Motor Unit Number Estimation (MUNE) and Quantitative EMG (Supplements to Clinical Neurophysiology, Vol. 60) Editors: M.B. Bromberg # 2009 Elsevier B.V. All rights reserved
The CMAP scan Gerhard H. Visser* and Joleen H. Blok Department of Clinical Neurophysiology, Erasmus MC, University Medical Center, Rotterdam, The Netherlands
1. Introduction An important aspect of the electrodiagnostic examination is gathering information with respect to the functioning and structure of motor units (MUs). The methods that are routinely used for this purpose are nerve conduction studies and concentric needle electromyography. In nerve conduction studies, nerve properties are usually studied using supramaximal nerve stimulation, which simultaneously activates all functioning axons and, subsequently, their associated muscle fibers. Therefore, the resulting maximal compound muscle action potential (CMAP) provides a measure of the number of functioning muscle fibers and – albeit indirectly – of the number of functioning motor neurons. In neurogenic disorders, loss of functioning motor neurons leads to a reduced number of MUs and denervation of associated muscle fibers. Loss of motor neurons, even when severe, may be masked by distal collateral sprouting of other, *
Correspondence to: Dr. G.H. Visser, Department of Clinical Neurophysiology, Erasmus MC, University Medical Center, P.O. Box 2040, 3000 CA Rotterdam, The Netherlands Tel.: þ 31.10.703 3751; Fax: þ 31.10.703 4621; E-mail: [email protected]
still intact nerve fibers. These sprouts reinnervate denervated muscle fibers. This compensatory process results in an increased MU size and ensures that the force output of affected muscles may, for a time at least, be fairly well maintained. Similarly, as long as collateral reinnervation is adequate, standard nerve conduction studies yield normal findings for CMAP amplitude. Although in this situation the CMAP still is an adequate indicator of the number of functioning muscle fibers, it fails as measure of motor neuron number. Needle electromyography (EMG) provides additional information regarding denervation and reinnervation processes. However, needle EMG has the disadvantage of being a relatively painful and invasive investigation. Moreover, its results can be adversely influenced by a sampling bias due to the relatively small uptake area of the needle tip: the needle EMG signal is determined by only a few muscle fibers that are close to the needle tip. Therefore, an enlarged MU is only indirectly detected from the larger concentration of its muscle fibers at the needle tip that results from type grouping. We have developed a new electrophysiological method that can address some of the above issues. It is based on the fact that motor units (MUs) differ with respect to the stimulus intensity that is
CMAP amplitude [uV]
9000 8000 7000 6000 5000 4000 3000 2000 1000 0 3
Stimulus intensity [mA]
Scan of a healthy subject, showing gradual recruitment of motor units of fairly homogeneous size.
required to electrically activate their motor neuron. These differing thresholds imply that if stimulus intensity is gradually increased from subthreshold to supramaximal values, all MUs in the muscle are successively activated. By plotting response size versus stimulus intensity (stimulus– response curve), information can be obtained about the MU potentials that are successively added to the recorded CMAP, and that ultimately build up the maximum CMAP as response from the entire muscle (Fig. 6.1). Such a plot, based on a limited number of stimuli (usually 30), is used as “scan” to assess stimulus intensity levels for statistical motor unit number estimation (MUNE) (Daube, 1995). We adopted the term scan, as the above approach yields a fairly global, quick assessment of the MU components in the CMAP response. In addition to an activation threshold, MUs have a range of stimulus intensities over which their firing probability gradually increases from 0 to 1 (Bergmans, 1970; Brown and MilnerBrown, 1976; Hales et al., 2004). If these ranges overlap between MUs, any combination of probabilistically active units can be activated upon successive stimuli with equal strength. This alternating behavior of MUs results in variability in
the recorded (submaximal) CMAP. Alternation explains why in Fig. 6.2 for a range of stimulus intensities (0.5–1 mA wide), two CMAP values may be observed: one generated by lower threshold MUs and one generated by those MUs plus the MU that is newly (but initially variably) recruited. Usually, the scan is composed of the contributions of tens to several hundreds of MUs. In those cases, alternation results in a band of CMAP variability around the gradually increasing mean. Apart from the statistical MUNE method (Daube, 1995), we could only determine a role for the scan in nerve excitability testing (threshold tracking) (Kiernan et al., 2000, 2002; Kuwabara et al., 2000; Vucic and Kiernan, 2006). In those applications, it is primarily employed to provide background information for more specialized tests. Applications that aim to extract clinically relevant information directly from the scan are rare (Svensson et al., 1995; Meulstee et al., 1997; Thomas et al., 2002; Ginanneschi et al., 2006; Henderson et al., 2006). A recent paper by Henderson et al. (2006) has confirmed our hypothesis that a scan which is recorded in sufficient detail (with adequate stimulus number) has much to reveal that can be of clinical interest.
CMAP amplitude [uV]
800 700 600 500 400 300 200 100 0 12.5
Stimulus intensity [mA]
Fig. 6.2 APB scan obtained from a patient with a severe form of Guillain–Barre´ syndrome. Four large steps are seen, representing MUs that were first activated at stimulus intensities of approximately 15, 22, 23, and 25 mA. Remaining superimposed variability at higher stimulus intensities might indicate a few more (alternating) small MUs, MU instability and/or noise.
2. Recording a scan To record an electrophysiological muscle scan, we use surface electrodes (cups with 10 mm diameter and Ten20W electrode paste) in a muscle belly-tendon derivation in conjunction with transcutaneous electrical stimulation of the efferent nerve. The scans that are presented in this chapter were mostly made from the abductor pollicis brevis (APB) muscle with median nerve stimulation at the wrist. Occasionally, recordings were made at other locations such as the abductor digiti minimi muscle (ADM; ulnar nerve stimulation). In the preparations for each scan, first the muscle belly electrode position is adjusted to maximize the negative CMAP peak amplitude. The ground electrode is placed nearby. The next step is to determine the optimal position for the stimulating electrodes by locating the point with lowest threshold. At this position, the stimulating electrodes are taped securely to the limb. Finally, subjects are asked to recline and relax without speaking.
Our scans were recorded using the MUNE500 program implemented on a Viking Select EMG machine (Nicolet Biomedical, Madison, WI, USA) (Henderson et al., 2003). This program stores the amplitude and area of each elicited CMAP to a file that can be imported in ExcelW (Microsoft, USA) for data analysis and to generate plots. Recordings started with the determination of the threshold of the lowestthreshold MU (S0) with the sensitivity set at 50 μV/division, followed by the determination of the lowest intensity at which a maximal CMAP could be recorded (S100). Next, a brief scan of 30 evenly spread stimuli between S0 and S100 was performed. Based on visual assessment of this scan, lower (S0) and upper (S100) limits were adjusted until the scan covered the entire CMAP range; that is, corrected intensities ranged from subthreshold to approximately 120% supramaximal values. With these final S0 and S100 settings, a more detailed scan was recorded using 300 stimuli that were equidistant in intensity (due to software limitations this full
68 3. Computer simulations as reference for scan patterns related to MU number The pattern, with which the CMAP increases, is determined by the number of MUs, their order of activation and their sizes, as well as by alternation. Further factors, which primarily affect CMAP variability, are noise and MU instability. To provide a better understanding of the effect of each of these variables on the scan, we have extended a previously described computer model of alternation (Blok et al., 2005) to generate scans (Blok et al., 2007). Figure 6.3 shows the results of the simulations for changing MU number. A comparison of recorded scans with these simulation results
CMAP (% OF CMAPmax)
set was collected in 10 successive series of 30 stimuli). Stimulus frequency was 2 Hz, stimulus duration was 0.1 ms. Additional recordings with sets of 30 stimuli were made in the intermediate CMAP range in subjects where rapid recruitment of new MUs occurred in this range. Furthermore, recordings were occasionally repeated at a particular stimulus intensity range if the full scan revealed some peculiarity that merited a more detailed study. Excluding preparations (which are usually done within the context of conventional nerve conduction studies), the test takes approximately 5 min. Further details on scan methodology can be found in the paper of Blok et al. (2007).
Stimulus Intensity (mA)
Fig. 6.3 Simulated scans (computer model) for a varying number of MUs present. Stimulus intensity at horizontal axis, percentage of maximum CMAP at vertical axis. Plots show irregular scan patterns for low MU numbers (up to about 50), while CMAP variability gradually decreases with increasing MU numbers, resulting in a narrowing of the scan.
69 can provide a rough estimate of the number of MUs present, particularly if this number is strongly reduced (Fig. 6.2).
such as multiple steps (discussed more extensively below). Semi-quantitative, by comparing the scan with “validated” scans (obtained from either simulations or extensive analysis) to obtain, for example, a rough estimate of the number of MUs present (Fig. 6.4). The scan may also form the basis for quantitative MU number estimation (MUNE) using a stochastic approach (Henderson et al., 2007; Ridall et al., 2007). In addition, a comparison between upward and downward scans may provide data on accommodation. In upward scans, MUs are newly recruited, implying that some accommodation may occur upon the first few stimuli after recruitment. In downward scans, initially all MUs are active and they gradually become inactive, so that effects of accommodation are absent.
4. Information in the scan, data analysis, and quantification Our current information extraction procedures roughly fall into one of two categories. The first aims to draw conclusions from the full CMAP pattern, the second uses feature extraction to provide detailed, quantitative data regarding a particular aspect of the scan. 4.1. CMAP pattern
Analysis of the CMAP pattern may be: Visual, determining if the shape of the scan curve appears normal or if there are irregularities 5 MUs
CMAP amplitude (uV)
CMAP amplitude (uV)
100 0 12.5
Stimulus Intensity (mA)
1000 0 15.9
Stimulus Intensity (mA)
Fig. 6.4 The comparison of the simulated scans (above) with measured patient scans (below) suggests a (semi-quantitative) MU number of about 5 for the GBS patient and about 15 for the post-polio patient (who had a near-normal maximal CMAP!).
70 4.2. Stimulus intensity (SI) Standard nerve conduction tests at best address axonal excitability globally at the site of stimulation. Yet, it is known that excitability measures may provide diagnostically relevant information (Kiernan et al., 2000; Nodera and Kaji, 2006). Part of this information can be collected with the scan. From each scan, we extracted the following features that are related to excitability: S0: SI marking the low end of the scan, which we defined as the SI at which the scan reaches the predefined, small value of 50 μV. S50: the SI at which the recorded CMAP equals 50% of the maximum CMAP. S100: SI marking the high end of the scan, where it reaches its maximum value or a high percentage of the maximum CMAP (e.g., 99%). S100–S0: SI range (absolute range width). (S100–S0)/S50: a corrected, relative SI range width. Changes in any of the above measures reflect changes in axonal excitability at the stimulus site. The relative width compensates to some extent for overall shifts in the curve that are due to
. . . ..
differences in nerve proximity, conductive properties of the intermediate tissue, or changes in stimulus duration. As such, it reduces the influence of suboptimal stimulation position on the SI range variable. Finally, in the case of a step that can be ascribed to a contribution of a single MU with reasonable certainty (such as in Fig. 6.2), the excitability properties of this MU can be studied separately. Its so-called recruitment range width is deduced from the SI at which the MU is always active (right end of the step “floor”) minus the SI at which the MU is first activated (the recruitment threshold and left end of the step “ceiling”). 4.3. Maximum CMAP
The maximum CMAP value reached by the scan, reflecting the total number of functioning muscle fibers Variability in the CMAP at supramaximal stimulus intensities. This variability provides a measure of the average within-MU variability from stimulus to stimulus (Fig. 6.5) and can be expressed as the standard deviation of the responses to a series of 30 supramaximal stimuli.
CMAP amplitude [uV]
1000 800 600 400 200 0 4
Stimulus intensity [mA]
Substantial CMAP variability at supramaximal stimulus intensity in a patient with ulnaropathy. Variability increases gradually with increasing stimulus intensities, suggesting within-MU variability.
71 Please note, at lower stimulus levels, MU instability cannot be differentiated from alternation. 4.4. Steps Steps are defined as clearly visible size differences between consecutive CMAPs (as in Figs. 6.2, 6.4 and 6.6). Steps are present if a MU size is large and/or MU number is severely decreased. In essence, steps are the surface potentials of a single MU. For that reason, they represent the number of muscle fibers in the MU. Because step size is based on a surface EMG assessment, effects of the distance between the muscle fibers and the recording electrode are relatively small compared to needle EMG. Steps can be parameterized as follows: Step size: absolute or relative (step size as percentage of the maximal CMAP); Step number (especially meaningful if linked to minimum step amplitude, e.g., 0.5 mV); Step %: percentage of the maximum CMAP made up by all detected steps together. If abnormal, all three features provide evidence for loss of MUs and reinnervation. If very few MUs are left, the summed step size (Step %) approaches 100% of the maximum CMAP, implying that the number of steps provides a good indication of the number of remaining MUs. If Step % is intermediate to large, it suggests moderate to severe MU loss. The quantitative analysis of steps can be automated, at least in part (Blok et al., 2007).
. . .
5. The scan in normal subjects To obtain a first impression of reference values, a muscle scan of the APB was obtained from 11 healthy subjects without neuromuscular complaints (5 men, 6 women; mean age 33 years). Their scans generally follow a smooth sigmoid course (Fig. 6.1). One scan showed a pattern that clearly deviated from the others, with a fairly large stimulus intensity range (absolute width of 18 mA) and (at least) five steps that together
formed 11% of the maximum CMAP. This scan was excluded in the determination of the following overall results. For the scans of the remaining 10 healthy subjects, the mean amplitude of the negative peak of the maximum CMAP was 10.2 mV (range 7.1–16.3). The mean SI threshold (S0) was 7.9 mA (range 5.0–11.0) and the maximum CMAP was obtained at a SI (S100) of 14.5 mA (range, 9.6–26.5). The average absolute SI range width was 6.6 mA (range 3.7–16.5) and the relative range width (S100–S0)/S50) was 0.6 (range 0.4–0.9). The SI which generates half of the maximum CMAP (S50) was 10.7 mA (range 7.0–19.2). As a rule of thumb, absolute stimulus range width is about equal or less than the threshold value (e.g., for a subject with a threshold of 12 mA, the S100 was 24 mA at most). This rule of thumb agrees with other published data (Henderson et al., 2006). Three out of 10 scans showed one step, accounting for 2–4% of the maximum CMAP. One additional scan showed two steps, which added up to 6% of the maximum CMAP. The maximum step size observed in these healthy subjects was 327 mV. 6. Examples to show the scan’s potential use In this section, we present a few illustrative cases to show the potential use of the scan. These examples were selected from a series of 43 scans that were recorded from 34 patients (20 men, 14 women) who were referred to our Clinical Neurophysiology department for an electrodiagnostic evaluation regarding a range of pathologies. The first example is derived from a patient with a suspected median nerve neuropathy and supports the interpretation of a marked scan step being caused by the presence of an enlarged MU. In response to a series of stimuli, only one (giant) F-potential of 800 μV was generated in the ABP, suggesting it to be a single axon response (Fig. 6.6A). The subsequent scan of this muscle showed a single large step of the same size, corresponding with a large MU, most likely the same as the one that generated the F-wave (Fig. 6.6B).
Patient with median nerve neuropathy. A: single “giant” F-wave of about 800 μV. B: scan showing large step of similar size, most likely originating from the same MU that generated the large F-wave.
The second example shows scans that were recorded from 2 different patients with Guillain–Barre´ syndrome (GBS), who were admitted to our hospital’s ICU. These patients were clinically equally affected with severe paresis; both required mechanical ventilation. Conventional nerve conduction studies showed similar results such as, for example, the abnormally low maximal APB CMAP amplitude of approximately 0.85 mV. However, the APB scans showed remarkable differences (Fig. 6.7). In the first patient (patient A; shown on the left), the pattern of the scan deviated only mildly from the normal s–shaped curve. Furthermore, there was just a single step, which can still be regarded as normal. By contrast, the second patient’s scan was markedly abnormal, consisting of little more than 4 relatively large steps (patient B; shown on the right). In this patient, Step % was 83%, suggesting a severely reduced MU number. Although we presently do not have sufficient information to interpret these differences in clinically meaningful terms, we hypothesize that they may be used to distinguish between different pathophysiological mechanisms. At least, this example demonstrates that the scan
offers additional information compared to conventional nerve conduction studies. The next example illustrates that certain changes in excitability during the course of GBS can be easily documented with the scan. For a third GBS patient, the stimulus intensity range (S0–S100) ran from nearly 15 to 55 mA in the acute phase of the disease (Fig. 6.8; left). This range is abnormally wide (see rule of thumb in previous paragraph on stimulus intensity). Six months later, these values had changed to approximately 25 and 45 mA (Fig. 6.8; right), implying that the range width had normalized. SI threshold (S0) remained increased. The initial scan might be interpreted as resulting from a mixture of axons with a (still) normal excitability and axons that were affected relatively early in the disease’s course and that had an increased stimulus threshold. More detailed follow-up studies are needed to clarify the time course of excitability changes during active disease and recovery, but we believe the potential use of the scan for such studies is clear. A last example comes from a patient with a suspected ulnar neuropathy. The distally evoked
Fig. 6.7 Two patients with Guillain–Barre´ syndrome, clinically equally severely paretic and comparable NCS results (both had a maximal CMAP of 0.85 mV). Scans of the APB were remarkably different. Patient B has considerably more MU loss than patient A.
Range < 15-55 mA
Range < 25-45 mA
Fig. 6.8 Guillain–Barre´ patient, showing the change in axonal exitibility during the course of the illness (left acute phase, right after 6 months).
CMAP had a normal baseline–peak amplitude of 6.5 mV, but inching around the elbow revealed a very focal conduction slowing and partial block (Fig. 6.9A). The concentric needle EMG investigation showed spontaneous muscle fiber activity and enlarged MU potentials up to 8 mV, suggesting active denervation and reinnervation, respectively (Fig. 6.9B). A scan of this patient’s hypothenar muscles showed normal stimulus intensities, consistent with normal axonal function at the (distal) site of stimulation. However, there was an abundance of steps (at least 17), suggestive of a decreased number of MUs.
Furthermore, the step sizes of up to 1 mV point to multiple enlarged MUs (Fig. 6.9C). In this patient, the actual need for needle EMG can be debated. Although signs of active denervation (such as fibrillation potentials) cannot be detected with a scan, this information is not always clinically relevant. More importantly, the scan will reveal all enlarged MUs present without sampling bias. The reinnervation and axonal loss that were evident on needle EMG were also shown by the scan, not only in this patient, but consistently in all patients that we have thus far studied.
Fig. 6.9 Patient with left ulnaropathy. A: inching around the elbow, showing focal slowing and a partial conduction block. B: needle EMG of the left abductor digit minimi muscle showing giant MUs up to 8 mV. C: scan of the same muscle with multiple steps, showing multiple enlarged MUs.
7. Discussion The scan is a rapid, efficient noninvasive method of assessing multiple parameters of nerve and motor unit function simultaneously. It is easy to record, relatively easy to interpret and it discloses information regarding the MU components of a CMAP. Specific patterns in or properties of the scan (steps, maximum CMAP, variability) provide clinically relevant information regarding MU number, size and stability. The major
disadvantage of the scan is the number of stimuli that it requires, including many at higher intensities. However, in our experience the test is well tolerated by both normal controls and patients, particularly in comparison to concentric needle examination. In this paper, we presented several cases that together demonstrate the ability of the scan to detect (subclinical) changes to MUs, which were not always revealed by conventional NCS or needle EMG. The scan appears to have several
75 advantages. compared to concentric needle EMG, and may replace this invasive investigation at least in part. Needle EMG is limited by the fact that the amplitude and morphology of the needle EMG signal are largely determined by only a few muscle fibers close to the needle tip. This introduces a sampling bias and necessitates sampling at multiple sites within a muscle. Furthermore, an increased MU potential amplitude measured with a concentric needle results from reduced average distance between fibers of a single MU (due to reinnervation within the surviving MU’s original territory). It only indirectly reflects an increased number of muscle fibers in the MU. For that reason, needle EMG potential size is not a good indicator of motor unit size. By contrast, the scan is built up from contributions of all activated MUs and their muscle fibers. For that reason, it will show any enlarged MU potentials present. There is no sample bias. In addition, as a surface EMG measure, a large step size actually reflects the large number of muscle fibers that constitute an enlarged MU (disregarding a relatively minor influence of MU depth on MU potential size for small distal muscles). Moreover, surface EMG MU potential size correlates fairly well with macro EMG, the gold standard for MU size estimation (Roeleveld et al., 1997; Sun et al., 1999). For these reasons, the scan provides a measure of MU size for each and all of the largest MUs (steps). Please note, however, that step size will often underestimate true MU size somewhat, due to alternation effects of smaller MUs with overlapping recruitment range, noise, and within-MU variability. Based on our experience thus far, the occurrence of many steps reflects a substantial loss of functional MUs. The size of the maximum step per se does not seem to hold much information, because large MUs are known to occur occasionally in normal controls (McComas et al., 1973; Feasby and Brown, 1974; Doherty and Brown, 1993). However, the number of steps and, particularly, Step % (cumulative step size as % of the maximum CMAP) appear to be clinically
informative. Although accurate age-dependent reference values are not available yet, in normal subjects Step % appears to be well below 10–15%, while in many patients we found Step % to be considerably larger. Particularly after reinnervation, the maximum CMAP cannot be used as accurate measure of the number of functioning axons. Assessments of MU number from the needle EMG (that is, from interference pattern analysis at maximum voluntary contraction) are qualitative at best. Quantitative estimation of the number of MUs (MUNE) requires running of a separate test, at least 10 min per investigated muscle, and special skills. The scan offers various possibilities to assess the number of MUs. A first visual assessment may show obvious steps or changes in CMAP variability as signs of a reduced MU number. In the case of “single patterns” on needle EMG, the number of MUs can often be simply counted in the scan and easily followed in time. By comparison of the scan pattern with “validated” scans (obtained from either simulations or extensive analysis), the number of MUs may be estimated roughly, especially if this number is markedly reduced. Finally, scan data may be used for a quantitative stochastic MUNE approach based on previously proposed principles (Henderson et al., 2007; Ridall et al., 2007). The scan provides more information on axonal excitability at the site of the stimulus than is usually documented with conventional nerve conduction studies. A large majority (13/15) of the GBS patients that we studied, showed an increase in S50 and/or the stimulus intensity range (Blok et al., 2007), suggesting changes in excitability at the site of stimulation. The observation that there is clinically and diagnostically relevant information in excitability data in GBS as well as in other conditions is in line with previous more detailed excitability studies that used threshold tracking (Brown et al., 1993; Meulstee et al., 1997; Kuwabara et al., 2002; Kaji, 2003; Nodera et al., 2004; Ginanneschi et al., 2006; Kanai et al., 2006; Nakata et al., 2006; Nodera and Kaji, 2006; Vucic
76 and Kiernan, 2006; Z’Graggen et al., 2006). Indeed, some of these have demonstrated that changes in excitability may even occur at a subclinical level. In principle, the data provided by the recordings at supramaximal intensities can also be used as input for conventional F-wave studies. They merely need separate plotting. In fact, the size and latency of the F-response could be provided for each of the stimuli that generated the scan. Thus, many more responses and, possibly, more reliable (minimal latency) information would become available than with the series of 20 stimuli that is commonly used for F-wave assessment. Similarly, the lower intensity stimuli of the scan may be used for H-reflex testing. In conclusion, the muscle scan provides much information in an easily accessible way. Because of its noninvasive nature, it may be more readily applied than needle EMG, for example, in children or in follow-up studies. The scan is especially useful to detect enlarged MUs and a (markedly) reduced number of MUs. Current results suggest that needle EMG may be restricted to the assessment of insertion activity/spontaneous activity at rest, and MU morphology at mild voluntary activation. These two parts of the needle EMG investigation tend to be better tolerated that maximal voluntary contractions. Finally, further exploration and development of the scan, particularly in a clinical setting, are essential to fully understand and employ its possibilities. References Bergmans, J. (1970) The Physiology of Single Human Nerve Fibers. Vander, Louvain. Blok, J.H., Visser, G.H., De Graaf, S., Zwarts, M.J. and Stegeman, D.F. (2005) Statistical motor number estimation assuming a binomial distribution. Muscle Nerve, 31: 182–191. Blok, J.H., Ruitenberg, A., Maathuis, E.M. and Visser, G.H. (2007) The electrophysiological muscle scan. Muscle Nerve, 36: 436–446. Brown, W.F. and Milner-Brown, H.S. (1976) Some electrical properties of motor units and their effects on the methods of estimating motor unit numbers. J. Neurol. Neurosurg. Psychiatry, 39: 249–257.
Brown, W.F., Feasby, T.E. and Hahn, A.F. (1993) Electrophysiological changes in the acute “axonal” form of Guillain– Barre´ syndrome. Muscle Nerve, 16: 200–205. Daube, J.R. (1995) Estimating the number of motor units in a muscle. J. Clin. Neurophysiol., 12: 585–594. Doherty, T.J. and Brown, W.F. (1993) The estimated numbers and relative sizes of thenar motor units as selected by multiple point stimulation in young and older adults. Muscle Nerve, 16: 355–366. Feasby, T.E. and Brown, W.F. (1974) Variation of motor unit size in the human extensor digitorum brevis and thenar muscles. J. Neurol. Neurosurg. Psychiatry, 37: 916–926. Ginanneschi, F., Mondelli, M., Dominici, F. and Rossi, A. (2006) Changes in motor axon recruitment in the median nerve in mild carpal tunnel syndrome. Clin. Neurophysiol., 117: 2467–2472. Hales, J.P., Lin, C.S.Y. and Bostock, H. (2004) Variations in excitability of single human motor axons, related to stochastic properties of nodal sodium channels. J. Physiol. (Lond.), 559: 953–964. Henderson, R.D., McClelland, R. and Daube, J.R. (2003) Effect of changing data collection parameters on statistical motor unit number estimates. Muscle Nerve, 27: 320–331. Henderson, R.D., Ridall, G.R., Pettitt, A.N., McCombe, P.A. and Daube, J.R. (2006) The stimulus–response curve and motor unit variability in normal subjects and subjects with amyotrophic lateral sclerosis. Muscle Nerve, 34: 34–43. Henderson, R.D., Ridall, P.G., Hutchinson, N.M., Pettitt, A.N. and McCombe, P.A. (2007) Bayesian statistical MUNE method. Muscle Nerve, 36: 206–213. Kaji, R. (2003) Physiology of conduction block in multifocal motor neuropathy and other demyelinating neuropathies. Muscle Nerve, 27: 285–296. Kanai, K., Kuwabara, S., Misawa, S., Tamura, N., Ogawara, K., Nakata, M., Sawai, S., Hattori, T. and Bostock, H. (2006) Altered axonal excitability properties in amyotrophic lateral sclerosis: impaired potassium channel function related to disease stage. Brain, 129: 953–962. Kiernan, M.C., Burke, D., Andersen, K.V. and Bostock, H. (2000) Multiple measures of axonal excitability: a new approach in clinical testing. Muscle Nerve, 23: 399–409. Kiernan, M.C., Guglielmi, J.M., Kaji, R., Murray, N.M.F. and Bostock, H. (2002) Evidence for axonal membrane hyperpolarization in multifocal motor neuropathy with conduction block. Brain, 125: 664–675. Kuwabara, S., Cappelen-Smith, C., Lin, C.S.Y., Mogyoros, I., Bostock, H. and Burke, D. (2000) Excitability properties of median and peroneal motor axons. Muscle Nerve, 23: 1365–1373. Kuwabara, S., Ogawara, K., Sung, J.Y., Mori, M., Kanai, K., Hattori, T., Yuki, N., Lin, C.S., Burke, D. and Bostock, H. (2002) Differences in membrane properties of axonal and demyelinating Guillain–Barre´ syndromes. Ann. Neurol., 52: 180–187. McComas, A.J., Sica, R.E.P. and Campbell, M.J. (1973) Numbers and sizes of human motor units in health and disease. In: J.E. Desmedt (Ed.), New Developments in
77 Electromyography and Clinical Neurophysiology, Vol. 1. Karger, Basel, pp. 55–63. Meulstee, J., Darbas, A., Van Doorn, P.A., Van Briemen, L. and Van der Meche, F.G. (1997) Decreased electrical excitability of peripheral nerves in demyelinating polyneuropathies. J. Neurol. Neurosurg. Psychiatry, 62: 398–400. Nakata, M., Kuwabara, S., Kanai, K., Misawa, S., Tamura, N., Sawai, S., Hattori, T. and Bostock, H. (2006) Distal excitability changes in motor axons in amyotrophic lateral sclerosis. Clin. Neurophysiol., 117: 1444–1448. Nodera, H. and Kaji, R. (2006) Nerve excitability testing and its clinical application to neuromuscular diseases. Clin. Neurophysiol., 117: 1902–1916. Nodera, H., Bostock, H., Kuwabara, S., Sakamoto, T., Asanuma, K., Jia-Ying, S., Ogawara, K., Hattori, N., Hirayama, M., Sobue, G. and Kaji, R. (2004) Nerve excitability properties in Charcot–Marie–Tooth disease type 1A. Brain, 127: 203–211. Ridall, P.G., Pettitt, A.N., Friel, N., McCombe, P.A. and Henderson, R.D. (2007) Motor unit number estimation using reversible jump Markov chain Monte Carlo methods. J. Roy. Statist. Soc., Ser. C: Appl. Statist., 56: 235–260.
Roeleveld, K., Stegeman, D.F., Falck, B. and Sta˚lberg, E.V. (1997) Motor unit size estimation: confrontation of surface EMG with macro EMG. Electroencephalogr. Clin. Neurophysiol., 105: 181–188. Sun, T.Y., Lin, T.S. and Chen, J.J. (1999) Multielectrode surface EMG for noninvasive estimation of motor unit size. Muscle Nerve, 22: 1063–1070. Svensson, P., Arendt-Nielsen, L., Nielsen, H. and Larsen, J.K. (1995) Effect of chronic and experimental jaw muscle pain on pain-pressure thresholds and stimulus-response curves. J. Orofac. Pain, 9: 347–356. Thomas, C.K., Nelson, G., Than, L. and Zijdewind, I. (2002) Motor unit activation order during electrically evoked contractions of paralyzed or partially paralyzed muscles. Muscle Nerve, 25: 797–804. Vucic, S. and Kiernan, M.C. (2006) Axonal excitability properties in amyotrophic lateral sclerosis. Clin. Neurophysiol., 117: 1458–1466. Z’Graggen, W.J., Lin, C.S., Howard, R.S., Beale, R.J. and Bostock, H. (2006) Nerve excitability changes in critical illness polyneuropathy. Brain, 129: 2461–2470.