Limitations of the spike-triggered averaging technique

The spike-triggered averaging technique, when used to obtain the twitch profile of a single motor unit, requires the maintenance by the subject of that unit's minimal rhythmic firing rate, which is typically between 8 and 12 pulses/sec. Literature on the rate-tension relationship of cat motor units shows that some twitches may start to fuse at stimulus rates as low as 2 pulses/sec. This fusion changes the mechanical characteristics of the observed apparent twitch from those of the unfused twitch. The present work has assessed the changes in contraction time, half-relaxation time, and twitch amplitude of slow versus fast twitch cat single motor units, when these units are demonstrating varying degrees of fusion of tension. As the stimulus rate applied to cat single motor units increased to those typically used during spike-triggered averaging, the values of all three twitch parameters decreased, with this decrease being more prominent for slow-contracting versus fast-contracting units.     

The estimation of motor unit twitch tensions in the human masseter muscle by spike-triggered averaging

Spike-triggered averaging (STA) has been used to extract twitch profiles of single motor units (SMU) within the human masseter muscle. However, the reported twitch tensions may have been biased by the voluntary firing frequency of the SMUs, the complex architecture of the muscle, and by the biomechanical linkage of the jaw. In this study, a rigid STA paradigm was used to record spike-triggered "measured tensions" (STMTs) for 32 SMUs in the masseter muscles of four subjects. STMTs were recorded at two different orientations of a force transducer placed between the incisor teeth. The STMTs produced by each unit were used to calculate jaw torque. STMTs were also recorded in 11 units with differing degrees of muscle coactivation. STMTs for each unit varied according to the orientation of the force transducer. However, no systematic changes in STMTs occurred with reciprocal changes in the jaw moment arm. STMTs could be altered significantly by different degrees of muscle co-activation. The use of STA as a method for determining SMU tension in the human masseter muscle appears to be highly task-dependent and in the presence of co-activation may be inappropriate.     

Factors affecting the stability of the spike-triggered averaged force in the human first dorsal interosseus muscle

The reproducibility of motor unit twitches obtained using spike-triggered averaging (STA) was examined in the human first dorsal interosseus. For each motor unit (30 total) a series of STA twitches was derived using a 30 s averaging window. Within each averaging window, eight independent measures characterizing motor unit discharge and whole muscle force properties were recorded. These included the mean and standard deviation (S.D.) of the interspike interval (ISI), the mean and S.D. of pre and post-trigger ISIs used in averaging, and mean and S.D. of whole muscle force. To determine the relative importance of the independent variables on twitch reproducibility, the variables were used in a multiple regression analysis performed on STA twitch peak force (PF), time to peak force (TTP) and time of half-force decay (HFD). It was found that PF was significantly correlated to the mean and S.D. of whole muscle force, and mean post-trigger ISI. TTP was significantly correlated to the S.D. of the post-trigger ISI and mean whole muscle force while HFD was related to the mean and S.D. of the pre-trigger ISI and the mean post-trigger ISI. It was concluded that by minimizing whole muscle force variability and the mean and S.D. of acceptable ISIs used in the STA process, the reproducibility of the STA twitch is improved.     

Effects of experimental muscle pain on mechanical properties of single motor units in human masseter.

OBJECTIVE: Muscle pain is known to influence muscle activity but the details of its effects on the mechanical properties of single motor units (SMU) have not been described. We have recently reported a decreased firing rate of SMU in the human masseter muscle during painful contractions with a constant force output. Force output can be modulated by the SMU discharge rate in relation to the contractile properties of SMU. Therefore, the objective of the present study was to measure the mechanical properties of SMU in the masseter to clarify the mechanism which underlies the decrease in SMU firing rate during jaw-muscle pain. METHODS: A spike-triggered averaging (STA) technique was used to determine the mechanical properties of low-threshold SMU in the masseter muscle recorded with fine wire electrodes during a voluntary isometric contraction. The twitch amplitude, contraction time, and half-relaxation time were determined from the averaged force records before and during experimental jaw-muscle pain induced by injection of 0.2 ml (100 microg/ml) capsaicin in 8 healthy subjects. Injections of 0.2 ml isotonic saline served as a non-painful control in 11 healthy subjects. RESULTS: The twitch amplitude was significantly increased during capsaicin-evoked muscle pain (P<0.001) without significant changes of half-relaxation time and contraction time. No significant changes in SMU twitch properties were observed during the control injections. CONCLUSIONS: Potentiation of twitch force could be a possible compensatory mechanism to maintain a constant force output during painful isometric contractions when SMU firing decreases. This finding therefore provides new information on the adaptation of motor function by muscle pain.     

Discharge variability and physiological properties of human masseter motor units

Intramuscular electrodes were used to study discharge variability in motor units of human masseter whose physiological properties were determined using spike-triggered averaging. Subjects voluntarily controlled the mean firing rate of a selected motor unit at 10 Hz for 15 min of continuous activation. Discharge variability was assessed at the beginning and end of this period. In 81% of units, the discharge variability at a mean interspike interval (ISI) of 100 ms increased after 15 min of continuous activity. There was a wide range of discharge variability within the population of masseter units studied, but no significant correlations were found between initial discharge variability and recruitment threshold, twitch tension or time-to-peak tension (TTP). There was, however, a significant correlation between motor unit fatigability and its initial discharge variability. This represents a link between the motoneuron and the functional properties of the muscle fibers in innervates.     

Computational methods for improving estimates of motor unit twitch contraction properties

Estimates of mechanical properties of human motor units have usually been made indirectly, using the technique of “spike-triggered averaging” (STA). In this method, a single motor unit action potential is used to synchronize the accumulation of an ensemble average of correlated force transients. However, under most realizable conditions, these transients are recorded during periods of sustained motor unit discharge, in which each motor unit is producing a partially fused tetanus. Therefore, the STA technique extracts the characteristics of the unfused force transient, instead of the desired single motor unit twitch. Although the STA method has been widely used, there is as yet no wellestablished relation between the force transient in the unfused tetanus, and the twitch contraction properties of the motor unit. To evaluate the accuracy of the STA as a measure of the motor unit mechanical properties, we applied two types of muscle models to the force transients recorded in an unfused tetanus, using data derived from experiments in which the response to a single twitch was also recorded. Our objective was to see whether accurate predictions of single motor unit mechanical characteristics are possible, working backward from the STA. The models chosen for this task were a linear second order model, and the distribution-moment (DM) model. These model predictions were then compared with the STA response, and with the twitch properties of the individual motor units. We also evaluated the utility of extrapolating the initial slope of the STA backward to improve the accuracy of the mechanical estimates. The results of our simulation suggest that there is no straightforward relation bwtween the characteristics of the unfused tetanus and the mechanical properties of the single twitch. Although our attempts to predict the properties of the single twitch from the STA were only partly successful, the results of the simulations were far more accurate than those derived from the STA alone. Because the errors in the use of the STA method were so substantial, we would urge that the STA technique be used with great caution as a measure of twitch contraction properties, unless accompanied by appropriate simulations of muscle mechanical behavior.© 1995 John Wiley &Sons, Inc.     

Extraocular muscle motor units characterized by spike-triggered averaging in alert monkey

Single-unit recording in macaque monkeys has been widely used to study extraocular motoneuron behavior during eye movements. However, primate extraocular motor units have only been studied using electrical stimulation in anesthetized animals. To study motor units in alert, behaving macaques, we combined chronic muscle force transducer (MFT) and single-unit extracellular motoneuron recordings. During steady fixation with low motoneuron firing rates, we used motoneuron spike-triggered averaging of MFT signals (STA-MFT) to extract individual motor unit twitches, thereby characterizing each motor unit in terms of twitch force and dynamics. It is then possible, as in conventional studies, to determine motoneuron activity during eye movements, but now with knowledge of underlying motor unit characteristics.We demonstrate the STA-MFT technique for medial rectus motor units. Recordings from 33 medial rectus motoneurons in three animals identified 20 motor units, which had peak twitch tensions of 0.5-5.25 mg, initial twitch delays averaging 2.4 ms, and time to peak contraction averaging 9.3 ms. These twitch tensions are consistent with those reported in unanesthetized rabbits, and with estimates of the total number of medial rectus motoneurons and twitch tension generated by whole-nerve stimulation in monkey, but are substantially lower than those reported for lateral rectus motor units in anesthetized squirrel monkey. Motor units were recruited in order of twitch tension magnitude with stronger motor units reaching threshold further in the muscle's ON-direction, showing that, as in other skeletal muscles, medial rectus motor units are recruited according to the “size principle”.     

A comparison of isometric contractions of the whole muscle with those of motor units in a fast-twitch muscle of the cat

Isometric twitches and tetani of motor units and of the whole muscle were observed at various muscle lengths in a fast-twitch muscle of the cat. Some of the motor units developed maximum twitch and tetanic tensions at similar muscle lengths which were close to the optimum lengths for whole muscle contractions, but for others the optimum for the twitch was much longer than that for the tetanus. Contrary to earlier reports the anomalous behavior of this second group of units was not related to the size of the motor unit, nor was it due to neuromuscular failure. Mechanical latencies were not prolonged below the motor unit twitch optimum. Nonlinear summation of single motor unit twitches with those of larger groups of muscle fibers was sometimes found. There was a deficit of one third in the summed twitches of all the motor units but little distortion of contraction times.     

Recruitment order, contractile characteristics, and firing patterns of motor units in the temporalis muscle of monkeys.

Action potentials of single motor units were recorded in the temporalis muscle of rhesus monkeys trained to hold static forces on a bite bar. The average relationship between mandibular force and motor unit firing rate was determined for many motor units using a computer program, and from this relationship the threshold force at which a motor unit began to fire was determined. Several aspects of motor unit firing rate were examined in relation to recruitment threshold. Computer averaging was used to determine twitch tension and twitch contraction times for many motor units. Motor units recruited at low forces had longer contraction times and produced smaller twitch tensions than higher-threshold motor units. Motor units with small action potentials were nearly always recruited at lower force levels than those with larger action potentials. These results indicate that the motor units of the monkey temporalis muscle were recruited in an orderly fashion, which is in accord with the predictions of the “size principle”.     

Correlation between tremor, voluntary contraction, and firing pattern of motor units in Parkinson''s disease

Patients with tremor of Parkinsonism show three characteristics of motor unit activity: rhythmic spontaneous resting discharge, abnormally low firing rates during voluntary contraction, and consistent differences in firing pattern between small and large motor units. Smaller units discharge once per tremor beat at weak contractions but change into bursts of two or three spikes per beat at stronger forces. Large units are later recruited and fire preferentially once per beat. The large tremor amplitudes can be partly explained by synchronization of unfused twitches of low frequency units which summate more powerfully than the partially fused contractions during physiological tremor, which is about twice as rapid. Tremor is strongly influenced by the force of voluntary contraction. It is strongest at rest or during weak muscular effort and with increasing force becomes continuously of higher frequency and smaller amplitude. Both changes are the consequence of increasing discharge rates of motoneurones at stronger contractions.     

Instability of motor unit firing rates during prolonged isometric contractions in human masseter

The firing patterns of up to 4 concurrently active masseter motor units were studied with intramuscular electrodes during a continuous isometric contraction of 15 min duration, in which the subject maintained the mean firing rate of one selected unit at 10 Hz. With this paradigm the net excitation (i.e. mean firing rate) of one unit in the muscle was controlled. This served as the reference for the functional state of other active units during the prolonged contraction. With the mean firing rate of one unit in the muscle fixed, 58% of other active units showed a slow, statistically-significant change in mean firing rate over the 15 min. The initial firing rate of the units did not influence the change in rate. The original firing rate hierarchy, which in short-term contractions reflects the recruitment order, was altered during the prolonged contraction. The explanation for these differential changes in motoneuron net excitation is not clear; they could be intrinsic to the motoneurons or perhaps mediated by reflex pathways. The selective facilitation or suppression of some motor units with continuous activation means that the original size-structured combination of motor units can be modified during a prolonged contraction.     

Motor unit numbers and contractile properties after spinal cord injury

The number of motor units in the thenar muscle group was estimated in 11 patients with cervical spinal cord injuries. The surface electromyogram and twitch force, in response to maximal stimulation of the median nerve was divided by the average surface electromyogram and twitch of single units. The average single unit size was obtained by intramuscular microstimulation of motor nerve branches and by graded whole nerve stimulation, which provided three independent estimates, two based on the electromyogram and one based on force. The motor unit estimates from the patients covered a wide range. Some had essentially normal motor units both in numbers and contractile properties, while others had varying reductions in numbers of units. Those patients who showed a large reduction in motor unit numbers also had greatly enlarged units, which produced an average of up to sixfold the normal force. These enlarged units summed to produce maximal compound action potentials and twitches that were sometimes indistinguishable from normal. Magnetic resonance imaging scans of the cervical spine obtained from some patients provided independent evidence that patients with low motor unit counts had sustained direct injury to the anterior aspect of the spinal cord at the relevant segmental levels. Some patients showed a normal number of motor units long after the injury. No evidence of transneuronal degeneration could be demonstrated in the thenar group in these patients with the current techniques.     

Antigravity posture for analysis of motor unit recruitment: the "45 degree test"

The maximum number of different motor unit action potentials (MUAPs), their firing rates, and total MUAP spikes/second recorded by monopolar needle electrode were determined for the biceps brachii muscle during 45-degree elbow flexion. There were 4.2 +/- 1.6 different MUAPs exceeding 100 microV. Mean firing rate was 10.0 +/- 1.7 Hz, and total MUAP spikes/second were 40.3 +/- 18. Recordings from 16 patients with neurogenic atrophy (NA) and just detectable weakness revealed corresponding values of 3.1 +/- 1.7 different MUAPs, a mean rate of 10.2 +/- 1.5 Hz and 30.6 +/- 19 total MUAP spikes/second, not different from normal. In these patients, increased force of muscle contraction was required to activate high threshold motor units firing at high rates. In each of 4 patients just able to hold the arm against gravity, 1 or 2 "overdriven" motor units firing at a mean rate greater than 20 Hz were recorded. In 8 patients with myopathy and just detectable weakness, greater than 100 total MUAP spikes/second were recorded. Antigravity posture as a reference level of innervation has the advantage that motor unit firing rate is set about that of physiologic tremor (10-13 Hz). Its application was helpful in quantifying recruitment.     

Functional compensation in partially denervated muscles

In patients with various types of chronic motor denervation, the numbers of surviving motor units have been compared with the twitch tensions developed by the same muscle (extensor digitorum brevis). It was found that functional compensation in partially denervated muscles was often marked; in most patients abnormally small twitches occurred only when fewer than 10% of motor axons remained. The factors responsible for this compensation are considered. The twitch speeds of partially denervated muscles differed markedly, even among patients with the same disorder; there was evidence to suggest that the twitches of some motor units might become slower than those found in normal muscles.     

Fast and slow twitch units in a human muscle

A study has been made of the isometric twitches of single motor units in the extensor hallucis brevis (EHB) muscle in man. The twitch contraction times ranged from 35 to 98 msec and the pooled results indicated the presence of at least two types of unit. The recorded twitch tensions also varied considerably (from 2 to 14 g) but were not related to the twitch speeds. The twitch tension developed by the whole EHB muscle depended on the initial length of the muscle and on the age of the subject. The EHB muscle was estimated to contain approximately 56 motor units.     

Properties of motor units in the first deep lumbrical muscle of the cat's foot.

Isometric contractions of single motor units were studied in the first deep lumbrical muscle of the cat's hind-foot. Motor units with short twitch contraction times (15-20 msec) generally differed from those with longer ones (23-50 msec; contraction time measured in unpotentiated twitches) in showing (1) a greater maximum tetanic tension, (2) a smaller resistance to fatigue, (3) more post-tetanic potentiation of twitch tension, and (4) no post-tetanic occurrence of repetitive activity in response to single nerve stimuli (such "post-tetanic repetitive activity" was seen in several of the slower units). The ratio between unpotentiated twitch tension and maximum tetanic tension was similar for units with brief and long contraction times. The peak-to-peak amplitude of a single motor unit spike, recorded with gross electrodes, tended to be directly proportional to the maximum tetanic tension of the same motor unit.     

Reliability of the interval death rate analysis for estimating the time course of the motoneurone afterhyperpolarization in humans

The reliability of the afterhyperpolarization (AHP) time course, as estimated by the interval death rate (IDR) analysis was evaluated both within and between investigators. The IDR analysis uses the firing history of a single motor unit train at low tonic firing rates to calculate an estimate of the AHP time course [Matthews PB. Relationship of firing intervals of human motor units to the trajectory of post-spike after-hyperpolarization and synaptic noise. J Physiol 1996;492:597-628]. Single motor unit trains were collected from the tibialis anterior (TA) to determine intra-rater reliability (within investigator). Data from the first dorsal interosseus (FDI), collected in a previous investigation [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 2003;552:657-64], were used to examine the inter-rater reliability (between investigators). The lead author was blinded to the original time constants and file identities for the re-analysis. The intra-rater reliability of the AHP time constant in the TA data was high (r(2)=0.88; p<0.001; ICC=0.91). The inter-rater reliability for the FDI data was also strong (r(2)=0.92; p<0.001; ICC=0.95). The standard error of measurement was 0.61 ms for the TA and 0.55 ms for FDI. It is concluded that the interval death rate analysis is a reliable tool for estimating the AHP time course with experienced investigators.     

Extraocular motor units: type classification and motoneuron stimulation frequency-muscle unit force relationships.

Extracellular and intracellular techniques were used to study single motor units of the abducens nucleus and lateral rectus muscle in the cat. Using a combination of two motor unit properties, the fusion frequency and an index of fatigability, the population of twitch motor units could be separated into 4 subgroups: fast fatigable (FF), fast fatigue resistant (FR), slow fatigable (SF) and slow fatigue resistant (S). Nontwitch motor units, a fifth subgroup (NT), formed 10% of the total studied population. The twitch tension and the maximum tetanic tension of the FF motor unit type were significantly stronger than all other motor unit types. The use of frequency varying stimulation patterns did not further differentiate the motor unit types. The relation between a series of single motoneuron stimulation frequencies and the resultant single muscle unit forces generated a slope defined as a motor unit's kt value. Motor units with low kt values had higher twitch tensions, higher maximum tetanic tensions, higher fusion frequencies and lower fatigue indices than motor units with high kt values. Motoneuron recruitment was tested by electrical stimulation of the medial rectus subdivision of the contralateral oculomotor nucleus. No correlations were seen between recruitment order and the mechanical parameters of the single abducens motor units.     

Comparison of conventional and decomposition-enhanced spike triggered averaging techniques.

OBJECTIVE: Spike triggered averaging (STA) is a technique to extract an estimate of a recurring motor unit potential from a complex electromyographic (EMG) signal. In conventional STA (C-STA), potentials related to the discharges of single intramuscular motor units are isolated and used to trigger an averager to obtain an individual surface-detected motor unit potential (S-MUP) from an EMG signal. In decomposition-enhanced STA (DE-STA), EMG signal decomposition algorithms determine discharges of a number of different motor units (4 to 10) that can be used to trigger an averager to obtain their corresponding S-MUPs. We tested the accuracy of extracting and averaging S-MUPs using DE-STA compared to C-STA for the same EMG signals. METHODS: We compared the intramuscular potentials used for triggering and the resultant averaged S-MUPs that were common in both techniques. RESULTS: We found no statistically significant differences in the metrics used to describe the triggering potentials and S-MUPs. CONCLUSIONS: We conclude that DE-STA is an accurate and efficient method to obtain a large number of intramuscular motor unit potentials and their corresponding S-MUP in proximal and distal muscles.     

Reliability of spike triggered averaging of the surface electromyogram for motor unit action potential estimation

Introduction: The reliability of estimated motor unit parameters using spike triggered averaging (STA) of the surface electromyogram (sEMG) has not been tested thoroughly. We investigated factors that may induce amplitude bias in estimated motor unit action potentials (MUAPs) and shape variations. Methods: An sEMG record was simulated. MUAPs were then estimated from the STA of the simulated EMG. Results: Variations in MUAP duration led to under‐estimation of real MUAP amplitude, while synchronized firing led to over‐estimation of amplitude. Spurious firing resulted in over‐estimation of the amplitude of small motor units but under‐estimation of the amplitude of large ones. Variability in amplitude and high firing rates had minimal influence on amplitude estimation. High firing rates and variation in MUAP duration led to large variations in MUAP shape. Estimation errors also correlated with shape variations. Conclusions: Recommendations to enhance the accuracy of the STA estimates have been proposed. Muscle Nerve 48 : 557–570, 2013