the right reach showed a phasic/tonic discharge during both the left and the right reach. Moreover 6/8 were also identified antidromically from the electrodes in the lumbar spinal cord showing that these were neurons capable of influencing both the fore- and hindlimbs.
The other common pattern of responses, observed in six cells (including 5 identified as RSNS by the electrodes in the lumbar spinal cord) during left reach, was a clear depression of activity in the hindlimb extensors (GL, VL, GlM) of the left limb and a small, but significant, facilitation of the activity in the lSrt. In 3/6 cells (all identified by stimulation of the lumbar spinal cord), there was facilitation of the lSrt and depression of the lGL and lGlM during both the left and right reach. One such example is illustrated in Fig. 14C. Note that the PSD in the lGlM is also visible in the control condition, at least for those trials in which the cat is instructed to make a right reach. In four cells, there was a depression of activity in the lTriL during left reach and in three cells, a facilitation of the activity in the rTriL during the right reach. Changes in activity in the BvC were seen in three cells and changes in LoD were seen in 3 different cells.
For the forelimb muscles, the average latency was 8.8 ± 2.4 ms (n = 43), and for the hindlimb muscles, it was 10.7 ± 3.2 ms (n = 17). We saw no evidence that cells producing postspike responses were preferentially localized in any one region of the PMRF.
Other cell types
In addition to those 67 cells that increased their discharge activity prior to the onset of the iClB activity, we also recorded 60 other reticular neurons during both the left and right reach. During the left reach, 14 cells showed decreased activity after the GO signal, 13 cells an increase in activity subsequent to the onset of the iClB, and 10 showed a mix of these two characteristics; a further 11 neurons showed no significant change in activity or were silent (Schepens and Drew 2004). During the right reach, 38/48 (79%) showed very similar patterns of activity to those observed during the left reach. For another 8/10 neurons, the major difference was the appearance of a significant increase in discharge after the GO signal (Fig. 15), similar to that illustrated in Fig. 6.
|
|
DISCUSSION |
|---|
|
) that the reticular formation contributes to the coordinated activity of posture and movement during voluntary movement. Consideration of the discharge activity of reticular neurons during reaches of both the left and right limbs provides strong support for a contribution of this structure to both the pAPAs, preceding movement, as well as to the complex patterns of muscle activity responsible for the movement and the coordinated patterns of postural support that accompany it. The present study extends our information concerning the contribution of the PMRF to the control of movement and posture by demonstrating that reticular neurons are active during reaches made with either forelimb, although discharge frequency was generally greater with reaches made with the left limb, ipsilateral to the recording site. In addition, there were some consistent differences between the patterns of discharge activity during the left and right reach, particularly with respect to the earliest discharge that was time-locked to the GO signal. Moreover, consideration of the results obtained from the regression analyses and from the STA suggests that the bilateral signal is asymmetric in nature and that the final expression of the descending signal is modified (gated) according to the limb moved. General characteristics
One of the most striking findings in this study was the overall similarity in the pattern and the frequency of the discharge of these cells despite the reciprocal nature of the overall behavioral strategy. As illustrated in Figs. 2 and 3, the reaching movements with the left and right forelimbs were characterized by reciprocal changes in activity in FV as well as in the major extensor muscles of each of the four limbs. Moreover, the major flexor muscles in the reaching limb were clearly phasically active only during the movements of the respective limb, left or right. Further, the movement of the center of vertical pressure is directed to the right side during movements of the left forelimb and to the left side during movements of the right limb (Schepens and Drew 2003a), and the center of mass is likely to show a parallel displacement (Ioffé et al. 1982). It is, therefore unlikely that the discharge activity of these cells is contributing to the control of any overall, global, variable of the movement or the postural responses that control them, at least in any simple manner.
As such, we suggest two primary, nonexclusive, explanations for the similarity in the discharge pattern of the cells. The first is that the activity of these cells is related to groups of muscles that show similar activity patterns during left and right reach. For example, Fig. 3 illustrates that some shoulder muscles (e.g., AcT) show broadly similar activity patterns during the reach of each limb as do the axial muscles. The second possibility is that the cell discharge contributes to the production of activity patterns in different muscle groups during the left and right reaches and that there is therefore a gating mechanism that regulates the synaptic efficacy of the discharge at the level of the spinal cord. Such a mechanism would be compatible with the results from studies during locomotion (see following text) as well as the differential effects obtained from the STA analysis.
Relationship to a single group of muscles
The simplest explanation of the general similarity of the pattern of discharge activity is that a given cell contributes to the activity of a given muscle, or group of temporally coincident synergistic muscles, during both the left and the right reach.
Figure 6, for example, suggests that such an explanation might be compatible with the overall pattern of discharge of those cells classified as exhibiting a reciprocal pattern of activity. These cells show a brief decrease in activity time locked to the GO signal during the left reach and an increase in activity, equally time locked to the GO signal, during the right reach. It is likely that this initial activity is related to the pAPA that precedes movement (Schepens and Drew 2004). Subsequently, these cells show a sustained increase during the left reach and a more phasic/tonic discharge during the right reach. These characteristics are similar to those exhibited by the rTriL and the rAcT during the left and right reach (Figs. 3 and 6). However, linear regression analyses rarely showed a strong relationship between cell discharge and either of these muscles during both the left and the right reach. As illustrated in Fig. 11, there was frequently a strong relationship between these reciprocal cells and the activity in the rTriL during the left reach but not during the right reach; a similar pattern was seen for the rAcT (not illustrated). Although the correlations with some other muscles, such as the lClB, were more equal during the left and right reach (not illustrated), it must be emphasized that the levels of activity in these muscles during the right reach were very low (Fig. 2). Taken together, the data suggest that there is little evidence that these reticular neurons contribute in any simple manner to the activation of any given muscle, or group of functionally synergistic muscles, during both the left and right reach.
Reciprocal relationship to homologous muscles
An alternative hypothesis is that the discharge activity of a given cell signals the EMG activity in one group of muscles during the left reach and the activity in a different group of muscles during the right reach. This is suggested, for example, by the data shown in Fig. 11A for the TriL. During the left reach, the discharge frequency of the cell is best related to the activity in the rTriL whereas during the right reach, it is best related to the activity in the lTriL. In other words, when considering the entire behavioral period, discharge frequency correlates with the period of activity in the coTriL during both the left and the right reach. In addition, the activity in the lTriL is reciprocal in that during the left reach, activity in this muscle decreases as discharge frequency increases, whereas during the right reach, it shows a parallel increase. The reciprocal nature between cell discharge and the behavior of the animal is also indicated by the relationship with the lateral shift in body weight during the reach (Fig. 11A). During the left reach, cell discharge increases as the body shifts to the right; during the right reach, cell discharge increases as body weight is shifted to the left (giving the negative relationship in Fig. 11A). This relationship was equally observed in the pretrigger period in which cell activity before the GO signal increased irrespective of the direction of the lateral shift of the body (Fig. 10).
The argument that cell discharge is related to activity in different groups of muscles depending on the limb moved is also supported by the temporal relationships illustrated in Fig. 9. During the left reach, there is clearly a relationship between the end of the phasic period of cell discharge and the end of the period of phasic activity in the prime flexor muscles such as the lBr and the lTrM during the dynamic period of the reach. During the right reach, however, there is clearly a strong temporal relationship between the end of the phasic period of discharge and the end of the period of activity in the rBr and the rTrM. The lack of a strong relationship between cell discharge frequency and the magnitude of the activity in different muscles reflects the complex nature of the discharge in these reticular cells and particularly the strong discharge that occurs during the pAPA. For example, Fig. 8 clearly shows a strong period of activity in this cell prior to any activation of the iBr. As such, linear regressions made from the onset of the GO signal to the end of the dynamic period (pAPA + dynamic, Fig. 11, B and C) are biased according to the relative strength of the discharge during the pAPA and during the reach itself. Although a confound to the analysis, this also serve to emphasize that there is no simple relationship between the discharge characteristics of reticular neurons and any one group of muscles.
Similar organization of reticulospinal influence during reaching and locomotion
The results from the multiple regression, illustrated in Fig. 12, suggest that the cells might encode the relationships between groups of muscles in different limbs. Indeed combining the activity in the coTriL with that in the iClB, for both the left and right reach, resulted in a substantial increase in the value of the coefficient of determination, especially for those cells showing a reciprocal pattern of activity. This type of organization is similar to that observed during locomotion, and a schematic representation of this organization, similar to that proposed for locomotion (Drew et al. 2004), is shown in Fig. 16. In this view, signals from the left and right motor cortex, responsible for the command to reach for the right and left limbs, respectively, impinge onto RSNs in the left PMRF. As stated in the Introduction, many of these neurons either branch to innervate both sides of the spinal cord (Matsuyama et al. 1988, 1997, 1999) or communicate with the contralateral side via commissural neurons (Jankowska et al. 2003; Matsuyama et al. 2004). They could therefore influence interneuronal networks on each side of the spinal cord. However, whereas during locomotion, the magnitude of the responses related to activity in the left and right limbs appears to be quite similar (Prentice and Drew 2001), the evidence from these experiments suggests a more asymmetric signal, especially for the reciprocal cells (see e.g., Fig. 12B).
We suggest that the interneuronal networks in the spinal cord that are activated by the descending signals from the reticular formation would include those that are normally activated during locomotion. Whereas, during locomotion, the gating of the signal is determined by both the rhythmical activation of the interneuronal networks and, possibly, a contingent signal from other descending pathways (Drew et al. 2004; Prentice and Drew 2001), we suggest that the gating during reaching is determined largely by the descending signal for movement, from the motor cortex and elsewhere. More specifically, we propose that the signal from the motor cortex, in addition to activating the interneuronal networks responsible for the reaching movement, would also gate the spinal interneuronal pathways onto which the reticulospinal axons impinge. Moreover, different pathways or different functional groups of cells might differentially modify these pathways at different times during the overall movement. In addition, it is possible that immediately after the initial pAPA, movement-related afferent input may also contribute to modify transmission in these interneuronal pathways.
The signal is gated
The strongest support for the idea of a gated signal comes from the results of the STA analysis (Figs. 13 and 14). In most cases, particularly for the PSD, the onset of the postspike responses identified by the STA was sharply demarcated and always began after the onset of the trigger spike. Moreover, the latency of these responses was compatible with the latencies of the responses evoked in fore- and hindlimb muscles by trains of stimuli applied to the reticular formation (Drew and Rossignol 1990b). This suggests that these postspike responses provide evidence of a causal relationship between the recorded neuron and the target muscle, as for cortico- and rubromotoneuronal cells (Cheney et al. 1991). None of the responses showing PSD included in our analysis began prior to, or just after, the trigger spike, and none exhibited a broad peak as would be expected if the postspike effect was due to synchronous discharge in adjacent populations of cells (Cheney et al. 1991; Smith and Fetz 1989). Nonetheless, we cannot rule out that some of the postspike facilitatory responses (e.g., Srt in Fig. 14C) might be the result of synchronous inputs. Although only a few cells provided positive evidence of a causal relationship, the most striking result was the depression in the activity of the extensor muscles of the left (ipsilateral to the recording site) limb during the right reach. During the left reach, there was no sign of either facilitation or depression. One of the interesting points in this result is that in all cells showing this pattern of activity the depression was present in the control period before the GO signal. This means that during the left reach, this inhibition was actively gated out. This provides clear evidence of a modulation of the descending signal during the reach, consistent with the hypothesis proposed with respect to Fig. 16. Moreover, it was equally striking that 5/8 of the cells in which this pattern of activity was observed discharged with a reciprocal pattern of activity, suggesting that this population of neurons might form a functionally homogenous subset of the whole population.
The depression of the activation of the ipsilateral extensor muscles is compatible with the results from locomotion studies in which microstimulation of the PMRF during locomotion in the intact cat frequently produces a depression of the activity of ipsilateral extensors when the stimulation is applied during left stance and therefore during right swing (Drew 1991); an analogous situation to the right reach. It is also similar to the results recently obtained by Davidson and Buford (2004), who have shown depression of activity in the ipsilateral extensors by single-pulse microstimulation during a reaching movement. At the same time, the result is somewhat paradoxical as activity in these cells depresses activity in the ipsilateral extensors at a period during which sustained activity is required for postural support (see Fig. 6). It is possible therefore that this signal should be seen as one providing inhibitory sculpting of the activity in the ipsilateral extensors. This would be compatible with one of the suggested functions for the cerebellum (Armstrong and Edgley 1984; Eccles et al. 1967), which, via the fastigial nucleus, provides strong monosynaptic input to RSNs in the PMRF (Eccles et al. 1975; Mori et al. 2000). It is possible that this functional subset corresponds to the population of RSNs described by Takakusaki et al. (1989, 2001, 2003) in the decerebrate cat as producing powerful inhibition of extensor muscle activity.
Asymmetrical signal
The fact that this response is only observed in the ipsilateral extensor muscles during reaches made with the limb contralateral to the recording site in the PMRF is a strong argument not only for gating of the descending signal but also provides more evidence of a strong asymmetry in the descending control to each limb. The data from the STA suggest that at least part of the inhibitory control for each limb comes only from the ipsilateral side. Microstimulation studies during locomotion (Drew 1991; Drew and Rossignol 1984) suggest that facilitation of the extensors on the ipsilateral side is likely to come at least in part from reticular neurons in the contralateral (right) PMRF. Examined from the other point of view, increased activity of the reticular neurons in the left PMRF is likely to contribute primarily to activity of the extensors of the right forelimb (as shown in Fig. 11). The fact that no sign of PSF was observed in the rTriL during the left reach suggests that this pathway may be less direct than that to the ipsilateral extensors, even though the latter is also likely to be at least disynaptic (Peterson 1979). These connections are illustrated schematically in Fig. 17. This figure also shows putative facilitatory connections from reticular neurons to ipsilateral extensors, as well as to ipsi- and contralateral flexors. These connections are suggested on the basis of the linear regression analyses (Fig. 11) and by the temporal analysis of Fig. 9. This is compatible with the results from our microstimulation studies during locomotion (Drew 1991).
|
In addition to the gating of the signal according to the limb to be moved, it is probable that there is also differential control of the initial discharge that is time locked to the GO signal and suggested to be related to the pAPA and the later movement-related signal. This suggestion is based in particular on consideration of those cells showing a reciprocal pattern of discharge time locked to the GO signal (e.g., Fig. 6) and of those cells showing decreased activity during the left reach and an increase time locked to the GO signal during the right reach (Fig. 15). In both of these cases, it seems clear that the activity related to the pAPA is modulated independently of the movement-related activity. This supports our previous suggestion (Fig. 16, Schepens and Drew 2004) that the neurons in the PMRF receive convergent input containing signals related either to the pAPA or to the movement and integrates these into a unified descending command signal to control posture and movement.
Signal related to control of the forelimbs and the hindlimbs
In the preceding text, we have addressed the functional significance of these discharge patterns primarily with respect to the changes in activity in the forelimbs, and this despite the fact that a number of these neurons were identified as projecting to the lumbar spinal cord (Schepens and Drew 2004). This is justified for several reasons. First, our previous analysis showed no functional distinctions between RSNs and unidentified neurons (Schepens and Drew 2004). Second, a majority of RSNs that project as far as the lumbar spinal cord also innervate the cervical spinal cord (Peterson et al. 1975). Third, the instructed task was for a voluntary movement of the forelimbs. Nevertheless, we do not intend to imply that these neurons will contribute only to the patterns of activity in the forelimbs. Indeed it is likely that many of these neurons will influence activity of muscles in both the fore- and hindlimbs and probable that some may preferentially activate hindlimb muscles. In these experiments, it is difficult to differentiate between these possibilities (although see next section) because the movement of the forelimb induces a coordinated pattern of postural support that is distributed between the four limbs (Schepens and Drew 2003a). This is implicit in finding of strong linear relationships between the cell activity and the extensors of the forelimbs (Fig. 11A) and hindlimbs (not illustrated). It is also explicitly demonstrated by the finding that cells that show a strong strong positive relationship with the rTriL equally show a strong positive relationship with the lVL, an extensor in the diagonally located hindlimb (Fig. 11D). A similar relationship occurs between the lTriL and the rVL (not illustrated).
A contribution of reticular neurons to postural modification in fore- and hindlimbs would be compatible with the finding that many RSNs increase their discharge as both the forelimbs and the hindlimbs step over an obstacle during locomotion (Prentice and Drew 2001). Moreover, it is compatible with the finding that 6/8 of the neurons that produced PSD in the ipsilateral forelimb extensors were positively identified as RSNs with an axon projecting to the lumbar spinal cord. A causal contribution to the control of postural activity in the hindlimbs was demonstrated in some cells by the expression of PSF and PSD in flexors and extensors, respectively, of both the ipsi- and contralateral hindlimbs during left and right reach. This might suggest a subpopulation of neurons that contribute preferentially to the coordinated change in hindlimb muscles during movements of the forelimbs. This would be compatible with our previous report that the postural changes in the fore- and hindlimbs might be independently controlled (Schepens and Drew 2003a). Nonetheless, as for the population of neurons producing PSD in the ipsilateral extensor muscles, it is possible that these neurons also influence other muscles, including those in the forelimb either through pathways with weaker synaptic efficacy or through polysynaptic connections.
Finally, although the results from the multiple regressions show that 60% of the total variance in cell discharge can be explained by combining flexor and extensor muscles from different limbs in a linear, combinatorial fashion (Fig. 12), it is possible that these reticular neurons may better encode higher-level features of the postural responses. For example, recent studies by Ting and MacPherson (2005) have suggested that the compensatory postural responses that result from horizontal displacement of a support surface might be the result of the activation of a limited number of synergies that activate multiple muscles through a weighting matrix. It is possible that some of the activity patterns that we observe here would be more compatible with the activation of these more abstract, or high-level, signals.
Several subpopulations of neurons
The data presented in this manuscript suggest the existence of at least three different subpopulations of neurons, each of which probably contributes to activity both during the aAPA and during the movement. The broadest division is between those neurons that have a tonic component and those that do not. We have previously suggested, based on the level of discharge activity in different parts of the behavior, that phasic and phasic/tonic neurons may fall on a continuum (Schepens and Drew 2004). However, the quantitative analyses in this study suggest that there may be real functional differences in these neurons that go beyond a simple comparison of the discharge pattern. This is particularly evident when inspecting the linear regressions compiled from the pAPA + dynamic period as illustrated in Fig.11, B and C. Whereas cells with a phasic/tonic pattern of activity increase their discharge with the lateral movement of body weight over the supporting limbs (Fig. 11B), those with a purely phasic period of activity increase their discharge in the inverse manner (Fig. 11C). As mentioned in preceding paragraphs, this difference is because of the relatively increased discharge activity of the phasic cells during the APA as the body weight is initially transferred over the reaching limb. As such, while contributing to the reach itself, as suggested by the later discharge activity, temporally related to the end of the period of activity in the flexor muscles (Fig. 9), these phasic cells may have a preferential role in contributing to the anticipatory postural adjustments preceding the movement. In contrast, although the phasic activity of the phasic/tonic cells undoubtedly contributes to the APA, the linear regressions suggest a preferential contribution to the flexor muscle activity during the reach and the postural activity that accompanies that movement.
Inspection of the activity patterns of those cells with a tonic component suggest at least two subpopulations, distinguished in particular by the initial changes in activity time locked to the GO signal. One population shows a reciprocal period of activity during the pAPA (Fig. 6) and the other a nonreciprocal pattern (Fig. 7). How this change in the pattern of activity relates to the changes in muscle activity is not clear. That the differences between these populations is real, however, is suggested by several other findings. For example, although both populations of cells show strong relationships with the rTriL during the left reach when the entire behavior is considered (Fig. 11A), the cells showing a nonreciprocal discharge show only a very weak contribution to this muscle when only the pAPA + the dynamic period is considered (Fig. 11B: blue lines). This might suggest the cells with a reciprocal discharge time locked to the GO signal might have a more preferential relationship to the dynamic part of the movement than the nonreciprocal cells. This is also supported by the multiple regression analysis (Fig. 12), which shows that, as a population, the reciprocal cells are much better related to muscle activity during the pAPA + dynamic period than during the static period, whereas the nonreciprocal cells show an almost equal relationship in the two phases.
Interestingly, the results from the STA showed that most of those cells showing PSF with the ipsilateral forelimb extensors exhibited reciprocal patterns of activity, whereas all three cells showing bilateral responses in the hindlimb muscles had a nonreciprocal pattern of activity. This might provide some further indication of the relative functions of these two populations. Cells with a reciprocal pattern of activity might preferentially contribute to forelimb activity, and those with a nonreciprocal pattern of activity might preferentially contribute to hindlimb activity. This would be compatible with the fact that the hindlimb extensor muscles do not show the same reciprocal modification of activity during the APA as do the forelimb muscles (see Fig. 3) and with our previous statements concerning independent control of the fore- and the hindlimbs (Schepens and Drew 2003a).
Concluding remarks
The results from these experiments provide further evidence that the PMRF contributes to the coordination of both movement and posture and that this activity is bilateral. This finding complements the results from locomotion studies that show discharge activity related to EMG activity in the left and right limbs both during unobstructed locomotion (Drew et al. 1986; Matsuyama and Drew 2000; Shimamura and Kogure 1983) and during voluntary modifications of gait (Prentice and Drew 2001). We have argued that the final expression of the descending signals from the reticular formation is gated by the rhythmical activity of the interneuronal networks on which these signals impinge (Drew 1991; Drew et al. 2004; Prentice and Drew 2001; see also Orlovsky 1972). The results from this study extend these arguments in several ways. First, the fact that the reaching movements are made from a stable and static standing posture allows us to make precise temporal correlations between cell activity and different behavioral events. This clearly shows that reticular cells contribute to multiple parts of the behavior including the postural responses preceding the movement. Second, the data clearly show that the descending signals during reach are asymmetric, discharging more strongly and showing better relationships with EMG activity during reaching movements made with the left (ipsilateral) limb than those with the right limb. Third, we have several lines of evidence suggesting that the signal is gated at the spinal level, including direct proof for a small subset of cells that produced PSD of the ipsilateral extensor muscles but only during reaches with the contralateral limb. This evidence provides strong support for arguments previously made only on the basis of more indirect evidence.
As for the mechanism involved in producing the gating, we suggest that the command for movement, originating in large part from the motor cortex, is responsible for modifying the transmission in interneuronal spinal pathways during the movement. A question remains, however, as to the signal responsible for the gating of the anticipatory postural responses that precede the movement and are time locked to the GO signal. We suggest two possibilities. One is that descending signals from the motor cortex are equally responsible for modifying the spinal networks before the onset of the reaching movement as well as during it. Motor cortical cells related to the GO signal in a similar task have been reported by Perfiliev (2005) although the author suggests that these cells are more involved in movement initiation and selection than in postural control. The other possibility is that some of the reticular neurons themselves, in particular those showing an initial reciprocal pattern of activity, may produce the gating. If such is the case, then these neurons may receive an input signaling the initiation of the pAPA either from the motor cortex (Birjukova et al. 1989; Kably and Drew 1998a,b; Matsuyama and Drew 1997) or other premotor cortical areas (Matsuyama and Drew 1997; Viallet et al. 1992).
|
|
GRANTS |
|---|
|
|
|
ACKNOWLEDGMENTS |
|---|
|
|
|
FOOTNOTES |
|---|
Address for reprint requests and other correspondence: T. Drew, Dept. of Physiology, Université de Montréal, PO Box 6128, Station "Centre-ville," Montréal, Qúebec H3C 3J7, Canada (E-mail: Trevor.Drew{at}umontreal.ca)
|
|
REFERENCES |
|---|
|
Armstrong DM and Edgley S. Discharges of nucleus interpositus neurones during locomotion in the cat. J Physiol 351: 411–432, 1984.
Birjukova EV, Dufossé M, Frolov AA, Ioffé ME, and Massion J. Role of the sensorimotor cortex in postural adjustments accompanying a conditioned paw lift in the standing cat. Exp Brain Res 78: 588–596, 1989.[Web of Science][Medline]
Chapman CE, Spidalieri G, and Lamarre Y. Activity of dentate neurons during arm movements triggered by visual, auditory, and somesthetic stimuli in the monkey. J Neurophysiol 55: 203–226, 1986.
Cheney PD, Fetz EF, and Mewes K. Neural mechanisms underlying corticospinal and rubrospinal control of limb movements. Prog Brain Res 87: 213–252, 1991.[Web of Science][Medline]
Davidson AG and Buford JA. Motor outputs from the primate reticular formation to shoulder muscles as revealed by stimulus-triggered averaging. J Neurophysiol 92: 83–95, 2004.
Drew T. Functional organization within the medullary reticular formation of the intact unanaesthetized cat. III. Microstimulation during locomotion. J Neurophysiol 66: 919–938, 1991.
Drew T, Dubuc R, and Rossignol S. Discharge patterns of reticulospinal and other reticular neurons in chronic, unrestrained cats walking on a treadmill. J Neurophysiol 55: 375–401, 1986.
Drew T, Prentice S, and Schepens B. Cortical and brain stem control of locomotion. In: Brain Mechanisms for the Integration of Posture and Movement, edited by Mori S, Stuart DG, and Wiesendanger M. New York: Elsevier, 2004, p. 251–261.
Drew T and Rossignol S. Phase dependent responses evoked in limb muscles by stimulation of the medullary reticular formation during locomotion in thalamic cats. J Neurophysiol 52: 653–675, 1984.
Drew T and Rossignol S. Functional organization within medullary reticular formation of intact, unanaesthetized cat. I. Movements evoked by microstimulation. J Neurophysiol 64: 767–781, 1990a.
Drew T and Rossignol S. Functional organization within medullary reticular formation of intact unanaesthetized cat. II. Electromyographic activity evoked by microstimulation. J Neurophysiol 64: 782–795, 1990b.
Eccles JC, Ito M, and Szentagothai J. The Cerebellum as a Neuronal Machine. New York: Springer-Verlag, 1967.
Eccles JC, Nicoll RA, Schwarz DWF, Taborikova H, and Willey TJ. Reticulospinal neurons with and without monosynaptic inputs from cerebellar nuclei. J Neurophysiol 38: 513–530, 1975.
Fetz EE and Cheney PD. Muscle fields and response properties of primate corticomotoneuronal cells. Prog Brain Res 50: 137–146, 1979.[Medline]
Fetz EE and Cheney PD. Postspike facilitation of forelimb muscle activity by primate corticomotoneuronal cells. J Neurophysiol 44: 751–772, 1980.
Ioffé ME, Frolov AA, Gahery Y, Frolov AG, Coulmance M, and Davydov VI. Biomechanical study of the mechanisms of postural adjustment accompanying learned and induced limb movements in cats and dogs. Acta Neurobiol Exp 42: 469–482, 1982.[Medline]
Jankowska E, Hammar I, Slawinska U, Maleszak K, and Edgley SA. Neuronal basis of crossed actions from the reticular formation on feline hindlimb motoneurons. J Neurosci 23: 1867–1878, 2003.
Kably B and Drew T. The corticoreticular pathway in the cat. I. Projection patterns and collaterization. J Neurophysiol 80: 389–405, 1998a.
Kably B and Drew T. The corticoreticular pathway in the cat. II. discharge characteristics of neurones in area 4 during voluntary gait modifications. J Neurophysiol 80: 406–424, 1998b.
Magoun HW. Bulbar inhibition and facilitation of motor activity. Science 100: 549–550, 1944.
Magoun HW and Rhines R. An inhibitory mechanism in the bulbar reticular formation. J Neurophysiol 9: 165–171, 1946.
Matsuyama K and Drew T. The organization of the projections from the pericruciate cortex to the pontomedullary brain stem of the cat: a study using the anterograde tracer, Phaseolus vulgaris leucoagglutinin. J Comp Neurol 389: 617–641, 1997.[CrossRef][Web of Science][Medline]
Matsuyama K and Drew T. Vestibulospinal and reticulospinal neuronal activity during locomotion in the intact cat. I. Walking on a level surface. J Neurophysiol 84: 2237–2256, 2000.
Matsuyama K, Mori F, Kuze B, and Mori S. Morphology of single pontine reticulospinal axons in the lumbar enlargement of the cat: a study using the anterograde tracer PHA-L. J Comp Neurol 410: 413–430, 1999.[CrossRef][Web of Science][Medline]
Matsuyama K, Nakajima K, Mori F, Aoki M, and Mori S. Lumbar commissural interneurons with reticulospinal inputs in the cat: morphology and discharge patterns during fictive locomotion. J Comp Neurol 474: 546–561, 2004.[CrossRef][Web of Science][Medline]
Matsuyama K, Ohta Y, and Mori S. Ascending and descending projections of the nucleus reticularis gigantocellularis in the cat demonstrated by the anterograde neural tracer, Phaseolus vulgaris leucoagglutinin (PHA-L). Brain Res 460: 124–141, 1988.[CrossRef][Web of Science][Medline]
Matsuyama K, Takakusaki K, Nakajima K, and Mori S. Multi-segmental innervation of single pontine reticulospinal axons in the cervico-thoracic region of the cat: anterograde PHA- L tracing study. J Comp Neurol 377: 234–250, 1997.[CrossRef][Web of Science][Medline]
Mori S, Matsui T, Mori F, Nakajima K, and Matsuyama K. Instigation and control of treadmill locomotion in high decerebrate cats by stimulation of the hook bundle of Russell in the cerebellum. Can J Physiol Pharmacol 78: 945–957, 2000.[CrossRef][Web of Science][Medline]
Orlovsky GN. The effect of different descending systems on flexor and extensor activity during locomotion. Brain Res 40: 359–371, 1972.[CrossRef][Web of Science][Medline]
Perfiliev S. Bilateral processing of motor commands in the motor cortex of the cat during target-reaching. J Neurophysiol 93: 2489–2506, 2005.
Perreault M-C, Rossignol S, and Drew T. Microstimulation of the medullary reticular formation during fictive locomotion. J Neurophysiol 71: 229–245, 1994.
Peterson BW. Reticulospinal projections to spinal motor nuclei. Annu Rev Physiol 41: 127–140, 1979.[CrossRef][Web of Science][Medline]
Peterson BW, Maunz RA, Pitts NG, and Mackel RG. Patterns of projection and branching of reticulospinal neurons. Exp Brain Res 23: 333–351, 1975.[Web of Science][Medline]
Prentice SD and Drew T. Contributions of the reticulospinal system to the postural adjustments occurring during voluntary gait modifications. J Neurophysiol 85: 679–698, 2001.
Schepens B and Drew T. Reticulospinal neuronal activity during a reaching task in intact cats. Soc Neurosci Abstr 460.7, 2000.
Schepens B and Drew T. Contribution of the reticulospinal system to the control of movement and posture during a reaching task in cats. Soc Neurosci Abstr 296.7, 2001.
Schepens B and Drew T. Strategies for the integration of posture and movement during reaching in the cat. J Neurophysiol 90: 3066–3086, 2003a.
Schepens B and Drew T. Independent and convergent pathways for the control of the integration of posture and movement during reaching in the cat. Soc Neurosci Abstr 70.5, 2003b.
Schepens B and Drew T. Independent and convergent signals from the pontomedullary reticular formation contribute to the control of posture and movement during reaching in the cat. J Neurophysiol 92: 2217–2238, 2004.
Shimamura M and Kogure I. Discharge patterns of reticulospinal neurons corresponding with quadrupedal leg movements in thalamic cats. Brain Res 260: 27–34, 1983.[CrossRef][Web of Science][Medline]
Smith WS and Fetz EF. Effects of synchrony between primate corticomotoneuronal cells on post-spike facilitation of muscles and motor units. Neurosci Lett 96: 76–81, 1989.[CrossRef][Web of Science][Medline]
Sprague JM and Chambers WW. Control of posture by reticular formation and cerebellum in the intact, anesthetized and unanesthetized and in the decerebrated cat. Am J Physiol 176: 52–64, 1954.[Web of Science][Medline]
Takakusaki K, Kohyama J, and Matsuyama K. Medullary reticulospinal tract mediating a generalized motor inhibition in cats. III. Functional organization of spinal interneurons in the lower lumbar segments. Neuroscience 121: 731–746, 2003.[CrossRef][Web of Science][Medline]
Takakusaki K, Kohyama J, Matsuyama K, and Mori S. Medullary reticulospinal tract mediating the generalized motor inhibition in cats: parallel inhibitory mechanisms acting on motoneurons and on interneuronal transmission in reflex pathways. Neuroscience 103: 511–527, 2001.[CrossRef][Web of Science][Medline]
Takakusaki K, Ohta Y, and Mori S. Single medullary reticulospinal neurons exert postsynaptic inhibitory effects via inhibitory interneurons upon alpha-motoneurons innervating cat hindlimb muscles. Exp Brain Res 74: 11–23, 1989.[Web of Science][Medline]
Ting LH and Macpherson J. A limited set of muscle synergies for force control during a postural task. J Neurophysiol 93: 609–613, 2005.
Viallet F, Massion J, Massarino R, and Khalil R. Coordination between posture and movement in a bimanual load lifting task: putative role of a medial frontal region including the supplementary motor area. Exp Brain Res 88: 674–684, 1992.[Web of Science][Medline]
Vicario DS, Martin JH, and Ghez C. Specialized subregions in the cat motor cortex: a single unit analysis in the behaving animal. Exp Brain Res 51: 351–367, 1983.
This article has been cited by other articles:
|
|
 |
|
  S. Yakovenko and T. Drew A Motor Cortical Contribution to the Anticipatory Postural Adjustments That Precede Reaching in the Cat J Neurophysiol, August 1, 2009; 102(2): 853 - 874. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 C. N. Riddle, S. A. Edgley, and S. N. Baker Direct and Indirect Connections with Upper Limb Motoneurons from the Primate Reticulospinal Tract J. Neurosci., April 15, 2009; 29(15): 4993 - 4999. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. A. Leonard, R. H. Brown, and P. J. Stapley Reaching to Multiple Targets When Standing: The Spatial Organization of Feedforward Postural Adjustments J Neurophysiol, April 1, 2009; 101(4): 2120 - 2133. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
P. J. Stapley and T. Drew The Pontomedullary Reticular Formation Contributes to the Compensatory Postural Responses Observed Following Removal of the Support Surface in the Standing Cat J Neurophysiol, March 1, 2009; 101(3): 1334 - 1350. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. N. Carlsen, R. Chua, J. T. Inglis, D. J. Sanderson, and I. M. Franks Differential Effects of Startle on Reaction Time for Finger and Arm Movements J Neurophysiol, January 1, 2009; 101(1): 306 - 314. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 K. Stecina, U. Slawinska, and E. Jankowska Ipsilateral actions from the feline red nucleus on hindlimb motoneurones J. Physiol., December 15, 2008; 586(24): 5865 - 5884. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 P. Crenna, I. Carpinella, L. Lopiano, A. Marzegan, M. Rabuffetti, M. Rizzone, M. Lanotte, and M. Ferrarin Influence of basal ganglia on upper limb locomotor synergies. Evidence from deep brain stimulation and L-DOPA treatment in Parkinson's disease Brain, December 1, 2008; 131(12): 3410 - 3420. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
B. Schepens, P. Stapley, and T. Drew Neurons in the Pontomedullary Reticular Formation Signal Posture and Movement Both as an Integrated Behavior and Independently J Neurophysiol, October 1, 2008; 100(4): 2235 - 2253. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
D. S. Soteropoulos and S. N. Baker Bilateral representation in the deep cerebellar nuclei J. Physiol., February 15, 2008; 586(4): 1117 - 1136. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. G. Davidson, M. H. Schieber, and J. A. Buford Bilateral Spike-Triggered Average Effects in Arm and Shoulder Muscles from the Monkey Pontomedullary Reticular Formation J. Neurosci., July 25, 2007; 27(30): 8053 - 8058. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Visit Other APS Journals Online |
TOP
ABSTRACT
INTRODUCTION
), tonic cells; (
), phasic/tonic cells. Cells illustrated in 
