Comparison of peripheral Ia and corticomotoneuronal composite EPSPs in human motoneurons

Transcription

1 ELSEVIER Electroencephalography and clinical Neurophysiology 101 (1996) Comparison of peripheral Ia and corticomotoneuronal composite EPSPs in human motoneurons Kelvin E. Jonesa,*, Blair CaIancieb, Anthony HaIlb, Parveen Bawaa 'School of Kinesiology, Simon Fraser University, Burnaby, B,C" V5A 1S6, Canada bmiami Project and Department of Neurological Surgery, University of Miami School of Medicine, 1600 NW 10th Avenue, Miami, FL, 33136, USA Accepted for publication: 21 June 1996 Abstract The effects of excitatory inputs arising from Ia afferent and corticomotoneuronal volleys on repetitively firing flexor carpi radialis (FCR) motoneurons were compared in normal human subjects. Peripheral (Ia) volleys were produced by transcutaneous electrical stimulation of the median nerve and by mechanical taps to the FCR tendon. Transcranial magnetic stimulation (TMS) was used to activate the corticomotoneuronal pathway. The duration of the excitatory response peaks measured from peri-stimulus time histograms (PSTHs) and the time course of the response trajectories were both taken to reflect the shapes of the underlying composite excitatory postsynaptic potentials (EPSP)s acting upon that motoneuron. The duration of excitatory response peaks for the H-reflex and the first sub-peak (SPJ of the motor unit's response to TMS were similar and were typically less than those arising from tendon taps. The response trajectories, which measure the excitability of the motoneuron during different phases of the afterhyperpolarization, overlapped for H-reflex and SP! responses, but were different for tendon tap inputs. Our results indicate that the SP] response of a motoneuron to TMS input and its response to near-synchronous Ia afferent activation are mediated by composite EPSPs with similar rise times. We suggesthat a similar spatial distribution of synaptic boutons for J>oth Ia and corticomotoneuronal input to motoneurons innervating FCR is likely. Keywords: Excitatory postsynaptic potentials; Ia afferent; Motoneuron 1. Introduction Since its introduction, the use of transcranial magnetic stimulation (TMS) to activate the human motor cortex has proliferated. With this technique have come questions about the central processes involved in the mediation of the excitatory and inhibitory effects of this descending volley on spinal motoneurons. Crosscorrelation of the TMS input with the spike train of a repetitively firing single motor unit (SMU) reveals that the descending volley produces multiple peaks of increased firing probability in the peri-stimulus time histogram (Day et al., 1989; Olivier et al., 1995). The multiple peaks are most likely due to multiple descending volleys in corticospinal tract neurons * Corresponding author. Goteborgs Universitet, Fysiologiska Institutionen, Medicinaregatan II, Goteborg, Sweden. Tel.: ; fax: ; kelvin@sfu.ca arising from a single magnetic stimulus (Edgley et ai., 1990; Burke et ai., 1993). However, whether the multiple volleys produced by TMS are caused by direct or indirect activation of pyramidal tract neurons remains an open question (Rothwell et ai., 1991). Irrespective of direct or indirect activation of the descending pathway, it is generally believed that the initial response of an SMU to TMS is mediated through corticomotoneurons which make monosynaptic connections with motoneurons (palmer and Ashby, 1992; A wiszus and Feistner, 1993; Priori et ai., 1993; Awiszus and Feistner, 1994b). Using spike train analysis techniques, researchers have estimated the parameters of the composite EPSP produced by TMS (palmer and Ashby, 1992; A wiszus and Feismer, 1994a). The aim of this paper is to present further data on the shape of the composite EPSP produced by TMS in a detailed comparison to the composite Ia EPSP in normal flexor carpi radialis (FCR) motoneurons. It 1996 Elsevier Science Ireland Ltd. All rights reserved PII SO92l-884X(96)95l78-4 EEM 95178

2 432 K.E. Jones ef al. / Electroencephalography and clinical Neurophysiology 101 (1996) would be helpful to know how these two monosynaptic connections compare when fonnulating possible roles for descending and peripheral inputs in the control of forearm muscles. A portion of this work has been presented as an abstract (Jones et al., 1995) and as part of a Pill thesis (Jones, 1995). 2. Methods Experiments were conducted on 8 right-handed subjects (6 male and 3 female, including all authors) ranging in age from 24 to 51 years. Experimental protocols were approved by the 'Human Ethics' committees at the two institutions, and all subjects gave their informed consent. SMU activity was recorded from the right flexor carpi radialis (FCR) muscle with higwy-selective, bipolar needle electrodes and surface EMG was recorded over the muscle belly using Ag-AgCl disc electrodes (Grass Instrument Co.). A subject was asked to recruit a low threshold SMU by making a weak, isometric contraction, and then maintain the unit's discharge at a relatively slow but steady rate using visual and auditory feedback. The motor unit shape was discriminated to generate a trigger pulse (BAK Electronics, Inc.). Spike-triggered averaging of the surface EMG from FCR was used to obtain the shape of the motor unit's potential to minimize the risk of re-sampling the same motor unit from a different needle electrode position (Jones and Bawa, 1995). Ia volleys were produced either by an electrical stimulus applied to the median nerve through a bipolar felt electrode placed at the cubital fossa, or by delivering a tap to the FCR tendon at the wrist via a moving coil linear motor (Model 310, Cambridge Technology, Inc.). Electrical stimuli to elicit the H-reflex were obtained from 0.5 ms square-wave pulses AC coupled to isolation and constant-current units (Grass Sill5 and CCU). The current was adjusted to elicit an H-reflex with minimal M-wave. The contact surface of the hammer percussing the FCR tendon was 5 mm wide and consistent positioning throughout the experiment was maintained by clamping the subject's forearm and hand in a fixed position. Taps were elicited in a length-controlled fashion with a square wave input of 3.0 ms duration and an amplitude adjusted so that the surface EMG reflex response (peak-to-peak) matched that elicited by electrical stimulation during a given experiment. Descending corticospinal volleys were gerated by transcranial magnetic stimulation (TMS) of the motor cortex using a circular coil of 9.5 cm outer diameter (Cadwell MES-lO). In all cases the coil was positioned along the midline, with the center approximately 1 cm anterior to the vertex (Cz). For all inputs, the minimal interval between successive stimuli was never less than 3.5 s. Peripheral and TMS inputs were delivered in one of two modes: (1) random with respect to the discharge of the SMU (random); or (2) at a fixed interval following the motor unit's discharge (triggered-delay). The delay in the latter mode was varied between 1 and 60 ms (see Jones and Bawa, 1995 for details). The average number of stimuli given for each trial was 150 in the random mode of stimulation and 45 in the triggered-delay mode. The sampling rate for data acquisition was 10kHz (Spike2; Cambridge Electronics, Inc.). The peri-stimulus time histograms (psths), bin widths of 0.1 ms, showed periods of increased firing probability (a peak) with onset latencies ranging from 12 to 25 ms following the stimulus. The onset and offset of the peak in the PSTH were defined as the first and last bins in a consecutive series of bins whose values exceeded the mean plus two standard deviations of the background firing probability. The onset of the PSTH peak was further confirmed by comparing it to the onset of the positive slope in the CUSUM of the PSTH. The total duration of increased firing probability, termed peak width, was determined from lhe onset and offset of the peak and compared for the two sources of excitatory inputs. Multiple subpeaks in the PSTH resulting from TMS input were easily distinguished using bin widths of 0.1 ms as the peaks were most often separated by 5 or more empty bins. Since we were using relatively large inputs, as estimated from response probability values, we were confident that the first subpeak was SP\. Only with much smaller inputs does the SP\ subpeak disappear (Olivier et al., 1995). Response probability was calculated as the number of counts in the peak above background, divided by the number of stimuli given. In the triggered-delay mode, response probability was plotted at sequential delays with respect to the arrival of the input volley following the motoneuron spike to construct a response trajectory (Jones and Bawa, 1995). To estimate the rising phase of the underlying EPSP, a cumulative bin integration was performed over all the bins of the peak obtained during random stimulation, starting with x = 1 as the first bin of the excitatory peak (Homma and Nakajima, 1979). The cumulative values at each bin were normalized with respect to the total area of the peak (number of counts in the peak) and the resulting normalized values were fitted, by the method of least mean square, to the following equation (Homma and Nakajima, 1979): where x is the bin number (0.1 ms per bin), s represents slope, RT is the time constant and Y(x) is the nonnalized cumulative value at bin x. The time constant, RT, determines the duration of the rising phase of the function and, therefore, approximates the rise time of the underlying EPSP. The integration of the PSTH peak, the peak width and the response trajectory were all taken as estimates of the composite EPSP acting upon a motoneuron. Statistical analysis of the data involved the use of paired t tests when comparing SMUs receiving excitatory input from two sources, and independent t tests when comparing

3 K.E. Jones et al. f Electroencephalography and clinical Neurophysiology 101 (1996) the populations of SMUs receiving either H-reflex or TMS input. Significance is reported for values of P < Results Data are reported for 31 flexor carpi radialis SMUs recorded from 8 subjects. Twelve of these SMUs were studied with the triggered-delay mode of stimulation with each motor unit receiving both H-reflex and TMS inputs. Six of these motor units were also studied with random stimulation: two received both H-reflex and TMS inputs, 3 received TMS only and one received H- reflex input alone. Thirteen units were studied with random stimulation alone with 11 of those receiving H-reflex and two receiving TMS input. An additional 6 SMUs were studied in the triggered-delay and random mode using both H-reflex and tendon tap input to each motor unit. The mean interspike interval (ISI) of SMUs receiving H-reflex input was 108.2:t 3.0 ms (:tsem) while those receiving corticomotoneuronal input had a mean ISI of :t 4.0 ms. The ISIs (inverse of firing rate) were not significantly different for the two inputs. During random stimulation, the onset latency of the excitatory peak with respect to the stimulus ranged ms (19.6 :t 1.4 ms) fortms, ms (21.0:t 0.9 ms) with H-reflex input and ms (27.8:t 3.7 ms) with tendon taps. '".J O",, 'ost-stimulus Time (ms 53 Bin Number (0 IllS, TMS '-'SP2. Fig. 1. PSTHs of a SMU in response to H-reflex and TMS inputs. The top part of the figure shows the PSTHs generated by H-reflex (left) with peak onset at 19.4 ms (n = 293 stimuli) and the TMS-evoked PSTH with peak onset at 16.8 ms (right; n = 326) during random stimulation. The SMU was firing with a mean ISI of 100 ms during both conditions. The peaks of the PSTHs are shown on an expanded time scale in the bottom panels of the figure showing the H-reflex peak width of 1.5 ms and the TMS peak width of 5.2 ms. The TMS peak is composed of two subpeaks, SP1 and SPb with the SP1 peak width of 1.5 ms. The open circles are normalized cumulative bin values for the peaks. The solid lines overlaying the peaks are an estimate of the rising phase of the EPSP obtained by fitting Eq. (1) to the open circles (see Section 2). The parameters for the equation are: H-reflex, s = 2.85, RT = 13.6 bins; SPh s = 2.80, RT = 14.0 bins. IU Q) 10.0( IU Q) '" 50 IU.. 0 A B H-reflex SPl of TMS Bin Number (0.1 rob/bin) 433 Fig. 2. Averaged estimate of the rising phase of the composite EPSP. Nonnalized cumulative data points for 14 peaks for H-refiex (A) and 7 peaks for SPt (B) are shown by the open circles. Eq. (I) was fit to these data points to obtain an average rising phase of the respective EPSP. The results of the curve fit are shown by the solid line which is plotted for values up to x = RT. For H-refiex input (A), the resulting fitted curve had the parameters of s = 2.9, RT = 13.5 bins where the curve fit had an r = 0.91 with a standard error of the estimate (SEE) = 14.9%. For the SP] component of the TMS input (B), the resulting fitted curve had the parameters of s = 2.6, RT = 15.1 bins where the curve fit had an r = 0.88 and an SEE = 16.1 %. Therefore, the fitted curves estimate the rise times of the composite EPSPs to be 1.35 and 1.51 ms for H-refiex and SPt, respectively Random stimulation Random stimulation resulted in short latency excitatory peaks in the PSTHs whose total widths varied between 0.8 and 1.6 ms (1.25:t 0.06 ms) for H-reflex, and between 3.6 and 5.2 ms (4.43:t 0.23 ms) for TMS. However, the excitatory peaks resulting from TMS were generally segmented into 2-3 subpeaks: SP\, SP2, SP3 (Day et al., 1989; Bawa and Lemon, 1993; Olivier et al., 1995). We limited our investigation to the SP\ response in comparing peripheral and corticomotoneuronal inputs. The mean peak width of the SP\ response was 1.33:t 0.08 ms and ranged between 1.0 and 1.6 ms. The values for peak widths in response to H-reflex input were not significantly different from those of the SP\ response. The response of a motor unit to both H-reflex and TMS input during random stimulation is shown in Fig. 1. The top part of the figure illustrates the level of background firing probability prior to the response peak. The motor unit's mean discharge rate was 10 imp/s prior to stimulation with either input. After a latency of 19.4 ms following H-reflex stimulation, the firing probability of the umt increased sharply, fonning a peak in the PSTH. e latency of the first subpeak following TMS was 16.8 ms for this same unit. In the bottom part of Fig. 1, the peaks of the PSTHs are shown on an expanded time scale illustrating a peak width of 1.5 ms for both the H-reflex and SP\ of the TMS input. A cumulative bin integration was perfonned and the area in each bin of the peak was plotted as a percentage of the total area of the peak (open circles). These data were then fitted by Eq. (1) as a means of estimating the shape of the underlying EPSP and the results

4 K.E. Jones et al. / Electroencephalography and clinical Neurophysiology 101 (1996) are illustrated as the lines overlaying the experimental data (Fig. 1, bottom). The solid curve represents an estimate of the rising phase of the composite EPSP acting upon this motoneuron. The subsequent broken line represents the solution of Eq. (1) beyond experimental points and is shown only to facilitate visualization of our data with respect to the profile of an EPSP as recorded in the cat. The SPz response to TMS has been ignored in the curve fit. Curve fitting with Eq. (1) resulted in similar estimates of the rising phase of the composite EPSP in this unit to H- reflex and SP1 of the TMS response. Eq. (1) was then fitted to all the data collected during random stimulation to estimate the average composite EPSP elicited by the two inputs. The data points (open circles) in Fig. 2A represent the cumulative bin integration for 14 SMUs in response to H-reflex stimulation, and in Fig. 2B for 7 SMUs in response to TMS input. The calculated time constant for the curve describing the H-reflex responses is RT = 13.5 bins while that for the curve describing the TMS input is RT = 15.1 bins. The rise time estimates of 1.35 ms and 1.51 ms for responses to H-reflex and TMS, respectively, are slightly higher than the mean peak widths for H-reflex and the SP\ as reported above, but fall within one standard deviation of the mean. Triggered delay stimulation Motor unit responses to the two inputs were then examined using triggered-delay stimulation. This allowed comparison of the response of a motor unit to the two inputs at various times during the afterhyperpolarization (AHP). At the longest delay tested, response probability and peak.. > H 11. QI 0. 5 III 0 0. III QI I.: 0.0 Fig. 3. Response trajectories and the influence of peak width. (A) Response trajectories in an SMU are shown for two inputs, tendon tap (closed circles) and H-reflex (open circles). The SMU was firing repetitly with an ISI of 105 ms during taps and 110 ms during H-reflex. The two trajectories started at approximately the same probability values at the longest delay. As the delay was shortened, the trajectories deviated from each other. Response probability fell sharply for the short dtion H-reflex peak (1.5 ms) compared to that for the longer duration peak (4.4 ms) with tendon taps. (B) Response trajectories of an SMU subjected to H-reflex (open circles) and TMS (open squares) input during repetitive firing with a mean ISI of 95 ms during both conditions. The trajectories for H-reflex and SPt start from a similar response probability at the longest delays and overlap throughout the course of the trajectory, suggesting similarities in the underlying EPSPs. Total delay is equal to the triggered-delay plus the time to onset of the response peak. width had their highest values. The mean peak widths at this delay, 1.23 :t 0.09 ms for H-reflex and 1.20 :t 0.10 ms for TMS (SPJinputs, were not significantly different for the two inputs. These values of peak width obtained with triggered-delay stimulation were compared to those obtained with the same input during random stimulation and no significant differences were found (n = 12). The relationship between the response probability to inputs induced at sequential delays during the AHP has been called the response trajectory (Jones and Bawa, 1995; Olivier et al., 1995). In the triggered-delay mode, the response probability of a motor unit decreased at shorter delays, regardless of the source of excitation (TMS, H-reflex or tendon tap; Fig. 3). This decrease was accompanied by an increased onset latency and narrowing of the PSTH peak (see Olivier et al., 1995, Fig. 2A). It was hypothesized that even if the EPSPs resulting from two inputs were similar at the longest delay, then as a result of differential effects of membrane conductance on EPSPs originating at different electrotonic distance from the soma, the response trajectories for the two inputs would not overlap. The corollary is that two inputs that give rise to overlapping response trajectories exhibit similarity under multiple test conditions. This hypothesis was tested with inputs produced by tendon taps and H-reflex in 6 SMUs. The peak widths with tendon taps were 3.70:t 0.46 ms and with H-reflexes were 1.25:t 0.15 ms. Fig. 3A shows the response trajectories of an SMU to tendon tap and H-reflex inputs. While both inputs resulted in response probabilities near 1.0 at the longest delay, the response probability fell more sharply for the narrower H-reflex peaks as the delays shortened. Applying the same logic, we compared the response trajectories for SP, and H-reflex in another motor unit. Fig. 3B shows that the time courses of the response trajectories overlap for the SP, response to TMS and the electrically induced Ia volley for this motor unit. Thus, it appeared as though the shape of the response trajectory could reveal differences in the underlying EPSP. The average response trajectories were then calculated from 12 SMUs receiving both H-reflex and TMS inputs and the results are illustrated in Fig. 4. The response probabilities for the two inputs are plotted on separate ordinates to allow for a graphical normalization of the data points at the longest delay. This normalization allows for a comparison of the shapes of the mean response trajectories resulting from the two inputs when the curves start from a common point. The figure illustrates that the proportional rate of change of response probability during the AHP is similar for both H-reflex and TMS input. 4. Discussion The primary aim of this study was to compare the monosynaptic component of the responses of repetitively firing FCR motoneurons to peripheral Ia and corticomotoneuro-

5 K.E. Jones et al. / Electroencephalography and clinical Neurophysiology 101 (1996) x OJ 0.-. '" OJ.. 0 " '".. 0.,., >..Q 3 E 0- Il...: 0 Po.. II 0: 0 60.tal Delay {ms' Fig. 4. Average response trajectories for H-reflex and SP1. The figure shows the mean and standard error from 12 SMUs for response probabilities at different total delays. The response probabilities have been plotted along separate ordinates so that both H-reflex and SP1 start from the same position at the longest delay. In this way differences in sizes of the inputs are normalized. The shapes of the two response trajectories overlap and suggest that the two inputs give rise to similar EPSPs. nal inputs. The inputs were delivered randomly with respect to the discharge of the motor units or at a fixed delay following the motor unit discharge. Results of the study suggest that the composite EPSPs which underlie the responses of motoneurons to H-reflex and SP1 of the TMS input have similar rise times. We restricted our comparison to the SP! response of the SMU to TMS as this is most likely due to monosynaptic excitation from corticomotoneuronal cells (Day et al., 1989; Palmer and Ashby, 1992; Awiszus and Feistner, 1993; Priori et al., 1993; Awiszus and Feistner, 1 994b). However, there is evidence from surface EMG responses in FCR that TMS produces disynaptic excitation of the motoneuron pool as early as 1 ms following the monosynaptic excitation (Gracies et al., 1994). This connection may, therefore, contribute to increased firing probability during SF!. It is interesting to note that the delay between the maximum facilitation of the FCR H-reflex from the monosynaptic and disynaptic pathways was greater than 2 ms (Gracies et al., 1994; Fig. 3A). In the present results the longest SP! peak width was 1.6 ms which leads to the question of why there was no response from the SMUs at a time when the propriospinal input is maximal. It may be that the disynaptic propriospinal pathway is contributing to the other subpeaks of the SMU response to TMS more than the SP! response. Thus, the SP! response of an SMU to TMS may be mediated primarily by the monosynaptic connection. As a means of furthering our understanding of the central processes involved in the mediation of the excitatory effects of this descending volley on spinal motoneurons, it seemed appropriate to compare it to another well documented monosynaptic connection to the motoneuron, that of the spindle la afferents. A prerequisite for such a com-

6 436 K.E. Jones et al.! Electroencephalography and clinical Neurophysiology 101 (1996) accounted for by recording responses to the two inputs (Ia and corticomotoneuronal) from the same motoneuron or the same population of low threshold motoneurons. The second factor, history of synaptic activation, would have little effect on the present results since the stimuli were delivered at slow rates «0.3 stimuli/s). With respect to factor 3, there is no evidence as yet which would suggest that the time course of post -synaptic conductance changes associated with Ia or CM input to motoneurons is different. We believe that the last two factors playa significant role in the interpretation of the present results. Different degrees of temporal dispersion of an input, factor 4 above, are known to cause differences in the rise times of composite EPSPs (Walmsley and Stuklis, 1989). We have demonstrated this effect by creating different amounts of temporal dispersion in the same pathway to the motoneuron by using two different stimuli to excite Ia afferents. Electrical stimulation of peripheral nerves produces nearsynchronous activation of the fast conducting Ia afferents. We expect that the temporal dispersion of this volley would be minimal (Segev et al., 1990). Edgley et al. (1990) have convincingly argued that SP\ results from the direct excitation of the CM cells resulting in a D- wave and these results have been supported in human subjects (Burke et al., 1993). On the other hand, if the dispersion within the D-wave with TMS was higher than that of the H-reflex we would expect longer rise time estimates, but this was not the case. It does not, on the other hand, seem that the temporal dispersion within the D-wave would be significantly less than the already minimal values obtained with H-reflex. Therefore, this factor is not likely to playa significant role in the comparison of rise times computed for H-reflex and the SP\ response with TMS input. The fifth factor affecting composite EPSP rise time is the distribution of the synaptic boutons on the soma-dendritic membrane. This information may be obtained either from anatomical reconstruction or electrophysiological estimates. To date, there have not been detailed anatomical studies on the distribution of synapses from the CM cells to motoneurons, in contrast to studies of Ia synaptic distribution (Burke et al., 1979). The electrophysiological estimates of the synaptic location rely upon the axiom that the further a current source is from the soma, the longer the rise time of the resulting voltage transient (Rall, 1967). The earliest work in baboon forearm motoneurons suggested similar time courses for both CM and Ia EPSPs (Clough et al., 1968). A similar conclusion was reached by Jankowska et al. (1975) for hindlimb motoneurons of Macaca irus monkeys. These authors further concluded that there was an overlap in distribution of synapses made by group Ia and CM cells on the somadendritic surface of the motoneuron (but cf. Porter and Hore, 1969). Given the above arguments, the present results are consistent with a similar distribution of synaptic boutons for both Ia and corticomotoneuronal inputs to the upper limb FCR motoneurons in man. This conclusion, however, is given cautiously in light of the many confounding factors that may be affecting the presumed composite monosynaptic EPSPs generated by the two inputs used in this study. The significance of these results to motor control theory is the importance of the interplay between the functional role of an input and its position on the soma-dendritic membrane. Since the CM pathway is the efferent limb of the transcortical servo loop (Phillips, 1969), then it seems reasonable that its synaptic input to the motoneuron overlaps with that of the segmental servo loop. The role of the CM pathway during voluntary contractions implied by the proximal placement of CM synapses in the present results, suggests that this pathway would be effective in directly controlling the discharge rate of the motoneurons. Distally distributed inputs, on the other hand, would be more effective in modulating the integrative properties of the motoneuron. Acknowledgements The authors are grateful to Joe Knight and Cadwell Laboratories for the generous loan of the MES-lO stimulator used in these studies. This work was supported by grants from the BCHRF, NSERC of Canada. K.E. Jones was supported by an NSERC postgraduate scholarship. References Awiszus, F. and Feistner, H. Abnonnal EPSPs evoked by magnetic brain stimulation in hand muscle motoneurons of patients with amyotrophic lateral sclerosis. Electroenceph. clin. Neurophysiol., 1993, 89: Awiszus, F. and Feistner, H. Correlations between size parameters and the amplitude of the excitatory postsynaptic potential evoked by magnetic brain stimulation in human hand muscle motoneurons. Exp. Brain Res., 1994a, 98: Awiszus, F. and Feistner, H. Quantification of D- and I-wave effects evoked by transcranial magnetic brain stimulation on the tibialis anterior motoneuron pool in man. Exp. Brain Res., 1994b, 101: Awiszus, F. and Feistner, H. Comparison of single motor unit responses to transcranial magnetic and peroneal nerve stimulation in the tibialis anterior muscle of patients with amyotrophic lateral sclerosis. Electroenceph. clin. Neurophysiol., 1995, 97: Bawa, P. and Lemon, R.N. Recruitment of motor units in response to transcranial magnetic stimulation in man. J. Pbysiol. (London), 1993, 471: Burke, R.E., Walmsley, B. and Hodgson, JoA. HRP anatomy of group Ia afferent contacts on alpha motoneurons. Brain Res., 1979, 160: Burke, D., Hicks, Ro, Gandevia, S.C., Stephen, J., Woodforth, I. and Crawford, Mo Direct comparison of corticospinal volleys in human subjects to transcranial magnetic and electrical stimulation. J. Physiol. (London), 1993,470: Clough, J.F.M., Kernell, D. and Phillips, C.G. The distribution of monosynaptic excitation from the pyramidal tract and from primary spindle afferents to motoneurons of the baboon's hand and forearm. J. Physiol. (London), 1968, 198: Day, BoL., Dressler, D., Maertens de Noordhout, A., Marsden, CoD., Nakashima, K., Rothwell, J.C., and Thompson, P.D. Electric and

7 K.E. Jones et al. I Electroencephalography and clinical Neurophysiology 101 (1996) magnetic stimulation of human motor cortex: surface EMG and single motor unit responses. J. Physiol. (London), 1989, 412: Edgley, S.A., Eyre, J.A., Lemon, R.N. and Miller, S. Excitation of the corticospinal tract by electromagnetic and electrical stimulation of the scalp in the macaque monkey. J. Physiol. (London), 1990,425: Fournier, E. and Pierrot-Deseilligny, E. Changes in transmission in some reflex pathways during movement in humans. News Physiol. Sci., 1989,4: Gracies, J.M., Meunier, S. and Pierrot-Deseilligny, E. Evidence for corticospinal excitation of presumed propriospinal neurones in man. J. Physiol. (London), 1994,475: Gustafsson, B. and Pinter, M.J. Influence of post-synaptic properties on the time course of synaptic potentials in different types of cat lumbar alpha-motoneurons. Neurosci. Len., 1984,51: Homma, S. and Nakajima, Y. Input-output relationship in spinal motoneurons in the stretch reflex. Prog. Brain Res., 1979, 50: Jankowska, E., Padel, Y. and Tanaka, R. Projections of pyramidal tract cells to a-motoneurons innervating hind-limb muscles in the monkey. J. Physiol. (London), 1975,249: Jones, K.E. The Physiology and Simulation of a-motoneurons in the Human Spinal Cord. Pill Thesis, Simon Fraser University, Burnaby, B.C., Canada, Jones, K.E. and Bawa, P. Responses of human motoneurons to Ia inputs: effects of background firing rate. Can. J. Physiol. Pharmacol., 1995, 73: Jones, K.E., Calancie, B., Hall, A. and Bawa. P. A comparison of descending and peripheral inputs on human motoneurons. Can. J. Physiol. Pharmacol., 1995,73: Axii-Axiii. Lev-Tov, A., Miller, J.P., Burke, R.E. and Rall, W. Factors that control amplitude of EPSPs in dendritic neurons. J. Neurophysiol., 1983, 50: Luscher, H.-R., Ruenzel, P. and Henneman, E. Composite EPSPs in motoneurons of different sizes before and during PTP: implications for transmission failure and its relief in Ia projections. J. Neurophysiol., 1983,49: Malmgren, K. and Pierrot-Deseilligny, E. Evidence for non-monosynaptic Ia excitation of human wrist flexor motoneurones, possibly via propriospinal neurones. J. Physiol. (London), 1988, 405: Olivier, E., Bawa, P. and Lemon, R.N. Responses to transcranial magnetic stimulation of human motoneurons at different firing rates. J. Physiol. (London), 1995, 485: Palmer, E., and Ashby, P. Corticospinal projections to upper limb motoneurones in humans. J. Physiol. (London), 1992, 448: Phillips, C.G. Motor apparatus of the baboon's hand. Proc. R. Soc. London B., 1969, 173: Pierrot-Deseilligny, E., Morrin, C., Bergego, C. and Tankov, N. Pattern of group I fibre projections from ankle flexor and extensor muscles in man. Exp. Brain Res., 1981,42: Porter, R. and Hore, J. Time course of minimal corticomotoneuronal excitatory postsynaptic potentials in lumbar motoneurons of the monkey. J. Neurophysiol., 1969,32: Priori, A., Bertolasi, L., Dressler, D., Rothwell, J.C., Day,BJ;".:Thompson, P.D. and Marsden, C.D. Transcranial electric and magnetic stimulation of the leg area of the human motor cortex: single motor unit and surface EMG responses in the tibialis anterior muscle. Electroenceph. clin. Neurophysiol., 1993,89: Rall, W. Distinguishing theoretical synaptic potentials computed for different soma-dendritic distributions of synaptic input. J. Neurophysiol., 1967,30: Rothwell, J.C., Thompson, P.D., Day, B.L., Boyd, S. and Marsden, C.D. Stimulation of the human motor cortex through the scalp. Exp. Physiol., 1991,76: Segev, I., Fleshman J.W. and Burke, R.E. Computer simulation of group Ia EPSPs using morphologically realistic models of cat a-motoneurons. J. Neurophysiol., 1990,64: Walmsley, B. and Stuklis, R. Effects of spatial and temporal dispersion of synaptic input on the time course of synaptic potentials. J. Neurophysiol., 1989,61: