Differential modulation of descending signals from the ...€¦ · Differential modulation of...

Differential modulation of descending signals from the reticulospinal systemduring reaching and locomotion

Kenneth S. Dyson,1,2* Jean-Philippe Miron,1* and Trevor Drew1,2

1Département de Neurosciences, Université de Montréal, Montreal, Quebec, Canada; and 2Groupe de recherche sur lesystème nerveux central (GRSNC), Université de Montréal, Montreal, Quebec, Canada

Submitted 10 March 2014; accepted in final form 18 August 2014

Dyson KS, Miron JP, Drew T. Differential modulation of de-scending signals from the reticulospinal system during reaching andlocomotion. J Neurophysiol 112: 2505–2528, 2014. First publishedAugust 20, 2014; doi:10.1152/jn.00188.2014.—We tested the hypoth-esis that the same spinal interneuronal pathways are activated by thereticulospinal system during locomotion and reaching. If such werethe case, we expected that microstimulation within the pontomedul-lary reticular formation (PMRF) would evoke qualitatively similarresponses in muscles active during both behaviors. To test this, westimulated in 47 sites within the PMRF during both tasks. Stimulationduring locomotion always produced a strongly phase-dependent, bi-lateral pattern of activity in which activity in muscles was generallyfacilitated or suppressed during one phase of activity (swing or stance)and was unaffected in the other. During reaching, stimulation gener-ally activated the same muscles as during locomotion, although themodulation of the magnitude of the evoked responses was less limbdependent than during locomotion. An exception was found for someforelimb flexor muscles that were strongly facilitated by stimulationduring the swing phase of locomotion but were not influenced bystimulation during the transport phase of the reach. We suggest thatduring locomotion the activity in interneuronal pathways mediatingsignals from the reticulospinal system is subject to strong modulationby the central pattern generator for locomotion. During reach, wesuggest that, for most muscles, the same spinal interneuronal path-ways are used to modify muscle activity but are not as strongly gatedaccording to limb use as during locomotion. Finally, we propose thatthe command for movement during discrete voluntary movementssuppresses the influence of the reticulospinal system on selectedforelimb flexor muscles, possibly to enhance fractionated control ofmovement.

cat; gating; locomotion; reaching; reticular formation

DESCENDING INPUTS from supraspinal structures to the spinalcord are subject to phase-dependent modulation during loco-motion. In the case of reticulospinal inputs, this phase-depen-dent modulation is generally expressed as a facilitation ofactivity in ipsilateral flexor muscles when a reticulospinalvolley is induced by microstimulation during the swing phaseof locomotion, together with a site-specific mixture of facili-tation or suppression in ipsilateral extensor muscles whenstimulation is applied during ipsilateral stance (Drew 1991;Drew and Rossignol 1984; Orlovsky 1972). In the contralaterallimbs, the responses are generally reciprocal, so that stimula-tion in ipsilateral swing produces facilitation in the contralat-eral extensor muscles and stimulation in ipsilateral stanceproduces facilitation of contralateral flexors. Similar effects are

seen in both fore- and hindlimb muscles (Drew 1991; Drewand Rossignol 1984). We, and others, have suggested that thisphase-dependent modulation is the result of changes in excit-ability in interneurons that either form part of or are influencedby the central pattern generator (CPG) for locomotion (Degt-yarenko et al. 1993; Drew 1991; Drew and Rossignol 1984;Floeter et al. 1993; Orlovsky 1972; Perreault et al. 1994;Shefchyk and Jordan 1985). Figure 1A (modified from Drew etal. 2004) illustrates our conceptual representation of this pro-cess. The signal from the pontomedullary reticular formation(PMRF) is illustrated as influencing interneuronal networks inboth the left and right limbs, with the CPG determining thephase-dependent, generally reciprocal pattern of activation indifferent muscle groups. Input from the cortex is illustratedwith dashed lines in Fig. 1A to indicate that this input isprobably facultative during unobstructed locomotion (see Arm-strong and Drew 1984a). Similar spinal mechanisms are sug-gested to ensure that changes in cell activity during voluntarygait modifications of one limb produce appropriate, phase-dependent, changes in muscle activity in the supporting limbs(Prentice and Drew 2001). In this latter case the cortical inputbecomes important as in our original illustration (Drew et al.2004).

Recently, we have suggested that the same mechanisms thatare responsible for coordinating activity in the limbs duringlocomotion may also be used to coordinate the postural re-sponses that occur during reaching movements with the leftand right forelimbs (see Schepens and Drew 2006). Duringreach, movement of one limb to a target is associated withpostural responses in the contralateral, supporting, limb. Thesepostural responses are organized such that one pattern ofmuscle activity is observed with movement of the left forelimband the reciprocal pattern is observed for movement of theright forelimb (Schepens and Drew 2003). However, record-ings of activity in reticulospinal neurons that we, and others,have suggested to be responsible for initiating and modulatinganticipatory postural activity (Drew et al. 2004; Luccarini et al.1990; Massion 1992; Sakamoto et al. 1991; Schepens andDrew 2004) show that many exhibit a generally similar, non-reciprocal, pattern of activity during voluntary movement ofeither forelimb (Schepens and Drew 2006). Clearly, the dis-charge activity in such cells cannot specify a reciprocal patternof postural activity, and we have suggested that the activity inthese cells signals primarily the timing and the magnitude ofthe postural activity. The expression of the postural responsesis then dependent on the excitability of the interneurons ontowhich the descending signal impinges.

* K. S. Dyson and J.-P. Miron contributed equally to this work.Address for reprint requests and other correspondence: T. Drew, Université

de Montréal, Pavillon Paul-G. Desmarais, C.P. 6128, Succursale Centre-ville,Montreal, QC, H3C 3J7 Canada (e-mail: [email protected]).

J Neurophysiol 112: 2505–2528, 2014.First published August 20, 2014; doi:10.1152/jn.00188.2014.

25050022-3077/14 Copyright © 2014 the American Physiological Societywww.jn.org

by 10.220.33.6 on Novem

ber 2, 2016http://jn.physiology.org/

Dow

nloaded from

mailto:[email protected]

http://jn.physiology.org/

We suggest that the interneurons mediating the coordinatedpostural responses during reaching include those that equallymediate the coordinated responses observed to PMRF stimu-lation during locomotion. During voluntary movements, wesuggest that the excitability of these interneurons is influencedby the descending signals responsible for the voluntary move-ment, including those transmitted by the corticospinal tract. Inour schematic representation of this circuit (Fig. 1B, modifiedfrom Schepens and Drew 2006), we have kept the samerepresentation of the spinal circuits as in Fig. 1A to emphasizethe hypothesis that the interneuronal circuits innervated bythe PMRF would maintain their intrinsic capacity to coordinateactivity between the two limbs. The representation indicatesthat some of these circuits would be activated, as a unit, by thecortical command for movement to produce the reciprocal andcoordinated postural responses observed in our previous be-havioral experiments (Schepens and Drew 2003). Corticalinput would be expected to modulate activity in both the PMRFand the spinal cord (see Schepens and Drew 2006).

If this hypothesis is correct, then one would expect thatmicrostimulation in the PMRF during voluntary reaches of theleft and right forelimbs would result in a limb-dependent

modulation of the responses evoked by stimulation of thePMRF in a manner similar to the phase-dependent responsesobserved during locomotion. Specifically, stimulation in agiven site in the PMRF during left reaching movements shouldevoke responses primarily in left forelimb flexors and rightforelimb extensors, while stimulation during right forelimbreaches would result in facilitation of flexors in the rightforelimb and either facilitation or suppression (site specific asin locomotion) of the extensor muscles in the left forelimb. Todetermine whether changes in excitability were the result ofchanges at the level of the spinal cord or within the brain stem,we also stimulated reticulospinal axons within the mediallongitudinal fasciculus (MLF) in one cat. While responsesproduced by stimulation of neuronal cell bodies may be subjectto task-related modulation of neuronal excitability at the site ofstimulation, those resulting from stimulation of axons arelargely independent of such changes (see, e.g., Baker et al.1995).

We therefore trained cats both to walk on a treadmill and toperform the same reaching task as used in our recent experi-ments (Schepens et al. 2008; Yakovenko et al. 2011; Yak-ovenko and Drew 2009). Stimulation in a given site was

A B gnihcaeRnoitomocoL

Fig. 1. Schema illustrating the mechanisms proposed to underlie the modulation of descending influences from the reticulospinal system during locomotion (A)and reaching (B): conceptual models taken from previous publications (A modified from Drew et al. 2004 with permission; B modified from Schepens and Drew2006 with permission) that illustrate the general premises on which the present study is based. A: during locomotion, many cells in the pontomedullary reticularformation (PMRF) discharge 2 periods of activity, especially during gait modifications. We have suggested (Drew 1991; Drew et al. 2004; Drew and Rossignol1984) that these 2 periods of activity have different effects on muscle activity depending on the excitability of spinal interneurons forming part of, or influencedby, the central pattern generator (CPG) for locomotion. The initial period of activity during ipsilateral swing (orange) is suggested to facilitate activity in theipsilateral flexors and the contralateral extensors (also in orange). The subsequent period of activity (blue), occurring during ipsilateral stance, will modify theactivity of the ipsilateral extensors and the contralateral flexors. B: during reaching, many reticulospinal cells discharge during both left and right reach, despitethe reciprocal requirements for postural support in the 2 conditions. We have suggested that cortical inputs will gate spinal circuits to modify reticulospinal inputin a manner similar to that observed during locomotion. Specifically, activity during left reach (orange) is suggested to modify ipsilateral flexor muscle andcontralateral extensor muscle activity. Activity during right reach will produce reciprocal responses. I, Ipsilateral; co, contralateral; ClB, cleidobrachialis; TriL,lateral head of triceps brachii; CST, corticospinal tract; RST, reticulospinal tract; E, extensor; F, flexor; IN, interneuron; MN, motoneuron.

2506 EFFECT OF PMRF STIMULATION DURING LOCOMOTION AND REACHING

J Neurophysiol • doi:10.1152/jn.00188.2014 • www.jn.org

by 10.220.33.6 on Novem


Dow

nloaded from


applied during locomotion as well as during reaching. Theresults support some aspects of our hypothesis while at thesame time emphasizing important differences in the contribu-tion of the PMRF to the regulation of locomotion and reaching.On the basis of these results, we modify our conceptualrepresentations of the interactions between the PMRF andspinal interneuronal pathways to incorporate our new findings.

METHODS

Task

Two male cats (weights 3.9 and 5.8 kg, respectively) were trainedin two tasks. First, they were trained to walk steadily on a treadmill forperiods of 5–10 min at a speed of 0.4 m/s, a speed at which cats walkcomfortably for relatively long periods. Second, the cats were trainedto stand quietly and, when instructed, to reach forward and to press alever with their left (ipsilateral) or right (contralateral) forelimb (Fig.2A). This task has been used in several of our previous studies(Schepens et al. 2008; Yakovenko et al. 2011; Yakovenko and Drew2009). In brief, a warning tone alerted the cat to the start of a trial, and1.5 s later a tone instructed the cat to reach with either the left(frequency 400 Hz) or the right (4 kHz) forelimb. This instruction tone

lasted a random period of 0.5–1.5 s, and at its cessation a shutteropened, giving the cat access to a lever that it was trained to press toobtain a food reward. Successful movements required the applicationof a force of 2 N applied through an angle of 5°.

Surgical Procedures

The cats were prepared for surgery in aseptic conditions and undergeneral anesthesia with a mixture of oxygen and isoflurane (2–3%). Acraniotomy was made over the occipital bone, and a rectangularstainless steel baseplate (internal diameter 8 � 6 mm) was positionedover the cerebellum to provide access to the left PMRF. A recordingchamber was formed by building four walls around the baseplate withdental acrylic. Pairs of braided stainless steel wires were inserted intoselected proximal and distal muscles of the four limbs of the cats andused to record electromyographic (EMG) activity. In one cat (RS27),we implanted 40 pairs of electrodes, 34 of which were both functionaland verified postmortem to be inserted into the targeted muscles. Inthe second cat (RS28), electrodes were inserted into 24 muscles, 18 ofwhich were included in the analysis. All surgical and experimentalmanipulations were approved by the local deontology committee atthe Université de Montréal and followed the guidelines of the Cana-dian Council for the Protection of Animals.

0.5s 1.5s 0.5-1.5s 5stone

left (ipsilateral) reach

right (contralateral) reach

cue signal Go-signalshutter open

instructeddelay

B C

A

2 3 4 5

GOSignal

pAPA

2 3 4 5

E1

E2

E3

E4

E5

GOSignal

-2000ms (Go)

-300ms (Go)

+130ms (Go)

+70ms (Lift)

+50ms (Lever)

lBr

lTriL

rTriL

rFV

lFV

lever

-500 -500500 5001000 1000

7 8 9 10 7 8 9 1016

Fig. 2. Reaching task and stimulation proto-col. A: schematic representation of the reach-ing task (see text for details) (adapted fromSchepens et al. 2008 with permission). B:stimuli were applied at 5 different times dur-ing the movement, as defined in METHODS.Figure shows approximate time of applica-tion of stimuli 2–5 superimposed on an ex-ample reach trial with the left forelimb. Graybar indicates the occurrence of the anticipa-tory postural response (pAPA) that precedesmovement. The onset of the transport phaseof the reach is defined by the sharp onset ofactivity in the left brachialis (lBr, top). Theonset of the lever press is indicated by theinitial decrease in the lever trace (bottom).Data are aligned to the onset of the Go signal.C: schematic showing the time period duringwhich control (open boxes, unstimulated) ac-tivity was collected for each of the 5 stimulusepochs, E1–E5 (shaded rectangles). Theequivalent epochs for reach of the contralat-eral limb were E6–E10 (indicated by graynumerical values in B and C, top). Note thatfor E1 the stimulated period was 2,000 msbefore the Go signal and is not illustrated toscale. Values on right indicate the time of theapplied stimulus with respect to the eventsused to trigger the stimulator (see text). FV,vertical ground reaction force; l, left; r, right;TriL, lateral head of triceps.

2507EFFECT OF PMRF STIMULATION DURING LOCOMOTION AND REACHING


by 10.220.33.6 on Novem


Dow

nloaded from


Protocol

In each experiment, an electrode held in a custom micromanipula-tor was driven manually through the cerebellum to the top of the brainstem while the cat was held lightly on the experimenter’s lap. Theelectrode was then advanced slowly, and trains of stimuli (cathodalcurrent, 11 pulses at 330 Hz, pulse duration 0.2 ms) were applied ata frequency of 1/s at every 0.5 mm while the movements and the EMGactivity evoked by the stimuli were examined. Stimulus intensity wasmaintained at 25 �A for all sites in both cats. As reported previously(Drew and Rossignol 1990a, 1990b), stimulation generally evokedface and neck movements in more dorsal locations and then producedipsilateral flexion and contralateral extension of the forelimb as theelectrode was lowered. Further advance of the electrode frequentlyevoked responses in both the forelimbs and the hindlimbs. Forstimulation during locomotion, the electrode was positioned either atthe depth at which fore- and hindlimb responses were evoked or, inthe absence of hindlimb responses, at the depth at which the strongestresponses in the forelimbs were evoked. All stimuli, in both cats, wereapplied on one side of the brain stem only. In both cats we stimulatedthe left brain stem, and reaches with the left limb, as well as allmuscles recorded on the left side, are referred to as ipsilateral.

The cat was then placed on the treadmill, and stimulation wasapplied during locomotion. Stimuli were triggered on the onset ofactivity in the ipsilateral (left) brachialis (iBr) and were applied atdifferent delays with respect to iBr onset in every third step cycle.Stimuli were applied in groups of 10–20 stimuli and in a fixed orderat delays of approximately 100, 0, 200, 400, 600, 800, 150, 300, 500,700, and 900 ms with respect to the onset of iBr activity. This orderensured that any long-term trends in the responses evoked by thestimulus would not bias the results. EMG activity was band-passfiltered between 100 Hz and 475 Hz, and a continuous record of thelocomotion data was digitized to computer at 1 kHz. In addition, aperiod of 50 ms prior to the stimulus and 150 ms subsequent to eachstimulus was digitized at 2 kHz and stored as a frame of data.

After this series of stimulation, the cat was transferred to thereaching apparatus and stimuli were applied in the same location andat the same strength (25 �A) at five different epochs of the reach foreach limb (Fig. 2, B and C) in blocks of 8 (cat RS27) or 10 (cat RS28)stimuli. Stimuli were applied at epochs 1) 2,000 ms prior to the Gosignal with the cat standing quietly before the onset of the instruction;2) 300 ms prior to the Go signal, during the instructed delay period;3) 130 ms after the Go signal, at the approximate time of theanticipatory postural adjustment preceding the movement (pAPA); 4)70 ms after the vertical force dropped below 3 N during the transportphase of the reach; and 5) 50 ms after the cat depressed the lever. Thecorresponding stimuli for the contralateral limb are referred to asepochs 6–10 (Fig. 2, B and C). For one complete series of stimulationduring reach we therefore have 10 epochs of stimulation. Eachexperimental series consisted of 90 trials in cat RS27 (8 stimuli ineach epoch � 5 epochs/limb � 2 limbs � 10 unstimulated cycles) and110 trials in cat RS28. The entire behavioral sequence of activity forthe reach for 2,500 ms prior to the Go signal and 3,000 ms subsequentto it was digitized to computer at 1 kHz. EMGs were band-passfiltered as for locomotion, and force traces were low-pass filtered at100 Hz. As for locomotion, we also digitized a frame of data at 2 kHzconsisting, in this task, of the period of 50 ms prior to a stimulus and200 ms subsequent to it.

In addition, for each trial, we also acquired additional frames ofunstimulated data at 2 kHz corresponding in time to the five stimu-lation epochs in each limb. For example, for a stimulus applied duringepoch 1, 2,000 ms prior to the Go signal (E1 in Fig. 2C), we alsodigitized a period of 50 ms before and 200 ms after the theoreticaltime of application of a stimulus in epochs 2–5, i.e., 300 ms before theGo signal (epoch 2), 150 ms after the Go signal (epoch 3), 70 ms afterlift (epoch 4), and 50 ms after lever press (epoch 5). These periods arerepresented by the open rectangles in Fig. 2C. By combining these

periods of unstimulated activity with the equivalent periods from the10 unstimulated trials, we were able to average the backgroundactivity from �25 trials for each epoch.

Data Analysis

Data obtained during locomotion were analyzed as detailed previ-ously (see, e.g., Bretzner and Drew 2005). In brief, because the timeof the (online) application of the stimuli during the experiment wasonly approximate, we first calculated, off-line, the true time of eachstimulus with respect to iBr onset. This value was converted to a phase(0–1.0) of the average duration of the unstimulated step cycle. Thestimuli were then divided into 10 equal groups, with group 1 repre-senting phases between 0.0 and 0.1 and group 10 those between 0.9and 1.0 (see Fig. 4A). The EMG responses evoked by all of the stimuliin each group were averaged together and superimposed on theactivity of the muscles at the same phase (average phase of all stimuliincluded in each group) but in the absence of any stimulus (see, e.g.,Fig. 5A). Similarly, during reach, the responses evoked by the 8 (catRS27) or 10 (cat RS28) stimuli applied in each epoch were averagedtogether and superimposed on the average of the EMG activity duringthe identical epoch from unstimulated trials (see, e.g., Fig. 7).

For both the reach and the locomotion data, we first manuallyidentified the onset and offset of the averaged evoked responsesevoked by the stimulation. We used the deviation of the stimulustraces from the interval of confidence (P � 0.01) of the standard error(SE) of the background traces as a guide to identify the response. Tobe included in the analysis, the latency of the onset of the response hadto be �50 ms and the duration �5 ms (Bretzner and Drew 2005). Inaddition, if there were more than two responses to the stimulus thatfulfilled these criteria, only the initial, primary, response was in-cluded. This analysis was applied to all EMGs (and forces for thereaching data) for all 10 groups during locomotion and for all 10behavioral epochs of the reach. For each cat, we then calculated theaveraged time of the onset and the offset of these responses, using thedata from all stimulated sites. These average onsets and offsets werethen used to define a window, which was used for the majority of ourquantitative analysis. For each trial, we calculated the integrated EMGactivity (sum of each 0.5-ms bin within the region identified as aresponse) for both stimulated and unstimulated data. The averageintegrated value for the unstimulated data was then subtracted fromeach individual stimulated trace to provide a net value for thestimulation for each trial. The net values for each stimulated trial in agiven phase or epoch were then averaged.

For the analysis of the force traces, illustrated in Fig. 10, we foundthat the variation in forces under the standing limbs prior to thestimulation was sufficiently large that it proved difficult to use thissame method to quantify changes in vertical force. For the force data,we therefore used the results obtained by simply averaging the datafor all stimuli in an epoch. We averaged the stimulated and theunstimulated traces and then displaced the stimulated traces so thatthe activity in the period before stimulus onset overlapped that in theunstimulated traces. The net change in activity was then calculatedfrom these averaged traces as above (see Fig. 10 for additionaldetails).

When averaging together the mean values from different tracks toobtain the overall mean activity in a given muscle (see, e.g., Fig. 11),we first removed any outliers as defined by Tukey’s method fordetermining the interquartile range (IQR). Outliers were defined asexceeding IQR � 1.5 of the upper and lower fences. This procedurewas performed on a file containing all of the data values for each trackand each group for the locomotion and each epoch for the reach.

Histology

At the end of the series of experiments, each cat was deeplyanesthetized and perfused per cardia with formaldehyde. The brain



by 10.220.33.6 on Novem


Dow

nloaded from


stem was blocked and sectioned in the sagittal plane at 40 �m beforebeing stained with cresyl violet. The location of the different electrodepenetrations was determined on the basis of marking lesions (30–40�A) placed in selected penetrations during the experiments. Thecalculated locations of the stimulus sites were transposed to sectionsof the brain stem based on standard sections from the atlas of Berman(1968).

Terminology

For stimuli during locomotion, we use the term “phase of stimu-lation” to refer to the general time of application of a stimulus (swingor stance) and the term “group” to refer to the specific moment ofstimulus application in 1 of the 10 groups into which the step cyclewas divided. During reach, we use the term “epoch” to refer to thespecific moment during the behavioral trial at which stimulation wasapplied. We use the term “transport phase” (equivalent to epochs 4and 9) to emphasize the major comparison that we wish to makebetween the swing phase of locomotion and the transport phase of thereach. Similarly, we use “phase dependent” in a general manner forboth locomotion and reach.

RESULTS

Database

Stimulation was applied in 19 sites in cat RS27 and 28 sitesin cat RS28. The locations of these 47 stimulus sites, all in theleft PMRF, are illustrated in Fig. 3. All 47 sites were includedin the PMRF, with the most rostral tracks being located in thenucleus reticularis pontis caudalis (NRpc) and the most caudalbeing located in the nucleus reticularis gigantocellularis(NRgc), rostral to the level of the inferior olive. All of thesestimulation sites were located between 0.3 and 1.8 mm of themidline, with the eight most medial stimulation sites in RS28lying in the MLF (Fig. 3).

Behavioral Activity

The behavior of the cats and the EMG activity recordedduring the locomotion and reaching tasks were similar to thosedetailed in previous publications from this laboratory (Drewand Rossignol 1987; Schepens and Drew 2003; Yakovenko etal. 2011). The pattern of EMG activity for selected ipsilateral(i) and contralateral (co) forelimb and hindlimb muscles duringunobstructed locomotion is illustrated in Fig. 4A (the anatom-ical location of these muscles is illustrated in Fig. 4D). In brief,most limb muscles exhibited a single period of activity duringlocomotion during either the swing (corresponding approxi-mately to stimulation groups 1–4 for muscles of the ipsilateralforelimb) or the stance (corresponding approximately to stim-ulation groups 5–10 for the ipsilateral forelimb) phase oflocomotion (Fig. 4A). Muscles active during ipsilateral fore-limb swing included muscles acting around the shoulder [spi-nodeltoideus (iSpD)], the elbow (iBr), and the wrist and digits[extensor digitorum communis (iEDC)]. Other recorded mus-cles that were active during the swing phase but that are notillustrated in Fig. 4A included the shoulder muscles teres major(iTrM) and latissimus dorsi (iLtD) and the elbow flexor andshoulder protractor cleidobrachialis (iClB). The two shouldermuscles discharged at the onset and the end of the swing period(see, e.g., Drew and Rossignol 1987; Krouchev et al. 2006).Recorded extensor muscles active during ipsilateral forelimbstance included the lateral and the long heads of triceps (iTriL

and iTri, respectively), acting primarily around the elbow, thewrist plantarflexor palmaris longus (iPaL), and the shouldermuscle acromiotrapezius (iAcT); the latter also had a burst ofactivity in swing. The period of activity in the flexor andextensor muscles in the contralateral (right) forelimb was 0.5out of phase with that of the ipsilateral forelimb (Fig. 4A).EMG activity in the flexor muscles of the ipsilateral hindlimb,such as the ankle flexor tibialis anterior (iTA), was slightlyphase-advanced with respect to the ipsilateral forelimb flexoriBr. Because of the cyclical nature of locomotion, the iTAactivity is represented as occurring in groups 8–10 in theillustration of Fig. 4A (equivalent to phases of �0.2 to 0.0).Muscles in the following text are referred to as flexors orextensors based on their predominant period of activity duringlocomotion (see also Drew and Rossignol 1987; Rho et al.1999).

During ipsilateral reach (Fig. 4B), most muscles, and partic-ularly the extensors, were tonically active during quiet standing(see e.g., iTriL and iAcT). After the Go signal, the initialchange was in the activity of the extensor muscles of theipsilateral and contralateral forelimb, corresponding to thepAPAs that precede the reach (Schepens and Drew 2003). InFig. 4B, this is most evident for the iTriL. As Fig. 4B issynchronized to the onset of the activity in the iBr, the pAPAsprecede the synchronization event. After the pAPA, there wasa brief, phasic increase in activity in all of the ipsilateralforelimb flexor muscles that resulted in the transport of thelimb forward, toward the lever. Note that the iAcT, which wasmostly active in stance during locomotion, shows a strongperiod of activity during the transport phase. There was also amore prolonged increase in activity in the contralateral fore-limb extensor muscles that anticipated, and compensated for,the loss of support under the reaching limb. Ipsilateral forelimbextensor muscles were reactivated after the reach in order todepress the lever. In Fig. 4, B and C, this is most evident for theiPaL during the ipsilateral reach and the coTriL during thecontralateral reach. The overall magnitude of the EMG activitywas slightly greater than during unobstructed locomotion,especially for the flexor muscles. Changes in the hindlimbmuscles were more tonic in nature.

The changes in the ground reaction forces under each paw(not illustrated here, but see Fig. 10) were similar to thosepreviously detailed (Schepens and Drew 2003), with an in-crease in vertical force in the supporting forelimb during thereach. Smaller changes were observed in the hindlimbs, withthe hindlimb diagonal to the supporting limb showing in-creased force and the other hindlimb decreased force.

Changes in EMG activity during contralateral forelimb reach(Fig. 4C) were reciprocal to those observed during ipsilateralforelimb reach. Note that forelimb flexor muscles were onlyphasically active during the transport phase of the reachinglimb (see iBr and coBr) in much the same way that they wereonly active during the swing phase of the respective limbduring locomotion (Fig. 4A).

Responses Evoked by PMRF Stimulation During Locomotion

The effects of stimulation of the PMRF on EMG activityduring locomotion were similar to those previously describedfor a more limited number of muscles (Drew 1991) and areonly briefly described here.



by 10.220.33.6 on Novem


Dow

nloaded from


An example of the responses evoked during swing andstance in one track from cat RS27 is shown in Fig. 5A.Stimulation during ipsilateral swing (group 1) produced large,brief, short-latency twitch responses in the ipsilateral forelimbflexor muscles (iBr, iSpD) as well as an increase in the iAcT.There was also an activation of the wrist and digit extensor (butphysiological flexor during locomotion), iEDC. There was asmall facilitation of the iTriL and a suppression of the activityin the coTriL. In the hindlimb, there was a facilitation of theiTA and a suppression of activity in the soleus (iSol). Note thatprominent responses were observed not only in proximal mus-cles but also in those with a more distal action (e.g., iEDC and

iTA). During ipsilateral stance (and contralateral swing)(group 6), responses were absent in most ipsilateral forelimbmuscles, with the exception of a suppression of activity in theiAcT, but a prominent facilitation was observed in the coBr.There was a suppression of activity in the iSol.

The phase-dependent nature of the responses is clearly seenin the graphs of Fig. 5B, which illustrate data for the musclesillustrated in Fig. 5A as well as other selected muscles. Ingeneral, EMG activity in flexor muscles in each of the fourlimbs (Fig. 5B, i–iv) was facilitated during the period ofactivity of the muscles and was unresponsive out of phase withthis activity. The responses in extensor muscles (Fig. 5B, v–x)

TRP

7G

1.26

12

DMV

DAO

MAOrMAOc

TB

CAE

0.8

Caudal

DAO

MAOrMAOc

TB

7G6

12

DMV

INTPH

Rostral

RM

1.6

7G

7G

DAO

MAOc

MAOrPO

Caudal Rostral

CAE

DAO

MAOrMAOc

TB

7G6

12

DMV

TRP

TB

7G 6

12

DMV

MAOrMAOc

TB

12

MAOrMAOc

0.3

0.8

TB

BA

INTPH

1.2

RM

DAO

RS28RS27

-10 0

NrgcNrpc

NrgcNrpc

Nrpc NRpc

NRgc NRgc

MLFMLF

MLF

TB TB

7 7

PT PT

BP BP

P4

P7

P4

P7

-10 0

7 0

DC

7 0

Fig. 3. Location of the stimulated sites. Circles indicate the location of the stimulated sites in the PMRF in cats RS27 (A and C) and RS28 (B and D) as determinedfrom histological reconstruction. A and B: stimulation sites are presented on sections based on the standard section from the atlas of Berman (1968) that is closestto the calculated laterality of each stimulation site. Filled circles on the most medial section (0.3) in RS28 indicate sites within the medial longitudinal fasciculus(MLF). C and D: locations of the stimulus sites are transposed onto transverse sections based on the atlas of Berman (1968). Stimulation sites in the nucleusreticularis pontis caudalis (NRpc) are illustrated on the sections at P4; those in the nucleus reticularis gigantocellularis (NRgc) are illustrated on the sections atP7. Sites in the MLF are illustrated according to their proximity to P4 or P7. Arrows in D point to the MLF. Vertical dashed lines in A and B indicate the levelsof P4 and P7. Scales in A–D are with respect to stereotaxic zero. 7G, genu of the facial nerve; BP, brachium pontis; IO, inferior olive; n, number of cells; PH,nucleus praepositus hypoglossi; PT, pyramidal tract; TB, trapezoid body.



by 10.220.33.6 on Novem


Dow

nloaded from


were more complex, consisting generally of either no responseor a small facilitation during the swing phase. There was a mixof facilitation and suppression during their period of activity instance. In the forelimb, both the ipsilateral and contralaterallong head of triceps (iTri and coTri, Fig. 5B, v and vi,respectively) were facilitated during the period of activity ofthe muscle in stance while activity in the lateral head (iTriLand coTriL, also Fig. 5B, v and vi) was suppressed at the endof the period of activity but showed a small facilitation duringswing. The iAcT and the ipsilateral spinotrapezius (iSpT) (Fig.5Bix) were facilitated during swing and exhibited strong sup-pression during stance. Activity in the ipsilateral hindlimbextensors was depressed during stance for most recorded mus-cles but facilitated for the gluteus medius (iGlM, Fig. 5Bvii).Responses were facilitatory in the contralateral gastrocnemius,lateral head (coGL) (Fig. 5Bviii), and there was mixed facili-tation and suppression in the contralateral vastus lateralis(coVL) (not illustrated).

The results of stimulation at all sites examined within thePMRF of cats RS27 and RS28 are shown in Fig. 6, A and B,respectively, for selected muscles. The two major observations

here are that 1) the pattern of evoked responses in any givenmuscle is very similar for all sites stimulated, albeit with somedifferences in the magnitude of the responses, and 2) responsesare observed in both proximal and distal muscles, including theEDC. For example, in the iBr and the iSpD, evoked responsesfrom all 19 stimulated sites in cat RS27 were maximal duringthe period of swing, when the muscles were active. Similarly,in the iEDC, the maximal responses were slightly later in thestep cycle (group 3) corresponding to the delayed period oflocomotor activity in the EDC compared with the iBr (Fig. 4A).Responses in the coBr and in the iTA were also maximalduring the period of activity of the muscle as illustrated for theindividual example in Fig. 5. Responses were more variablein some of the extensor muscles. For example, althoughstimulation most commonly elicited suppression of the ac-tivity in the iTriL, some sites produced facilitation or amixture of facilitation and suppression. Similarly, in thecoTriL, there were two periods in which facilitatory re-sponses were produced, at the beginning and end of theperiod of activity, and these were separated by a smallperiod in which there was either no response or a small

Ipsilateral Reach

lever

-500 0 500 1000

Locomotion

iBr

iSpD

iEDC

iTriL

iAcT

iPaL

coBr

coTriL

iSol

iTA

coTA

Averages step cycle =1040 ms1.0

N=118 N=36

Contralateral Reach

lever

-500 0 500 1000

N=21

pAPA pAPAA B C

5 10GroupsPhase .1 .5

DiBr onset coBr onset

R27Tr10

1

Forelimb Hindlimb

smsm

bmileroF

bmildni

H

Fig. 4. Averaged EMG activity during loco-motion and reach. A: EMG activity recordedfrom selected ipsilateral (i) and contralateral(co) forelimb and hindlimb muscles duringunobstructed locomotion. Activity is syn-chronized to the onset of activity in the iBrand is displayed for 1 full step cycle (until theonset of the next burst of activity in iBr).Stimulation was applied at 10 different partsof the step cycle as indicated at bottom. Ver-tical dashed lines indicate the time of appli-cation of stimuli in groups 1 and 6 (note thatgroup 1 falls between phase 0.0 and 0.1, etc.).B and C: EMG activity during ipsilateral (B)and contralateral (C) reach. Data are alignedto the onset of activity in the respective Br(vertical dashed line). Gray bars precedingthe onset of the Br illustrate the EMG activityduring the anticipatory postural adjustmentsthat precede the reach (pAPA). The scale foreach individual EMG is the same in A–C. N,number of trials in each group of averages. D:illustrations of the anatomical arrangement ofthe muscles recorded in this study (adaptedfrom Krouchev et al. 2006 with permission).Colors in this figure are used only to helpdistinguish one muscle from another. Fore-limb: AcD, acromiodeltoideus; AcT, acro-miotrapezius; ECR, extensor carpi radialis;EDC, extensor digitorum communis; LtD,latissimus dorsi; PaL, palmaris longus; SpD,spinodeltoideus; SpT spinotrapezius; Tri, tri-ceps brachii, long head; TriL, triceps brachii,lateral head; TrM, teres major. Hindlimb: BF,biceps femoris; GlM, gluteus medius; GL,gastrocnemius, lateral head; GM, gastrocne-mius, medial head; Sol, soleus; Srt, Sartorius;St, semitendinosus; TA, tibialis anterior; VL,vastus lateralis (see Crouch 1969 for furtherinformation on functions).



by 10.220.33.6 on Novem


Dow

nloaded from


0 1 2 3 4 5 6 7 8 9 10 0 1 2 3 4 5 6 7 8 9 10 0 1 2 3 4 5 6 7 8 9 10 0 1 2 3 4 5 6 7 8 9 10

0 1 2 3 4 5 6 7 8 9 10

0 1 2 3 4 5 6 7 8 9 10

0 1 2 3 4 5 6 7 8 9 10

0 1 2 3 4 5 6 7 8 9 10

0 1 2 3 4 5 6 7 8 9 10

Group 1: Ipsilateral Swing Group 6: Ipsilateral Stance

iBr

iTriL

coBr

coTriL

iTA

iSoL

iSpD

iEDC

0

100

50

-50

0

100

50

-50

0

100

50

-50

0

100

50

-50

0

100

50

-50

0

100

50

-50

-100

0

100

50

-50

-100

0

100

50

-50

-100

0

100

50

-50

-100

iFL coFL iHL coHL

Stim Stim-50 150ms

iBriSpDiEDC

coBrcoClB

iTA coTA

iTriLiTri coTriL

coTri

0 1 2 3 4 5 6 7 8 9 10

0

100

50

-50

-100

iGMiSoliGM

coGL

iGlMiVL

iAcTiSpT

A

B

iAcT

iClB

RS27Tr10

50 100 -50 15050 100ms

egatnecrePegatnecreP

egatnecr eP

i ii iii iv

viiiviiviv

ix x

puorGpuorGpuorG Group

puorGpuorGpuorG Group

puorGpuorGFig. 5. Responses evoked by PMRF stimulation during locomotion. A: averaged responses evoked in selected muscles by stimulation of the PMRF duringipsilateral swing (group 1, left) and stance (group 6, right). As indicated in Fig. 4, group 1 corresponds to phases of 0.0 – 0.1 and group 6 correspondsto phases of 0.5– 0.6. Thick lines indicate responses during stimulation, and thin lines indicate activity in unstimulated cycles. EMG traces for each muscleare scaled to the same arbitrary units for swing and stance. Shaded rectangles indicate the approximate region from which we measured responses. B, i–x:for each muscle the responses evoked in each group are plotted as % of the maximum response (100%) evoked in that muscle. Traces above each plotshow the averaged and normalized activity of the indicated muscle during locomotion (scaled to minimum and maximum levels) as calculated from the118 steps illustrated in Fig. 4A. FL, forelimb; HL, hindlimb.



by 10.220.33.6 on Novem


Dow

nloaded from


suppression. On the other hand, all sites produced profoundsuppression of the activity in the iSol throughout its periodof activity during hindlimb stance. Qualitatively similarresponses were observed in the muscles recorded in catRS28 (Fig. 6B), although the suppression in the iTriL wasmore pronounced during stance than in cat RS27. Note thatthe responses evoked by stimulation of the reticulospinalaxons within the MLF in cat RS28 (Fig. 6B, red traces) areintermingled with those evoked from more lateral sites.

Responses Evoked by PMRF Stimulation During QuietStanding and Reach

The responses evoked in selected muscles by PMRFstimulation at different times of the reaching movement areshown in Fig. 7 for the same stimulation site as in Fig. 5.

Figure 7, A–E, illustrate, respectively, the responses evokedin different muscles by stimulation at each of the five epochsof the ipsilateral reach, as illustrated in Fig. 2B. In general,the evoked responses showed several similarities during allfive epochs, even though these included static support of thebody (Fig. 7, A and B), the pAPA (Fig. 7C), the transportphase of the reach (primarily flexion, Fig. 7D), and leverpress (dynamic extension, Fig. 7E).

For stimulation during quiet standing, 2,000 ms before theGo signal (Pre-Cue, Fig. 7A), the responses included a weakfacilitation of the activity in the ipsilateral flexor muscles iBrand iEDC (but not the iSpD) together with a suppression of theiTriL (not visible at the scale illustrated). On the contralateralside, there was no response in the coBr, a suppression of thecoTriL, and facilitation of the coTri. In the ipsilateral hindlimb,

LocomotioniBr

iBr

iSpDiEDC

iTriL

iTriL

coBr

coBr

coTriL

coTriL

iTA iSol

0 1 2 3 4 5 6 7 8 9 10 0 1 2 3 4 5 6 7 8 9 10 0 1 2 3 4 5 6 7 8 9 10 0

0

0

0

0

0

0

1 2 3 4 5 6 7 8 9 10

0 1 2 3 4 5 6 7 8 9 100 1 2 3 4 5 6 7 8 9 10 0 1 2 3 4 5 6 7 8 9 10

0 1 2 3 4 5 6 7 8 9 10 0 1 2 3 4 5 6 7 8 9 10 0 1 2 3 4 5 6 7 8 9 10

0 1 2 3 4 5 6 7 8 9 10

7000

5000

5000

5000

5000

-5000

-10000

-15000

2000

2000

2000

20003000

2000

-1000-1000-2000

-2000

-1000 -1000

1200

3400

5600

7800

10000

-1000

1200

3400

5600

7800

10000

1000

-2000

-4000

8000 50004000

4000

4000

40006000

60005000

5000

-5000

-10000

3000

3000

3000

10001000

1000

-1000 -1000

A

B

4000

2000

-2000

-4000

-60000 1 2 3 4 5 6 7 8 9 10

Group Group Group Group

0000000000000000000000000000

00000000

stinU yrartibr

Astin

U yrartibrA

stinU y rar tibr

A

* * *

*

Fig. 6. Averaged responses evoked in selected muscles in cats RS27 (A) and RS28 (B) during locomotion. A: activity evoked in 8 selected muscles from PMRFstimulation of each of the 19 stimulus sites examined in cat RS27 as a function of the phase of stimulation. Responses are plotted as arbitrary units and are scaledto the largest response evoked in a given muscle (at any stimulated site) by stimulation either during locomotion or during reach. Traces above each plot showthe averaged activity of the indicated muscle during locomotion as calculated from 5 experiments (682 step cycles). Thick red traces indicate the responses evokedat the stimulus site illustrated in Fig. 5. The outlying trace (asterisk) in the plots for the flexor muscles in RS27 was recorded from the most caudal location shownin Fig 3A. B: data from 4 selected muscles in RS28 plotted in the same manner for all 28 stimulated sites. Thin red traces indicate the responses from thosepenetrations within the MLF. Averaged EMG traces above the plots were calculated from 945 step cycles in 7 experiments. A value of 0 for any trace indicatesthat there was no evoked response.



by 10.220.33.6 on Novem


Dow

nloaded from


there was an increase in the iTA and suppression of the iSol.A similar pattern was observed for stimulation following thecue (Fig. 7B) and during the pAPA (Fig. 7C) epochs, except forthe appearance of a small response in the iSpD (Fig. 7C). Thetransport phase of the reach (Fig. 7D), which we consider the

analog of the swing phase of locomotion, was marked by acomplete abolition of the short-latency response in the iBr.This is in contrast to the robust evoked responses observedduring the swing phase of locomotion (Figs. 5 and 6). Therewas also a marked increase in the size of the facilitation in the

Stand: Pre-Cue (1) Stand:Cue (2)

Ipsilateral pAPA(3) Ipsilateral Reach (4)

Ipsilateral Lever (5) Contralateral Reach (9)

iBriSpD

iEDCiTriL

coBr

coTriL

iTA

iSol

iTri

coTri

iBriSpD

iEDCiTriL

coBr

coTriiTA

iSol

iTri

coTriL

iBriSpD

iEDCiTriL

coBr

coTriiTA

iSol

iTri

coTriL

-50 0 150

Stim

-50 0 150

Stim

A

C

E

B

D

F

iBriSpD

iEDCiTriL

coBr

coTriL

iTA

iSol

iTri

coTri

iBriSpD

iEDCiTriL

coBr

coTriiTA

iSol

iTri

coTriL

iBriSpD

iEDC

coBr

coTriiTA

iSol

iTri

coTriL

RS27Tr10

iTriL

smsm

Fig. 7. Examples of responses evoked by PMRF stimulation during different periods of the reach. A–E: thick traces indicate responses evoked in selected musclesduring different epochs of the ipsilateral reach: quiet standing (A), during the cue period (B), during the pAPA (C), during the transport phase of the reach (D),and during the lever press (E). F: similar display for stimulation applied during the transport phase of the contralateral reach at the same site. Thin traces indicatethe averages of unstimulated trials. Shaded bar in A–F emphasizes the period from which the evoked responses were generally calculated (see METHODS). Numbersafter the headers for each panel indicate the epoch of each stimulation as illustrated in Fig. 2B. Each EMG is scaled identically in A–F. Data are taken from thesame experiment as illustrated in Fig. 5, and the same muscles are illustrated, with the exception of the absence of the iAcT and the inclusion of the iTri andthe coTri.



by 10.220.33.6 on Novem


Dow

nloaded from


iSpD. During the lever press (Fig. 7E), the responses in allmuscles were similar to the basic pattern described for Fig. 7A.During the transport phase of the contralateral reach, theresponses in the ipsilateral forelimb muscles were small, orabsent. In those muscles of the contralateral forelimb and theipsilateral hindlimb in which responses were clearly visible,both the sign and the magnitude of the responses were similarto those observed during the ipsilateral reach. Thus, surpris-ingly, a suppression of the coTriL and a facilitation of the coTriwere observed regardless of whether the contralateral forelimbwas the support limb, as in Fig. 7D, or the moving limb, as inFig. 7F.

The most striking observation in these results is thecomplete lack of an evoked response in the iBr muscleduring the transport phase of the ipsilateral forelimb (Fig.7D), and this despite the existence of a large response insome other ipsilateral muscles active during flexion, such asthe iSpD. That this task-related gating of the response wasa consistent result is illustrated in Fig. 8, A–C, for threestimulation sites in cat RS27 and one in cat RS28. Figure 8Ashows that stimulation during the swing phase of the stepcycle evoked clear facilitatory responses in the iBr from allfour illustrated sites, in both cats. Stimulation during quietstanding equally evoked clear responses for three of the foursites. When stimulation was applied during the transportphase of the reach (Fig. 8A, right), however, no clearfacilitatory responses were evoked. In contrast, the samestimulation, in the same four sites, produced strong, short-latency responses in the iSpD during both reaching andlocomotion, although no responses during quiet standing(Fig. 8B). Responses in the coBr (Fig. 8C) showed aresponse pattern qualitatively similar to those in the iBr.Stimulation during contralateral swing evoked responsesfrom all four of the illustrated sites. In contrast, the samestimulation during the transport phase of the contralateralreach evoked facilitatory responses in none of the four sitesand, in three sites, evoked instead a clear suppression of thebackground activity. A similar gating of activity was ob-served in the ClB (Fig. 8D), which, unlike the Br, remainsactive throughout the transport phase of the reach (Schepensand Drew 2003). During locomotion, stimulation evokedfacilitation of iClB and coClB during the swing phase of therespective limb. During the transport phase of the same,respective, limb, a small facilitatory response was evokedfrom one of the illustrated sites and in the other two sitesthere was either no response or a suppression of activity. Inthis situation it is important to note that the backgroundmuscle activity during the stimulated epoch of the reach iscomparable to that observed during the stimulation duringthe swing phase of locomotion.

The summary of the responses for selected musclesshown in Fig. 9 serves to illustrate several general pointsconcerning the relative organization of the responses evokedby PMRF stimulation during locomotion and reach. First,the lack of any short-latency increase in the iBr during thetransport phase of the reach (epoch 4 in Fig. 9), illustratedfor selected tracks in Figs. 7 and 8, is confirmed for many(13/19) of the stimulated sites for cat RS27 (Fig. 9A) and all28 sites in cat RS28 (Fig. 9B). A few sites in RS27 (n � 6)did evoke facilitatory responses during the transport phase,but these responses were generally smaller than those ob-

served during locomotion. Four of these sites were withinthe NRpc, while the other two were in the NRgc. In RS28,there were large responses when stimulation was applied inepoch 3 (pAPA) but not during the transport phase of theipsilateral reach (epoch 4). Moreover, these responses wereseen not only during the pAPA preceding the ipsilateralforelimb reach but also during the pAPA (epoch 8) preced-ing the contralateral forelimb reach. Responses were absentduring the lever press period of each limb (epochs 5 and 10).Thus the effects in the iBr can be described as a generalizedfacilitation during standing and the pAPA, together withsuppression during reach and lever press. In contrast to theresults for the iBr, there were pronounced facilitatory re-sponses in the iSpD and the iEDC during the transport phaseof the reach and these were of a similar, or larger, magnitudeto those observed during locomotion (see Figs. 5 and 8). Asfor the iBr, the responses in the coBr during the transportphase of the contralateral limb, in both cats RS27 and RS28,were either absent or much smaller than those observed inlocomotion.

Responses in the iTriL exhibited a similar pattern ofactivation in both cats RS27 and RS28. In general, activityin this muscle was suppressed during all stages of the reachexcept for the transport phase, and that for both the ipsilat-eral (epoch 4) and the contralateral (epoch 9) limb. In catRS27 stimulation produced facilitation of activity frommany sites during the transport phase of the ipsilateral andcontralateral reach, while in cat RS28 responses during thisepoch were absent (magnitude � 0), i.e., the suppressionwas absent but was not replaced by facilitation. In both catsthere was suppression of the coTriL during the transportphase of the ipsilateral reach (epoch 4) and weak facilitationin the transport phase of the contralateral reach (epoch 9). Inthe iTA, facilitatory responses were evoked at all epochs ofthe forelimb reaching movements. As during locomotion,stimulation during the reach consistently evoked suppres-sion of activity in the iSol. Moreover, suppression wasobserved at all epochs.

Note again that responses evoked from the MLF in cat RS28(Fig. 9B) were intermingled with those from more lateral sites.

Ground Reaction Forces to PMRF Stimulation DuringReach

The behavioral expression of the stimulation on the fore-limbs in the form of ground reaction forces in the vertical planeis illustrated in Fig. 10A for the same stimulation site asillustrated in Fig. 7, while the responses from all sites areillustrated in Fig. 10B. As already described, stimulation dur-ing quiet standing (Fig. 10A) produced suppression of theactivity in the iTriL and the coTriL, together with facilitationof the iTri and coTri. Despite the reciprocal responses in thetwo heads of the triceps, stimulation evoked a decrease in thevertical force under the ipsilateral forelimb and an increase invertical force under the contralateral forelimb. Evoked changesin vertical force under the hindlimbs were smaller and morevariable (not illustrated).

Overall, stimulation at the majority of sites in cat RS27 (Fig.10B) produced decreases in vertical force in the ipsilateralforelimb and increases in the contralateral forelimb to stimu-lation during all epochs. During the transport phase of the



by 10.220.33.6 on Novem


Dow

nloaded from


ipsilateral reach (epoch 4), the changes in vertical force underthe contralateral forelimb were significantly decreased withrespect to quiet standing (ANOVA and t-test with Bonferronicorrection) and in some cases reversed to a small suppression.Similarly, during the transport phase of the contralateral limb

(epoch 9) the suppression of vertical force in the ipsilateralforelimb was significantly less than during quiet standing andin some cases was reversed to a facilitation. Similar changeswere observed in cat RS28 during the transport and lever pressperiods.

StandLocomotioniBr

iSpD

A

B

Reach

RS28:Tr10

RS27:Tr07

RS27:Tr10

RS27:Tr18

-25 100Stim

50 -25 100Stim

50 -25 100Stim

50)sm()sm( (ms)

coBrC

-25 100Stim

50 -25 100Stim

50 -25 100Stim

50)sm()sm( (ms)

-25 100Stim

50 -25 100Stim

50 -25 100Stim

50)sm()sm( (ms)

RS28:Tr10

RS27:Tr07

RS27:Tr10

RS27:Tr18

RS28:Tr10

RS27:Tr07

RS27:Tr10

RS27:Tr18

iClB

coClB

RS28:Tr10

RS28:Tr10

RS27:Tr10

-25 100Stim

50 -25 100Stim

50 -25 100Stim

50)sm()sm( (ms)

D

Fig. 8. Task-related gating of responses: responses (thick traces) evoked by PMRF stimulation during the swing phase of locomotion (left), during standing(center), and during the transport phase of the reach (right) for the iBr (A), the iSpD (B), the coBr (C), and the ClB (D). Responses in A–C are illustrated for3 stimulation sites in cat RS27 and 1 stimulation site in cat RS28; in D 1 stimulus site is illustrated for cat RS27 and 2 for cat RS28. The responses for eachmuscle in each cat are scaled to the same value across tasks and within cats. The shaded rectangles in the traces during reach indicate the period at which ashort-latency response would be expected. Thin lines indicate the control (unstimulated) level of activity. Stimuli in A and B and for the iClB in D were appliedduring ipsilateral swing and reach; those in C and for the coClB in D were applied during contralateral swing and reach. Arrows indicate responses during reachthat were decreased with respect to control.



by 10.220.33.6 on Novem


Dow

nloaded from


Comparison of Responses During Locomotion and Reach:Summary

To summarize these data we calculated the overall averagedactivity of the responses for each phase or epoch of stimulationfor each muscle, after removal of outliers for each phase (seeMETHODS). These averages are illustrated in Fig. 11 for fourselected muscles and in Figs. 12 and 13 for most of theforelimb and hindlimb muscles, respectively, recorded in catRS27. The overall average response in the iBr was largest at theonset of swing (group 1) and declined progressively through-out swing. In the contralateral flexor, coBr (Fig. 11B), theevoked responses were maximal in group 6, i.e., 50% out ofphase with the ipsilateral response; again the evoked responsesduring stance were minimal. The responses in the iSpD (Fig.11C) were slightly more heterogeneous but showed the samebasic organization as for the iBr, with the largest responsesoccurring in group 1. The coTri (Fig. 11D) was facilitatedthroughout the period of the normal activity of the muscle,being effectively antiphase with the responses evoked in thecoBr.

During the reach, the magnitude of the overall average of theevoked responses in the iBr (Fig. 11E) during the transportphase of the reach (epoch 4) was less than that observed during

the swing phase of locomotion (reduced by 78%). This wasequally true when only considering those few stimulation sitesthat produced facilitation during the reach (Fig. 11E, see alsoFig. 9). Moreover, the responses evoked during the transportphase of the ipsilateral reach were only slightly larger thanthose observed during quiet standing (epoch 1) or during thecontralateral transport phase (epoch 9). This is quite differentfrom the strongly phase-dependent responses observed duringlocomotion. A similar result was observed for the coBr, withthe responses during the transport phase of the contralateralreach being substantially smaller (reduced by 96%) than thoseobserved during locomotion. In contrast, the averaged maximalresponse in the iSpD during the reach was substantially largerthan that observed during locomotion (increased by 123%). Inthis muscle the evoked responses were also substantially largerthan those evoked during quiet standing or during the contralat-eral reach. Finally, in the coTri, the average response duringthe transport phase of the ipsilateral reach, when the coTri isinvolved in weight support (see Fig. 4C), was facilitatory andwas slightly larger (increased by 21%) than both the maximalresponse during locomotion (group 2, Fig. 11D) and theresponse evoked during quiet standing (epoch 1, Fig. 11H).The evoked response was maximal during the lever press

iBr iSpD iTriL

coBr coTriL iTA iSol0 1 2 3 4 5 6 7 8 9 10 0 1 2 3 4 5 6 7 8 9 10 0

0

0

0

0

1 2 3 4 5 6 7 8 9 10

0 1 2 3 4 5 6 7 8 9 100 1 2 3 4 5 6 7 8 9 10 0 1 2 3 4 5 6 7 8 9 10 0 1 2 3 4 5 6 7 8 9 10

7000

5000

5000

3000

5000

-5000

-10000

-15000

1000

2000

2000

2000

-1000

-1000

-2000

-2000

-2000

-4000

5000

50004000

4000

40006000

iBr iTriL coBr coTriL

0

0

0 1 2 3 4 5 6 7 8 9 10 0 1 2 3 4 5 6 7 8 9 10 0 1 2 3 4 5 6 7 8 9 10

20003000

-1000 -1000

1200

3400

5600

7800

10000

1000

4000

60005000

5000

-5000

-10000

3000

3000

3000

10001000

1000

-1000 -1000

A

B

iEDC

0 1 2 3 4 5 6 7 8 9 10

5000

2000

-1000

8000

4000

2000

0

-2000

-4000

-60000 1 2 3 4 5 6 7 8 9 10

Epoch Epoch Epoch Epoch

REACH

000000000000000000000000000000

000000000000000000000000000000000000

stinU yrartibr

Astin

U yrartibrA

stinU yr artibr

A

Fig. 9. Averaged responses evoked in selected muscles in cats RS27 (A) and RS28 (B) by PMRF stimulation during reach. Data are plotted in a mannersuperficially similar to those illustrated in Fig. 6 for all 19 stimulation sites in RS27 and all 28 sites in RS28. However, here the x-axis indicates the epoch ofthe stimulation. The 2 shaded rectangles in each graph indicate the transport phase of the ipsilateral (epoch 4) and contralateral (epoch 9) reach. We illustratethe same muscles as in Fig. 6, and the responses are scaled to the same values as in Fig. 6 to allow direct comparison of the magnitude of the evoked responsesduring locomotion and in reach. Thick red traces in A indicate the data illustrated in Fig. 7. Thin red traces in B indicate the effects of stimulation sites withinthe MLF.



by 10.220.33.6 on Novem


Dow

nloaded from


(epoch 5) when the cat was exerting a downward force todepress the lever with the ipsilateral limb. Note, however, thatfacilitatory responses were also observed during the transportphase of the contralateral reach (epoch 9), when that limb is notsupporting the body.

Figure 12 extends this summary by comparing the averageresponses during reach and locomotion for several of theforelimb muscles recorded from cat RS27. Note that in thisfigure we illustrate the overall average percent change, calcu-lated as the average of the percent change observed in eachgroup (locomotion), or epoch (reach), from all 19 sites in catRS27. Plotting the average of the percent change allows us toplot the responses from different muscles on a single graph. Inthis representation, a muscle in which the largest response isalways evoked in group 1 during locomotion will have anoverall average of 100% (during group 1). In contrast, a musclethat shows its largest response at variable times during the gaitcycle (in different experiments) will have a maximum value of�100%.

The average responses evoked in all of the ipsilateral fore-limb flexor muscles during locomotion were very consistent(Fig. 12A). Facilitatory responses were maximal (and close to100%) at the onset of swing (group 1) in most muscles andwere smaller or absent during stance in the majority. In theEDC, responses were maximal in group 3, again with amagnitude of 100%. During the transport phase of the ipsilat-

eral reach (epoch 4 in Fig. 12B), responses were absent for theiClB and the iBr (as illustrated in Fig. 11) but were larger thanduring locomotion for the iEDC, the iSpD, and the iLtD. Theresponses in the contralateral flexors coBr and coClB duringlocomotion were generally reciprocal to those observed in theipsilateral flexor muscles, and similar to those illustrated forthe coBr in Fig. 11. Similarly, responses were weak or absentduring reach (not illustrated).

Evoked responses in the forelimb extensor muscles weremore variable. Mixed responses (facilitation and suppression)were evoked in most muscles (Fig. 12C). Activity in the iAcTand iSpT (latter not illustrated) was suppressed throughout thestance phase of locomotion, while activity in the iTriL (and theiPaL) was suppressed only at the end of stance. Facilitationwas the major effect observed in the iTri. All of the recordedextensors, except for the iPaL, were weakly facilitated bystimulation during swing. During quiet standing (Fig. 12D),activity in all five of the recorded forelimb extensors, exceptfor the iTri, was suppressed. In contrast, during the transportphase of the ipsilateral reach (epoch 4), most muscles werefacilitated. During the lever press (epoch 5), when the extensormuscles were recruited, activity in all five recorded extensormuscles was again suppressed. During the transport phase ofthe contralateral reach (epoch 9), when the ipsilateral extensorsmuscles were active in a postural context to support the cat’sweight, the responses in the iAcT (and the iSpT and iPaL) was

LHFZ

RHFZ

Stand: pre-cue (1)

18

1116

9

iFL (Fv)

coFL (Fv)

coTri

coTriL

iTri

iTriL

-50 50 100 150 200

A B 2000

1000

0

-1000

-20001 2 3 4 5 6 7 8 9 10

iFL

coFL

1 2 3 4 5 6 7 8 9 10

2000

1000

0

-1000

-2000Stim

ms

Fig. 10. Changes in ground reaction forces produced by the PMRF stimulation. A: averaged EMG activity of ipsilateral and contralateral elbow extensors togetherwith the changes in the vertical component (FV) of the ground reaction force for stimulation applied during quiet standing. Thick traces indicate activity instimulated trials and thin traces in the absence of stimulation. Force traces have been adjusted to superimpose the traces prior to the stimulation (see text). Forcevalues are in Newtons (N); scales for the EMGs are arbitrary. Force traces are scaled to show the same relative magnitude of force in each condition (e.g., 7N for ipsilateral and contralateral forelimb force traces). B: averaged evoked changes in vertical force produced in all experiments in cat RS27. Oval symbolsshow the averaged values of force from individual experiments, while bars show the overall averaged activity (�SE). Note that during ipsilateral and contralateralreach and lever press (epochs 4 and 5 and 9 and 10, respectively), the force exerted against the ground by the reaching forelimb is, by definition, equal to 0. Scalesshow changes in force as N·ms, where ms is the duration of the response (a maximum duration of 100 ms was defined to allow for traces that did not recrossthe control trace).



by 10.220.33.6 on Novem


Dow

nloaded from


equally suppressed while activity in the iTri and iTriL wasfacilitated to approximately the same level as that evokedduring the ipsilateral reach.

In the contralateral extensor muscles that we recorded (coTriand coTriL), responses analogous to those for the ipsilateralmuscles were observed during locomotion (Fig. 12E). ThecoTri was, on average, facilitated during the period of activityof the muscle, and the coTriL was generally suppressed.During reach (Fig. 12F) the coTri was facilitated at all epochsin both limbs. Activity in the coTriL was generally suppressed,including during the transport phase of the ipsilateral limb, butwas facilitated during the transport phase of the contralateralreach as well as during lever press.

Stimulation in the PMRF caused consistent phase-dependentfacilitatory responses in hindlimb flexor muscles during theirperiod of activity (Fig. 13A) similar to those evoked in the

forelimb muscles. Phase-dependent, but muscle-specific, facili-tatory and suppressive responses were observed in the ipsilateralextensor muscles during the stance phase of the hindlimb gaitcycle (Fig. 13C, groups 2–8). During the reach, the hindlimbmuscles show postural changes during the movements of theforelimbs (Fig. 4 and Schepens and Drew 2003). In the flexormuscles of the ipsilateral hindlimb, the magnitude of the re-sponses evoked during standing and reaching (Fig. 13B) wassubstantially larger than that observed during locomotion (Fig.13A). There was a supplementary increase in the level of activityduring the transport phase of the reach (epoch 4) and during thelever press (epoch 5). A similar overall increase during reach wasobserved in the contralateral sartorius (coSrt) as well as in theipsilateral sartorius (iSrt) and semitendinosus (iSt) in cat RS28(not illustrated). The ipsilateral hindlimb extensor muscles gener-ally showed smaller changes during reach (Fig. 13D) but of the

Locomotion Reach

223

4000

3000

2000

1000

0

-10001 2 3 4 5 6 7 8 9 10

A

B

C

D

E

F

G

H

Rs27: iBr

Rs27: coBr

Rs27: iSpD

100

+123%

100

22

-78%

100

4

-96%

1 2 3 4 5 6 7 8 9 10

1 2 3 4 5 6 7 8 9 10 1 2 3 4 5 6 7 8 9 10

1 2 3 4 5 6 7 8 9 10 1 2 3 4 5 6 7 8 9 10

1 2 3 4 5 6 7 8 9 10 1 2 3 4 5 6 7 8 9 10

Group Epoch

4000

3000

2000

1000

0

-1000

4000

3000

2000

1000

0

-1000

4000

3000

2000

1000

0

-1000

-1000

0

2000

4000

6000

-1000

0

2000

4000

6000

100

121+21%

10000

8000

6000

4000

2000

0

-2000

10000

8000

6000

4000

2000

0

-2000

Rs27: coTri

stinU yrartibr

Astin

U yrartibrA

stinU y rartibr

Astin

U yr artibrA

Fig. 11. Averaged magnitude of responsesevoked by PMRF stimulation during loco-motion and reach for selected muscles: barcharts of the averaged magnitude (�1 SE) ofthe evoked response at each epoch of stim-ulation during locomotion (A–D) and duringdifferent parts of the reach (E–H) for 4 se-lected muscles from cat RS27. Histogramsrepresent the overall average of the meanresponses evoked in each track after removalof outliers. Solid line connects the averagedactivity for each epoch. For the iBr duringreach, dotted line indicates the average ofonly the facilitatory responses. Black circlesindicate the values during locomotion andreach that were used to calculate the %changes in activity (values on right of the 4muscles). Values express the responses dur-ing the transport phase of the ipsilateral (E,F, and H, group 4) or the contralateral (G,group 9) reach as % of the maximum re-sponse during locomotion (group 1 in A andC, group 2 in D, and group 6 in B).



by 10.220.33.6 on Novem


Dow

nloaded from


same sign as during locomotion (Fig. 13C) and showed no clearmodulation according to the stimulation phase in either RS27 (Fig.13D) or RS28 (not illustrated). Contralateral hindlimb extensorsshowed a similar, unmodulated response pattern during reach (notillustrated).

The responses evoked in cat RS28 were generally similar tothose illustrated for cat RS27 as described in the followingsection.

Comparison of Responses During Locomotion and Reach:Synthesis

The relative magnitude of the PMRF-evoked responses inshoulder and more distal flexor muscles in the transport phaseof reach compared with the swing phase of locomotion isillustrated in Fig. 14 for both the ipsilateral and contralateralforelimb and for both cats RS27 and RS28. For this analysis weused the maximum values obtained during the swing phase oflocomotion and during the respective reach (epoch 4 or 9). Forexample, for the iBr the responses during epoch 4 werecompared to those during group 1 during locomotion, for theiEDC responses in epoch 4 were compared to group 3, and forthe coBr responses in epoch 9 were compared to group 6 (seeFig. 9). In some muscles, as shown in Fig. 12B, there was anincrease in the relative magnitude of the response. This is

clearest for the shoulder muscles (iSpT, iSpD, iLtD, iAcT) butwas also observed in the distal muscle, the iEDC. Mostpertinent, however, is the consistently smaller level of activityevoked in some of the other flexor muscles, particularly thoseinvolved in flexion of the elbow and retraction of the shoulder.In the Br, ClB, and TrM in both the ipsilateral and contralateralforelimb and in both cats RS27 and RS28, the amplitude of theevoked responses in eight of nine muscles during reach was�25% of the amplitude evoked during locomotion; the excep-tion was the iTrM (58.9%).

The variability in the magnitude of the responses evokedfrom different sites (Figs. 6 and 9) raises the question as towhether the magnitude of the responses during reach is corre-lated to the magnitude of the responses during locomotion. Inessence, does the presence of a large facilitatory response in agiven muscle from a given site during locomotion predict alarge response during reach, and is this relationship graded andsignificant? Such is clearly not the case for the iBr in cat RS27as can be seen in the scatterplot of Fig. 15A. As one mightexpect from the data illustrated in Figs. 6 and 9, responsemagnitude during reach was clearly independent of magnitudeduring locomotion and responses during reach were depressedfrom all sites, including those producing both large and smallfacilitatory responses during locomotion. As a result, all of the

A 200

0

B

C D

iBriClB iSpD

Locomotion

0

200

0 2 3 4 5 6 7 8 9 101

iLtD

iTriLiTri

iAcT

coTriLcoTri

100 100

0 2 3 4 5 6 7 8 9 101

Reach

0 2 3 4 5 6 7 8 9 101 0 2 3 4 5 6 7 8 9 101

300

200

100

0

-100

300

200

100

0

-100

0 2 3 4 5 6 7 8 9 101 0 2 3 4 5 6 7 8 9 101

200

100

0

-100

200

100

0

-100

E F

egatn ecrePegatnecreP

egatnecr eP

iEDC

Fig. 12. Comparison of evoked responses in mul-tiple forelimb muscles during locomotion andreach. A, C, and E: EMG responses evoked inmultiple forelimb muscles of cat RS27 as a func-tion of the phase of the locomotion for musclesactive during ipsilateral forelimb swing (A), dur-ing ipsilateral forelimb stance (C), and duringcontralateral forelimb stance (E). The magnitudeof each muscle is plotted as % of the largestaverage % change in that muscle (average of themean responses from each experiment). As such,the largest possible value for each muscle is100%. Thicker cyan traces in C and E illustrate themean activity of the iTriL and the coTriL, respec-tively, during locomotion (same traces as in Fig.6). B, D, and F: the responses during reach areplotted as % of the magnitude of the responsesduring locomotion. As such, responses can belarger or smaller than those observed during loco-motion. Horizontal lines indicate values of 0 (noevoked response) and 100% (maximum possibleevoked response during locomotion). Note that forthe extensor muscles, the values during reach werecalculated as % of the largest positive (facili-tatory) values during locomotion.



by 10.220.33.6 on Novem


Dow

nloaded from


points lie beneath the line of equivalence in Fig. 15A. Thelinear regression was not significant (R2 � 0.07, P � 0.27), andthe slope of the relationship was 0.29. The slope was thereforenot significantly different from 0. Overall, for those flexormuscles (Br, ClB, and TrM) in which the responses weresmaller during reach than during locomotion, all of the corre-lations, in both limbs and in both cats, returned slopes of �0.5and none of these correlations was significant (Fig. 15B).

In contrast, many other muscles had slopes of �0.5 thatwere significantly different from 0. For example, the coTri(Fig. 15C) showed a significant linear relationship (R2 � 0.71,P � 0.001) between the responses evoked in the muscle duringcontralateral stance during locomotion and those evoked dur-ing the transport phase of the ipsilateral reach (when thecontralateral limb is weight-supporting), with a slope close tounity (1.05). Similarly, all of the ipsilateral forelimb shouldermuscles in cat RS27 had slopes of �0.5 and many of thesewere significant, including the iAcD, iAcT, iLtD, and iSpT(Fig. 15D). A significant relationship was also observed for theiSpD in cat RS28. Significant slopes between the evokedresponses in ipsilateral reach and ipsilateral swing (contralat-eral stance) were observed in both contralateral elbow extensormuscles (coTri, coTriL) in both cats (Fig. 15D). Note that thecontralateral extensors are naturally active at both of theseperiods to support the weight of the cat. In addition, slopes forthe relationships of the ipsilateral extensor muscles (iTri andiTriL in cat RS27 and iTri in cat RS28) for ipsilateral swing andipsilateral reach were also significant (Fig. 15D). A relation-ship for the evoked responses of the ipsilateral extensors duringipsilateral stance and contralateral reach (muscles active duringboth periods) was not calculated because of the inhibitoryresponses observed during stance. Significant slopes �0.5were also seen in the three ipsilateral hindlimb flexors (iSrt, iSt,iTA) recorded in cat RS28 and in the iSrt in cat RS27 (Fig.15D). The slope of the activity in the ipsilateral biventercervicus (iBvC), recorded only in cat RS28, was also close tounity (Fig. 15D).

Several other muscles showed nonsignificant relationships inwhich the evoked responses during reach were mostly abovethe line of equivalence. For example, evoked responses in theiSpD from cat RS27 (Fig. 15E) showed larger responses duringthe transport phase of reaching than during the swing phase of

A B

C D

500

400

300

200

100

01 2 3 4 5 6 7 8 9 10

iSrtiTA

100 500

400

300

200

100

0

200

100

0

-100

-200

Locomotion Reach

200

100

0

-100

-200

1 2 3 4 5 6 7 8 9 10

1 2 3 4 5 6 7 8 9 10 1 2 3 4 5 6 7 8 9 10

iGlMiVL

iSol

Group Epoch

egatnecrePegatnecreP

Fig. 13. Comparison of evoked responses inmultiple hindlimb muscles during locomotionand reach. Graphs plot the EMG responsesevoked in the ipsilateral hindlimb of cat RS27as a function of the phase of the locomotion (Aand C) or the epoch of reach (B and D) formuscles active during swing (A and B) andstance (C and D). Calculations and organiza-tion of the illustration as for Fig. 12.

iFL coFL coFLiFLRs27 Rs28

400

300

200

100

0

-100iSpT iSpD

iEDCiLtD

iAcT

iECRiA

cDiTrM iB

riC

lBco

TrMcoBr

coClB iB

riTrM co

Br

tnecreP

Fig. 14. Magnitude of the responses evoked by PMRF stimulation in forelimbflexor muscles during reach as a function of the magnitude during locomotion.Average magnitude of the evoked responses during reach is plotted as % of theaverage magnitude during locomotion for all shoulder and forelimb flexormuscles in the iFL and coFL in cats RS27 and RS28. Responses of equalmagnitude would be plotted as 100%. As examples, the averaged evokedresponse in the iSpD in cat RS27 during reach is 223% of the response evokedduring locomotion and the response in the iBr of cat RS27 during reach is 22%of that in locomotion (see Fig. 11). Responses evoked during reach withmagnitudes �60% of those during locomotion are shown in gray; those withevoked responses � 60% are shown in black.



by 10.220.33.6 on Novem


Dow

nloaded from


locomotion (see Figs. 11 and 14). However, the relationshipwas not significant (R2 � 0.05, P � 0.40), indicating norelationship between the magnitude of the responses duringlocomotion and those during reach. Indeed, the plot shows amuch larger variation in the responses during reach comparedwith those observed during locomotion. Similar relationshipswere observed in several other muscles, including the iTAfrom cat RS27 (Fig. 15F) and the iTri in cat RS28 (Fig. 15G).

DISCUSSION

We confirm the phase-dependent nature of the modulation ofdescending volleys from the PMRF during locomotion asreported previously (Drew 1991; Drew and Rossignol 1984).The present results extend these previous studies during loco-motion by recording from a larger range of muscles. Mostimportantly, the results illustrate that the evoked responsesproduced by stimulation of the PMRF during reaching main-tain many of the attributes of those produced during locomo-tion. The responses evoked in most muscles during reach wereof sign and magnitude similar to those evoked during locomo-tion. They were equally distributed to flexors and extensors ofall four limbs, including both the moving limb and those actingas postural support. However, the results also show somedifferences in the organization of the responses. For example,in contrast to locomotion, during which responses during theswing phase of the ipsilateral and contralateral limbs were

highly reciprocal, the magnitude of the evoked responsesduring reaching in many muscles showed much weaker mod-ulation according to the limb that was used. Most significantly,while stimulation of the PMRF during the swing phase oflocomotion evoked strong facilitatory responses in all flexormuscles, the same stimulation during the transport phase of thereach was much less effective in producing activity in anumber of these flexor muscles. This strongly suggests somedifferential modulation of reticulospinal signals during reach-ing and locomotion.

Modulation of Muscle Activity During Locomotion bystimulation of PMRF

The results obtained in this study confirm and extend thoseobtained in our previous publications. First, we show that thephase-dependent nature of the responses in the limb muscles istrue for a wide variety of muscles acting around shoulder,elbow, and wrist in the forelimb and hip, knee, and ankle in thehindlimb. Second, we show that these responses are verystereotypical, being almost identical in form in any one muscleover the entire range of the brain stem examined in this study.As in our previous publications (Drew 1991; Drew and Ros-signol 1984), we suggest that the phase-dependent modulationis determined more by the activity of the spinal circuits than bychanges in excitability of reticulospinal neurons in the vicinityof the stimulating electrode. This suggestion is supported by

iBr (Locomotion)-1000 1000 3000 5000 7000

7000

5000

3000

1000

-1000

)hcaeR( r

Bi

0 2000 4000 6000 80000

2000

4000

6000

8000

iSpD (Locomotion)

)hcaeR(

DpSi

-500 0 500 1000 1500

1500

1000

500

0

-500

coTriL (Locomotion)

)hcaeR( irToc

A

E

R =0.07m =0.29p =0.27

2

R =0.05p =0.40

2

R =0.71m =1.05p <0.001

2

iBr

iClB

iTrM coBr

coClB

coTrM iB

riTrM co

Br

iSpDiA

cD iLtDiSpT

iAcT

coTriL

coTri iTri

iTriLco

TriLco

TriiTriL iSrt iSrt iTA iSt

iBvC

2.0

1.5

1.0

0.5

0.0

-0.5

Rs27Rs28

Slop

e

2.0

1.5

1.0

0.5

0.0

-0.5

Slop

e

B

D 8.42.33.52.62.4C

150100500

150

100

50

0

)hcaeR(

ATi

iTA (Locomotion)

R =0.03p =0.51

F

2

400

310

220

130

40

-50-50 40 130 220 310 400

iTri (Locomotion))hcae

R( irTi R =0.10p =0.23

G

2

Fig. 15. Correlations between the averagedevoked responses during reach and duringlocomotion. A, C, and E–G: scatterplots ofthe relationship of the magnitude of the re-sponses evoked in 5 representative musclesduring reach as a function of the magnitudeevoked during locomotion. Each point indi-cates the average response during the maxi-mum phase of activity during locomotionand during the transport phase of the respec-tive reach for each site. Dotted diagonal lineindicates the line of equivalence; solid line inA and C indicates the linear regression. Dot-ted horizontal line in A indicates the lack ofa response (magnitude � 0). N, number ofpoints; R2, coefficient of determination, m,slope; p, probability. Values are in arbitraryunits. B: summary of the slopes for the 9forelimb flexor muscles that decreased to�60% of the value during reach (black barsin Fig. 14). D: muscles similar to the coTri inC for which the relationship was significantat the P � 0.05 level. Values above the barsin D provide the slopes of those relationshipsthat were �2.0.



by 10.220.33.6 on Novem


Dow

nloaded from


the finding that the responses evoked by stimulation of theMLF (comprised predominantly of reticulospinal axons) weresimilar to those evoked from within the PMRF itself (see alsoDrew 1991). Third, we show that stimulation of the PMRFevokes strong responses in two distal muscles, the wrist dor-siflexor extensor carpi radialis (ECR) and the wrist and digitdorsiflexor EDC. These distal muscles were not examined inour previous study (Drew 1991), and the present result bothconfirms the findings of Baker (Riddle et al. 2009; see alsoDavidson and Buford 2006) of a strong descending input fromthe PMRF to muscles acting on distal muscles of the forelimband shows that the presence of such connections is not re-stricted to the primate. Although it is possible that someresponses may have been produced, in part, by activation ofcorticospinal collaterals (see, e.g., Kably and Drew 1998;Matsuyama and Drew 1997), the fact that similar responseswere observed from stimulation of the MLF would suggest thatmost of these responses are truly produced by the reticulospinalsystem. Moreover, the predominance of the large responses inproximal muscles would not be expected from stimulation ofthe corticospinal tract.

Sampling from a larger number of extensor muscles than inour previous study in the intact cat (Drew 1991) serves toemphasize the more complex pattern of activity in these mus-cles than in the flexor muscles. In particular, the resultsemphasize the mixed nature of the facilitatory and suppressiveresponses evoked not only in each limb, but also in eachmuscle. In the ipsilateral limb, stimulation evoked suppressionof activity in the majority of muscles, especially toward the endof the stance phase. In the contralateral limb, we confirmed thepresence of facilitatory responses in the coTri during its periodof activity but were surprised to observe suppression in thecoTriL. This difference may reflect the different biomechanicalfunctions of these two muscles, with the coTriL acting onlyaround the elbow while the coTri acts also to retract theshoulder (see Drew and Rossignol 1987).

Finally, it is worth noting that the general suppressive effectof PMRF stimulation on ipsilateral forelimb extensor musclesis also observed in the hindlimb. As illustrated in Fig. 13,stimulation of the PMRF produced a suppression of activity inmost of the hindlimb extensor muscles from which we re-corded, including the iGL, the ipsilateral gastrocnemius, me-dial head (iGM), the iSol, and the iVL. However, as in theforelimb, facilitatory responses were observed in other mus-cles, including the GlM.

Overall, the results from this and the other studies stronglysupport the suggestion that the functional expression of aninduced descending reticulospinal volley as a complex patternof facilitatory and suppressive responses in different muscles isthe result of the excitability of the spinal circuits and is almostcertainly determined, in great part, by the activity of the CPG(Fig. 1A) as emphasized in several other studies (Degtyarenkoet al. 1998; Drew 1991; Drew and Rossignol 1984; Floeter etal. 1993; Gossard et al. 1996; Orlovsky 1972; Perreault et al.1994). The responses are generally evoked only in one phase orthe other, swing or stance, of the behavior, with a pronounceddifference in the magnitude of the responses in these twophases. However, the modulation is not homogeneous. Re-sponses vary in different groups of muscles and even withinone phase of locomotion (see, for example, the iTriL). Thissuggests that there exist multiple interneuronal pathways that

are activated by the PMRF and that each of these pathwaysmay be subject to a differential modulation.

Modulation Of Muscle Activity During Reach by Stimulationof PMRF

Forelimb muscles: similarities between locomotion andreach. Anatomically, it is evident that the interneuronal path-ways available for transmitting commands for the reticulospi-nal system during reaching are the same as those available totransmit commands during locomotion. However, the questionis whether these anatomical substrates are equally availableduring different behaviors or whether some are available, orused, only during one behavior or the other. In other words, arethe anatomically available pathways differentially gated duringthe tasks? In our original hypothesis, as detailed in the intro-duction, we proposed that the same interneuronal pathwayswould be used in both behaviors. The responses evoked bystimulation of the PMRF during the transport phase of theipsilateral reach should therefore be similar to those evoked bystimulation during the swing phase of the ipsilateral limbduring locomotion. Similarly, those evoked during contralat-eral reach should be similar to those observed during contralat-eral swing (ipsilateral stance). Our results support this proposalfor many muscle groups but contradict them for some others.

The factors supporting the hypothesis include the fact thatresponses of the same sign are observed in most muscles actingaround the shoulder during locomotion and reach. Moreover,the magnitude of the averaged maximal responses evoked inmany of the shoulder muscles during reach was proportional tothat observed during locomotion. Some sites produced smallresponses in both behaviors, while others produced large re-sponses in both behaviors (Fig. 15D). This suggests that thesame pathways mediated transmission from the reticulospinalpathways to the shoulder motoneurons. However, for someshoulder muscles the magnitude of the responses was greaterduring reach than during locomotion, representative of in-creased facilitation of transmission produced by the stimulationduring reaching.

Evoked responses in the extensor muscles were also similarduring reach and locomotion. For example, most of the ipsi-lateral extensor muscles were suppressed during quiet standingand during the lever press, and they were facilitated (ordisinhibited) during the transport phase of reach, in much thesame way that some of these muscles were facilitated duringthe swing phase of locomotion. Although these responses weresimilar in both behaviors they are also, to some extent, incon-gruous with the behavioral requirements of the task. Forexample, both during ipsilateral stance and during the con-tralateral transport phase, activity in the ipsilateral muscles,which are weight-bearing at these times, was suppressed. Thereason for this suppression is not clear, but it should beemphasized that the reticulospinal system has bilateral effectsand that the effect from the contralateral brain stem in theipsilateral muscles would be one of facilitation. It is possiblethat the overall expression of muscle activity in the ipsilateralextensors during load bearing is a balance of suppression fromthe ipsilateral brain stem and facilitation from the contralateralside.

The presence of facilitatory responses in the ipsilateralextensors during the transport phase of ipsilateral reach andduring swing is equally incongruous with the expected behav-



by 10.220.33.6 on Novem


Dow

nloaded from


ior of these muscles. However, it should be noted that mechan-ical perturbation of the limb during the swing phase of loco-motion (Drew and Rossignol 1987) also produces out-of-phaseresponses in extensors. In this situation, the extensor muscleactivity has a clear functional importance in braking the for-ward movement of the limb to prevent it pushing against theobstruction. The facilitation of the ipsilateral extensors by thePMRF stimulation during swing and stance may reflect acontribution of the PMRF to this type of response, perhaps aspart of a spino-bulbo-spinal pathway (Drew et al. 1996; Shi-mamura 1973; Shimamura and Kogure 1979; Stapley andDrew 2004). The facilitatory pathway to the ipsilateral exten-sors may be activated by removal of the limb from the ground.This is supported by the fact that the facilitatory responseduring the transport phase of the limb reverted to suppressionupon contact of the paw with the lever.

Forelimb muscles: differences between locomotion andreach. The most important difference in the responses evokedduring locomotion and reach was the almost complete absence(gating) of the responses evoked in some of the forelimb flexormuscles during the reach, namely the Br, ClB, and TrM. Thiswas a robust finding observed for these flexor muscles in eachlimb and in both cats. We believe that the lack of responsereflects an active gating of the input from the PMRF rather thanbeing the result of differences in the background activity ofthese flexor muscles. For example, absence, or even suppres-sion, of responses in the iClB was observed despite a level ofbackground activity similar to that during locomotion (Fig.8D). The fact that background activity during reach was lessthan the magnitude of the response observed by stimulationduring locomotion suggests that occlusion was not a majorfactor. This indicates the presence of a strong suppressivegating of some of these pathways during the transport phase ofthe reach, perhaps by the command for movement transmittedby descending pathways, including the corticospinal tract.Such a targeted suppression might serve to ensure that the lessspecific control signals observed in reticulospinal neurons(broad discharge patterns throughout the transport phase) donot conflict with the more precise and fractionated dischargepatterns observed in the motor cortex during the same task(Yakovenko et al. 2011).

Even in those forelimb muscles in which responses wereevoked in both behaviors there were some important differ-ences. Among these was the phase or limb dependence of theresponses. During locomotion, the evoked responses in mostmuscles were strongly phase (and thus limb) dependent. Par-ticularly in the flexor muscles, responses were maximal in theperiod of activity of a given muscle during the swing phase ofthe respective limb and mostly silent during the stance phase.As a correlate of this, there was no activity during the swingphase of the limb contralateral to the analyzed muscle. How-ever, during reaching, the magnitude of the responses duringthe transport phase of either limb was not as modulated asduring locomotion. This suggests that the interneuronal path-ways are more strongly gated during the swing and stancephases of locomotion than during the different behavioralepochs of the reach.

Hindlimb muscles. While the forelimbs are actively dis-placed during both locomotion and reaching, the hindlimbs areused quite differently in the two tasks. During locomotion theyare active rhythmically, while during the reach they provide

only postural support. The responses evoked in hindlimb mus-cles by PMRF stimulation during locomotion and reach reflectthis difference in function. While the evoked responses duringlocomotion were strongly phase dependent, those evoked dur-ing reach varied little throughout all 10 epochs of stimulation.This suggests that there is little phase-dependent or limb-dependent gating of the reticulospinal volley in the lumbo-sacral cord during reach. However, it should be noted that thesign of the responses, particularly in the hindlimb extensormuscles, is the same during both locomotion and reaching,suggesting that the same pathways are used during locomotionand during postural support. In some of the ipsilateral hindlimbflexor muscles, however, there was a greatly increased activityduring the reach compared with locomotion. This suggests anactive facilitation of these pathways over and above thatobserved during the swing phase of locomotion. The fact thatthe activity in the forelimb muscles is strongly modulatedwhile that in the hindlimb shows little phase dependence againargues for a spinal rather than a brain stem origin for phase-and limb-dependent gating.

Similar Mechanisms for Locomotion and Reaching?

A synthesis of the points made in the previous paragraphsleads us to suggest that, to a large extent, the same interneu-ronal pathways are being used in both behaviors but that theyare differentially modulated in the two tasks. We have there-fore modified the schema that we introduced in Fig. 1 to betterreflect the organization suggested by the present results. First,our results support the view that the major gating of the evokedresponses during locomotion, expressed in the form of phase-dependent modulation of the descending volley, originates inthe rhythmic activity produced by the CPG. However, we havemodified our original conceptual model (Fig. 1A) by empha-sizing the presence of reticulospinal inputs to the interneuronalgroups that lie outside the CPG (Fig. 16A). Whether bothrhythm- and pattern-generating parts of the CPG, as proposedby McCrea et al. (McCrea and Rybak 2007; Rybak et al. 2006),should be included in this representation of the CPG or whetherour interneuronal pathways may indeed represent pattern-gen-erating pathways is not known. We have maintained a projec-tion to the CPG to reflect the fact that stimulation of the PMRFcan produce changes in cycle timing and may also reset thestep cycle (Drew and Rossignol 1984; Perreault et al. 1994).Although we collapse the complexity of the circuit into a singleconceptual model, we emphasize that the gating signals fromthe CPG are differentially modulated according to musclegroup. This is clear in the complexity of the phase-dependentmodulation observed in some muscles, and particularly in theextensors. The input from the cortex is again illustrated with adashed line indicating its nonessential contribution to unob-structed locomotion. Whether the cortex modifies the outputfrom the PMRF during unobstructed locomotion is, however,unknown.

During reaching, the reticulospinal pathways project to thesame interneuronal groups, which, in this task, are obviouslynot influenced by the rhythm generator. Instead, we proposethat during reaching the gating of activity is the result of amodulation of interneurons by the descending command formovement, including the corticospinal tract. The fact that thereis a specific suppression of activity in the forelimb flexor



by 10.220.33.6 on Novem


Dow

nloaded from


muscles while other muscles, such as some shoulder muscles,are facilitated suggests that this gating is likely to be at thespinal level. In Fig. 16B, the gating of the flexor muscles isrepresented by an inhibitory interneuron. The differential ef-fects observed in different groups of muscles also suggestmodulation of activity in specific interneuronal pathways pro-jecting to different groups of muscles. Nonetheless, the largevariability in the responses of some of the shoulder musclesduring reaching (Fig. 15, E–G), despite the relative invarianceof the behavior, equally raises the possibility that there is somefacilitation of the activity of reticulospinal neurons in the brainstem.

In contrast to the situation during locomotion, the resultssuggest that apart from the strong and specific inhibition of thepathways to the flexor motoneurons, the gating to most path-ways is not as strong as during locomotion. This is implicit inthe results showing that the magnitude of the responses is notas clearly dependent on the activity of a given limb as duringlocomotion. In addition to the gating from descending path-ways, it is probable that feedback from the paws is alsoimportant in modulating activity as discussed above for the

effects observed during locomotion. This activity might beparticularly important in modifying the responses to the exten-sors depending on whether the limb is in contact, or not, witha support surface.

Comparison with Other Studies

There are few other studies that have examined gating ofsupraspinal activity during reaching, and only one other groupthat we are aware of has examined this in the reticulospinalsystem, and that in the primate. Davidson and Buford (2004,2006) used stimulus-triggered averaging to examine the patternof activity evoked in arm muscles by stimulation of the PMRF.They found two results that are pertinent to our study. First,they reported that stimulation at different stages of the move-ment, or when movements were made in different directions(Davidson and Buford 2004) or even with different limbs(Davidson and Buford 2006), made no difference to the re-sponses that were evoked. This is different from what wereport, although as we emphasize the relative difference in thepattern of the responses to stimulation at different periods is

Locomotion ReachA B

Fig. 16. Conceptual models illustrating possible gating controls during locomotion and reach. A: modified version of the model illustrated in Fig. 1A emphasizingthe existence of reticulospinal projections that influence muscle activity both via the CPG as well as through interneuronal pathways (contained within therectangle) that are influenced by, but are not part of, the CPG. The interneuronal pathways include separate channels to arm flexor muscles and other shouldermuscles, but these are influenced similarly during locomotion. B: modified version of the model illustrated in Fig. 1B for the situation during reaching. This modelcontains only the interneuronal pathways that are outside the CPG. During reach, there is input from the RST to these interneuronal pathways as well as inputfrom the CST. This latter tract is suggested to suppress activity in flexor interneurons influenced by the RST and to facilitate other interneurons influenced bythe RST. In this schema, the CST initiates and regulates limb movement via a separate population of interneurons that is not influenced by the RST. See textfor other explanations. Sh, shoulder.



by 10.220.33.6 on Novem


Dow

nloaded from


much less than during locomotion. As pointed out by Davidsonand Buford (2006), this may reflect the difference in posturesand context, particularly the need for load bearing during bothlocomotion and our reaching task. However, it probably alsoreflects differences in the protocols in that we specificallydesigned these experiments to detail the responses at specificphases of the task. The second finding was that the weakestresponses were in the forelimb flexor muscles that they studied.Very few sites produced facilitation of any of the ipsilateralarm flexor muscles studied, namely, the ipsilateral and con-tralateral biceps brachii and iBr, and none produced facilitationof the coBr (Davidson and Buford 2006). This is very similarto what is reported in our study and supports our view thatsignals from the PMRF to the arm flexor muscles are sup-pressed during voluntary movements. In contrast, they reportedstrong responses in shoulder muscles, as we also found in ourstudy. Together, these findings strongly support our contentionthat there is a specific suppression of the pathway from thePMRF to some of the arm flexor muscles during the reach.

The responses evoked by stimulation of the PMRF can beusefully contrasted with those observed to motor cortical stim-ulation. For example, Cheney (Park et al. 2004) showed thatstimulus-triggered averaging of the responses produced bysingle stimuli to the primate monkey cortex during a complexreaching task produced strong and frequent facilitation of theBr. Similarly, we have found that stimulation of the motorcortex during the transport phase of the reach frequentlyevokes responses in the coBr (Drew et al., unpublished obser-vations). Lemon (Lemon et al. 1995) showed that transcranialmagnetic stimulation during different phases of a reach andgrasp movement in humans also produced strong facilitation ofarm and hand muscles, including the forearm flexor brachio-radialis.

Relationship to Cell Activity During Behavior

The discharge characteristics of reticulospinal neurons havebeen detailed both in intact cats during locomotion (Drew et al.1986; Matsuyama and Drew 2000) and reaching tasks (Schep-ens et al. 2008; Schepens and Drew 2004, 2006) as well asduring reaching in primates (Buford and Davidson 2004; Sot-eropoulos et al. 2012). All of these studies have emphasized thecomplexity of the discharge patterns of reticulospinal neurons,and the results show that few cells discharge in discrete burstsof activity restricted to specific parts of the transport phase inthe same way as motor cortical neurons (Drew 1993; Griffin etal. 2008; Krouchev and Drew 2013; Lavoie and Drew 2002;Yakovenko et al. 2011). Moreover, while most pyramidal tractneurons discharge only, or predominantly, during movement ofa single limb, many reticulospinal neurons discharge duringmovements of multiple limbs (Drew et al. 1986; Prentice andDrew 2001; Shimamura and Kogure 1983).

The results from the present stimulation study are fullycompatible with the conclusions drawn from our previous unitrecording studies. Cells active during reach of both the ipsilat-eral and contralateral forelimbs are likely to influence muscleactivity on both sides of the body, and the sign and magnitudeof this interaction will vary with the limb moved and the epoch ofthe movement under study. A given cell active during reach ofa given limb will therefore likely produce a differential andbilateral modification of the activity of a wide range of mus-

cles. However, given the results obtained in this stimulationstudy, cell activity during reach may be prevented from influ-encing the activity of the muscles involved in flexing the arm.This would suggest that the primary effect of reticulospinalactivity during a reach is on the more postural activity (includ-ing both joint stabilization and equilibrium) than on the mus-cles involved primarily in controlling the active movement.

We emphasize, however, that these results do not imply acomplete lack of any specificity in the information content ofreticulospinal cells. Stimulation in the PMRF is likely toactivate a quite diverse population of cells, in contrast tostimulation in the motor cortex, where the stimulation willactivate a more homogeneous population. In addition, thestimulation will also activate axons of passage, possibly lead-ing to activation of cells distant from the site of the stimulation(see Davidson and Buford 2004 and Drew and Rossignol1990a for a fuller discussion of this point). As such, we haveto consider that the stimulation effects demonstrated in thisstudy may, to some extent, reflect a combination of morespecific and heterogeneous response effects or may mask morespecific effects. For example, during both locomotion andreaching some reticulospinal neurons do have quite restricteddischarge patterns and the termination patterns of reticulospi-nal axons are quite variable (see, e.g., Matsuyama et al. 1988,1993, 1997). In this respect our spike-triggered averagingstudies (Schepens and Drew 2006) are important in showingthat activity in some cells does directly suppress activity inipsilateral extensor muscles, consistent with the results fromthis study. Moreover, this suppression is observed only duringquiet standing and contralateral reach and is not seen (i.e., isgated) during the ipsilateral reach, as are the stimulationeffects. This latter observation would be compatible with theproposed role of peripheral afferent input in gating PMRFresponses from suppression during standing to facilitationduring reach.

Summary and Conclusions

In this study we have compared the effects of transientlyactivating the reticulospinal system on the natural muscleactivity underlying locomotion and reaching. The results em-phasize the importance of the CPG in producing the stronglyphase-dependent responses that are generally observed in flex-ors and extensors during swing and stance. These responses arestrongly reciprocal, generally being present in one phase andlargely absent in the other. In contrast, the responses duringreaching, with the exception of some arm flexors, are notsubject to the same rigorous modulation, with similar levels ofactivity frequently being observed during quiet standing andduring transport of either limb. We suggest that this providesmore flexibility for the PMRF to influence motor activity indifferent contexts during discrete voluntary movements thanduring locomotion, in which behavior the response is morestrictly regulated by the CPG. At the same time, we suggestthat the expression of the response is dependent to a certainextent on the excitability of the interneurons onto which theaxons of the cells impinge, and we further suggest that thisexcitability is modified by activity in other descending path-ways, including the corticospinal tract. In the case of some ofthe forelimb flexor muscles, the modulation of the interneuro-nal excitability is sufficient to completely suppress the influ-



by 10.220.33.6 on Novem


Dow

nloaded from


ence of the reticulospinal system during the transport phase ofthe limb. We suggest that this is a mechanism to ensure precisecontrol over the targeted muscles while still permitting moregeneralized, postural responses in the muscles of the support-ing limbs. It is to be noted that shoulder flexor muscles areunaffected or even facilitated by the PMRF during reaching,perhaps reflecting the relative lack of control over shoulder andaxial muscles by the motor cortex in the cat (Armstrong andDrew 1984b, 1985; Yakovenko et al. 2011) and the relativeimportance of the PMRF for their control.

Finally, it is interesting to speculate that if reticulospinalactivation of these forelimb muscles is suppressed by thecorticospinal system, loss of this suppression, for example, bycortical infarct, should release this pathway to reticulospinalcontrol. Such an effect is supported by the findings of Zaaimiet al. (2012) of increased reticulospinal efficacy followingpyramidal tract lesion in the primate.

ACKNOWLEDGMENTS

We thank M. Bourdeau, N. De Sylva, P. Drapeau, C. Gauthier, F. Lebel,and J. Soucy for technical assistance in the performance and analysis of theseexperiments. We thank Drs. Elaine Chapman and Serge Rossignol for helpfulcomments on this manuscript.

Present address of K. S. Dyson: Centre de recherche de l’Institut Univer-sitaire de Gériatrie de Montréal, 4565 Chemin Queen-Mary, Montreal, QCH3W 1W5, Canada.

GRANTS

This work was supported by an operating grant (MOP-9578) from theCanadian Institutes of Health Research (CIHR) and an infrastructure grantfrom the Fonds de la Recherche en Santé du Québec (FRSQ).

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the author(s).

AUTHOR CONTRIBUTIONS

Author contributions: K.S.D., J.-P.M., and T.D. analyzed data; K.S.D.,J.-P.M., and T.D. interpreted results of experiments; K.S.D., J.-P.M., and T.D.prepared figures; K.S.D. drafted manuscript; K.S.D., J.-P.M., and T.D. ap-proved final version of manuscript; J.-P.M. and T.D. performed experiments;T.D. conception and design of research; T.D. edited and revised manuscript.

REFERENCES

Armstrong DM, Drew T. Discharges of pyramidal tract and other motorcortical neurones during locomotion in the cat. J Physiol 346: 471–495,1984a.

Armstrong DM, Drew T. Topographical localisation in the motor cortex ofthe cat for somatic afferent responses and evoked movements. J Physiol 350:33–54, 1984b.

Armstrong DM, Drew T. Electromyographic responses evoked in muscles ofthe forelimb by intracortical stimulation in the cat. J Physiol 367: 309–326,1985.

Baker SN, Olivier E, Lemon RN. Task-related variation in corticospinaloutput evoked by transcranial magnetic stimulation in the macaque monkey.J Physiol 488: 795–801, 1995.

Berman AJ. The Brainstem of the Cat. A Cytoarchitectonic Atlas withStereotaxic Coordinates. Madison, WI: Univ of Wisconsin Press, 1968.

Bretzner F, Drew T. Contribution of the motor cortex to the structure and thetiming of hindlimb locomotion in the cat: a microstimulation study. JNeurophysiol 94: 657–672, 2005.

Buford JA, Davidson AG. Movement-related and preparatory activity in thereticulospinal system of the monkey. Exp Brain Res 159: 284–300, 2004.

Crouch JE. Text-Atlas of Cat Anatomy. Philadelphia, PA: Lea and Febiger,1969.

Davidson AG, Buford JA. Motor outputs from the primate reticular formationto shoulder muscles as revealed by stimulus-triggered averaging. J Neuro-physiol 92: 83–95, 2004.

Davidson AG, Buford JA. Bilateral actions of the reticulospinal tract on armand shoulder muscles in the monkey: stimulus triggered averaging. ExpBrain Res 173: 25–39, 2006.

Degtyarenko AM, Simon ES, Burke RE. Locomotor modulation of disyn-aptic EPSPs from the mesencephalic locomotor region in cat motoneurons.J Neurophysiol 80: 3284–3296, 1998.

Degtyarenko AM, Zavadskaya TV, Baev KV. Mechanisms of supraspinalcorrection of locomotor activity generator. Neuroscience 52: 323–332, 1993.

Drew T. Functional organization within the medullary reticular formation ofthe intact unanesthetized cat. III. Microstimulation during locomotion. JNeurophysiol 66: 919–938, 1991.

Drew T. Motor cortical activity during voluntary gait modifications in the cat.I. Cells related to the forelimbs. J Neurophysiol 70: 179–199, 1993.

Drew T, Cabana T, Rossignol S. Responses of medullary reticulospinalneurones to stimulation of cutaneous limb nerves during locomotion inintact cats. Exp Brain Res 111: 153–168, 1996.

Drew T, Dubuc R, Rossignol S. Discharge patterns of reticulospinal and otherreticular neurons in chronic, unrestrained cats walking on a treadmill. JNeurophysiol 55: 375–401, 1986.

Drew T, Prentice S, Schepens B. Cortical and brainstem control of locomo-tion. Prog Brain Res 143: 251–261, 2004.

Drew T, Rossignol S. Phase-dependent responses evoked in limb muscles bystimulation of medullary reticular formation during locomotion in thalamiccats. J Neurophysiol 52: 653–675, 1984.

Drew T, Rossignol S. A kinematic and electromyographic study of cutaneousreflexes evoked from the forelimb of unrestrained walking cats. J Neuro-physiol 57: 1160–1184, 1987.

Drew T, Rossignol S. Functional organization within the medullary reticularformation of intact unanesthetized cat. II. Electromyographic activityevoked by microstimulation. J Neurophysiol 64: 782–795, 1990a.

Drew T, Rossignol S. Functional organization within the medullary reticularformation of intact unanesthetized cat. I. Movements evoked by microstimu-lation. J Neurophysiol 64: 767–781, 1990b.

Floeter MK, Sholomenko GN, Gossard JP, Burke RE. Disynaptic excita-tion from the medial longitudinal fasciculus to lumbosacral motoneurons:modulation by repetitive activation, descending pathways, and locomotion.Exp Brain Res 92: 407–419, 1993.

Gossard JP, Floeter MK, Degtyarenko AM, Simon ES, Burke RE. Disyn-aptic vestibulospinal and reticulospinal excitation in cat lumbosacral mo-toneurons: modulation during fictive locomotion. Exp Brain Res 109:277–288, 1996.

Griffin DM, Hudson HM, Belhaj-Saif A, McKiernan BJ, Cheney PD. Docorticomotoneuronal cells predict target muscle EMG activity? J Neuro-physiol 99: 1169–1186, 2008.

Kably B, Drew T. The corticoreticular pathway in the cat. I. Projectionpatterns and collaterization. J Neurophysiol 80: 389–405, 1998.

Krouchev N, Drew T. Motor cortical regulation of sparse synergies providesa framework for the flexible control of precision walking. Front ComputNeurosci 7: 83, 2013.

Krouchev N, Kalaska J, Drew T. Sequential activation of muscle synergiesduring locomotion in the intact cat as revealed by cluster analysis and directdecomposition. J Neurophysiol 96: 1991–2010, 2006.

Lavoie S, Drew T. Discharge characteristics of neurons in the red nucleusduring voluntary gait modifications: a comparison with the motor cortex. JNeurophysiol 88: 1791–1814, 2002.

Lemon RN, Johansson RS, Westling G. Corticospinal control during reach,grasp, and precision lift in man. J Neurosci 15: 6145–6156, 1995.

Luccarini P, Gahery Y, Pompeiano O. Cholinoceptive pontine reticularstructures modify the postural adjustments during the limb movementsinduced by cortical stimulation. Arch Ital Biol 128: 19–45, 1990.

Massion J. Movement, posture and equilibrium: interaction and coordination.Prog Neurobiol 38: 35–56, 1992.

Matsuyama K, Drew T. The organization of the projections from thepericruciate cortex to the pontomedullary brainstem of the cat: a study usingthe anterograde tracer, Phaseolus vulgaris leucoagglutinin. J Comp Neurol389: 617–641, 1997.

Matsuyama K, Drew T. Discharge characteristics of vestibulospinal andreticulospinal neurons during locomotion in the intact cat. I. Walking on alevel surface. J Neurophysiol 84: 2237–2256, 2000.

Matsuyama K, Kobayashi Y, Takakusaki K, Mori S, Kimura H. Termi-nation mode and branching patterns of reticuloreticular and reticulospinal



by 10.220.33.6 on Novem


Dow

nloaded from


fibers of the nucleus reticularis pontis oralis in the cat: an anterogradePHA-L tracing study. Neurosci Res 17: 9–21, 1993.

Matsuyama K, Ohta Y, Mori S. Ascending and descending projections of thenucleus reticularis gigantocellularis in the cat demonstrated by the antero-grade neural tracer, Phaseolus vulgaris leucoagglutinin (PHA-L). Brain Res460: 124–141, 1988.

Matsuyama K, Takakusaki K, Nakajima K, Mori S. Multi-segmentalinnervation of single pontine reticulospinal axons in the cervico-thoracicregion of the cat: anterograde PHA-L tracing study. J Comp Neurol 377:234–250, 1997.

McCrea DA, Rybak IA. Modeling the mammalian locomotor CPG: insightsfrom mistakes and perturbations. Prog Brain Res 165: 235–253, 2007.

Orlovsky GN. The effect of different descending systems on flexor andextensor activity during locomotion. Brain Res 40: 359–371, 1972.

Park MC, Belhaj-Saif A, Cheney PD. Properties of primary motor cortexoutput to forelimb muscles in rhesus macaques. J Neurophysiol 92: 2968–2984, 2004.

Perreault MC, Rossignol S, Drew T. Microstimulation of the medullaryreticular formation during fictive locomotion. J Neurophysiol 71: 229–245,1994.

Prentice SD, Drew T. Contributions of the reticulospinal system to thepostural adjustments occurring during voluntary gait modifications. J Neu-rophysiol 85: 679–698, 2001.

Rho MJ, Lavoie S, Drew T. Effects of red nucleus microstimulation on thelocomotor pattern and timing in the intact cat: a comparison with the motorcortex. J Neurophysiol 81: 2297–2315, 1999.

Riddle CN, Edgley SA, Baker SN. Direct and indirect connections with upperlimb motoneurons from the primate reticulospinal tract. J Neurosci 29:4993–4999, 2009.

Rybak IA, Shevtsova NA, Lafreniere-Roula M, McCrea DA. Modellingspinal circuitry involved in locomotor pattern generation: insights fromdeletions during fictive locomotion. J Physiol 577: 617–639, 2006.

Sakamoto T, Gahery Y, Mori S. Effects of bethanecol injection into pontinereticular formation upon postural changes accompanying a food retrievaltask by a forelimb in a standing cat. In: Neurobiological Basis of HumanLocomotion, edited by Shimamura M, Grillner S, Edgerton VR. Tokyo:Japan Scientific Societies Press, 1991, p. 45–50.

Schepens B, Drew T. Strategies for the integration of posture and movementduring reaching in the cat. J Neurophysiol 90: 3066–3086, 2003.

Schepens B, Drew T. Independent and convergent signals from the pontomed-ullary reticular formation contribute to the control of posture and movementduring reaching in the cat. J Neurophysiol 92: 2217–2238, 2004.

Schepens B, Drew T. Descending signals from the pontomedullary reticularformation are bilateral, asymmetric, and gated during reaching movementsin the cat. J Neurophysiol 96: 2229–2252, 2006.

Schepens B, Stapley PJ, Drew T. Neurons in the pontomedullary reticularformation signal posture and movement both as an integrated behavior andindependently. J Neurophysiol 100: 2235–2253, 2008.

Shefchyk SJ, Jordan LM. Excitatory and inhibitory postsynaptic potentials in�-motoneurons produced during fictive locomotion by stimulation of themesencephalic locomotor region. J Neurophysiol 53: 1345–1355, 1985.

Shimamura M. Spino-bulbo-spinal and propriospinal reflexes in variousvertebrates. Brain Res 64: 141–165, 1973.

Shimamura M, Kogure I. Reticulospinal tracts involved in the spino-bulbo-spinal reflex in cats. Brain Res 172: 13–21, 1979.

Shimamura M, Kogure I. Discharge patterns of reticulospinal neuronscorresponding with quadrupedal leg movements in thalamic cats. Brain Res260: 27–34, 1983.

Soteropoulos DS, Williams ER, Baker SN. Cells in the monkey ponto-medullary reticular formation modulate their activity with slow fingermovements. J Physiol 590: 4011–4027, 2012.

Stapley P, Drew T. Postural adjustments to sudden vertical displacements ofthe supporting limbs in the cat (Abstract). Soc Neurosci Abst 2004: 306.–308, 2004.

Yakovenko S, Drew T. A motor cortical contribution to the anticipatorypostural adjustments that precede reaching in the cat. J Neurophysiol 102:853–874, 2009.

Yakovenko S, Krouchev N, Drew T. Sequential activation of motor corticalneurons contributes to intralimb coordination during reaching in the cat bymodulating muscle synergies. J Neurophysiol 105: 388–409, 2011.

Zaaimi B, Edgley SA, Soteropoulos DS, Baker SN. Changes in descendingmotor pathway connectivity after corticospinal tract lesion in macaquemonkey. Brain 135: 2277–2289, 2012.



by 10.220.33.6 on Novem


Dow

nloaded from


Differential modulation of descending signals from the ...€¦ · Differential modulation of...

Documents

Transcript of Differential modulation of descending signals from the ...€¦ · Differential modulation of...