Active Dendrites in Motor Neurons

University of Colorado

Boulder, Colorado
June 24-26, 2004

The Meeting was supported by an award from the National Institute of Neurological Disorders and Stroke

Back to the Meeting page.


Introduction to Motor Neuron Properties and the Generation of Firing Patterns (C. J. Heckman)

Motoneurons are unique: they are the only neurons in the CNS whose firing patterns can be directly measured in human subjects. Accordingly, the first goal of this meeting was to present results on motoneuron electrical properties from animal preparations and on motor unit firing patterns in human subjects. The second goal was then to discuss how motor unit firing patterns emerge from the electrical properties of motoneurons.

The key element that has revolutionized our understanding of how recruitment and rate modulation are generated in motoneurons is the persistent inward current (PIC). PICs were discovered in motoneurons by Schwindt and Crill in Seattle, who relied on blockers of potassium channels. They showed that PICs could produce bistable behavior (i.e., self-sustained firing toggled on and off by brief periods of synaptic input). Shortly thereafter, the laboratories of Hultborn and Hounsgaard in Copenhagen discovered that naturally occurring neurotransmitters (the monoamines serotonin and norepinephrine) can also produce bistable behavior. The subsequent work in vivo in the decerebrate preparation in the Hultborn lab and in vitro in the turtle spinal slice by the Hounsgaard lab established the now classic findings on which all subsequent work is based. As the summaries from the presentations at this meeting show, new results continue the revolution and our understanding of motoneurons and their firing patterns evolves at a rapid pace.

Ionotropic vs neuromodulatory inputs

Patterns of motor unit recruitment and rate modulation result from the processing of synaptic inputs to motoneurons by their intrinsic electrical properties. The most important point is that there are two types of synaptic input, ionotropic and metabotropic (also known as neuromodulatory). It is likely that motor commands to produce volitional movements and also reflex responses to perturbations are mediated by ionotropic inputs. Ionotropic inputs are mediated by classic EPSPs and IPSPs, which depolarize or hyperpolarize neurons by opening channels to allow synaptic currents to flow. In contrast, metabotropic (neuromodulatory) inputs act via G-protein coupled receptors that do not open channels directly. Instead, these neuromodulatory inputs activate various intracellular cascades to modify the behavior of the voltage-sensitive ion channels that set the intrinsic electrical properties of the cells (neuromodulatory inputs can achieve other actions as well, even changing gene transcription, but the studies of motoneurons thus far have focused on their actions on voltage sensitive channels). The terms metabotropic, neuromodulatory, and G-protein coupled are often used interchangeably. These inputs determine how the motoneuron responds to a given amount of synaptic input. Thus, the firing pattern produced by the ionotropic input from a motor command or a reflex can be dramatically different, depending on the level of neuromodulatory input to the motoneuron.

Voltage-sensitive channels

The behavior of each voltage-sensitive channel behavior can be described by 3 parameters: activation, deactivation, and inactivation. Each of these parameter typically varies as a function of membrane voltage; moreover, the time course for activation, deactivation, and inactivation also varies with voltage. Deactivation and inactivation are very different. Once a step change in membrane potential activates a channel, the offset of this step deactivates this channel. However, all channels also inactivate to one degree or another; that is, they tend to close even though the voltage step used to activate them is maintained. The classic example is the sodium channel underlying the action potential; this channel rapidly activates (over a time course of less than 1 ms), but then almost as rapidly inactivates (time course slightly greater than 1 ms). The inactivation plays a major role in keeping the spike brief. A key set of channels involving both sodium and calcium in motoneurons produce persistent inward currents (PICs). In these channels, inactivation is slow and may take many seconds. Once activated, PICs produce prolonged depolarization until shut off by an inhibitory input. Much of the work in recent years on motoneurons has focused on PICs.

Voltage clamp

The behavior of ion channels is usually studied using the technique of voltage clamp. In this method, the electrode in the cell is connected to an amplifier that produces negative feedback, such that current is injected into the cell to cancel any change in membrane potential. The injected current is a faithful replica of the current generated by the cell's voltage-sensitive channels or by its synaptic inputs. By holding voltage constant and measuring current, the conductance changes generated by these channels can be calculated. In contrast, during unclamped conditions, often referred to as "current clamp", voltage, current, and conductance all change at the same time. This is the normal mode of function of the cell, but does not allow assessment of channel behavior from channel conductance. A key issue for understanding voltage clamp in motoneurons is that the electrode is almost always at or near the soma of the cell. The voltage clamp is only effective at keeping the voltage constant near the electrode. Much of the vast dendritic tree of motoneurons is thus unlikely to be well clamped — this is known as lack of space clamp. Poor dendritic space clamp from an electrode in the soma can be used to advantage. When a synaptic input is applied to the cell, voltage at the soma remains unchanged so that voltage-sensitive currents at the soma are not activated or deactivated (in this technique, the clamp is initiated long enough before onset of synaptic input to allow all somatic voltage-sensitive currents to reach their steady state levels of activation and inactivation). However, the synaptic input can depolarize or, if it is inhibitory, hyperpolarize, the dendrites and change the activation of dendritic voltage sensitive conductances. In this way, the effect of voltage-sensitive conductances in the dendrites on synaptic input can be identified.

The classic neurotransmitters are glutamate (EPSPs) and glycine and GABA (IPSPs). These are all amino acids. However, there are many other neurotransmitters now known to have actions in the CNS. Several small molecules are particularly important in the spinal cord: norepinephrine (NE), serotonin (5-HT), and acetylcholine (ACh). Two other neurotransmitters in this class, dopamine and histamine, appear to be less important in the mammalian spinal cord, but further studies are warranted. In addition, there are over 50 peptides that act as neurotransmitters, such as substance P, thyroid releasing hormone (TRH), and enkephalin. A key point is that all neurotransmitters act via a variety of receptor subtypes. For example, there are 14 different 5HT receptor subtypes. Furthermore, all neurotransmitters studied in any detail thus far have neuromodulatory receptor subtypes. Again using serotonin as an example, there is one ionotropic subtype (5HT3); all others, including the 5HT2 subtype that predominates on motoneurons, are neuromodulatory and are thus coupled to G-proteins. The only exception thus far appears to be glycine, which as yet is only known to act via ionotropic receptors.

For motoneurons, the most important neuromodulatory receptors are the following:

Most studies have focused on actions of 5HT and NE; studies of GABAergic receptors, muscarinic receptors, metabotropic glutamate receptors and the receptors for the numerous other peptide receptors are just beginning. The bias toward 5HT and NE is perfectly understandable because these monoamines have especially profound effects on motoneurons and, as demonstrated in several of the presentations in this meeting, likely play an essential role in normal motoneuron behavior.

Intrinsic electrical properties of motoneurons in the absence of neuromodulators

The electrical properties of motoneurons can be succinctly summarized by two functions. The first is the frequency-current function (F-I), in which increasing amplitudes of injected current bring the cell to firing threshold and then drive it to progressively higher rates. The threshold current of the F-I function is the primary determinant of the recruitment threshold of the motoneuron. Its slope plays a major role is determining the steepness of rate modulation. In voltage clamp, a progressive increase in voltage command produces a current response. This current-voltage relation (I-V) quantitfies the net currents flowing into the cell. In the absence of neuromodulatory inputs both the F-I and I-V functions are nearly linear. However, the F-I function has a high slope region at high levels of input; the I-V function has some downward curvature near voltages where the spike is initiated. Both of these nonlinearities are likely due to small amplitude PICs.

The effects of 5HT and NE on motoneurons

The monoamines 5HT and NE are released by axons originating in the brainstem (the caudal raphe nuclei for 5HT and the locus coeruleus for NE). 5HT and NE have similar and profound effects on motoneuron excitability. These effects are summarized in Figure 1. Perhaps the most important effect is to markedly facilitate the PIC. In voltage clamp, the PIC is evident as a downward deflection in the I-V function (by convention in voltage clamp, inward currents that depolarize the cell are shown as going downward because what is actually measured is the current injected by the electrode to oppose the current in the neuron). The stronger the monoaminergic input to the cord, the larger the PIC (Fig. 1 A). In addition, the monoamines depolarize the motoneurons, bringing them close to threshold (this is evident as the downward shift of the I-V functions in Fig. 1A). These effects also profoundly influence the F-I function and, hence, recruitment and rate modulation (see Fig. 1B). As the amplitude of the PIC increases, the high gain portion of the F-I function shifts to lower input levels, until it begins at or below threshold. This high-gain region probably corresponds to the secondary range initially identified by Kernell and colleagues.

Figure 1: The effects of increasing monoaminergic input on two basic electrical properties of motoneurons. Low monoaminergic drive corresponds to the state seen in deeply anesthetized animals, which may approximate the state of sleep (blue in A and B). Medium-drive effects are based on studies from the decerebrate cat preparation, which may approximate a state of tonic motor output (green). High-drive effects are based on adding an exogenous NE agonist to the decerebrate state, and may correspond to the extremely high state of arousal occurring in emergencies (red). A: Current-voltage (I-V) relations during voltage clamp. Arrows indicate the onset of the PIC. B: Frequency-current (F-I) relations generated by current injected via the microelectrode. For low monoaminergic input, the small PIC is only activated at high input levels, initiating a secondary range (s) above the primary range (p) At medium levels, activation of the PIC gives a secondary range at a lower input output level but then full PIC activation generates a tertiary (t) range with a low gain. At high levels, the PIC usually activates near threshold for firing and only the secondary and tertiary ranges are present. These figures are cartoons based on average data for low threshold, presumably type S, motoneurons from the studies of Lee and Heckman. These studies were all carried out in the cat; firing rates in human subjects tend to be lower (see below). Note also that PIC onsets in voltage and current for firing are substantially lower when using synaptic as opposed to injected current (see below).

The above I-V changes have really only become clear in recent years. As noted above, the initial studies tended to focus on bistable behavior and plateau potentials. Basically, the PIC produces both phenomena, as illustrated in Figure 2, where the PIC evoked by a brief tendon vibration is shown during voltage clamp in the bottom traces (to match the current-clamp records above, the voltage-clamp currents are inverted from their standard convention), during current clamp with spikes blocked in the middle trace (giving the plateau potential) and with spike present in the top trace (bistable behavior, self-sustained firing). In the past several years, the emphasis has shifted to the enhancement or amplification the PIC provides for synaptic input. This enhancement is evident in the bottom trace of Fig. 2 as the large difference between the Ia synaptic current with (red trace) and without (blue trace) the PIC. The PIC can be prevented from influencing the synaptic input simply by holding the cell hyperpolarized, so that the synaptic input does not depolarize the dendrites sufficiently to reach the voltage threshold for the PIC.

Overall, it is clear that the descending monoaminergic input in particular, and neuromodulatory input in general, set the state of excitability of the motoneuron. The stronger the monoaminergic input to the motoneuron, the greater the effect of a given amount of synaptic input.

Figure 2: Persistent currents facilitated by descending monoaminergic inputs can generate plateau potentials and self-sustained firing in motoneurons. In each panel, the blue trace is taken while the cell was held hyperpolarized and indicates the time course of the applied synaptic input, and the red trace was taken at more depolarized levels where the PIC can be activated (though shifted to line up with the blue trace prior to stimulation). This ionotropic, monosynaptic excitatory input was generated by sustained activation of muscle spindle Ia afferents by tendon vibration. Lower panel: the synaptic input activates a strong persistent inward (PIC) current when the cell is voltage clamped at a level where spikes are generated in unclamped conditions. Middle panel: When spikes are blocked in current clamp (here by intracellular injection of QX-314), the PIC generates a sustained plateau potential. Upper panel: When the cell is allowed to fire normally, the persistent inward current drives self-sustained firing of the motoneuron.

System-level effects

A final point to consider is that the motoneuron does not operate in isolation, but as part of a pool of motoneurons innervating a single muscle (the potential for activating of groups of motoneurons that may not correspond to a clearly defined anatomical muscle does not violate the assumptions made in the following discussion; the same rules discussed here would apply within such a group). The well known systematic differences in motor units that result in the size principle of recruitment produce a pattern of recruitment and rate modulation that is in fact very sensitive to the state of neuromodulatory input. First it should be emphasized that the descending monoaminergic inputs continue to support the normal order of recruitment from units of small to large forces (or, in terms of motor unit types, S before FR before FF). PICs appear to be approximately equally enhanced in S, FR, and FF units and all motoneurons undergo significant subthreshold depolarization (the relative changes in this depolarization among S, FR and FF units have yet to be quantified but it appears that the normal order prevails). What then happens to the pattern of recruitment and rate modulation when the input-output gain of the motoneurons is increased by monoamines? This increase in gain produces a similar increase in the overall input-output function of the motor pool and the muscle it innervates. The overall input-output function has a sigmoidal form; thus, monoamines shift this sigmoid to lower input levels and increase its slope.

Figure 3: Computer simulations of firing rate versus whole muscle force. Simulations based on the cat medial gastrocnemius muscle. The blue rate-force functions were derived from perfectly linear motoneuron F-I functions. The red rate-force functions used the same F-I functions, but with threshold currents reduced by 50%.

It is also important to realize that the classic method of plotting motor unit firing patterns ( i.e. plotting firing rate vs. muscle force or joint torque) necessarily includes an inverted form of this sigmoidal function. In other words, the firing rate-force plot includes the transformation from synaptic input to force. As a result, even if motoneuron F-I functions are completely linear, the low threshold units will exhibit a downward curvature and the highest threshold units a slight upward curvature, both due to the corresponding curved portions of the sigmoidal input-output function of the motoneuron pool and muscle. This is illustrated in Figure 3. Moreover, the relative slopes of the rate-force functions are sensitive to changes in threshold or gain of the motoneurons. The red rate-force functions in Fig. 3 show the effect of simply halving the gains of all motoneuron F-I functions while maintaining their linear form. These result is a decrease in rate-force slope. Ultimately, quantitative interpretations of motor unit firing patterns will need to be run through accurate computer simulations to determine the actual motoneuron F-I functions and synaptic inputs from which the firing patterns are generated.

Recent reviews

In the past 5 years or so, there have been a number of excellent reviews of motoneurons, PICs and monoamines. These are listed below according to the main theme of the review.

1. Powers RK, Binder MD. Input-output functions of mammalian motoneurons. Rev Physiol Biochem Pharmacol 143: 137-263, 2001.

2. Rekling JC, Funk GD, Bayliss DA, Dong XW, Feldman JL. Synaptic control of motoneuronal excitability. Physiol Rev 80(2): 767-852, 2000.

3. Alaburda A, Perrier JF, Hounsgaard J. Mechanisms causing plateau potentials in spinal motoneurones. Adv Exp Med Biol 508: 219-226, 2002.

4. Perrier JF, Alaburda A, Hounsgaard J. Spinal plasticity mediated by postsynaptic L-type Ca2+ channels. Brain Res Brain Res Rev 40(1-3): 223-229, 2002.

5. Hounsgaard J. Motoneurons do what motoneurons have to do. J Physiol 538(Pt 1): 4, 2002

6. Binder MD. Integration of synaptic and intrinsic dendritic currents in cat spinal motoneurons. Brain Res Brain Res Rev 40(1-3): 1-8, 2002.

7. Binder MD. Intrinsic dendritic currents make a major contribution to the control of motoneurone discharge. J Physiol 552(Pt 3): 665, 2003.

8. Heckman CJ, Lee RH, Brownstone RM. Hyperexcitable dendrites in motoneurons and their neuromodulatory control during motor behavior. Trends Neurosci 26(12): 688-695, 2003.

9. Hultborn H, Brownstone RB, Toth TI, Gossard JP. Key mechanisms for setting the input-output gain across the motoneuron pool. Prog Brain Res 143: 77-95, 2004.



Multi-photon fluorescence imaging of calcium fluxes in the dendrites of rat hypoglossal motoneurons

Marc D. Binder, Christopher Davenport, David Margolis, and Randall K. Powers

University of Washington, Seattle, USA

Several lines of indirect evidence implicate dendritic calcium channels as the predominant source of persistent inward currents (PICs) in vertebrate motoneurons (reviewed in Powers & Binder Rev Physiol Biochem Pharmacol 143: 2001; Heckman et al TINS 26:2003; Hultborn et al Prog Brain Res 143: 2004). We have now used multi-photon fluorescent imaging to make direct observations of Ca2+ fluxes in soma and dendrites of rat hypoglossal motoneurons before, during, and after PICs evoked by somatic current injection. We made 300mm transverse brainstem slices at the level of the hypoglossal nucleus from the brainstems of 10-20 day old rats. The slices were maintained at room temperature in ACSF. Hypoglossal (HG) motoneurons were identified and targeted for whole-cell patch clamp analysis using infrared video microscopy. Electrical recordings were made using an Axopatch 200B amplifier (Axon Instruments). Changes in Ca2+-dependent fluorescence were measured by including 140 mM of the indicator dyes Oregon Green Bapta (OGB)1 or OGB 5N (high and low affinity, respectively) in the patch pipette. Imaging was done with a custom built multi-photon microscope with a mode-locked Ti-sapphire laser tuned to 930 nm as the excitation source. After inducing a partial pharmacological blockade of Na+ and K+ channels (using internal CsCl and external TTX), large (> 1 nA) Ca2+-PICs appear in many cells in response to depolarizing voltage-clamp steps or triangular ramps applied to the soma. Often these PICs outlast the period of somatic depolarization by several seconds (prolonged tail currents). Coincident with the persistent inward Ca2+ currents recorded at the soma, changes in Ca2+-fluorescence were observed at the soma and at all dendritic locations (up to 400 µm from soma). The changes in Ca2+-fluorescence in both the soma and dendrites were graded with the level of depolarization, as shown in the figure, which presents somatic voltage-clamp commands and currents (top traces) and changes in dendritic Ca2+-fluorescence (bottom traces). However, large persistent tail currents were not accompanied by proportional changes in Ca2+-fluorescence in the soma. Changes in Ca2+-fluorescence appear to be activated at lower voltages in the dendrites, before either fluorescence or Ca2+ currents are observed at the soma. These observations are consistent with the hypothesis that voltage-gated Ca2+ channels on the dendrites of rat HG motoneurons make a prominent contribution to PICs.

Supported by NIH grants NS26840, EY02048 and RR15769


Dendritic plateau potentials in functionally mature motoneurons in vitro

Friederike Bergmann and Bernhard U. Keller

Zentrum Physiologie, Universitat Gottingen, Humboldtallee 23, 37073 Gottingen, Germany

Motoneurons (MNs) are particularly affected by inhibition of mitochondrial metabolism, which has been linked to their selective vulnerability during pathophysiological states like hypoxia and amyotrophic lateral sclerosis, a fatal neurodegenerative disorder. To elucidate underlying events, we utilized sodium cyanide (CN) as a pharmacological inhibitor of complex IV of the mitochondrial respiratory chain ("chemical hypoxia") and investigated the cellular response in vulnerable and resistant MN types. Bath application of 2mM CN activated TTX-insensitive Na+ conductances in vulnerable hypoglossal MNs, which depolarised MNs by 10.2 ± 1.1 mV and increased their action potential activity. This response was mimicked by sodium azide (2mM) and largely prevented by pre-incubation with the antioxidants ascorbic acid (1mM) and trolox’ (750µM), indicating an involvement of reactive oxygen species (ROS) in the activation mechanism. CN also elevated cytosolic [Ca2+] levels through i) Ca2+ release from mitochondria-controlled stores, ii) significant retardation of cytosolic Ca2+ clearance rates, even when cytosolic ATP levels were held constant during whole-cell recording and iii) secondary Ca2+ influx during elevated firing rates. Blocking mitochondrial ATP production additionally raised cytosolic Ca2+ levels and prolonged recovery of Ca2+ transients with a delay of 5-6 minutes. Comparative studies on hypoglossal MNs, facial MNs and dorsal vagal neurons suggested that CN-responses were dominated by activation of K+ conductances in resistant motoneurons, thus reducing the excitability during mitochondrial inhibition. In summary, our observations therefore support a model where selective MN vulnerability results from a synergistic accumulation of risk factors, including low cytosolic Ca2+ buffering, strong mitochondrial impact on [Ca2+]i, and a mitochondria-controlled increase in electrical excitability during metabolic disturbances

Figure legend. Events following inhibition of mitochondrial metabolism in vulnerable hypoglossal MNs. CN (similar to azide, hypoxia and SOD1) inhibits complex IV (cytochrome c oxidase) of the respiratory chain. This increases the formation of reactive oxygen species, which potentially activate a depolarising Na+ current (ICN). Alternatively, the Na+ current is activated by a direct redox mechanism. The Na+ influx enhances the neuronal excitability and promotes Ca2+ influx during elevated firing rates (action potentials, AP). Inhibition of the respiratory chain furthermore decreases the mitochondrial potential gradient (Dy), leading to reduced Ca2+ uptake into the mitochondrial matrix and release of Ca2+ that was taken up during preceding activity. Mitochondrial inhibition additionally decreases cellular ATP levels, and this further enhances accumulation of intracellular Ca2+.

Bergmann F, Keller BU. Impact of mitochondrial inhibition on excitability and cytosolic calcium levels in brain stem motoneurons from mouse. J Physiol 555: 45 - 59, 2004

Ladewig T, Kloppenburg P, Lalley PM, Zipfel WR, Webb WW, Keller BU. Spatial profiles of store-dependent calcium release in motoneurones of the nucleus hypoglossus from newborn mouse. J Physiol 547: 775-787, 2003.

Lips MB, Keller BU. Endogenous calcium buffering in motoneurones of the nucleus hypoglossus from mouse. J Physiol 511: 105-117, 1998.

Lips MB, Keller BU. Activity-related calcium dynamics in motoneurons of the nucleus hypoglossus from mouse. J Neurophysiol 82: 2936-2946, 1999.

Vanselow B, Keller BU. Calcium dynamics and buffering in oculomotor neurons that are selctively resistant in amyotrophic lateral sclerosis (ALS)-related motoneuron disease. J Physiol 525: 433 - 445, 2000.


Kv2.1 ion channel expression in motoneurons: relation to specific synaptic sites

Robert E. W. Fyffe and Elizabeth A. L. Muennich

Wright State University, Dayton, OH, USA

Delayed rectifier K+ currents are involved in the control of motoneuron excitability, but the precise spatial distribution and organization of the membrane ion channels that contribute to these currents have not been defined. We used immunohistochemistry and confocal and electron microscopy to determine the organization of voltage gated Kv2.1 channels in motoneurons and other spinal neurons. Kv2.1 channels in motoneurons are clustered in a punctate fashion over the soma and dendritic membrane. Prominent large disk-shaped clusters are selectively located postsynaptic to large cholinergic C-terminal boutons. These Kv2.1 channels colocalize with m2 muscarinc receptors. Smaller kv2.1 clusters, which make up more than 90% of all the individual punctae, are localized postsynaptic to excitatory S-terminals. In motoneurons, Kv2.1 channels are not associated with inhibitory synapses and are found only occasionally in non-synaptic membrane. Clusters vary widely in size, and the number of large clusters is approximately the same as the number of C-terminal synapses on motoneuron cell bodies and proximal dendrites. The figure shows large and small Kv2.1 immunoreactive punctae in the surface membrane of adjacent motoneurons, and electronmicroscopic confirmation of localization at C-terminal synapses. The results were discussed in terms of differential regulation of Kv2.1 channels based on subcellular location, and point to novel roles for these channels in control of motoneuron excitability.


What do the integrative functions of te dendritic trees of motoneurons and the selling of real estate have in common? Location, Location, Location!

Ken Rose, Tuan Bui, Maria ter Mikaelian, Diane Bedrossian, and John Grande

Queens University, Kingston, Canada

Part I: Frequency and distribution of contacts between vestibulospinal axons and the dendritic trees of splenius motoneurons

Splenius motoneurons are one of several groups of neck motoneurons that receive monosynaptic, excitatory, connections from cells in the contralateral descending and medial vestibular nuclei. Due to powerful connections from semi-circular canal afferents, these connections are ideally suited to generate a compensatory movement of the head in response to sudden perturbations of the head or body. The goal of the present experiments was to determine the distribution of contacts formed between boutons of vestibulospinal axons, stained with anterograde tracer, PHA-L, and dendrites of contralateral splenius motoneurons, intracellularly stained with Neurobiotin. Careful mapping of the entire dendritic tree revealed that contacts are not uniformly distributed throughout the dendritic tree. Instead, most contacts are located on dendrites medial to the soma. Based on the frequency of contacts per motoneuron (13 to 25) and estimates of the proportion of contralaterally projecting vestibulospinal axons stained with PHA-L (5%), we calculated that each splenius motoneuron receives approximately 400 synapses from cells in the contralateral descending and medial vestibular nuclei.

Part II: Effective synaptic current generated by the 400 synapses on the dendritic trees of motoneurons

It is difficult to experimentally measure the effective synaptic current generated by an entire descending system without co-activating anatomically adjacent pathways. We therefore constructed compartmental models based on anatomical measurements of intracellularly stained neck motoneurons. Each model was equipped with physiologically synapses. The effective synaptic current (ala Binder, Powers and Heckman) reaching the soma, measured in response to activation of 400 synapses at 100 Hz (with a release probability of 50%), was 1.0 nA. Based on current-frequency relations of motoneurons in pentobarbital anaesthetized animals, an effective synaptic current of this magnitude would increase motoneuron firing by only 1 to 2 spikes/s. Thus, we appear to have identified a highly ordered innervation of the motoneuron dendritic tree where the input produces a paradoxically minor change in motor activity. In the context of this conference, the answer to this paradox is obvious, persistent inwards currents (PICs).

To determine the amplification provided by PICs, we redesigned the compartmental models to include L-type calcium channels. The distribution of these channels was determined using an analytical strategy. This strategy compared the distribution of voltage changes throughout the dendritic tree in response to synaptic inputs and somatic current injections that mimicked synaptic-activity induced changes in threshold for plateau potentials (Bennett, Hultborn, Fedirchuk and Gorassini, J. Neurophysiol. 80: 2023-2037, 1998). Based on the results of this strategy, we assigned L-type calcium channels to 20 to 35 "hot-spots" (the number depended on the motoneuron), located on secondary and tertiary dendritic branches, 200 to 350 µm from the soma. Each "hot-spot" was arbitrarily assigned a length of 100 µm. The density of L-type calcium channels was adjusted until we mimicked the "at rest" response of motoneurons to a current ramp injected at the soma, as described by Bennett et al. Changing the background synaptic activity of these models led to systematic changes in the threshold for activation of plateau potentials that matched the experimental observations reported by Bennett et al. Thus, the "hot-spot" distribution model fulfills at least one key experimental feature of "real" motoneurons.

The final step in this study examined the response of the "hot-spot" distribution model to increasing levels of excitatory synaptic activity. Unexpectedly, the "hot-spots" were not activated in an all-or-none fashion. Instead, there was an orderly recruitment of hot-spots as the number of active synapses increased, leading to a linear relationship between the number of active synapses and the effective synaptic current reaching the soma. Moreover, the additional current provided by the L-type calcium channels caused a 4-fold amplification of the current evoked by activating 400 synapses.


The title of this presentation begins as a question: What do the integrative functions of the dendritic trees of motoneurons and the selling of real estate have in common? and ends with an answer: location, location, location. The results of our anatomical and simulation studies support the validity of this answer. At least some inputs to motoneurons, vestibulospinal is one example, are not distributed throughout the dendritic tree. Instead, these inputs appear to seek out a specific region of the dendritic tree. Similarly, the results of our modeling studies support the conclusion that some of the channels responsible for PICs are restricted to hot spots, located 200 to 350 microns from the soma. Hence, we have at least 2 examples of location dependent properties. Nevertheless, there remains a critical missing piece. We have yet to examine the consequences of the combination of these 2 location dependent features. The interaction between these properties may provide the means of solving what we believe is a critical outstanding question. Even with the activation of PICs, the total current delivered by the vestibular spinal system, 4 nA, is not adequate to significantly alter motoneuron activity and hence control head position. Could it be that the selective innervation pattern that we have found, in combination with localized hot-spots of the voltage dependent calcium channels, provides another source of amplification, one that would drive the motoneuron to functionally meaningful levels of activity?



Roles of dendritic and somatic components of the persistent inward currents in controlling motoneuron firing

C. J. Heckman, J. J. Kuo, E. W. Ballou, M. D. Johnson, and A. J. Hyngstrom

Northwestern University, Chicago, USA

The amplitude of the persistent inward current (PICs) in the dendrites of spinal motoneurons is proportional to the intensity of monoaminergic input from the brainstem (Heckman et al. 2003). However, inhibition by local spinal interneurons also has been recently shown to have a strong effect on PIC amplitude (Hultborn et al. 2003; Kuo et al. 2003). At present, it is not clear to what degree this inhibitory suppression of the PIC is ionotropic or metabotropic, but there is no doubt that even a large PIC in a preparation with strong descending monoaminergic drive can be completely suppressed by synaptic inhibition. Moreover, the suppression of PIC amplitude is linearly proportional to the amplitude of the synaptic inhibition. This suggests that excitation and inhibition may interact linearly in dendrites with strong PICs. This hypothesis is presently being tested. A second prediction is that the PIC should become sensitive to inhibition that is linked to muscle length, i.e. the reciprocal inhibition produced by stretch of antagonist muscles. Preliminary data were presented from 5 cells in the decerebrate preparation that showed that simply rotating the ankle joint to stretch the muscles antagonistic to the one innervated by the motoneuron being studied could result in a decrease in PIC amplitude by as much as 80%. This suggests that the overall excitability of the motoneuron is not only dependent on descending motoneuron drive but also on relative muscle lengths.

Although most of the PIC is generated in dendritic regions, a significant component also exists at or near the soma and initial segment. The somatic PIC can assist in synaptic amplification but may also play a more fundamental role in spike initiation. A substantial portion of the PIC is generated by persistent Na current (NaPIC). We tested the hypothesis that NaPIC not only assists in depolarizing the cell to spike threshold, but is in fact essential for initiation of action potentials during slowly rising inputs. Experiments were carried out using whole cell patch clamp in motoneurons cultured from embryonic mice and grown for about 20 days. These cells exhibit a high density of at least one type of Na channel (NaV 1.6), as shown in Figure 1. Note that the two portions of the proximal dendrites in this cell do exhibit immunofluorescence for NaV 1.6, but that the immunofluorescence in the somatic region is especially dense. These cultured cells also exhibit good rhythmic firing to prolonged current steps, as shown in Figure 2 (blue trace). The drug riluzole was bath applied to sharply reduce NaPIC but leave the transient Na current generating the action potential relatively unaffected (Urbani and Belluzzi, 2000). At a concentration of 2 to 5 microM, rhythmic firing was blocked and NaPIC was reduced by 40 to 60%. The red trace in Fig. 1 shows the resulting response in this cell: a strong spike is initiated at step onset, but thereafter spike initiation fails. Note the AHP following the initial spike matches the control AHP well, but no spike occurs. This failure was not due to insufficient depolarization, as increasing step size did not restore firing (green trace in Figure 1). The presence of the spike at step onset during riluzole administration shows that the transient Na current was strong. Similar results have been obtained in over 20 cells. These results provide strong support for the hypothesis that NaPIC at the soma is essential for spike generation. Its likely role is to provide a rapid depolarization as the AHP decays, producing a rate of change of membrane potential fast enough for the activation of the transient Na current to escape its inactivation (Lee and Heckman, 2001). Thus control of NaPIC by monoamines and other neuromodulators can not only influence synaptic input but also control the spike initiation mechanism.

Figure 1: Immunofluorescence for the Na channel subtype NaV 1.6 on a portion of a cultured motoneuron from the mouse. The upper panel shows the soma and two proximal dendrites of the motoneuron (red) and all immunofluorescence for NaV 1.6 (green) in this 7.3 micron thick set of images. The middle panel shows only the NaV 1.6 immunofluorescence that is localized within <0.5 microns of the cell membrane. The bottom panel shows this colocalized fluorescence with the cell. Note that much of the NaV 1.6 fluorescence is on the bottom of the soma in this view.

Figure 2: Response of a cultured mouse motoneuron to steps of injected current in control conditions (black trace) and following bath administration of 10 microM riluzole. The red trace shows the response after riluzole to a step of exactly the same amplitude as for the control. The green trace shows the post-riluzole response to a larger amplitude step.

Heckman CJ, Lee RH, Brownstone RM. Hyperexcitable dendrites in motoneurons and their neuromodulatory control during motor behavior. Trends Neurosci 26:688-695, 2003.

Hultborn H, Denton ME, Wienecke J, Nielsen JB. Variable amplification of synaptic input to cat spinal motoneurones by dendritic persistent inward current. J Physiol 552:945-952, 2003.

Kuo JJ, Lee RH, Johnson MD, Heckman HM, Heckman CJ. Active dendritic integration of inhibitory synaptic inputs in vivo. J Neurophysiol 90:3617-3624, 2003.

Lee RH, Heckman CJ. Essential role of a fast persistent inward current in action potential initiation and control of rhythmic firing. J Neurophysiol 85:472-475, 2001.

Urbani A, Belluzzi O. Riluzole inhibits the persistent sodium current in mammalian CNS neurons. Eur J Neurosci 12:3567-3574, 2000.


Variable amplification of synaptic inputs by dendritic persistent inward currents

Hans Hultborn

University of Copenhagen, Denmark

Electrophysiological and computational evidence indicate that the excitatory current from the synapses on the somato-dendritic membrane is not large enough to drive the motoneurones to the firing frequencies actually attained under normal motor activity (Rose & Cushing, 1999; Powers & Binder, 2001). It was proposed that this paradox could be explained if the voltage-dependent persistent inward currents (PICs) present in the dendrites of motoneurones served to amplify synaptic excitation. According to this hypothesis it would be expected that the PICs are graded rather than of ³all-or-none character.

As described in recent reviews (Powers & Binder, 2001; Hultborn et al. 2004) plateau potentials in motoneurones were recorded as all-or-none phenomena in current clamp mode in the original publications. Correspondingly, in voltage clamp mode a strong negative resistance slope in the I-V curve was seen to have an amplitude that would be regenerative and lead to a stable membrane potential (the plateau potential) in current clamp mode. Much later the effect of tonic synaptic excitation and inhibition on the threshold of the plateau potentials was investigated (evoked by current injection to the soma; Bennett et al. 1998). With synaptic excitation the plateau threshold was reduced, as seen from the recording microelectrode (presumably positioned in the cell's soma), while synaptic inhibition had the opposite effect. This was interpreted as being due to the dendritic localization of the PICs - the membrane depolarization and hyperpolarization by the synaptic input were closer to the PIC channels than the somatic electrode. In most studies up to this date the experimental protocol in which the plateau currents were elicited by current through the microelectrode (soma) are likely to have had a tendency to over-emphasize the “all-or-none” character of the current producing the plateau potential. But are plateau potentials normally seen as an all-or none phenomenon? It seems functionally more realistic to imagine that a local synaptic excitation (predominately in the dendrites) would activate PICs only in its immediate environment - and lasting only as long as the local active synaptic excitation.

Recent results (Hultborn et al 2003) strongly support this view of a gradual activation. A constant synaptic excitation causes gradually increasing additional discharge frequencies along the course of the “primary range”. This was interpreted as follows: with stronger depolarizing currents through the recording microelectrode (and higher firing frequencies) an increasing number of PIC channels are recruited around the excitatory synapses - but still far below the threshold for recruiting the plateau current in an all-or-none manner. In line with this reasoning it was suggested synaptic inhibition would prove to be more effective in reducing a discharge evoked by synaptic excitation as compared to injected current. In order to test this prediction, the efficiency of recurrent inhibition in reducing firing was tested for comparable firing frequencies evoked either by a current pulse or by synaptic excitation. The prediction was supported by the results as illustrated in Fig 1.

The results summarized above would be explained if motoneuronal discharge by synaptic excitation - but not by current injection in the soma - is always supported by dendritic PICs. We conclude that dendritic PICs contribute dynamically to the transformation of synaptic input into a motoneuronal frequency code.

Figure 1. Recurrent inhibition is more effective in reducing firing evoked by synaptic excitation than by current injection Recordings from a semitendinousus motoneurone. A and B, repetitive firing elicited by a rectangular current pulse alone (A) and together with recurrent inhibition (B). C and D, repetitive firing evoked by a train of impulses to the corticospinal tract (CST) alone (C) and together with recurrent inhibition (D). Each alternative was repeated 20 times and the means and S.D. are shown as a bar graph in F. E, the recurrent inhibition (90 stimuli at 100 Hz, 5 xT) and the EPSP evoked by the CST stimulation (14 stimuli at 200 Hz) used in B-D, but here recorded at a membrane potential close to firing threshold. (From Hultborn et al 2003).

Bennett DJ, Hultborn H, Fedirchuk B, Gorassini M . Synaptic activation of plateaus in hindlimb motoneurons of decerebrate cats. J Neurophysiol 80, 2023-2037, 1998.

Hultborn H, Brownstone RB, Toth TI, Gossard J-P. Key mechanisms for setting the input-output gain across the motoneuron pool. In Brain Mechanisms for the Integration of Posture and Movement, ed. Mori S, Stuart DG & Wiesendanger M Progr Brain Res 143, 77-942, 2003.

Hultborn H, Enrķquez Denton M, Wienecke J, Nielsen JB. Variable amplification of synaptic input to cat spinal motoneurones by dendritic persistent inward current. J Physiol 552, 945-95, 2003.

Powers RK, Binder MD. Input-output functions of mammalian motoneurons. Rev Physiol Biochem Pharmacol 143, 137-263, 2001.

Rose PK, Cushing S. Non-linear summation of synaptic currents on spinal motoneurons: lessons from simulations of the behaviour of anatomically realistic models. Prog Brain Res 123, 99-107, 1999.


Mechanisms and control of bistability vs. amplification

Robert Lee

Emory University, Atlanta, USA

Currently, it is widely held that synaptic amplification and bistability result from a single mechanism, namely dendritic plateau potentials based on L-type calcium conductances. However, this raises several issues: 1) Plateaus would need to be smoothly gradable to exhibit proper synaptic amplification; 2) both amplification and bistability have been observed simultaneously; and 3) plateaus are inherently slow to activate. Of these three issues, one has yet to be tested experimentally, namely the slow kinetics of dendritically located L-type calcium-based plateau potentials.

Theoretically, a slowly activating/deactivating plateau potential would be capable of amplifying only "slow" inputs. We examined this issue experimentally by measuring the fast amplification of the IA synaptic input by vibrating the Achilles tendon at 180 Hz. Assessment of fast amplification was made by examining the amplitude of the 180 Hz component of the measured synaptic current arriving at the soma and its voltage dependence during somatic voltage clamp. We found that the amplification at 180 Hz was indeed present and similar in voltage profile as the synaptic amplification previously described. We conclude, therefore, that synaptic amplification is too fast to be due to gradable plateau potentials and therefore must be due to a separate mechanism. Furthermore, we propose a new mechanism for synaptic amplification based largely on dendritic persistent sodium currents.

Preliminary modeling work indicates that a dendritic fast amplification mechanism can be constructed from a careful balance of persistent sodium and delayed rectifier potassium conductances that effectively reduces the electrotonic length of the dendrites over a modest voltage range. This reduction of electrotonic length reduces the passive attenuation resulting in an apparent amplification of synaptic input as the dendritic voltage rises into the proper range. Experimental studies are underway to verify this hypothesized mechanism for synaptic amplification.



Lowering the spike threshold of spinal neurons: neural activation is facilitated during motor output

Brent Fedirchuk and Jonathan Gilmore

University of Manitoba, Winnipeg, Canada

Previous work from our group has shown there is an increase in motoneuron excitability produced by hyperpolarization (i.e. "lowering") of the threshold potential at which an action potential is elicited (Vth) at the onset, and throughout brainstem-induced fictive locomotion in the decerebrate cat (Krawitz et al. 2001). This represents a transient facilitation of motoneuronal activation dependent on the presence of fictive locomotion.

In recent work we have shown that an analogous hyperpolarization of Vth is demonstrable in ventral horn neurons of the neonatal rat. Bath application of the monoamines serotonin (5-HT) or noradrenaline (NA) to the spinal cords isolated from postnatal day 1 to 5 neonatal rats caused a hyperpolarization of Vth in the absence of locomotor activity (see Fedirchuk and Dai, 2004). This lowering of Vth was assessed using a voltage clamp protocol where the threshold holding potential able to elicit a fast inward current was assessed prior to and during the application of monoamine. The effect on Vth was reversed by washing out the monoamine, and repeatable with subsequent administration of 5-HT. These results demonstrate that Vth hyperpolarization occurs in rats, can be elicited in early neonates, and can be induced by monoamines.

Hyperpolarization of Vth can also be induced by endogenous systems in the neonatal rat (Gilmore and Fedirchuk, 2004). In a brainstem/spinal cord preparation isolated from 1 to 5 day old neonatal rats, electrical stimulation of the ventromedial medulla can elicit locomotor-like patterns of ventral root activity (indicated by the bar in Figure 1A) that is accompanied by hyperpolarization of Vth for spinal motoneurons and unidentified ventral horn neurons (Figure 1B, summary of 11 cells in Figure 1C). This facilitation of neuronal firing is of a similar amplitude (-2 to -18 mV, see Figure 1C) and therefore appears analogous to the Vth hyperpolarization seen during fictive locomotion in the decerebrate cat. However, Vth hyperpolarization was also seen during electrical brainstem stimulation that evoked alternating, rhythmic, or tonic VR activity; or failed to evoke VR activity. Comparable hyperpolarization of Vth was seen in antidromically identified lumbar motoneurons and unidentified lumbar ventral horn neurons during electrical stimulation of the brainstem. The hyperpolarization of Vth and VR activity induced by brainstem stimulation was reversibly blocked by cooling of the cervical cord, indicating it is mediated by descending fibers. Application of the serotonergic antagonist ketanserin to the spinal cord was effectively able to block the brainstem-evoked hyperpolarization of Vth.

These results demonstrate that the descending serotonergic system is able to facilitate recruitment of spinal motoneurons by inducing a hyperpolarization of Vth. Whether modulation of spike threshold is due to monoaminergic facilitation of a persistent inward current, direct modulation of voltage dependent sodium channels or is accomplished by different modulatory effects in different cell types remains to be determined.

Supported by the Canadian Institutes of Health Research and the Canadian Neurotrauma Research Program

Fedirchuk B, Dai Y. Monoamines increase the excitability of spinal neurones in the neonatal rat by hyperpolarizing the threshold for action potential production. Journal of Physiology 557: 355-361, 2004.

Gilmore J, Fedirchuk B. The excitability of lumbar motoneurones in the neonatal rat is increased by a hyperpolarization of their voltage threshold for activation by descending serotonergic fibers. Journal of Physiology 558: 213-224, 2004.

Krawitz S, Fedirchuk B, Dai Y, Jordan LM, McCrea DA. State-dependent hyperpolarization of voltage threshold enhances motoneurone excitability during fictive locomotion in the cat. Journal of Physiology 522: 271-281, 2001.


Contributions of the input signal and prior activation history on the discharge behavior of motoneurons

Randall K. Powers, Yue Dai, Bradley Dell, Donald Percival and Marc D. Binder

University of Washington, Seattle, USA

The probability of spike occurrence in neurons is determined by the time course and magnitude of the total current reaching the spike initiation zone. The features of this current that are most effective in evoking spikes can be determined by injecting a Gaussian current waveform into a neuron and using spike-triggered reverse correlation to calculate the average current trajectory (ACT) preceding spikes. The time course of this ACT (and the related first-order Wiener kernel) provides a general description of a neuron's response to dynamic stimuli (Bryant & Segundo, 1976; Polikov, Powers & Binder, 1997). In many different neurons, the ACT is characterized by a shallow hyperpolarizing trough followed by a more rapid depolarizing peak immediately preceding the spike. The hyperpolarizing phase is thought to reflect an enhancement of excitability by partial removal of sodium inactivation. However, an alternative explanation is that during repetitive discharge, interspike intervals that are longer than average result from a period of membrane hyperpolarization to delay spike occurrence, and that the ACT computed over all spikes reflects this requirement. We have used a combination of experimental and simulation techniques to provide evidence supporting this alternative explanation.

We made sharp electrode recordings from rat hypoglossal motoneurons in brainstem slices. Repetitive discharge was elicited by a combination of a current step and noise, and we calculated ACTs by averaging the noise preceding spikes. We found that the hyperpolarizing trough is larger in ACTs calculated from spikes preceded by long interspike intervals and minimal or absent in those based on short interspike intervals. This difference may reflect the influence of the post-spike afterhyperpolarization (AHP) on firing probability. During the initial part of the AHP, the membrane potential is far from threshold and spikes can only be evoked by large positive noise excursions with no need for preceding hyperpolarization. Toward the end of the AHP the membrane is close to threshold and long intervals can only occur if a hyperpolarization precedes a positive noise excursion. Simulations with threshold-crossing models support this interpretation. The left panel of the figure shows ACTs calculated from noise-driven discharge of two threshold-crossing models, one with an AHP (black) and one without an AHP (red). The trough was only present for ACTs calculated from the discharge of the model with the AHP. We show that it is possible to represent noise-driven discharge as an autoregressive-moving average (ARMA) process (Box et al. 1994). The ARMA model predicts discharge based on the sum of a feedback kernel (AR, autoregressive) and a stimulus kernel (MA, moving average), as shown in right panel of the figure. The feedback kernel reflects the influence of the AHP, and it increases in amplitude at increasing firing rates (black: 12 imp/s, blue: 17 imp/s, red: 20 imp/s), and when AHP amplitude is increased by pharmacological manipulations. Finally, we found that the predictions of the ARMA model are virtually identical to those based on the first-order Wiener kernel. This suggests that the Wiener kernels derived from standard white-noise analysis of noise-driven discharge in neurons actually reflect the effects of both stimulus and discharge history.

Supported in part by grants from the NSF (IBN-9986167) and NIH (NS26840)

Bryant HL, Segundo JP. Spike initiation by transmembrane current: a white-noise analysis. J Physiol 260: 279-314, 1976.

Box GEP, Jenkins GM, Reinsel GC. Time Series Analysis: Forecasting and Control. 3rd edn. Prentice Hall, New York, 1994.

Poliakov AV, Powers RK, Binder MD. Functional identification of the input-output transforms of motoneurones in the rat and cat. J Physiol 504: 401-424, 1997.


A system intrinsic to the spinal cord which regulates motoneurone excitability?

Rob Brownstone

Dalhousie University, Halifax, Canada

(Conradi and Skoglund, 1969) first described the ultrastructure of various types of synaptic terminals on motoneurones, including large C-type terminals on the somata. These large and abundant C-terminals were later shown to be cholinergic (Nagy et al. 1993; Li et al. 1995; Arvidsson et al. 1997). More recently C-terminals have been shown to be associated with postsynaptic m2 receptors (Hellstrom et al. 2003; Wilson et al. 2004) and KV2.1 delayed rectiŽer channels (Muennich and Fyffe, 2004; Wilson et al. 2004). It remains unclear, however, where these C-terminals come from and what their physiological role is. We therefore examined the origin of C-terminals and characterised muscarinic effects on spinal MNs.

Molecular biological and immunohistochemical analysis demonstrated that C-terminals likely originate from large ChAT+/Dbx1+/nNOS- spinal interneurons found near the central canal (unpublished data). We therefore filled large cells in this region with neurobiotin during patch-clamp recordings to investigate their axonal projections. Following fixation, confocal microscopy demonstrated that large ChAT+ cells near the central canal extended axons into MN pools.

Next, to examine the effects of cholinergic receptor activation on MNs, muscarine (50 µM) was bath applied to lumbar spinal cord slices (250 µm) from P8 - P14 mice during whole-cell patch-clamp recordings of spinal MNs. Muscarine applications evoked inward currents (0.67 ± 0.21 pA/pF; n = 3) in voltage clamp and depolarisations (7.2 ± 2.6 mV; n = 5) in current clamp mode. Using brief current injections (10 ms) to elicit single action potentials, muscarine was also shown to reduce the amplitude of action potential afterhyperpolarisations (AHPs, n = 5). This inhibition likely reflects direct actions on Ca2+-dependent K+ channels since muscarine had no effect on the magnitude of voltage-activated Ca2+ currents (n = 3).

Figure. Longer current injections (1 s) revealed that muscarine also increased the frequency of repetitive firing, and increased the action potential threshold. The increased excitability under muscarine was quantified as a 33 ± 13 % (n = 10) increase in the slope of frequency-current plots, and was occluded by application of apamin. Similar effects were seen with application of the m2 agonist, oxotremorine.



Role of monoamines in controlling motoneuron PICs in chronic spinal rats

David Bennett

University of Alberta, Edmonton, Canada

Following long-term spinal cord injury motoneuron properties change and ultimately lead to spasticity. In particular, large persistent sodium and calcium currents (Na PIC and Ca PIC) emerge that are not prominent immediately after injury. I discusssed how these large PICs are regulated by monoamines, and how, in particular, residual monoamines in the spinal cord still regulated these PICs after injury, even though levels of 5HT and NE reduce to 2-10% of normal. PICs are still affected by these small amounts of monoamines because the motoneurons and PICs develop a supersensitive response to these monoamines (responding at 10 to 100 times lower doses to monoamines). Furthermore, combined application of antagonists to the three specific monoamine receptors 5HT2a, 5HT2c, and NE alpha 1 completely eliminate spontaneously occurring Na PICs, while leaving the Ca PICs unchanged. So, these three Gq protein-coupled receptors regulate the Na PIC, but not the Ca PIC. With the loss of the Na PIC in the antagonists, motoneurons can only fire phasically, and are difficult to recruit during slow current ramp injections. We do not yet know what specific receptors regulate the Ca PIC, although this Ca PIC is regulated by 5HT and NE. These results ultimately suggest that a cocktail of specific monoamine antagonists may prove to be effective antispastic agents.


Changes in chronic activation patterns evoke learning and unlearning responses in rat motoneurones

Phillip Gardiner, Eric Beaumont, Bruno Cormery, and Kristine Csukly

University of Manitoba, Winnipeg, Canada

Previous research from this laboratory has provided convincing evidence that the biophysical properties of alpha-motoneurones innervating ankle extensors, measured in situ in ketamine/xylazine anesthetized rats, are altered when the normal activity patterns of the animal are changed. For example, increased daily locomotor activity resulting from unlimited access of rats to voluntary exercise wheels (Beaumont & Gardiner, 2002), or from 2 hours per day of forced treadmill training (Beaumont & Gardiner, 2003), results in hyperpolarization of the resting membrane potential and the spike threshold (by approx. 5 mV in both cases), an increased rate of rise of the orthodromic and antidromic spikes, and larger-amplitude afterhyperpolarizations (AHPs). These changes occur without any systematic changes in rheobase, input resistance, cell capacitance, or AHP time-course, or in the proportions of "fast" (AHP _ decay time < 20 ms) and "slow" (AHP > 20 ms) motoneurones in the tibial nerve. These changes are restricted to the motoneurones which are "overloaded" (ie, only in low-threshold motoneurones following increased voluntary exercise, where intensity of exercise is relatively mild), and occur in the opposite direction 4 weeks after T10-11 spinal cord transection (Beaumont et al., 2004). Furthermore, daily "passive" exercise of the hindlimbs of rats following spinal cord transection, using a motorized ergonometer, which generates alternating locomotor-like movements of the hindlimbs, prevents the transection-associated changes in these properties, thus demonstrating that peripheral afferent activation of motoneurones is an important modulator of these properties. Many of these changes are similar to those seen during learning in lower animals, such as Aplysia (Cleary et al., 1998). Modeling of these changes using a 5-compartment motoneurone model (Dai et al., 2002) suggests that such changes might involve increased sodium, leak, and AHP conductances.

We have very recently turned our attention to a model of decreased usage, hindlimb unweighting (HU). In this model, which is a preferred and much-used ground-based model of the weightlessness experienced in space, each rat has its hindquarters suspended above the floor of the cage by attachment of the tail to a puller system which allows otherwise free movement around the cage using the forelimbs. We have found evidence that many of the changes in biophysical properties following increased usage are opposite after 2 weeks of HU (decreased AHP amplitudes, depolarized spike thresholds). We have extended our analysis to include motoneurone responses to 500-ms periods of square-wave depolarizing current injections through the penetrating microelectrode, in order to determine the effects of HU on motoneurone frequency/current (f/I) relationships. Our results demonstrate that the f/I relationships for both "fast" and "slow" motoneurones are shifted to the right after HU (ie, less excitable, see Figure 1). In addition, f/I relationships for slow motoneurones, in which the minimum rhythmic firing frequencies and the currents required to evoke these are normally significantly lower than in fast motoneurones, become similar to fast motoneurones after HU. This result is consistent with previous findings that slow muscle fibers show evidence of changing towards fast muscle fiber phenotypes during reduced weight-bearing. Model simulations suggest possible decreases in sodium and L-type calcium conductances, as well as reduced persistent inward currents, as bases for these f/I shifts.

We are continuing to examine chronic activity-related changes in motoneurone properties, using the decerebrate preparation to avoid the influence of anesthetics on active motoneurone properties. Our immediate future goals will be to further characterize these changes, attempt to determine the phenotypic changes which underlie them (by measuring the concentrations and gene expression of various ion channels), and to outline the metabolic signaling mechanisms that precede these changes (involvement of calcium? CREB? MAPK? neurotrophins?).

Figure legend. Hind limb unweighting (HU) for 2 weeks shifted frequency/current relationships to the right, with no significant influence on the slope. In addition the relationship for slow motoneurones also shifted "up" following HU, such that minimum rhythmic firing frequencies and the currents required to evoke them resembled those of fast motoneurones.

Beaumont E, and Gardiner P. Effects of daily spontaneous running on the electrophysiological properties of hind limb motoneurones in rats. J Physiol 540: 129-138, 2002.

Beaumont E, Gardiner P. Endurance training alters the biophysical properties of hind limb motoneurones in rats. Muscle & Nerve 27: 228-236, 2003.

Beaumont E, Houle J, Peterson C, Gardiner P. Passive exercise and fetal spinal cord transplant both help to restore motoneuronal properties after spinal cord transaction in rats. Muscle & Nerve 29: 234-242, 2004.

Cleary L, Lee W, Byrne J. Cellular correlates of long-term sensitization in Aplysia. J Neurosci 18: 5988-5998, 1998.

Dai Y, Jones K, Ferirchuk B, McCrea D, Jordan L. A modeling study of locomotion-induced hyperpolarization of voltage threshold in cat lumbar motoneurones. J Physiol 544: 521-536, 2002.

This research was supported by grants from NSERC, CIHR, and CSA. The assistance of Tanguy Marqueste and Duane Button is gratefully acknowledged.


Nerve injury and central plasticity

Tim Cope, Edyta Bichler, Stan Nakanishi, and Marty Pinter

Emory University, Atlanta, USA

Here we report our attempts to identify mechanisms responsible for inducing neural plasticity (sustained transformation in neural excitability) in response to injury. Cutting or crushing a muscle nerve results in profound changes in properties of the motoneurons whose axons are severed. For example, the intrinsic excitability of motoneurons with severed axons decreases nearly 2-fold (Foehring et al. 1986) and the short-latency excitatory potentials (ESPSs) produced in these motoneurons by primary spindle afferents (Miyata and Yasuda, 1988; Seburn and Cope, 1998) increases by 2-fold within days after peripheral nerve crush. These changes might be induced by any of several effects of peripheral nerve injury, including cessation of neuromuscular transmission and interruption of axonal transport between muscle and central nervous system. Here we report findings of studies designed to test the role of these injury effects in signaling changes in motoneuron excitability and synaptic function. Data were taken from intracellular records of spinal motoneurons obtained from barbiturate-anesthetized Wistar rats in vivo. Chronic treatment effects described below were identified by comparing rats in the treatment groups with untreated controls. Findings are interpreted in terms of negative versus positive injury signals (review, see Perlson et al. 2004), which induce neural changes, respectively, either by reducing regulatory signals that are normally present or by introducing novel signals that are not normally present.

Reduction in rheobase current (Irh)

Axotomy of motoneurons interrupts transmission at the neuromuscular junction (NMJ), resulting in the elimination of acetylcholine (ACh) binding to its postsynaptic receptors on muscle. We reasoned that if ACh binding regulates the intrinsic excitability of motoneurons, then blocking binding should decrease Irh. To this end, the medial gastrocnemius (MG) muscle in rats was injected with the non-competitive ACh antagonist alpha-bungarotoxin (BTX). Rapid turnover of ACh receptors required repetition of the BTX injections on each of days 1 through 4. Saline injections were given the sham group. Fig. 1 shows that 5 days after BTX treatment began, Irh decreased dramatically and in a way similar to that produced by axotomy.. From these data we speculate that the reduction in Irh caused by axotomy is induced by a negative injury signal, and that maintenance of the intrinsic excitability of motoneurons requires ACh binding to its postsynaptic receptor and/or events occurring downstream to binding. See Pinter et al. (1991) for additional discussion.

Figure 1 Blocking acetylcholine receptor binding alone reduces motoneuron rheobase current. Cumulative distributions of motoneuron rheobase current pooled from untreated rats and from rats with MG nerve crush or MG muscle repeatedly injected either with alpha-bungarotoxin.

Enlargement of short latency EPSPs

Axotomy of motoneurons and of the sensory neurons that traverse a crushed muscle nerve, interrupts transport of molecules between the periphery and the CNS. If axotomy has its effect on central synaptic function by blocking axon transport, then EPSPs should be increased by blocking transport alone. In order to block transport, the microtubule poison colchicine was applied to the tibial nerve for 20 minutes 3 days prior to measuring EPSPs produced in tibial motoneurons by electrical stimulation of the tibial nerve proximal to the treatment site. Block was verified by the accumulation of the vesicular protein SV2 in the tibial nerve using Western blot analysis. Blocking transport in the intact nerve did not enlarge EPSP amplitude, demonstrating that molecular signaling between CNS and periphery does not regulate central synaptic function, at least in the short term. However, blocking transport at a site proximal to nerve crush did prevent the enlargement of EPSPs observed with crush alone (Fig. 2). We interpret the latter finding to mean that EPSP enlargement is induced by a positive injury signal introduced by the injury either at or distal to the injury site. Moreover, this signal is carried by axon transport, and not by electrical activity, e.g. injury discharge, which is unaffected by colchicine. Ongoing experiments are aimed at identifying the signal, which we currently speculate to be associated with the inflammatory response induced by nerve injury.

Figure 2 Blocking axon transport prevents injury-induced EPSP increase. Cumulative distributions of short-latency EPSPs pooled from untreated rats and from rats with tibial nerve either crushed alone or treated with colchicine proximal to the crush site.

Foehring RC, Sypert GW, Munson JB. Properties of self-reinnervated motor units of medial gastrocnemius of cat. II. Axotomized motoneurons and time course of recovery. J Neurophysiol 55: 947-965, 1986.

Perlson E, Hanz S, Medzihradszky KF, Burlingame AL, Fainzilber M. From snails to sciatic nerve: retrograde injury signaling from axon to soma in lesioned neurons. J Neurobiol 58: 287-294, 2004.

Pinter MJ, Vanden Noven S, Wallace N. Axotomy-like changes in cat motoneuron electrical properties elicited by botulinum toxin depend on the complete elimination of neuromuscular transmission. J Neurosci 11: 657-666, 1991.

Seburn KL, Cope TC. Short-term afferent axotomy increases both strength and depression at Ia-motoneuron synapses in rat. J Neurosci 18: 1142-1147, 1998.

Supported by NIH grants NS R01 38693 and NS P01 40405.


Introduction to Motor Units (Roger Enoka)

The introduction described the techniques used to record and discriminate the discharge of motor units in humans, the properties that are used to characterize motor units, the discharge patterns exhibited by motor units, and some issues related to populations of motor units.

Measurement Techniques

The most common technique used to record motor unit potentials is an intramuscular electrode, which can comprise either a set of fine wires or a needle. Fine wire electrodes consist of 2-4 wires (10-100 µm in diameter) that are insulated (nylon or Formvar) and inserted into the wire with a hypodermic needle (25 or 27 gauge). Typically, the cut ends of the wires are used as the recording surface. When 4 wires are inserted, the investigator has 3 bipolar recordings available. Needle electrodes are available in a variety of configurations and generally consist of a recording medium inside the cannula with different numbers and sizes of recording surfaces. One configuration, known as the Macro EMG, comprises a small surface to detect nearby motor unit potentials and trigger a distant larger surface. The amplitude of the Macro EMG provides a reasonable index of motor unit size.

In addition to intramuscular electrodes, motor unit potentials are also recorded with subcutaneous electrodes, intraneural electrodes, and surface electrode arrays. Subcutaneous electrodes, which were developed by Gydikov and colleagues, are inserted between the muscle and the subcutaneous tissue. The insulation is removed from the middle of the two wires and placed directly above the active muscle. Subcutaneous electrodes are stable and can provide recordings during movement and fatiguing contractions. Intraneural electrodes were developed for microneurography by Vallbo and colleagues and are inserted into peripheral nerves either to stimulate single axons or to record action potentials propagating along motor and sensory axons. Electrode arrays are placed on the skin above the muscle and can provide a non-invasive measurement the conduction of motor unit potentials along the muscle fibers.

The technical limitation in this field is the ability to discriminate the potentials of single motor units in multi-unit recordings. The hardware available for this process has improved considerably in recent years. Two popular options are Spike2 by Cambridge Electronics Design and the Precision Decomposition technique. Although the applications of these two systems has merged, the original applications were to use Spike2 and other such systems to discriminate recordings that contained only a few motor unit potentials, whereas the decomposition algorithms specialize in decomposing multi-unit interference records into constitute trains of motor unit potentials. Each approach has its strengths and weaknesses.

Motor Unit Characteristics

Motor units are characterized by the measurement of recruitment threshold, discharge rate, and twitch properties. Recruitment threshold is most often expressed as the force at which the motor unit begins to discharge action potentials repetitively during a gradual increase in muscle force during an isometric contraction. An alternative measure is the minimal discharge rate of the unit, but this value is probably confounded by the active properties of the dendrites. For most muscles, the upper limit of motor unit recruitment is about 90% of the maximal voluntary contraction (MVC) force. For some hand muscles, however, the upper limit of recruitment is about 50% of MVC force.

There is a substantial range of minimal and maximal discharge rates for motor units in a single muscle. For example, minimal discharge rates for motor units in tibialis anterior range from 5 to 15 Hz and maximal discharge rates range from 15 to 65 Hz. The modal values were 9 and 35 Hz, respectively. The functional significance of these rates depends on the force-frequency relation. For the toe extensor muscles, for example, the steep region of the curve extended from 5 to 16 Hz. Despite frequent reports that maximal discharge rates are greater for low-threshold motor units, the measurement of peak discharge rates appears to depend on the protocol used to determine these values.

As first applied to motor units by Stein and colleagues, spike-triggered averaging is the most common approach used to evaluate the contractile properties of the motor units. As demonstrated by several groups, however, spike-triggered averaging provides a biased assessment of the contractile properties. Due to the partial fusion of twitches at low discharge rates, spike-triggered averaging cannot measure the contraction time of low-threshold motor units. Conversely, the presence of motor unit synchronization enhances the apparent twitch force, especially for high-threshold motor units. Nonetheless, spike-triggered averages performed on several hundred motor units provide significant information about motor units, including the continuous distributions of twitch force and time to peak force. These measurements demonstrate that human motor units cannot be categorized into slow and fast units, as is common for motor units in experimental animals and muscle fibers in humans. Rather, investigators who study motor units in humans prefer to distinguish between low- and high-threshold motor units.

Discharge Patterns

Over the range from resting to the upper limit of motor unit recruitment, the gradation of muscle force is accomplished by concurrent variation in recruitment and discharge rate. Above the upper limit of recruitment, further increases in muscle force are accomplished solely by increases in discharge rate. The discharge of action potentials by motor units exhibits a significant amount of variability, with a nominal value for the coefficient of variation of about 20%. The coefficient of variation decreases as mean discharge rate increases and it usually increases during fatiguing contractions. Motor units sometimes discharge a few action potentials at a high rate, such as at the onset of a contraction and during rapid contractions. The brief times between these action potentials has a marked effect on muscle force and is amenable to adaptation with a program of physical activity.

The trains of action potentials discharged by motor units are often correlated. The presence of correlated discharge times by pairs of motor units, which produces a peak in the cross-correlation histogram, is known as motor unit synchronization and is typically interpreted as an index of the relative proportion of common inputs received by the motor neurons. The shape of the cross-correlation histogram determines the location of the peak in the coherence spectrum for the discharge times of the two motor units. Motor unit synchronization usually produces a peak in the 16-32 Hz band of the coherence spectrum. An alternative measure of correlated activity, known as common drive, is the low-frequency modulation (1-2 Hz) of discharge times by motor units. The magnitude of the peak in the spectrum, however, varies linearly with discharge rate variability. Both measurements (motor unit synchronization and common drive) assume a certain degree of fidelity between modulation of synaptic input and motor neuron output; that is, the motor neuron does not alter the input-output modulation. Active dendritic properties, however, probably exert a significant influence on this relation.

Population Issues

Although there is a reasonable amount of information on average motor unit properties in the literature, our understanding of motor unit behavior is limited by the near-absence of data on the range of properties within populations and across muscles. Even the number of motor units in different muscles is uncertain. For example, electrophysiological techniques suggest that biceps brachii comprises about 100 motor units, whereas a labeling study indicated that the number is more likely 1000. Similarly, the range of innervation numbers from the smallest to the largest motor unit in a population is largely unknown and is often inferred from the range of tetanic forces. The absence of these data impairs our ability to develop conceptual schemes about motor unit activity. Because only a few motor units can be recorded in each experiment, the experimental work must be performed in conjunction with computational studies. The lack of data on population characteristics hinders the development of these models.

Furthermore, the translation of information about motor unit activity to human performance is hindered by the involvement of multiple muscles in each action. For this reason, many studies on human motor units are performed on relatively simple motor systems (hand and foot muscles) with the assumption that the observed general principles can be applied on a broader scale.



Effect of experimental muscle pain on motor unit control properties

Dario Farina, Lars Arendt-Nielsen, Roberto Merletti, and Thomas Graven-Nielsen

Politecnico di Torino, Italy and Aalboorg University, Aalborg, Denmark

The decline of motor unit firing during fatigue has partly been explained by a progressive activity in thin caliber afferents during a fatiguing contraction [1]. If this explanation is valid, then a progressive motor unit firing inhibition in non-fatiguing contractions should be present when exciting group III and IV afferents by experimental muscle pain. In this short report we will describe the results obtained from two human studies aimed at verifying the hypothesis that experimental muscle pain decreases motoneuron firings.

In both experiments, single motor unit activities were analyzed with concomitant recording of intramuscular and surface electromyographic (EMG) signals. Intramuscular signals were collected from a location about 15 mm proximal with respect to the most distal innervation zone of the muscle. The intramuscular signals were used to detect single motor unit action potentials. Surface EMG signals were recorded with a linear array of 16 equi-spaced electrodes placed between the most distal innervation zone and the distal tendon region. The surface EMG signals were averaged with the intramuscularly detected potentials as triggers. From the averaged multi-channel surface EMG signals, single motor unit conduction velocity was estimated. Thus, both motor unit control and membrane properties were assessed. The method has been validated in previous work [2]. In both experiments the muscle investigated was the tibialis anterior.

In the first experiment, in 12 healthy subjects nociceptive afferents were stimulated by three intramuscular injections of hypertonic saline (0.2 ml, 0.5 ml, and 0.9 ml) separated by 140 s. These injections caused a gradual increase in the level of pain. The subject performed six isometric contractions (20-s long) at 10% of the maximal voluntary contraction during the experimental muscle pain. The pain intensity was scored on a visual analog scale. The first contraction started immediately after the infusion of the first bolus of hypertonic saline, the third after the second infusion, and the fifth after the third infusion.

The same set of six contractions was performed without any infusion before the painful condition on the right leg. Moreover, the procedure was repeated for the left leg with 12 contractions, the last six with infusion of isotonic (non-painful) saline. Single motor unit firing rate and conduction velocity were assessed in the different conditions with the method described above.

The firing rate of the active motor units did not significantly change in the three control conditions (without infusion for the right and left leg, and with infusion of isotonic saline in the left leg). There was, on the contrary, a significant decrease (on average, mean ± SE, 1.03 ± 0.21 pps) of the firing rates in the painful condition (Figure 1). Moreover, MU firing rates were inversely correlated (linear regression analysis, R = -0.45) with the subjective scores of pain. Single motor unit conduction velocity was the same in the four conditions (3.88 ± 0.03 m/s), i.e., the injection of hypertonic saline did not alter the muscle fiber membrane properties of the motor units under study.

Figure. Mean (±SE) of the firing rate estimated from the detected motor units in the 12 subjects of the first experiment for the right and left leg (N = 55 and N = 49 motor units, respectively). For each leg, the values obtained in the 12 contractions are shown

In the second experiment, 10 subjects performed four isometric contractions of 4 min each at 25% MVC with both legs. For each leg, the first contraction was without any infusion, while the second corresponded to the infusion of 0.5 ml of hypertonic (right leg) or isotonic (left leg) saline. The initial motor unit firing rate was not different in the three control contractions (without infusion for the right and left leg, and with infusion of isotonic saline in the left leg) (11.2 ± 0.1 pps) while it was slightly but significantly lower in the painful contraction (10.7 ± 0.2 pps). The firing rate significantly decreased over time in all conditions. Its rate of change (second order polynomial interpolation) was significantly lower (in absolute value) in the painful contraction with respect to the control contractions ((-4.1 ± 1.7)×10-3 pps/s vs (-9.2 ± 0.7)×10-3 pps/s). As for the first experiment, initial single motor unit conduction velocity was not different in the four conditions (4.3 ± 0.1 m/s). Moreover, conduction velocity significantly changed over time with the sustained contraction but its rate of change was not different in the four conditions ((-3.9 ± 1.0)×10-3 m/s2). The results of both experiments indicated that nociceptive stimulation induces a decrease in motor unit firing rates which is modulated by the intensity of noxious stimuli. This decrease is not likely a consequence of changes in membrane fiber properties since conduction velocity (and its rate of change over time with sustained activation) of the investigated motor units did not change with pain. On the contrary, the detected decrease may reflect a reflex inhibition of motor unit firings, as a consequence of the activity of group III and IV nociceptive muscle afferents.

In conclusion, these studies indicate a reflex inhibitory effect of pain with an efficacy correlated to the amount of nociceptive activity. Moreover, they constitute further proof that thin afferents may indeed have a role in motor unit firing inhibition caused by fatigue.

[1] Woods JJ, Furbush F, Bigland-Ritchie B. Evidence for a fatigue-induced reflex inhibition of motoneuron firing rates. J Neurophysiol 58: 125-37, 1987.

[2] Farina D, Arendt-Nielsen L, Merletti R, Graven-Nielsen T. Assessment of single motor unit conduction velocity during sustained contractions of the tibialis anterior muscle with advanced spike triggered averaging. J Neurosci Methods 115(1): 1-12, 2002.


Cortically induced inhibitory effects on motoneurone firing time

A. Schmied, S. Attarian, B. Mattei, and J. P. Vedel

CNRS, Marseilles, France.

As soon as technical progress has made it possible to activate non invasively cortical areas first by transcranial electric stimulation and later on by transcranial magnetic stimulation, striking evidence for inhibitory effects has been reported: Depending on its location, transcranial stimulation may induce behavioural impairments powerful enough to suggest that transcranial magnetic stimulation might be used as a tool to induce a virtual lesion (Day et al., 1989; Priori et al., 1993; Amassian et al, 1989; Pascual-Leone et al., 1991).

In conscious humans, stimulating the motor cortex at an intensity subthreshold for excitatory effects induces a rather short-lasting depression of EMG activity (Davey et al. 1994; Petersen et al., 2001). When supra-threshold intensity is used, a strong and long-lasting depression usually follows the excitatory response. The late part of this depression is considered to have a cortical origin (Chen et al. 1999).

Single motor unit studies have confirmed the presence of inhibitory-like components in the responses induced by stimulating the motor cortex. The first evidence has been provided in a pioneering study performed with transcranial electric stimulation (Calancie et al. 1987). At a sub-threshold intensity, stimulation of the motor cortex was found to induce a lengthening of the interspike interval in which the stimulation was delivered with an effectiveness that was enhanced together with the excitatory effects, which appeared as the stimulation intensity was increased.

Most of the many studies performed with transcranial magnetic stimulation, have used supra-threshold intensities together with peristimulus time histogram (PSTH) analysis. In these conditions, pure inhibitory responses were never observed and the silent period that followed the excitatory responses could not be dissociated from the various inhibitory effects that might have resulted of the synchronous activation of the motoneurones (after-hyper-polarization summation, spinal or cortical recurrent inhibition, re-afferent loop..). The presence of pure inhibitory-like responses in PSTHs has been confirmed, however, in the absence of early excitation in studies using near-threshold intensities (Palmer &Ashby, 1992, Brouwer & Qiao, 1995; Boniface et al. 1994). These inhibitory responses were observed with quite variable latencies, ranging from a few ms up to a few tens of ms later than the early excitatory responses and heterogeneous durations spreading from few tens of ms up to 1s. Two studies using interspike intervals (ISIs) have reported that, as previously observed with electric stimulation, transcranial magnetic stimulation was able to delay the next firing time in trials intermingled with those in which a spike had been triggered by the stimulus (Classen & Benecke, 1995; Garland & Miles, 1997).

Cortically induced inhibitory effects have been first attributed to the postsynaptic action of inhibitory neurons, which might be located at various subcortical levels. More recently, strong arguments have favoured the pre-synaptic action of inhibitory networks that might remove part of the excitatory drive of the motoneurones. Cortical inhibitory interneurones are thought to play a major role in this disfacilitatory process (Kujirai et al. 1993).

The hypothesis has been proposed that a dysfunction of inhibitory networks may promote hyper-excitability processes leading ultimately to cell death in neuro-degenerative disease,s such as amyotrophic lateral sclerosis, which affects both the spinal motoneurones and the cortico-motoneurone cells that innervate them (Eisen & Swash 2001). The evidence for a deficit in cortical inhibition is based on data obtained at rest with EMG recording using paired pulse transcranial magnetic stimulation (Ziemann et al. 1997).). The data obtained with this method are not fully consistent (see Salerno and Georgesco, 1998), and might be partly contaminated by refractory components (Chen et al.).

The first aim of our study was to implement in healthy subjects a method for assessing the strength of the inhibitory effects induced by single pulse transcortical magnetic stimulation on the firing time of tonically discharging motor units. This was done by measuring the changes in the duration of successive ISIs (peri-stimulus ISI analysis) depending on whether or not a spike was triggered by the stimulation (cf. Calancie et al. 1987).

Our second aim was to determine the dependence of the cortically induced inhibition on the delay between the stimulation and the previous firing of the motoneurone. This was done by computing phase response curves, which give an integrated representation of the time course and strength of the cortically induced changes in ISIs (lengthening or shortening in the case of inhibitory or excitatory effects, respectively) as a function of the post-spike timing of the stimulation. This type of analysis stems out from theoretical and experimental studies that attempt to provide a mathematical description of the changes induced by a perturbation (synaptic input) on the phase (ISI) of a an oscillator (regular spiking motoneuron) depending on the timing of the perturbation in the phase (Perkel et al. 1964; Reyes and Fetz, 1993).


Single motor unit activity was recorded in ECR muscles by a metal microelectrode as the subject performed an isometric wrist extension at less than 20% of the maximal voluntary force. The wrist area of the motor cortex was stimulated through a figure-eight coil located on the scalp above the region from which a response had been selectively induced at rest in the ECR at the lowest intensity (resting threshold intensity). The subjects were provided auditory and visual feedback of the motor unit discharge and were asked to keep the unit steadily discharging during several minutes while transcranial magnetic stimulation was delivered at a frequency of 0.3 Hz at the resting threshold intensity; all analyses were performed off-line. Two samples of 39 motor units were tested in 7 healthy subjects with a minimum of 130 transcranial magnetic stimulation trials with intensity ranging from 35% to 50 % of the maximal output of the stimulator (Magstim 200). The inhibitory (lengthening) and excitatory (shortening) effects of the stimulation on successive post-stimulus firing times were determined by assessing the duration of the ISI during which the stimulation was delivered and 7 ISIs thereafter. The mean of 3 ISIs preceding the stimulation was taken as reference and all changes were expressed in percent of the pre-stimulus mean.

The specificities of this analysis were threefold:

  1. The non-physiological ISI lengthening that occurs whenever a spike has been missed due to the stimulus artefact was taken into account. This was done by introducing a virtual artefact lasting a few ms as the real artefact and associated to a virtual stimulation marker located 1.5 s before the real stimulus. Any spike occurring in the virtual artefact window was deleted.
  2. The monosynaptic conduction time from the cortex up to the muscle in which the motor unit activity was recorded was taken into account in order to analyse the effects of the synaptic potentials induced by the stimulation at the motoneurone level. The latency of the monosynaptic like-response in the PSTH was taken as an estimate of the conduction time form the cortex to the muscle. The spike train were accordingly shifted backward by this amount of time in such way that in the PSTH computed after the shift, the excitatory response occurred less than 1 ms after the stimulus as if it has been recorded in the spinal cord.
  3. The trials were organized into two groups (response- and no response-trials, respectively) depending on whether or not a spike had been generated within the limits of the mono-synaptic like response. A peri-stimulus ISI analysis was performed on each trial group in order to differentiate the pure inhibitory changes occurring without excitation from those that followed the excitatory response of the motoneurone tested.


The PSTH analysis revealed the presence of monosynaptic-like responses in 25 out of 39 motoneurones tested (64%). Pure inhibitory responses were observed in only 2 motoneurones (5 %).

The peristimulus ISI analysis revealed a consistent pattern in the distribution of the excitatory and inhibitory effects on the successive post-stimulation firing times. In the group of no-response-trials, 28 of the 39 units tested presented a significant inhibitory effect with a delay of the first firing time after the stimulation. This was occasionally followed by an excitatory effect that advanced the second post-stimulus firing time. In the group of response-trials, the excitatory effect advanced the first firing time after the stimulation, which was followed by a conspicuous inhibitory effect with a delay of the second post-stimulus firing time. In the whole sample, no consistent effect could be detected after the fourth post-stimulus ISI.

In the present experimental conditions, the strongest effect of transcranial magnetic stimulation appeared to be inhibitory in terms of instantaneous firing rate or firing time. This effect was particularly conspicuous on the ISI that followed the excitatory response. The strongest excitatory effects in the response-trials tended to be associated with the strongest inhibitory effects in the no response-trials of the same unit. Finally, the inhibitory effect on the second post-stimulus firing was positively correlated with the excitatory effect on the first post-stimulus firing. This favours the contribution of various post-excitatory processes among which summation of the motoneurone AHP, cortical as well as spinal recurrent inhibition, and peripheral inhibitory feedback.

t is well known that transcranial magnetic stimulation of moderate intensity is much more effective in triggering a spike when the stimulus is delivered in the second half of the ISI toward the end of the ISI (Boniface et al., 1991, Olivier et al.1995, Garland et Miles, 1997). The question arises as to whether a similar dependence exists for the inhibitory effects. This was investigated by computing a phase response curve following a procedure illustrated Fig. 1.

Figure 1A shows the PSTH computed with the discharge of a motoneurone tested with an intensity of 50 % of the maximal stimulator output. In Fig. 1 B , each dot represents a trial. The timing of the stimulus regarding the previous spike is plotted on the abscissa (post-stimulus delays) and the duration of the ISI during which the stimulation was delivered is plotted in the ordinate. The values obtained with the real and virtual stimulation trials are shown with black dots and grey circles, respectively. In the trials in which a spike has been triggered by the stimulation, the ISI values (black) are aligned below those obtained in the virtual stimulation trials (grey), in keeping with an excitatory effect (ISI shortening). In this example, the triggering effect can be detected 30 ms after the motoneurone spike. In the trials in which no spike has been triggered, the ISIs are located clearly above the ISIs of the virtual stimulation trials in keeping with an inhibitory effect. The inhibitory effect was detectable right after the motoneurone spike and lasted several tens of ms. For post-spike delays ranging from 30 to 50 ms, both excitatory and inhibitory effects occurred concurrently.

In Fig. 1 C, the phase response curve shows the net effect of transcranial magnetic stimulation on the next firing time. It was computed by pooling the trials on the basis of the stimulus post-spike delays in successive 20-ms bins (abscissa) and averaging the ISI values in each bin expressed in percent of the pre-stimulus ISI mean value and plotted on the ordinate. The black and grey curves were computed with the real and virtual stimulation trials, respectively. Up to 20 ms after the spike (bin 1), the inhibitory effects clearly predominated leading to a significant ISI lengthening compared with the virtual value. The excitatory effect became predominant when the stimulation was delivered with a post-spike delay of 60 ms (bin 3).

Biphasic phase response curves have been observed quite consistently in the sample of 39 motoneurones tested. In some cases, the inhibitory effect (ISI lengthening) observed with the earliest post-spike delays was prolonged for several bins (post-spike delays of 60 to 80 sm) and could mask weak excitatory effects ascertained by the presence of a small peak in the PSTH with the second ISI after the stimulation. In many cases, particularly when strong inhibitory and excitatory were detected in the first ISI, the phase-response curve computed with the second ISI was an inverted image of what occurred in the first ISI: when the stimulus was delivered within early post delays, the strong inhibitory effects that delayed the first post-stimulus firing time was followed by a powerful excitatory effect that advanced the next firing. Conversely, when the stimulation was delivered with late post-spike delays (more than 60 ms after a spike), the excitatory effect on the first post-stimulus firing time was followed by a strong inhibitory effect revealed by the marked lengthening of the next ISIs.

Finally, in preliminary data obtained with a sample of 15 motor units tested in the same muscle in a completely deafferented patient, no inhibitory effects could be detected in the peri-stimulus ISI analysis and in the phase response curves whatever the post-spike timing of the stimulation. An example is shown in Fig 1 D, E, F. Although synaptic reorganization may have occurred in this patient, the striking disappearance of the cortically induced inhibitory effects on motoneurone firing time can be taken as an indication that part of these effects might involve the contribution of sensory feedback generated by the motor response. This is in keeping with data obtained in a recent study combining functional MRI and transcranial magnetic stimulation (Bestmann et al. 2004).


Using transcranial magnetic stimulation at a near-threshold intensity, we found a consistent delay of the motoneurone next firing time that could be detected in trials in which no spike had been triggered by the stimulation. This inhibitory effect was prominent when the stimulation was delivered very early after a spike and could be prolonged for several tens of ms. The inhibitory processes involved must be either late or long-lasting.

In intracellular recording of motoneurones in anesthetized monkeys, microstimulation of motor cortex typically generates mixed responses that include an EPSP attributed to the monosynaptic cortico-motoneuronal pathway followed by an IPSP lasting up to 40 ms, which may actually curtail the preceding EPSP (Lemon et al. 1986). Assuming that transcanial magnetic stimulation induces similar postsynaptic responses (probably stronger), it can be expected that with very early post-spike delays, the EPSP would be able to take the membrane potential up to the firing threshold, whereas the following IPSP might be effective if sufficiently prolonged. With longer post-spike delays, the same combination of EPSP followed by long lasting IPSP may account for the duality of inhibitory and excitatory effects, which appeared to be intermingled (fig 1 B). In the middle of the ISI, when the membrane potential is still below the firing threshold, the effectiveness of the EPSP in triggering a spike would fluctuate depending on the synaptic noise and the IPSP would be effective whenever the EPSP has failed. When the membrane potential is nearing the firing threshold at the end of the ISI, the EPSP triggering effect would fully obliterate the following IPSPs. To explain our data, however, the post-excitatory IPSP should last up to 80 ms or more, which seems much longer than the available evidence from motoneurone intracellular recordings (Lemon et al. 1986).

Very long lasting IPSPs attributed to GABA B mechanisms have been described in the cortex. A popular hypothesis is that these long lasting ISPSs might temporarily reduce the excitatory cortical drive that contribute to the voluntary activation of the motoneurones. This does not preclude, however, the contribution of subcortical disfacilitation processes, such as presynaptic inhibition of proprioceptive inputs. This would be in good agreement with the disappearance of cortically induced ISI inhibitory changes which seems to occur in the de-afferented patient.

AMASSIAN VE, CRACCO RQ, MACCABEE PJ, CRACCO JB, RUDELL A, EBERLE L. Suppression of visual perception by magnetic coil stimulation of human occipital cortex. Electroencephalog. Clin Neurophysiol 74: 458-462, 1989.

BESTMANN S, BAUDEWIG J, SIEBNER H, ROTHWELL J, FRAHM J. Functionnal MRI of the immediate impact of transcranial magnetic stimulation on cortical and subcortical motor circuits. Eur.J Neuroscience 19: 1950-1962, 2004.

BONIFACE S, SCHUBERT M, MILLS K. Supression and long latency excitation of single spinal motoneurons by transcranial magnetic stimulation in health, multiple sclerosis and stroke. Muscle & Nerve 17: 642-646, 1994.

BONIFACE S, MILLS K, SCHUBERT M. Responses of single spinal motoneurons to magnetic brain stimulation in healthy subjects and patients with multiple sclerosis. Brain 114: 643-662, 1991.

BROUWER B, QIOA J Characteristics and varaibility of lower limb motoenuron responses to transcranial magnetic stimulation. EEG. Clin. Neurophysiol 97: 49-54, 1995

CALANCIE B, NORDIN M, WALLIN U, HAGBARTH KE. Motor-unit responses in human wrist flexor and extensor muscles to transcranial cortical stimuli. J Neurophysiol 58: 1168-1185, 1987.

CHAN J, LIN C, PIERROT-DESEILLIGNY E, BURKE D. Excitability changes in human peripheral nerve axons in a paradim mimicking paired-pulse transcranial magnetic stimulation. J Physiol (lond) 542: 951-961, 2002.

CHEN R, LOZANO AM, ASHBY P. Mechanism of the silent period following transcranial magnetic stimulation. Evidence from epidural recordings. Exp Brain Res 128: 539-542, 1999.

CLASSEN J, BENECKE R. Inhibitory phenomena in individual motor units induced by transcranial magnetic stimulation. Electroencephalogr.Clin.Neurophysiol 97: 264-274, 1995.

DAVEY NJ, ROMAIGUERE P, MASKILL DW, ELLAWAY PH. Suppression of voluntary motor activity revealed using transcranial magnetic stimulation of the motor cortex in man. J Physiol 477: 223-235, 1994.

DAY BL, ROTHWELL JC,. THOMPSON PD, MAERTENS DE NOORDHOUT A, NAKASHIMA K. SHANNON K, MARSDEN CD. Delay in the execution of voluntary movement by electrical or magnetic brain stimulation in intact man. Evidence for the storage of motor programs in the brain. Brain 112: 649-66, 19893.

EISEN A, SWASH M. Clinical neurophsyiology of ALS. Clin Neurophysiol 112: 2190-2201, 1997.

GARLAND SJ, MILES TS. Responses of human single motor units to transcranial magnetic stimulation. Electroencephalog. Clin Neurophysiol 105: 94-101, 2001.

KUJIRAI T, CARAMIA MD, ROTHWELL JC, DAY BL, THOMPSON PD, FERBERT A, WROE S, ASSELMAN P, MARSDEN CD. Corticocortical inhibition in human motor cortex. J Physiol 471: 501-519, 1993.

MATTEI B, SCHMIED A. Delayed and prolonged effects of a near threshold EPSP on the firing time of human a-motoneurones. 2002.

OLIVIER E. BAWA P, LEMON RN. Excitability of human upper limb motoneurones during rhythmic discharge tested with transcranial magnetic stimulation. J Physiol 485: 257-269, 1995.

PASCUAL-LEONE A, GATES JR, DHUNA A. Induction of speech arrest and counting errors with rapid-rate transcranial magnetic stimulation. Neurology 41: 697-702, 1991.

PERKEL DH, SCHULMAN JH, BULLOCK TH, MOORE A. Pacemaker neurons: effect of regularly spaced synaptic input Science 145: 61-63, 1964.

PETERSEN NT, BUTLER JE, MARCHAND-PAUVERT V, FISHER R. LEDEBTA, PYNDT HS, HANSEN NL, NIELSEN JB. Suppression of EMG activity by transcranial magnetic stimulation in human subjects during walking. J Physiol 537: 651-656, 2001.

PRIORI A, BERTOLASIL, ROTHWELL JC, DAY BL, MARSDEN CD. Some saccadic eye movements can be delayed by transcranial magnetic stimulation of the cerebral cortex in man. Brain 116: 355-367, 1993.

SALERNO A, GEORGESCO M. Double magnetic stimulation of the motor cortex in amyotrophci lateral sclerosis. EEG. Clin. Neurophysiol 107: 133-139, 1998.

ZIEMANN U, WINTER M, REIMERS CD, REIMERS K, TERGAU F, PAULUS W. Impaired motor cortex inhibition in patients with amyotrophic lateral sclerosis. Evidence from paired transcranial magnetic stimulation. Neurology 49: 1292-1298, 1997.


Human "respiratory" motorneurones

Simon Gandevia

Prince of Wales Medical Research Institute, Sydney, Australia

Human respiration is a complex process involving integration of descending inputs to motor neurones from ponto-medullary rhythm and pattern generators, and from motor cortical and other cortical areas. The presentation reviewed some of the evidence that in volitional tasks, such as arm or leg movement, there can be pre-movement activity in trunk and respiratory muscles. This has been demonstrated for the diaphragm, which is the principal muscle responsible for displacement of volume during inspiration (Hodges et al. 1997). A second activity in which volitional drives to inspiratory motor neurones are active is during "targetted" breathing at different flows to different inspired lung volumes. Recordings of single motor unit activity in the costal diaphragm have revealed that there can be task-dependent adjustments in the recruitment order when a specific pattern of inspiration is adopted (Butler et al. 1999). These results highlight the difficulties in nomenclature for inspiratory muscles, many of which have different mechanical actions on the rib cage.

The second part of the presentation formed a link with some of the earlier presentations dealing with the properties of hypoglossal motoneurones in slice preparations. There have been few recordings of the activity of single motor units in the valve muscles of the upper airway. These muscles are important as they supply the rhythmic tone and changes in dimension that accompany breathing. Without such activity, the upper airway can be sucked closed during inspiration, particularly when supine. In current studies, we have examined the coordinated output to inspiratory pump muscles (acting on the thorax) and to one upper airway muscle (genioglossus) (Saboisky et al. in preparation). For the pump muscles (costal diaphragm, parasternal intercostals, scalenes, and 3rd and 5th dorsal external intercostal muscle) there is recruitment throughout quiet inspiration. The rate of recruitment of units is greatest for the diaphragm and scalene muscles, with the majority of units being recruited within the first 350 ms. The slowest recruitment occurred for the lower dorsal external intercostal muscles. The firing rates of motor units during quiet breathing also differed amongst the pump muscles, with the highest rate being observed for the costal diaphragm. Tonic firing with a superimposed inspiratory modulation was rarely found, except in the lower dorsal external intercostal muscles, where nearly half the units showed tonic inspiratory firing. The cause of this tonic discharge is unknown.

Finally, preliminary results were shown for recordings genioglossus muscle activity during quiet breathing in the supine posture. Here, 5 patterns of activity could be discerned: phasic inspiratory, tonic inspiratory, phasic expiratory, tonic expiratory, and purely tonic firing. When compared with pump muscles, firing rates were high during quiet breathing, with rates of 20-30 Hz being common. Overall, it is likely that there is a net inspiratory action on the genioglossus, which begins well before airflow and serves to open the inspiratory valve maximally by early inspiration. Further studies will be required to discern the complex drives which this set of hypoglossal motoneurones receive. Overall, there are differences in the timing, distribution, and amount of drive expressed by the different respiratory motoneurone pools.

Acknowledgements: I am grateful to a number of colleagues involved in the current studies including: Jane Butler, André De Troyer, Rob Fogel, Robert Gorman, Julian Saboisky, Janet Taylor, John Trinder, David White. This work is supported by the National Health and Medical Research Council.

BUTLER JE, MCKENZIE DK, GANDEVIA SC. Discharge properties and recruitment of human diaphragmatic motor units during voluntary inspiratory tasks. Journal of Physiology 518: 907-920, 1999.

HODGES PW, BUTLER JE, MCKENZIE DK, GANDEVIA SC. Contraction of the human diaphragm during rapid postural adjustments. Journal of Physiology 505: 539-548, 1997.


Comparison of motoneuron responses to high frequency inputs via Ia afferents and corticomotoneuronal pathways in humans

Parveen Bawa1 and Blair Calancie2

1Simon Fraser University, Vancouver, Canada and 2State University of New York, Syracuse, NY.


Stimulation of the mammalian motor cortex has been used for many years for studies of central motor control. It has also been used on humans in the operating room for identifying motor cortex during resection of tumor or epileptogenic tissue. Stimulus patterns consisting of single-pulse, repetitive high-frequency (~ 50 Hz), and brief trains of very high frequency (~ 300 - 500 Hz) pulses are generally employed.

Properties of EPSPs in spinal motoneurons in response to excitatory inputs suggest that higher-frequency stimulation would be more likely to cause facilitation of the motoneuron response and recruitment. This conclusion is supported by an examination of neuronal discharge rates associated with the execution and consequences of voluntary movements. For example, corticomotoneurons (CMN) have been shown to discharge very high frequency (> 500 Hz) bursts of action potentials during goal-directed movements in the forelimb. Similar discharge rates have been seen in Ia afferents under certain conditions.

While it seems clear that very high frequency inputs lead to a higher probability of motoneuron recruitment, it is not clear whether discharge frequencies are passed along to the level of muscle fibers. The muscle unit is known to show a markedly non-linear (i.e. facilitated) force response to a short high frequency burst from the motoneuron. One question we asked, then, is to what extent very high frequency motoneuron inputs were reflected by comparable output high frequency discharge rates in motoneurons? In the process of addressing this question, we also hoped to establish the stimulation parameters that produce maximal motoneuron facilitation, and to gain a better understanding of the site(s) where this facilitation might occur.

Our approach was to examine motoneuron responses to very high frequency stimulation of either: 1) the motor cortex using transcranial magnetic stimulation; or 2) Ia afferents following peripheral nerve stimulation. Both pathways are known to make monosynaptic connections onto spinal motoneurons that are roughly comparable in density and position, hence should lead to recruitment patterns that are themselves comparable, unless other factors might come into play.


Both population-based responses (via surface EMG) and single-cell responses (via single motor unit discharge) were studied for muscles in upper and lower limbs. For TMS inputs, we examined flexor carpi radialis (FCR), extensor carpi radialis (ECR), and tibialis anterior (TA). For Ia inputs, responses were recorded from FCR or soleus following stimulation of the median or tibial nerves, respectively. The stimulus consisted of either a single pulse, or a brief train of pulses described by m@n, where m is the number of pulses in the train, and n is the interpulse interval (IPI) within a stimulus train. The range for m was 2- 4 pulses, and 1-7 ms for n. Within a pulse train, the stimulus strength was constant. For TMS, we used a ‘Pyramid’ stimulator (four ‘Magstim 200’ units working through 3 ‘BiStim’ units; Magstim Co., Wales) and skull-conforming double-cone coils for arm- or leg-area activation. A Digitimer DS7A was used for peripheral nerve stimulation (pulse duration 0.5 ms for the median nerve and 1.0 for the posterior tibial nerve); stimuli were applied through saline-soaked pads within a bar electrode.


Surface EMG: For all three muscles examined with TMS at supra-threshold intensity, surface EMG responses were much greater with trains than with single pulses. This was true whether the subject was resting, or was making a weak background contraction in the muscle at the time of TMS delivery. Within any one train, the total area of EMG in the rectified and averaged record increased with increasing numbers of pulses in the train, as seen to the top of Figure 1. Keeping the number of pulses constant and varying the interpulse interval, the total EMG tended to be largest for the longest IPI values (e.g. Figure 1, bottom).

With electrical stimulation of Ia afferents, although the total response to a pulse train input was larger than with a single pulse, the absolute magnitude of the contributions from sequential pulses in the train declined. This decrement was more pronounced in soleus than in FCR.

Single motor units: The response probability of each of the single motor units examined was higher for a TMS pulse train than for a single TMS pulse. This effect was especially pronounced for a 4-pulse train with an IPI > 3 ms. At these longer IPI intervals, the response peaks tended to be more distinct and with timing between peak onsets identical to the IPI value applied. This timing suggests that each PSTH peak was created exclusively by one of the pulses in the train, with no other volley or activity outside of these IPI values influencing the discharge probability of the motoneuron. This was true for all motor units tested from FCR, ECR and TA motor units.

Electrical stimulation with constant intensity of Ia afferents with trains also increased the response probability of each motor unit over and above that with a single stimulus pulse and for every IPI. However, clear multiple PSTH peaks were observed only with very long IPIs (i.e. 6 ms or longer) for these inputs, as shown in Figure 3. By causing the second Ia stimulus to be stronger than the first, it was possible to force the motoneuron to discharge at shorter intervals, as demonstrated to the right of Figure 3.


Spinal motoneurons are capable of responding to any one of and up to 4 closely-spaced excitatory inputs, whether from corticomotoneuronal or Ia afferent sources. The multiple-pulse train causes marked facilitation of the response. For the CMN-MN pathway, this facilitation could occur at the cortical level, at the synaptic level (i.e. synaptic bouton), or at the motoneuronal level depending on the IPI. For the Ia-MN pathway, transmission does not seem to occur for IPI < 4 ms, not because the motoneuron is unable to respond to this input at this frequency. Rather, the Ia afferents recruited by the first sub-maximal stimulus pulse appear to be refractory during this brief time period; recruitment of other Ia afferents with a stimulus pulse stronger than the first can restore the second peak in the PSTH record, as seen to the right of Figure 3.

This work was supported by Natural Sciences and Engineering Research Council of Canada and the State University of New York, Syracuse, NY.



Firing patterns of spontaneously active motor units in spinal cord injured subjects

Inge Zijdewind, Lillian Peterson*, Christine K. Thomas*

Groningen University, Groningen, The Netherlands and *University of Miami School of Medicine, Florida, USA

In spinal cord injured subjects, paralyzed muscles are not always quiescent. We have recorded this involuntary motor unit activity over 30 minute periods while the subjects relaxed. Intramuscular and surface EMGs were recorded from the thenar muscles, together with force in the directions of thumb flexion and abduction. The mean rate, standard deviation and coefficient of variation were used to describe the basic motor unit firing properties. With the use of spike triggered averaging we have estimated the force and the contraction time of the units. Finally, we have estimated an important intrinsic motoneuron property, the afterhyperpolarization using two different methods, that described by Matthews (1996) and the method described by Person and Kudina (1972).

Spontaneous motor unit activity was seen in the thenar muscles of all subjects. In general, the units could be divided into two distinct groups: i) units that were only sporadically active. The few spikes that these units generated were followed by long periods of rest, and ii) tonically firing motor units (see also Zijdewind and Thomas, 2001). Sometimes sporadically active units began to fire tonically. Overall, 28 tonically firing units were followed for more than 3 minutes. The average (± SD) number of recorded action potentials per unit was 4325 ± 2610. The average of the mean firing rate of these tonically active units was 6.5 ± 1.5 Hz (range of the mean firing frequencies: 3.8-9.6 Hz). The variability in the firing could be very different between units. Some motor units had a highly variable firing frequency, while the firing behaviour of other motor units was very regular.

We have divided the motor units into two groups based on the variability in their firing frequency. Motor units that fired with a variability >0.20 (sd/mean frequency) were considered to be irregularly firing units, while units with a variability < 0.20 were labelled as regularly firing units. The mean firing frequency did not differ between the two groups of motor units (6.4 ± 1.9 Hz versus 6.6 ± 1.3 Hz, respectively). The range of mean motor unit firing rates was also comparable (3.8-9.1 Hz versus 4.5-9.6 Hz). However, the mean forces for the two groups, estimated when the units were firing at low frequencies to minimise the amount of twitch summation, were significantly different from each other (96 ± 108 mN versus 14 ± 13 mN); the irregularly firing units being the strongest.

For the regularly firing units, the estimates of the afterhyperpolarization (AHP) from using the methods described by Matthews (1996) and Person and Kudina (1972) showed a significant correlation (R2=0.45). The mean value for the estimation of the time-constant of the afterhyperpolarization (42 ± 7 ms) was comparable to the values obtained from control subjects for soleus units (Matthews 1996) and for low threshold motor units in first dorsal interosseous (Gossen et al. 2003). The mean duration of the AHP estimated by the method of Person and Kudina (1972) was 167 ± 27 ms. These values were lower than data obtained in healthy subjects for the biceps brachii and soleus muscles (Tokizane and Shimazu, 1964). No significant correlations were found between unit force- or speed-related twitch parameters and the estimations of the AHP.

In conclusion, thenar motor units that were tonically active in spinal cord injured subjects differed in their firing variability. Motor units with high firing variability were significantly stronger than units with low firing variability. These results suggest that the irregularly firing units are higher threshold units. The estimated durations and time-courses of the afterhyperpolarizations of the regularly firing units suggest that these units are comparable to the low threshold, relatively slow motor units found in muscles of uninjured individuals.

Gossen ER, Ivanova TD, Garland SJ. The time course of the motoneurone afterhyperpolarization is related to motor unit twitch speed in human skeletal muscle. J Physiol 552: 657-664, 2003.

Matthews PB. Relationship of firing intervals of human motor units to The trajectory of post-spike after-hyperpolarization and synaptic noise. J Physiol 492: 597-628, 1996.

Person RS, Kudina LP. Discharge frequency and discharge pattern of human motor units during voluntary contraction of muscle. Electroencephalogr Clin Neurophysiol 32: 471-483, 1972.

Tokizane T, Shimazu H. Functional differentiation of human skeletal muscle. Charles C Thomas, Springfield, 1964.

Zijdewind I, Thomas CK. Spontaneous motor unit behavior in human thenar muscles after spinal cord injury. Muscle Nerve 24: 952-962, 2001.


Saturation of Motor Unit Firing Rate

Andrew J. Fuglevand and Richard K. Johns

University of Arizona, Tucson, USA

During voluntary contraction, the firing rates of individual motor units saturate at frequencies substantially lower than that needed to elicit maximal force responses using electrical stimulation (Johns and Fuglevand, Soc Neurosci Abstr 2003). Such saturation of motor unit discharge may result from limitations in the voluntary excitatory drive, from intrinsic properties of the motor neurons, or both. Excitatory drive to a motor neuron pool can be augmented artificially by activation of muscle spindle Ia afferents through tendon vibration. Therefore, we examined the effects of tendon vibration on the change in firing rate of individual MUs that were voluntarily activated at various rates below saturation and at saturation. Activity of 103 MUs from biceps brachii were recorded using intramuscular microelectrodes in 10 human subjects during isometric contractions. Vibratory stimuli (80 Hz for 4 s) were administered distally at the biceps tendon while subjects maintained motor unit discharge at various rates. Tendon vibration proved an effective excitatory stimulus as it led to an increase in force output of the elbow flexors in all trials. The vibration-induced changes in firing rate of individual motor units, however, were affected differentially depending on the rate of maintained voluntary discharge immediately prior to the application of vibration. The change in firing rate associated with vibration was negatively correlated (P<0.001) with the initial firing rate. For example, when motor units discharged at ~ 6 Hz, firing rate increased in response to vibration on average by about 4 Hz. However, when motor units discharged at frequencies from 12 - 16 Hz, the same vibratory stimulus had virtually no effect on firing rate. Therefore, these finding suggest that motor neurons discharging at saturation frequencies are unresponsive to additional synaptic excitation. The intrinsic cellular mechanisms that limit firing rates of motor units during voluntary effort remain to be identified..


The role of cutaneous afferents in motor control

Penelope McNulty, Kemal Türker*, James Fallon, Leah Bent, and Vaughan Macefield

Prince of Wales Medical Research Institute, Sydney, Australia and *University of Adelaide, Adelaide, Australia

The central nervous system continuously integrates feedback from peripheral somatosensory receptors to help regulate fine motor control of the human body. The discharge of an already active motoneurone pool can be facilitated by the synchronised input from a population of cutaneous (e.g., Caccia et al. 1973) or muscle spindle (Gandevia et al. 1986) afferents. We have shown that the input of a single cutaneous low-threshold mechanoreceptor in the glabrous skin of the hand is sufficient to modulate ongoing motoneurone activity in muscles acting on the hand (McNulty et al. 1999). However there is no evidence of a similar effect from the input of a single muscle spindle afferent located either in the contracting parent muscle (homonymous projection) or a nearby quiescent or synergistic muscle (heteronymous projection) and regardless of whether the afferent is a primary or secondary spindle ending (McNulty & Macefield, 2001). Based on latency measurements the observed effects on the discharge of the motoneurone pool are thought to operate via a spinally mediated oligosynaptic reflex pathway. We recently demonstrated this strong synaptic connection also operates in the lower limb (Fallon et al. 2004).

The discharge of single cutaneous mechanoreceptors was recorded using an insulated tungsten microelectrode inserted percutaneously into a peripheral nerve enabling the receptor end organ to be identified and its receptive field to be stimulated appropriately. Subjects performed a constant weak (<10% MVC) isometric contraction while the receptor was activated. EMG data were recorded using surface electrodes and then sampled with reference to every neural discharge using spike triggered averaging, highlighting any event in the EMG time locked to the neural output.

Our results demonstrated that the activity of motoneurones in both the upper and lower extremity can be influenced by the input from single cutaneous low threshold mechanoreceptors. The pattern of motoneurone modulation depends on the pattern of afferent input. When an afferent discharges irregularly (such as a rapidly ("fast") adapting type II, FA II; or slowly adapting type I, SA I) a single peak projects beyond the background activity in the averaged EMG signal. However if the afferent discharges regularly (slowly adapting type II receptor, SA II) or irregularly but in regularly occurring bursts (rapidly adapting type I, FA I) a cyclic pattern of modulation is seen in the averaged EMG. In this case the periodicity of the cyclic modulation in the averaged EMG matched the mean interspike interval of the afferents' discharge with a correlation coefficient of 0.97. This suggests that each neural discharge is modulating the pattern of activity in the motoneurone pool reflecting a strong spinal synaptic connection between the two. These patterns of activation are illustrated below.

The pattern of motoneurone modulation depends on the pattern of neural input. Left hand panel: averaged data from an SA II receptor located in the nail bed of digit II of the hand, the regular discharge is highlighted in the autocorrelogram (top) and produced a cyclic modulation in the averaged EMG of the first dorsal interosseus (FDI) muscle (bottom). Right hand panel: averaged data from an FA II receptor located in the webspace between digits II and III. The irregular discharge highlighted in the top panel is associated with a single excitatory peak in the averaged EMG (bottom).

In the lower limb recordings we were also able to discriminate single motor unit activity within the compound EMG signal. The activity of the single motor unit was cross correlated to the neural discharge. The compound muscle reflex response can but does not necessarily reflect the modulation of its constituent single motor units. In some recordings a significant modulation of the single motor unit was observed while no such change was observed in the parent muscle activity. Alternatively, significant inflections were either observed in both the single motor unit and the compound muscle activity or in neither.

These results suggest that the activity of motoneurones can be influenced by the input from some low threshold cutaneous mechanoreceptors in the glabrous skin of the hand and foot. The functional significance of such modified behaviour may relate to the maintenance and control of posture during quiet standing and movement in the lower limb and in fine motor control of the hand during exploration of the external environment or manipulation of held objects.

CACCIA MR, MCCOMAS AJ, UPTON AR, BLOGG, T. Cutaneous reflexes in small muscles of the hand. Journal of Neurology Neurosurgery and Psychiatry 36: 960-977, 1973.

FALLON JB, BENT LR, MCNULTY PA, MACEFIELD VG. Reflex coupling between single low threshold mechanoreceptors and spinal motoneurones of human foot muscles. Proceedings of the Australian Neuroscience Society 15: 91, 2004.

GANDEVIA SC, BURKE D, MCKEON B. Coupling between human muscle spindle endings and motor units assessed using spike-triggered averaging. Neuroscience Letters 71: 181-186, 1986.

MCNULTY PA, MACEFIELD VG. Modulation of ongoing EMG by different classes of low-threshold mechanoreceptors in the human hand. Journal of Physiology 537: 1021-1032, 2001.

MCNULTY PA, TÜRKER KS, MACEFIELD VG. Evidence for strong synaptic coupling between single tactile afferents and motoneurones supplying the human hand. Journal of Physiology 518: 883-893, 1999.



Discharge pattern of single motor units during shortening and lengthening contractions

Benjamin Pasquet, Alain Carpentier, and Jacques Duchateau

Universite Libre de Bruxelles, Brussels, Belgium

Conflicting results have been reported regarding motor units (MUs) recruitment patterns during lengthening (LEN) contractions. Although some authors suggested that LEN contractions differ from shortening (SHO) contractions and involve a selective recruitment of high-threshold MUs (Nardone et al., 1989; see also Howel et al., 1995), others investigators showed no difference in the recruitment order (Sogaard et al. 1996; Laidlaw et al., 2000; Stotz and Bawa, 2001). In addition, there is a lack of consistent observations regarding MUs discharge rate during these two contraction types. The objective of this study was to compare the recruitment pattern and discharge rate of MUs during standardized LEN and SHO contractions.

The experiments were performed on 6 subjects (22-48 years). During the experimental session, they sat on a chair with one foot strapped to a footplate fixed to the rotational axis of a computer-controlled ankle ergometer (Pasquet et al., 2000). The torque of the dorsiflexors and the surface electromyogram (EMG) of the tibialis anterior were monitored during the different contraction modalities. In addition, single MU action potentials were recorded by a selective wires electrode (50µm in diameter) inserted into the muscle belly. The task consisted for the subjects to match a target isometric torque and to keep a steady firing of a selected MU for at least 5s. Thereafter, SHO or LEN contractions at constant speed (10°/s over a 20° range of motion) were imposed while the subject maintained a given activation. For each MU, instantaneous discharge rate was recorded during the task and its recruitment threshold was determined during isometric ramp contractions.

A total of 59 MUs were investigated. No systematic recruitment of additional high-threshold MUs and derecruitment of low-threshold MUs was observed during LEN contractions. In contrast, additional MUs were occasionally recruited during SHO contractions, while the previously active units maintained their discharge rate (fig.1; left panel). A different behaviour was observed in LEN contractions, since MUs with the highest recruitment threshold usually stopped to discharge before the end of the movement (fig.1; right panel). During SHO contractions, the typical discharge pattern observed at the beginning of the movement was a long first interspike interval (ISI) followed by a quick increase in discharge rate (cf fig.1). In the second part of the movement, the discharge rate was progressively enhanced. In contrast, during LEN contractions, 2-3 shorts initial ISI were followed by a more stable discharge rate throughout the movement.

In line with previous investigations (Sogaard et al. 1996; Laidlaw et al., 2000; Stotz and Bawa, 2001), our study does not support the viewpoint of a preferential recruitment of high-threshold MUs or possible change in recruitment order during LEN contractions, compared with SHO ones (Nardone et al., 1989). If some MUs recruitment or derecruitment occurred during SHO and LEN contractions, respectively, this behaviour happened in accordance with the size principle (Henneman et al., 1965). The discharge pattern, however, differed between LEN and SHO contractions. In the former condition, the lengthening of the muscle presumably have induced an increased discharge of the muscle spindles and consequently intensified muscle activation. In contrast, the unloading of the spindles at the onset of muscle shortening can explain the decrease in MUs discharge rate. The derecruitment of some MUs at the end of the LEN contraction should be due to increased passive and active force during muscle lengthening (Pasquet et al., 2000), and therefore the need of a reduced number of active MUs to maintain the same net force. It is concluded that although the discharge pattern differs in SHO and LEN contractions, in both conditions MUs are recruited and derecruited in an orderly fashion, from small to large ones and vice versa, respectively.

Henneman E, Somjen G, Carpenter DO. J Neurophysiol 28 : 560-580, 1965.
Howell JN, Fuglevand AJ, Walsh ML, Bigland-Ritchie B. J Neurophysiol 74 : 901-904, 1995.
Laidlaw DH, Bilodeau M, Enoka RM. Muscle Nerve 23 : 600-612, 2000.
Nardone A, Romano C, Schieppati M. J Physiol 409 : 451-471, 1989.
Pasquet B, Carpentier A, Duchateau J, Hainaut K. Muscle Nerve 23 : 1727-1735, 2000.
Sogaard K, Christensen H, Jensen BR, Finsen L, Sjogaard G. EEG Clin Neurophysiol 101 : 453-460, 1996.
Stotz PJ, Bawa P. Muscle Nerve. 24 : 1535-1541, 2001.

Figure.: Discharge pattern of two MUs with different recruitment threshold (MU1 : 6.4%; MU2 : 26.3% of maximal voluntary contraction) during SHO (left panel) and LEN (right panel) contractions. In each condition, position (a), torque (b), intramuscular EMG (c) and instantaneous MU discharge rates (d) are illustrated. The arrows indicate the recruitment and derecruitment of MU2.


Bilateral motor unit activation during postural tasks

S. Jayne Garland, George Mochizuki, and Tanya D. Ivanova

University of Western Ontario, London, Canada


Postural tasks require the simultaneous activation of bilateral leg muscles. Postural activity in the leg muscles in patients following stroke has been shown to be depressed in both legs even though the hemiparesis was unilateral (Stevenson & Garland, 1996). This led to the speculation that the control of bilateral muscles in the legs during postural tasks involved common descending control. The control of simultaneously active motor units can be detected using the cross-correlation histogram technique (Schmied et al., 2000) or alternatively, common drive analysis (DeLuca et al. 1982). A short duration peak in the cross-correlation histogram reflects branching of the last-order presynaptic inputs and is commonly described as short-term synchronization. Broad duration peaks reflect synchronization of independent presynaptic inputs. Gibbs et al. (1995), using multi-unit surface EMG recordings, found that 4/10 subjects had short-term synchronization of motor units in medial gastrocnemius muscles bilaterally in standing which increased to 7/10 in a difficult balancing task. Common drive analysis examines the extent to which the firing rate of motor units modulates in parallel. The strength of common modulation is measured on a scale from ­1.0 (motor unit firing rate oscillations are negatively correlated) to 1.0 (motor unit firing rates modulate in unison). The purpose of the study was to examine whether motor units in bilateral soleus muscles shared common inputs during standing postural tasks.


Eight subjects (5 male and 3 female) participated in 16 experiments. Subjects were asked to stand quietly for 10 minutes with each foot on a separate force platform. Two tasks were performed: eyes open and eyes closed, in random order. Intramuscular fine wire electrodes were inserted into the lateral aspect of the soleus muscle of each leg. The classified spike trains (Spike2) were assessed during epochs of data in which there were no missed spikes and there was certainty of the accuracy of the motor unit classification.


There was little evidence of synchronization of soleus motor units between legs. However, there was evidence of common drive between legs. Ten motor unit pairs were recorded in both tasks. The common drive correlation coefficient was r = 0.59 ± 0.14 in the eyes open task and this decreased to r = 0.50 ± 0.16 in the eyes closed task.

To determine the influence of peripheral feedback on the level of common drive, a separate experiment was performed in which the Achilles tendon on one leg was vibrated for 5 minutes (100 Hz) so the sensory inputs would differ between legs. Figure 1 depicts the decrease in the common drive correlation function during vibration (red) in the eyes open (black) and eyes closed (blue) tasks in a single subject; the solid lines represent the data before vibration and the dashed lines depict the common drive after vibration ceased.


There was a moderate amount of common modulation in motor unit firing rate in soleus motor units between legs that was reduced in the eyes closed condition. Preliminary data suggest that the strength of the common modulation during postural tasks is influenced by peripheral feedback.

De Luca CJ, LeFever RS, McCue MP, Xenakis AP. Control scheme governing concurrently active human motor units during voluntary contractions. J Physiol 329: 129-142, 1982.

Gibbs J, Harrison LM, Stephens JA. Organization of inputs to motoneurone pools in man. J Physiol 485: 245-256, 1995.

Schmied A, Pagni S, Sturm H, Vedel JP. Selective enhancement of motoneurone short-term synchrony during an attention-demanding task. Exp Brain Res 133: 377-390, 2000.

Stevenson TJ, Garland SJ. Standing balance during internally produced perturbations in subjects with hemiplegia: validation of the balance scale. Arch Phys Med Rehabil 77: 656-662, 1996.


Motor unit activity differs with load type during fatiguing contractions

Carol J. Mottram, Jennifer M. Jakobi, John G. Semmler, and Roger M. Enoka

University of Colorado, Boulder, USA

As a strategy to identify the neural adjustments that occur during fatiguing contractions, we compared the performance of individuals on two tasks: a force task and a position task. The arm was rigidly connected to a force transducer during the force task and the subject was required to sustain a submaximal force (15% of maximum) with the elbow flexor muscles for as long as possible. The position task involved the subject supporting an equivalent inertial load and maintaining a constant elbow angle for as long as possible. The net muscle torque exerted by each subject was identical for the two tasks. Previous studies have demonstrated that the time to failure was about two times longer for the force task compared with the position task. Although the rate of increase in the average EMG of the elbow flexor muscles was similar for the two tasks, the rates of increase in perceived exertion, mean arterial pressure, and fluctuations in motor output were greater for the position task. The purpose of our study was to compare the discharge characteristics of the same motor unit (subcutaneous electrode) in biceps brachii during the two types of fatiguing contractions.

Subjects performed both tasks on the same day (random order) for a duration that was based on the recruitment threshold of the isolated motor unit. Thirty-two motor units were recorded during the force and position tasks. Both tasks were performed at 3.5% of maximum above the recruitment threshold of the isolated motor unit. The mean target force was 22.2 ± 13.4% MVC (range 3-49 % MVC). Contraction time was identical for both tasks (mean 2.7 min; range 0.75-5.5 min). Subjects were provided with visual feedback of force during the force task, and elbow angle during the position task. Although the rates of increase in EMG for the elbow flexor (P = 0.60) and antagonist triceps brachii muscles (P = 0.78) was similar for the two tasks (P > 0.05), the rates of increase in perceived exertion, mean arterial pressure, and fluctuations in motor output were greater for the position task (P < 0.05), despite the tasks not being performed until failure.

The discharge characteristics of the same motor units differed during the two tasks. Mean discharge rate declined at a greater rate during the position task (P = 0.28). Furthermore, discharge rate variability (coefficient of variation) did not change during the force task, whereas it increased progressively during the position task (P = 0.018). This difference in rate coding between the two tasks was accompanied by variable amounts of motor unit recruitment. In addition to the 32 motor units recorded during both tasks, the subcutaneous electrode detected the recruitment of another 32 units during the force task and another 46 units during the position task. Twenty-six of these units were recruited during both tasks. These data demonstrate that motor units in biceps brachii experience greater adjustments in discharge activity during the position task compared with the force task, despite similar increases in the average EMG during the two tasks. In conclusion, the rate of increase in central neural activity was more rapid for the position task even though the the two tasks required a similar net muscle torque.

Figure legend. Change in discharge rates of the same motor units (n = 32) during the force and position tasks. Mean discharge rate declined more (left) and discharge rate variability (coefficient of variation) increased more (right) during the position task compared with the force task.



Effects of strength training on motor unit control properties in the elderly

Zeynep Erim1,2, Kathleen M. Brady3, Jens Meyer3, Alexander Adam3, David T. Burke4, Carlo J. De Luca3, and Winsean Lin1

1Rehabilitation Institute of Chicago, 2Northwestern University, 3Boston University, and 4Harvard University

Exercise-induced strength increases that far exceed the levels that can be explained by hypertrophy in elderly individuals have emphasized the role of neural adaptation following strength training in this population. The goal of this study was to determine whether age-related changes in the common firing behavior of motor units observed previously in the First Dorsal Interosseous (FDI) muscle was a neural parameter significant in determining the force output, and whether this parameter was ³corrected² via exercise, in turn contributing to the reported strength gains.

18 healthy subjects (aged 67 to 75 years) participated in a 10-week strength training program that included isokinetic exercise at 80% of 1 repetition maximum (1RM) 3 times/week. At baseline and at the end of the 10-weeks, intramuscular recordings were made to assess motor unit control properties. Intramuscular EMG was recorded in the Vastus Lateralis (VL) muscle during isometric knee extensions with trapezoidal trajectories reaching either 20%- or 50%-MVC. Additionally, surface EMG was recorded from the VL, Vastus Medialis, Rectus Femoris, Biceps Femoris.

Isokinetic force was significantly increased, while the variability in the isometric force was significantly decreased following the training program. Among the mean interfiring interval, the coefficient of variation of the interfering interval, the commonality of firing patterns as measured by the common drive index (peak of the cross-correlation function between the smoothed mean firing rates), and motor unit synchronization, the common drive index was the only parameter that displayed a significant training-related difference that accompanied the increases in strength and smoothness of force output. Antagonist activation was not a major contributor to the force gains as the Biceps did not display any considerable activity either before or after the training. Likewise, unaltered patterns of overall muscle activation in the VL, VM, and RF following the training program suggested that synergist cocontraction was not a means of improving force or its smoothness. A parallel study which revealed that in the VL the common drive index was increased in the elderly as compared to the young led to the conclusion that the training-related decrease in the common drive index effectively moved the motor unit firing patterns closer to those seen in the young.


Persistent inward currents in human motoneurons and its possible facilitation in monoamines

Monica Gorassini and Jennifer Nevett-Duchcherer

University of Alberta, Edmonton, Canada

Monoamines, such as serotonin (5-HT) and nor-adrenaline (NA) have been shown in animal models to decrease the threshold for and increase the amplitude of persistent inward currents (PICs) in motoneurons (Heckman et al. 2003). In this presentation, we provide evidence which suggests that similar facilitation of motoneuron PICs by increasing the endogenous release of NA via oral dextro-amphetamines can also occur in humans. To estimate the contribution of PICs to motoneuron activation we employed a paired motor unit technique where the difference in synaptic drive required to recruit a test motor unit during a voluntary contraction was compared to the synaptic drive just prior to its de-recruitment (Gorassini et al. 2002). This difference in synaptic drive to the test unit, or DF, was provided by the firing rate profile of a slightly lower-threshold control motor unit firing throughout the activation period of the test motor unit. In 3 subjects tested to date, following 25mg ingestion of dextro-amphetamine, DF (or our estimation of PIC amplitude), increased from 4.4 Hz on average before the drug to 7.2 Hz after drug intake. In addition, in one subject, decreases in firing variability during very low firing rates (5Hz) also occurred after drug intake, suggesting increases in activation of sub-threshold sodium-dependent PICs which has been shown in rat tail motoneurons to provide stable, low frequency self-sustained firing of motoneurons (Li et al. 2004).

Gorassini M, Yang JF, Siu M, Bennett DJ. Intrinsic activation of human motoneurons: possible contribution to motor unit excitation. J Neurophysiol 87: 1850-1858, 2002

Heckman CJ, Lee RH, Brownstone RM. Hyperexcitable dendrites in motoneurons and their neuromodulatory control during motor behavior. Trends Neurosci 26: 688-695, 2003.

Li Y, Gorassini MA, Bennett DJ. Role of persistent sodium and calcium currents in motoneuron firing and spasticity in chronic spinal rats. J Neurophysiol 91: 767-783, 2004


Motor unit behavior during spasms of thenar muscles paralyzed by spinal cord injury

Christine Thomas and Inge Zijdewind*

University of Miami, Florida, USA and *University of Groningen, The Netherlands

Trivial stimuli such as a change in limb position or light touch of the skin commonly trigger spasms in paralyzed muscles of individuals with spinal cord injury (SCI). Our aim was to compare modulation of motor unit firing rates during thenar muscle spasms to that seen during voluntary contractions performed by individuals with cervical SCI and uninjured control subjects. Women and men (21-55 years) with chronic (>1 yr) SCI at C4, C5, C6 or C7 were evaluated. Subjects sat in their wheelchair or on a chair with the test forearm resting in a vacuum cast and on a support. The hand was stabilized in Theraputty. Surface and intramuscular EMG was recorded from the thenar muscles. Force was recorded in the abduction and flexion directions at right angles. Resultant force was calculated. Muscle spasms were triggered by stretching the contralateral shoulder, a change in body position, light touch of the skin or forced respiratory expulsion. The strength of these involuntary contractions was normalized to the force evoked by supramaximal stimulation of the median nerve at 50 Hz.

Thenar muscle spasms typically lasted a few seconds and involved a rise and fall in thenar muscle surface EMG and force. Significant positive correlations were found between thenar surface EMG and force. During spasms, motor units were recruited up to high forces (>95 % of maximal muscle force). Motor unit recruitment also occurred to high forces in thenar muscles that could be controlled voluntarily after cervical SCI (Zijdewind and Thomas, J Neurophysiol 89: 2065-2071, 2003), whereas most unit recruitment is complete by 30-40 % maximal voluntary force in intrinsic hand muscles of control subjects (Kukulka and Clamann, Brain Res 219: 45-55, 1981). During spasms, the force at which units were recruited was higher than the derecruitment force for 52 % of units. In contrast, most units (74 %) had higher recruitment frequencies than derecruitment frequencies, in part because some units fired for prolonged periods at low frequencies (5-6 Hz) after the decline in surface EMG and force.

The spasms varied in intensity, ranging from 4 to 115 % of the maximal evoked thenar force, the additional force presumably arising from activation of other muscles. In general, motor unit firing rates increased as the strength of the spasm rose, but the extent of rate modulation varied for different units. For example, three motor units recruited at < 20 % maximal force during the same spasm modulated their firing rates by 16 %, 109 % or 187 % by the time the spasm reached maximal muscle force. During weak, medium and strong spasms (forces reached 5 %, 25 % or >95 % maximal force), mean (± SD) firing rates were 5 ± 2 Hz, 10 ± 4 Hz and 15 ± 5 Hz, respectively. In contrast, minimal modulation of motor unit firing rates occurred during voluntary thenar contractions performed by other individuals with cervical SCI, with mean rates of 8 ± 2 Hz and 9 ± 3 Hz achieved during 25 % and 100 % maximal voluntary contractions, respectively (Zijdewind and Thomas, J Neurophysiol 89: 2065-2071, 2003). In control subjects, units fired at 8-12 Hz during the weak contractions used to average unit forces by spike triggered averaging, but at 34 ± 10 Hz during maximal voluntary contractions (Thomas, J Electmyogr Kinesiol 7: 15-26, 1997). Thus, rate modulation during thenar muscle spasms exceeds that seen during voluntary contractions of thenar muscles influenced by spinal cord injury, but is reduced compared to that seen in uninjured subjects.

When pairs of units were coactive during spasms, the lower threshold unit (control unit) was derecruited before the higher threshold unit (test unit) in 13 % of cases (recruitment thresholds of these pairs of units varied from 2-26 % maximum force). In the remaining unit pairs, the control unit was firing at a lower frequency when the test unit was derecruited compared to recruited in 53 % of cases. The difference in the recruitment versus derecruitment frequency of the control unit ranged from 0.2 Hz to 7.9 Hz. These data are consistent with the idea that sustained depolarization in the test unit prolonged its firing. In the other 47 % of cases, the control unit was firing at a higher frequency when the test unit was derecruited versus recruited (frequency difference from -0.1 Hz to -5.8 Hz).

The higher motor unit firing rates attained during maximal spasms compared to maximal voluntary contractions of thenar muscles influenced by chronic cervical spinal cord injury suggest that the motoneurons are capable of responding to inputs and that the voluntary deficits relate to weak descending drive. Less motor unit rate modulation was seen during maximal spasms and voluntary contractions of spinal cord injured subjects than during maximal voluntary contractions performed by control subjects, emphasizing the importance of motor unit recruitment, sometimes to high forces, in both the involuntary and voluntary contractions of thenar muscles after chronic cervical spinal cord injury.

Funded by USPHS grant NS-30226, The Miami Project to Cure Paralysis and the University of Groningen.


Poster Presentations


Short- and long-latency reflex responses in young and elderly subjects

Malgorzata Klass, Stéphane Baudry, and Jacques Duchateau

Université Libre de Bruxelles, Belgium

Ageing is associated with a loss of motor control (Vandervoort, 2002; Enoka et al, 2003). These alterations cause instability during locomotion, with an increased risk of falling due to less efficient regulation of equilibrium. Since the early muscle reaction to postural disturbance has a reflex origin (Schieppati & Nardone, 1997), the latency and magnitude of the reflex responses are important parameters. The aim of this study was to investigate the short- (SL) and long- (LL) latency reflex responses to stretch in the tibialis anterior and to compare the relative changes in SL amplitude with those of the Hoffmann (H) reflex. The latter, bypassing the muscle spindle, allows the investigation of the neural circuitry of the reflex loop.

The experiments were carried out on 17 young (20-35 yr) and 14 elderly subjects (70-87 yr). During the experiments, the subjects sat on a seat with the dominant foot strapped (ankle joint at 90°) to a footplate fixed on the rotational axis of a motorised ergometer. The torque produced during isometric voluntary contractions of the dorsiflexors and corresponding electromyographic activity of the tibial anterior were recorded. The degree of muscle activation during a maximal voluntary contraction (MVC) was assessed by the twitch interpolation method (paired stimulation). H reflexes, evoked by electrical stimulation of the common peroneal nerve (intensity 15% above motor threshold), and reflex responses to a quick stretch (range of motion : 10°; angular velocity : 200°/s) were recorded in the tibialis anterior. Both reflexes were recorded during a sustained contraction of 10% MVC. To compare the young and elderly subjects, all reflex activities were averaged over time (40-50 responses), the background EMG activity (bgEMG) was substracted from the reflex responses to obtain the reflex activity (reflexEMG). The value was then normalised to the bgEMG (reflexEMG / bgEMG).

Elderly and young subjects were able to fully activate the ankle dorsiflexors muscle since the interpolated method did not induce any force increment. The normalised SL and LL reflex activities of the tibialis anterior to a quick stretch were affected differently with ageing (Fig. 1, B). SL and LL were reduced by 55 % (P<0.001) and 29 % (P<0.05), respectively. The H reflex (Fig. 1, B) amplitude decreased to a similar extent compared with SL (48%; P<0.001). These changes in amplitude were associated with an increase in latencies (Fig.1, A). Compared with the young participants, the elderly subjects displayed a significant (P<0.05) and similar increase in latency (12% and 11% for H and SL respectively). The latency of the LL component was only prolonged by 6 % (P<0.05).

Figure 1: Normalised reflex amplitudes and latencies in young and elderly subjects

In agreement with the study of Corden and Lippold (1996), the amplitude of the SL response to stretch changed to a greater extent with ageing, compared with the LL response. This observation suggests that the supraspinal component of the stretch reflex is less prone to the effects of ageing. The finding that the average amplitude of H and SL reflex responses were reduced to a similar extent with ageing indicates that the decrease in reflex amplitude is not due to a change in muscle spindle responsiveness, but is rather related to an impairment of the neural mechanisms in the spinal reflex loop. Since muscle activation was maximal in young and elderly subjects (see also De Serres and Enoka, 1998), the main mechanisms should be located at a pre-motoneuronal level and might be due to a less effective synaptic transmission and/or increased presynaptic inhibition (Morita et al, 1995). The longer reflex latencies in elderly subjects should be related to a slowing of nerve conduction velocity (Vandervoort, 2002) and a less effective synaptic transmission (Scaglioni et al, 2002).

In conclusion, although both spinal and supraspinal reflex mediated-inputs onto the motoneurone pool were reduced with ageing, it appears that the alterations result mainly from pre-motoneuronal mechanisms located in the spinal reflex loop.

Corden and Lippold, J Neurophysiol 76 :2701-2706, 1996
De Serres and Enoka, J Appl Physiol 84: 284-291, 1998
Enoka et al, J Electromyogr Kinesiol 13: 1-12, 2003
Morita et al, Exp Brain Res 104 : 167-170, 1995
Scaglioni et al, J Appl Physiol 92 : 2292-2302, 2002
Schieppati and Nardone, J Physiol (Lond) 503 : 691-698, 1997
Vandervoort, Muscle Nerve 25 : 17-25, 2002


Fatigue of human thenar motor units paralyzed by spinal cord injury

CS Klein, CK Häger-Ross*, CK Thomas

University of Miami, Florida and *Umeå University, Sweden

Thenar muscles paralyzed chronically by cervical spinal cord injury are more fatigable than muscles of able-bodied individuals (1). No previous studies have examined this greater fatigability at the level of the motor unit. In this study our aim was to examine the changes in the surface electromyographic activity (EMG) and force in paralyzed thenar units using intraneural motor axon stimulation (2). Seven subjects (2 women, 5 men; mean ± SE age; 36 ± 4 years) participated in the study. All had chronic cervical spinal cord injury (mean injury duration: 10 ± 2 years) and complete paralysis of the thenar muscles. Subjects reclined on a bed. The arm and hand were supported, and the thumb was positioned against a transducer to measure isometric forces in abduction and flexion directions. Resultant force was calculated. EMG was recorded with surface electrodes over the proximal and distal thenar muscles. Units were fatigued with 13 pulses at 40 Hz 1/s for 2 min. The 40 Hz force, half-relaxation time, and EMG parameters of the first and last pulses in a train (latency, duration, amplitude and area) were measured every 20 s. Before and after fatigue, twitches and tetanic forces at frequencies of 5-100 Hz were recorded.

After 20 s of the fatigue test, the mean latency, duration, amplitude and area of the first EMG potential of the train increased (P < 0.001, n = 17 units). Similarly, the mean latency and duration of the last EMG potential in the train increased during fatigue. However, the mean amplitude and area of the last EMG potential in the train decreased significantly (in 16 of 17 units and 12 of 17 units, respectively). Despite this within-train decrement, the rest period between trains was adequate for the first potential of the next train to recover or potentiate. Force decreased in all units, from an initial mean ± (SE) of 66.7 ± 16.8 mN to 24.1 ± 7.3 mN at 120 s (P < 0.001). The mean force fatigue index (final force/initial force) was 0.34 ± 0.04 (range; 0.08-0.60), which was significantly lower than the fatigue index recorded in able-bodied subjects (0.78 ± 0.03; P < 0.001) (3). The fatigue indices for the paralyzed units were positively correlated with initial twitch contraction times (R2 = 0.35) and half-relaxation times (R2 = 0.22) (P <0.05). Thus, units with shorter contraction and half-relaxation times were more fatigable. The force fatigue index was not significantly correlated with fatigue indices (120 s/0 s values) for EMG latency, duration, amplitude or area (first or last potentials of the train), or the initial twitch or tetanic forces, axon conduction velocity, normalized tetanic contraction or relaxation rates.

Twitch force was reduced, and twitch contraction and half-relaxation times were increased, in most units after fatigue. Tetanic forces were reduced at all stimulus frequencies in 8 of the 10 units in which measurements could be made. The frequency necessary to produce 50 % of maximal force increased from an initial mean of 9.8 ± 1.9 Hz to 13.1 ± 1.5 Hz post-fatigue, reflecting an increase in 7 of 10 units. Thus, in most units, higher stimulus frequencies were needed to evoke the same relative force after just 2 min of stimulation.

These results show that 2 min of stimulation in paralyzed units alters the EMG potential, both over time and within a train. However, the changes in EMG latency, duration, amplitude or area were not significantly correlated with the reduction in force during fatigue. The decrease in force combined with potentiation of the first EMG potential in the train during fatigue, and the altered contractile properties post-fatigue, suggest that excitation-contraction uncoupling, contractile failure, or both processes contribute to the excessive fatigability. The lower fatigue index in paralyzed compared to control units (3) suggests that chronic paralysis results in greater intrinsic fatigability of the muscle fibers.

1. Thomas et al. J Neurophysiol 89;2055-2064, 2003
2. Westling et al. J Neurophysiol 64:1331-1338, 1990
3. Thomas et al. J Neurophysiol 65:1509-1516, 1991

Funded by USPHS grant NS-30226, The Miami Project to Cure Paralysis, the Swedish Medical Research Council, and the University of Umeå.


Discussion Session

Summary Discussion of Animal Studies by Marc Binder (University of Washington, Seattle, USA)

What do we think we know about the active properties of motoneuron dendrites?

  • Dendrites are richly endowed with voltage-gated Na+ and Ca2+ channels that generate persistent inward currents (PICs).
  • PICs are triggered by ionotropic excitatory synaptic inputs.
  • PICs can generate a substantial portion of the depolarizing drive to motoneurons required for repetitive firing.
  • The channels mediating the PICs are modulated by monoamines.
  • The PICs can be turned off by ionotropic inhibitory synaptic inputs.

What don't we know about the active properties of motoneuron dendrites?

  • Where are the voltage-gated channels located?
  • Can the dendrites be activated in a graded fashion?
  • Can the dendrites be turned on and off rapidly?
  • What "integration modes" can dendrites support?

How will we progress?

  • Immuno-fluorescent labeling of channels
  • Opto-physiological-pharmacology
  • Modeling and simulation

Other input-output problems

  • Role and modulation of transient and persistent Na+ channels in control of motoneuron activation threshold, repetitive firing and spike-frequency adaptation.
  • Resolving the contribution of discharge history to motoneuron input-output functions.
  • Calcium buffering-sequestering: how can motoneurons use Ca2+ for depolarizing drive and not lose control of Ca2+-mediated intracellular cascades? (Do the dendrites use mitochondria as the muscle cell uses the lateral cisternae of the SR?)

Understanding plasicity

  • Discover components of the signal-transduction cascades linking altered motor activity to changes in motoneuron intrinsic properties and synaptic input efficacy.
  • Discover components of the signal-transduction cascades linking changes in the inputs to motoneurons to changes in their excitability and discharge properties.

How can our joint meetings facilitate the work?

  • What measurements made from reduced animal preparations would be most helpful to those working on human subjects? (e. g., testing the Matthews' transform; Powers & Binder, J. Physiol. 28:2000)
  • What kinds of experiments performed in man would be most interesting to those working on reduced preparations? (e. g., firing rate saturation; Fuglevand & Johns, this meeting)

Summary Discussion of Human Studies by Hans Hultborn (University of Copenhagen, Denmark)

What is so special with motoneurones in relation to all other central neurones?

  • We know their function - and can interpret their firing relation to muscle properties and muscle contraction.
  • Keep the relation to “motor control” - CNS and muscle!
  • And keep the bridge between “animals and man”!
  • Studies in humans are possible as the activity of individual motoneurones can be studied as motor unit action potentials

At the 1st Boulder meeting Peter Matthews and Daniel Kernell were here - much reference to their work was done at this meeting. Some additional citation by a recent publication by Peter Matthews (2004) deserves quotation here:

  • “Effective science depends on tackling problems that are soluble with the means at hand”
  • “In human studies, collection of the data is easy but the interpretation is fraught with pitfalls”
  • In human studies: “ ..the shared input analysis has been studied exhaustively to little enduring purpose” (!!!)
  • “Transcranial magnetic stimulation to measure ‘cortical excitability’ represents a sad neglect of the underlying complexities” (!!!)

Rate limitation in humans and cats: in humans there is a strong saturation of motor unit firing rate - even at frequencies lower than needed to elicit maximal force responses (see Fuglevand & Johns). Why? Differences to the cat?

  • It was agreed that this paradoxical difference should be solved before the 3rd Boulder meeting.
  • In human subjects it would be worthwhile to study whether additional excitation by corticospinal volleys (transcranial stimulation) is occluded.
  • In animal experiments it should be tested whether the frequency increase to intracellular current injections are “occluded” when given during firing evoked by synaptic excitation - as during fictive locomotion or scratching, when the “plateau current” is likely to be activated

Active dendrites in humans? What are the criteria?

  • Evidence for plateau potentials in human motoneurones is obtained by “paired motor unit recordings” (Kiehn O, Eken T. J Neurophysiol. 1997;78:3061-8. , Gorassini M, Yang JF, Siu M, Bennett DJ. J Neurophysiol. 2002;87:1850-8)
  • Gorassini et al. (2002a,b) have recently considerably extended work on human MNs by quantifying the paradigm of paired motor-unit recording.
  • Problems?: (1) Common synaptic drive to test and control units??; (2) And the PICs are graded - not all-or-non. How is that affecting the experimental paradigm? (3) Is rise and fall of the triangular isometric contraction the same - just with opposite sign?? Certainly much more difficult to relax than to contract.

Do plastic changes in motoneuron properties explain spasticity in humans?

  • The work by Bennett, Gorassini and collaborators in the chronic spinal rat (paralyzed and spastic tail) is very convincing.
  • It has high priority to use the indirect and circumstantial methods now available to give evidence on whether human spasticity may also involve enhanced “plateau currents” in the motoneurones.

Return to the Meetings page.

Some of the photographs: