| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
1Department of Orthodontics, School of Dental Medicine, University of Freiburg im Breisgau, Freiburg i.Br., Germany; 2 Institute of Neurology, Department of Clinical Neurophysiology, Radboud University Nijmegen Medical Centre and Interuniversity Institute for Fundamental and Clinical Human Movement Sciences (IFKB), Amsterdam/Nijmegen; and 3FC Donders Centre for Cognitive Neuroimaging, University of Nijmegen, Nijmegen, The Netherlands
Submitted 11 March 2005; accepted in final form 6 July 2005
 |
ABSTRACT |
|---|
|
|
|
INTRODUCTION |
|---|
|
; Braus and Elze 1954; Kennedy and Abbs 1979; Nairn 1975; Salmons 1995). Facial movements are mediated by an integrated network that includes neural systems mediating both voluntary and emotional drive from the cortex and the facial nucleus (Hopf et al. 1992; Morecraft et al. 2001).
|
). This is especially true for systematic electrophysiological investigations at a single motor unit (MU) level. In the few electromyographic (EMG) studies on facial MU characteristics, firing behavior (Blair 1988; Kamen and De Luca 1992), recruitment patterns, and reflex responses (Lansing et al. 1991; Mateika et al. 1998; McClean 1991; McClean and Smith 1982; Smith et al. 1981; Valls-Sole et al. 1992; Wohlert 1996b) were examined mostly with needle electrodes (the standard technique for electrophysiological studies at the MU level). Until now, other basic properties of facial MUs, such as their spatial spread and orientation, and the location of the endplate zones, were not specified (except a rough determination of MU territories in the human lip musculature; Goffman and Smith 1994). High-density surface electromyography (sEMG; Blok et al. 2002; Farina et al. 2000; Huppertz et al. 1997; Masuda and Sadoyama 1988; Prutchi 1995; Wood et al. 2001; Yamada et al. 1987) is the obvious method of choice to fill that gap of knowledge. However, this technique could not be applied to the facial musculature because of technical limitations and lack of appropriate signal processing tools, combined with the methodological demands for EMG studies in the facial area (Cole et al. 1983; Lapatki et al. 2003). The latter are related to 1) uneven skin contours in the face requiring highly flexible electrode arrays; 2) skin distortions during facial muscle contractions and flow of saliva, both increasing the risk of sensor attachment loss; and 3) general problems in attaching an electrode array to the face. These problems have been solved by the development of thin and highly flexible multielectrode grids together with a suitable skin attachment technique for this type of sensor (Lapatki et al. 2004). This study was undertaken to systematically determine topographical MU characteristics (i.e., their location, spatial orientation, and endplate zone) of the facial musculature by means of high-density sEMG, and to compare the results to basic anatomical knowledge and data obtained in histochemical studies. It is expected that the results will significantly contribute to our basic knowledge on human (neuro-) physiology, and also enable informed guidelines to be derived for the placement of conventional (surface or needle) EMG electrodes. In this article, results of measurements in the lower face are presented. In a subsequent paper we will publish the results from the mid- and upper face.
|
|
METHODS |
|---|
|
The study group consisted of 13 subjects (six men and seven women, mean age 27.2 yr, range 21–43 yr) without known neurological or general health disorders. Nine of the subjects were trumpeters (seven music students and two professionals), who were expected to have good facial motor control and to be motivated to improve on that. In the run-up to the measurements, a special training program was designed and implemented to improve the ability of the participants in performing selective contractions (i.e., isolated activations of the individual facial muscles). The Ethics Commission of the University Medical Centre Nijmegen (The Netherlands) approved the protocol of the study.
Sensors for high-density sEMG measurements
The recently developed multielectrode sEMG grids (Digiraster Tetzner GmbH, Stuttgart, Germany) applied in this study were manufactured using flexprint techniques. They consist of regularly arranged, chlorided silver electrodes 2 mm in diameter with an interelectrode distance (IED) of 4 mm (center to center) in both directions. This IED value allowed a sufficient spatial sampling for the facial musculature. These electrode grids (Fig. 2A) were attached to the skin using double-sided adhesive tape (1522 medical double-coated tape, 3M, St. Paul, MN) that had been prepared for this application by creating regular perforation patterns.
|
). Data acquisition
Electrical connections to the amplifiers were made with the help of a headset equipped with flexible arms and connectors for the electrodes (Fig. 2B). Data were acquired "monopolarly," i.e., referred to a remote reference electrode (material: Ag/AgCl, diameter: 4 mm) attached on the dorsum nasi. A second reference electrode provided a common mode signal (CMS). In addition, a driven right leg (DRL) electrode was attached at the forehead (Metting van Rijn et al. 1990).
The system used for data acquisition (Mark-6, BioSemi, Amsterdam, The Netherlands) has an input impedance >100 M
and a common mode rejection ratio (CMRR) >120 dB. Signals were band-pass filtered (3.2–400 Hz; high-pass: first-order Bessel; low-pass: fourth-order Bessel) and synchronously sampled at 2,000 Hz with a resolution of 0.5 µV/bit over a range of ±16 mV (16 bits). The acquisition software allows the experiments to be controlled by on-line inspection of the raw monopolar or bipolar data of selected electrode rows or columns (Blok et al. 2002). Because contractions of facial muscles at a constant level are difficult to perform, we implemented an sEMG amplitude feedback tool in the acquisition software that visually displayed the normalized root mean square (RMS) value of selected bipolar signals (and so the activity level of the underlying muscle) to the subject by the PC monitor. The RMS was calculated for data windows of 500 ms; normalization was made to a maximal reference contraction of the muscle.
Setup for the measurements and electrode positioning
We examined the subjects' lower facial musculature unilaterally either on the right (seven subjects) or the left (six subjects) side of the face. Two grids of 6 x 10 electrodes were positioned side by side, but vertically displaced by one electrode row according to criteria explained in the legend to Fig. 3.
|
After electrically connecting the grids, subjects were again instructed and trained in the performance of the experimental tasks now controlled by on-line inspection of selected signal amplitudes (i.e., the visual amplitude feedback derived above). When the selective motor control in the lower face was considered to be sufficient and the subjects were accustomed to the recording conditions, data (120 channels) were acquired while the depressor anguli oris (DAO), the depressor labii inferioris (DLI), the mentalis (MEN), or the orbicularis oris inferior (OOI) muscle was selectively activated at different levels. The facial poses corresponding to the selective contractions of these four lower facial muscles are shown in Fig. 4. In each muscle we performed two recordings per activity level, each lasting about 20 s. The first recordings were made at the activity level at which clear MU firings could be determined in the bipolar signals over the area of the corresponding muscle; this level corresponded most often with 2–3% of maximal voluntary contraction (MVC). Then, the subjects were asked to perform contractions at 5, 10, 15, 20, 25, and 30% of MVC so that in total 14 x 20 s of data were recorded per subject and muscle.
|
Data were analyzed off-line using algorithms programmed in Matlab, Version 6.5 (The MathWorks, Natick, MA). After extracting MUAPs from all recorded data (decomposition) we selected for each subject and muscle a data file containing a representative set of MUAPs; the physiological parameters of interest, the MU endplate positions and muscle fiber orientations, were then extracted from these MUAPs; finally, to allow the comparison and combination of the individual results, the extracted topographical MU parameters were normalized.
Decomposition of the sEMG pattern
MU action potentials were extracted from the first 16 s of the recorded data on the basis of the specific topographical profiles and amplitude characteristics of the peaks. The decomposition procedure was based on a previously described method (Kleine et al. 2000) and primarily consisted of the following three components: 1) peak detection in three selected bipolar signals, 2) classification of the data around the detected peaks according to their differential spatiotemporal amplitude characteristics, and 3) averaging broader data windows around the classified peaks, for each peak group, and for all recording channels. Most of these steps were largely automated.
Peak detection was performed in three bipolar signals recorded in distinct locations along the fibers of the contracted muscle (indicated as an example by the three small gray rectangles in the left grid of Fig. 3). Thus we took into account that facial MUs can lie in distinct areas of the muscle. The algorithm guaranteed that peaks within one channel occurring <8 ms apart were eliminated from the determined list of peak-occurrence times, and that peaks resulting from a specific MU firing, detected in several channels, were considered only once.
Peak classification was based on the Ward's clustering algorithm (Everitt et al. 2001) using "Euclidean distances" between the spatiotemporal amplitude characteristics of 16 ms of data around the detected peaks. One important issue in this classification is the consideration of the differential amplitude topographies of the peaks. Preliminary observation of data from the facial muscles revealed that the amplitude topographies of facial MUAPs could be significantly distinguished in both the dimensions perpendicular and parallel to the muscle fiber direction. As a consequence, we selected an extended set of input channels for the peak classification procedure that represented distinct topographical regions of the corresponding muscle in both dimensions (indicated by the three rows of four bipolar signals illustrated in Fig. 3). The time shift between peaks from distinct channels, resulting from signal propagation, was accounted for by determining the peak latency (using a cross-correlation technique) and aligning the signals accordingly. The stopping criterion of the hierarchical clustering algorithm requires that the number of clusters be predefined. We chose this number to be approximately four times the number of active MUs expected. The latter number was estimated from the length of the evaluated data segment, the number of detected peaks, and the average firing rates of facial MUs (Blair 1988; Kamen and De Luca 1992). In this manner we ensured that there was a low probability that peaks with significantly distinct amplitude topographies (i.e., peaks belonging to distinct MUs) were classified into one group.
After performing spike-triggered averaging of the peaks for each classified peak group and recording channel (here we used 80 ms of data around the classified peaks to include the whole width of the MUAP), we obtained a set of low-noise MUAP templates for each activity level and repetition, which could be regarded as representative "fingerprints" of the MUs that contributed to the contraction. From this data set, MUAPs averaged from a small number of peaks were removed.
Selection of a representative MUAP template set
All decomposed data were inspected to select one set of MUAP templates (belonging to a certain recording level and repetition, respectively) per subject and muscle that best represented the distinct MUAP topographies observed in the template sets obtained from all recording levels and repetitions. To facilitate the differentiation of the MUAPs' distinct amplitude topographies, we scanned the template sets in their bipolar montage (see also Fig. 6). Topographical MU parameters were determined for only the selected MUAP template sets. The corresponding methods used the MUAP data in a monopolar montage.
|
The location of the innervation zone and the main muscle fiber orientation of a MU correspond to the location where the MUAP starts and in which direction it propagates. From the spatial and temporal information contained in an averaged monopolar MUAP, which can be illustrated either as a template or as a time sequence of amplitude maps (see Fig. 5, A or D), both of these parameters can be derived. Our method included the following steps: 1) two-dimensional interpolation of the monopolar MUAP data; 2) determination of the initial phase of the MUAP (MUAP initiation); 3) localization of the standing maximal amplitude area in the initial MUAP phase (and endplate zone, respectively) on the electrode grid; 4) determination of the latencies (in both directions of signal propagation) when the propagation terminated or became diffuse (MUAP termination); and 5) localization of the maximal amplitude area on the electrode grid at the terminal MUAP latency (in both directions of propagation) and subsequent calculation of the main muscle fiber orientations.
|
The method for determining the initial phase of the MUAP was based on the fact that the main peak of the monopolar MUAP appears first in the grid area around the endplates. Another important aspect is that the monopolar amplitude values in this area are usually high compared with the amplitudes in other grid areas and thus contribute significantly to the average of the absolute monopolar MUAP signals over the whole grid (called "single time series"; see Fig. 5C). For these reasons, the MUAP initiation can be determined by means of a predefined amplitude threshold in the single time series. We used a threshold of 40% of the maximum level: this value was on the one hand usually above the average noise and smaller peaks in the single time series and, on the other hand, below the average value of the signals around the endplate (so that their leading edge was detected). It is important to note that not the real start of the MUAP but a slightly longer latency was determined by our algorithm. However, because there is no apparent signal propagation in the initial MUAP stage, this had only a negligible influence on the accuracy of localizing the endplate zone.
The MUAP initiation was topographically characterized by searching for the (interpolated) electrode grid location of maximal amplitude in the corresponding interpolated monopolar amplitude map (Fig. 5D, latency 19). To reduce the influence of inhomogeneities in the area of maximal MUAP amplitude, the algorithm did not localize the single maximum amplitude value, but a maximal amplitude area (see black frame in Fig. 5D, latency 19).
The two latencies indicating the terminal stage of signal propagation away from the endplate zone in both (opposite) directions were determined by observing the complete amplitude map sequences of the corresponding MUAP. The main criteria for detecting the terminal MUAP stage included finding either a drastic reduction in amplitude [i.e., a decrease of about 50% within one IED (Fig. 5D, upper fiber direction)], a beginning diffuse displacement of the propagating high-amplitude zone (Fig. 5D, lower fiber direction), or a stationary amplitude maximum.
The position on the electrode grid corresponding to MUAP termination was localized in the same way as described for MUAP initiation, i.e., we also determined the position of the zone of maximal amplitude in the interpolated monopolar amplitude maps of the corresponding latencies (as shown by the black frames in Fig. 5D, latencies 23.5 and 24.5). The orientation of the muscle fibers could then be characterized as two vectors pointing from the endplate position to the (two) positions where the MUAP termination was localized.
Accurate MUAP endplate location and muscle fiber orientation were verified by a final, careful observation of the complete amplitude map sequence (or "MUAP movie") with the superimposed results. If necessary, determination of these parameters was repeated with adapted threshold values and latencies, respectively.
Normalization
The electrode grids were positioned according to the criteria specified in the legend of Fig. 3. The distances from the electrode grids to the facial midline, to the corner of the mouth, as well as to the median and lateral lower mandibular border were quantitatively determined after the sEMG recordings directly on the subject's face. Using the known interrelations between these specified facial landmarks and the electrodes it was possible to define the endplate positions and muscle fiber orientations (previously determined in electrode coordinates) in an individual facial coordinate system. Separately for each subject, we then applied a spatial scaling of the data to account for the different sizes and shapes of the subjects' faces; the individual facial dimensions (characterized by the center and corner of the mouth as well as the lower mandibular border at the facial midline and at the mediolateral level of the corner of the mouth) were warped so that they agreed with the average facial dimensions of the study group. The (constant) horizontal scaling factor for this warp was determined by the relation between the individual and average distance between the corner and the center of the mouth. In the vertical dimension, separate scaling factors were calculated at the facial midline and at the corner of the mouth. The vertical scaling at other mediolateral levels varied linearly between these two scaling factors. After applying this spatial normalization, we could combine the data of all subjects in one figure and perform interindividual comparisons. Because no systematical differences between endplate positions and fiber orientations determined on the right and left face were found in any muscle, we finally performed a "mirror-image transformation" of the parameters obtained from measurements on the left facial side to facilitate interindividual comparison.
|
|
RESULTS |
|---|
|
The correct localization of the endplate (our method used the monopolar MUAPs) can be verified in the bipolar templates (Fig. 6, C and D), in which the endplate position is indicated by low signal amplitude (in between signals of high amplitude) or reversal of signal polarity. The MUAP with higher mono- and bipolar amplitudes (Fig. 6, A and C) represents the majority of the templates decomposed from this recording and has its endplates in the craniocaudal (i.e., vertical) center of the grid. The graphic in the middle of Fig. 6 shows that this location roughly corresponds with the craniocaudal center of the muscle. The MUAP presented on the right side (in monopolar and bipolar montage) has a smaller-amplitude territory and its endplate zone has a more cranial and medial location.
Figure 7 shows the endplate zones and DAO muscle fiber orientations of two representative subjects (Fig. 7, A and B) and those parameters for all subjects included in our study (Fig. 7, C and D). The data illustrated in Fig. 7A were derived from the MUAP template set from which two examples were presented in Fig. 6. The occurrence of a dominant endplate zone could be observed in 12 of the 13 subjects (one subject could not voluntarily contract the DAO muscle at all). In nine of these 12 subjects, a second endplate cluster (or single additional endplates) was found either in the cranial–medial direction (in eight subjects; see examples in Fig. 7, A and B), or in the caudal–lateral direction (in one subject). With the exception of one subject, the MUs of this additional cluster had a smaller electrical size (i.e., the MUAPs had lower-amplitude magnitudes) and smaller territories (indicated by the somewhat shorter muscle orientation lines).
|
The data from two representative subjects from the DLI muscle (Fig. 8, A and B) demonstrate the occurrence of MU endplates only in the lower portion of the muscle, typically in a lateral and medial cluster. The fibers above the endplates are mainly orientated toward the medial portion of the respective lower lip quadrant. Fiber orientation in the caudal portion of the DLI muscle could not be determined for most decomposed MUAPs because of the short distance from the endplates to the origin.
|
Figure 9, A and B exemplifies that MEN MU endplates could be found in both superior and inferior muscle portions, with a predominance of the superior location in most of the subjects. Typically, the superior endplates were concentrated in a relatively delimited area; a few inferior endplates were more widely distributed in the muscle's caudal portion (see as an example Fig. 9A).
|
Figure 10, A and B, which shows individual results from the OOI muscle, exemplifies the distribution of endplates on the whole area of the muscle and the presence of horizontal and short skew OOI fibers, as well as stationary (i.e., nonpropagating) MUAPs.
|
|
|
DISCUSSION |
|---|
|
High signal quality was achieved with the multielectrode sEMG grids especially developed for this study (Lapatki et al. 2004). Because of the high mechanical flexibility of the new sensors, good electrical connections of the single electrodes could be achieved even in subjects with very uneven contours in the recording area, such as those attributed to a deep mentolabial sulcus. Moreover, the flexible electrode grids could, to a large extent, follow the skin movements resulting from the lower facial muscle contractions. As a result the sensors did not significantly hinder the performance of the experimental tasks.
Two-dimensional topographical information about the electrical activity of a muscle allows the extraction of single MU potentials from the sEMG interference pattern (Masuda et al. 1985; Rau and Disselhorst-Klug 1997; Stegeman et al. 2004). Because we applied a largely automated decomposition method and we did not consider the information about the firing moments of the MUs, the signal decomposition in our study was inherently incomplete. The risk that peaks of MUs with distinct amplitude profiles could have been grouped together and averaged respectively was largely minimized by classifying the detected peaks into a much greater number of groups than the number of expected active MUs. Thus the only possible compromises for our results were 1) either the averaging of more or less identical MUAP peaks belonging to distinct, but electrophysiologically (and probably anatomically) similar MUs, or 2) the subdivision of the peaks belonging to a single MU into two (or more) different peak groups. In both cases, the consequences for the determined endplate zone distributions and muscle fiber orientations were only of minor significance with regard to the mapping of these topographical MU parameters and the interpretation of the results with respect to anatomical features.
Besides histological and histochemical techniques (Aquilonius et al. 1984; Christensen 1959; Desmedt 1958; Snobl et al. 1998), high-density sEMG techniques (Hilfiker and Meyer 1984; Masuda and Sadoyama 1988; Masuda et al. 1983) and EMG techniques with several inserted invasive electrodes (Lateva et al. 2002) have already proven in previous studies to be suitable to determine the locations of the MU endplates and their distribution on the muscle. The underlying hypothesis that the MUAP amplitude waveforms are closely related to aspects of MU (and muscle fiber) morphology is well studied and an important issue in electrodiagnostic medicine (Dumitru 2000; Lateva and McGill 2001). In the present study, we focused on two main aspects of MU topography: the endplate zone location and the main muscle fiber orientation. With respect to the latter parameter, it is important to note that we determined the muscle fiber orientation in the main (central) area of each MU and did not always trace the propagating signal to the end of the muscle fibers (e.g., in the case of diffuse terminal propagation). The only possibility to localize fiber ends with an electrophysiological technique would rely on the detection of "end effects" in the MUAP (Dumitru 2000; Dumitru and King 1999; Stegeman et al. 1997), which were in most instances not visible in our averaged facial MUAPs. End effects may not be visible from these data because of relatively short muscle fiber lengths (i.e., the end effects would be covered by the main peak of the MUAP) and aspects of facial muscle architecture (i.e., the diffuse interdigitation and interweaving of muscles at their origins and insertions; see Fig. 1), which reduce the "sharpness" of the end effect.
A general problem in the interpretation of sEMG data in anatomical respects is connected with the differential movement of the muscle tissue and the electrodes during the contraction. This problem can be especially relevant in recordings with multielectrode sEMG grids that reduce the mobility of the skin in the relatively large recording area and therefore may impede the skin in following the contracting muscle tissue. In our study, the MUAP template sets selected for further evaluation were decomposed from contractions between 5 and 15% of MVC. In an exemplary analysis in the DAO muscle (contraction of this muscle is accompanied by relatively large skin movements) in three subjects and three MUs per subject, we estimated an average relative movement between the electrode grid and the underlying muscle when the contraction level was increased from 5 to 15% MVC of about 1 mm. This value indicates the possible, relatively small bias of our results arising from this effect. Accurate determination of topographical MU characteristics requires that the recording area completely spans the skin area above the examined muscles. With the 128 available recording channels and an IED of 4 mm, the recording area of the two attached electrode grids was about 20 cm2 in size. Because we optimally adapted the shape of this area to the average anatomical configuration of the underlying facial musculature, it was rare that the MU territory occurred at the border or even out of the recording area. An exception in this respect were the OOI MUs; in this muscle, endplate locations above the upper border of the medial electrode grid were unavoidable because sEMG grids cannot be placed on the lip vermilion where the density of the OOI fibers is the highest (Blair 1986). To minimize this "vertical" bias in our results as far as possible we positioned the sensors directly at the lower lip's vermilion border.
Discussion of results in an anatomical and physiological context
Topographical facial MU parameters (i.e., endplate positions and muscle fiber orientations) were imaged and quantitatively determined in relation to facial landmarks. The characteristic spatial profile allowed the decomposed MUAPs to be classified as belonging to certain facial muscle subcomponents. Thus it can be concluded that the complex facial muscle architecture could be successfully mapped with the applied techniques. Discussion of our results is possible in the context of both macro- and microanatomical knowledge.
In general, we found the facial muscle positions (estimated from the mono- and bipolar amplitude profiles and endplate locations) and fiber orientations in good agreement with descriptions and drawings (see, as an example, Fig. 1) in anatomical textbooks (Braus and Elze 1954; Nairn 1975; Salmons 1995; Sobotta 2000), as well as quantitative macroanatomical data (Kennedy and Abbs 1979). In anatomical literature, facial muscle architecture has often been described as having considerable interindividual variability compared with that of other musculature. An interesting finding in our study was that interindividual variability, mainly that of the MU endplate positions, could be significantly reduced by the spatial scaling of the data on the basis of the individual and average facial dimensions (see Fig. 7). Based on this finding, it is possible that a great part of the variation in facial muscle anatomy discussed in the literature is attributed to complex interindividual variations in facial size and shape.
Detailed histological studies on the myoneural junctions revealed that normal adult mammalian skeletal muscles are innervated by a single motoneuron (i.e., consist of fibers with a single MU endplate in the middle) and that the MU endplates are usually arranged to form a narrow band crossing the muscle's central region, described as the "motor band" (Aquilonius et al. 1984; Christensen 1959). This dogma of classical anatomy and physiology has been revised in several respects. There is growing histological and electrophysiological evidence that some mammalian muscles—including the human laryngeal muscles (Perie et al. 1997; Rossi 1990)—may contain fibers with more than one endplate (Lateva et al. 2002, 2003). Further studies with sophisticated electrophysiological techniques indicate that a variety of architectural organizations of MUs, including the presence of single and multiple endplate zones in central and noncentral locations (Lateva and McGill 2001), are present in normal human limb muscles. Such differences from the classical arrangement of MU endplates in one motor zone or band could also be observed in human facial muscles. Histochemical studies revealed that myoneural junctions are distributed on the facial muscles in round or oval-shaped clusters (Happak et al. 1997). Histological examinations provided evidence that most investigated facial muscles show a small number of clusters corresponding to the number of innervating terminal motor nerve branches. Exceptions were the orbicularis oris, orbicularis occuli, and buccinator muscles, in which MU endplates were found to be spread over the whole muscle resulting in a greater number of small motor zones. Other characteristic findings in these studies were a predominance of one MU endplate cluster (in most of the examined muscles) and eccentric positions of the MU endplates on the corresponding muscle fiber bundles. These distinctive characteristics of facial MU topography could also be observed in our electrophysiologic study.
In the DAO muscle, which is innervated by two nerve branches from the lateral margin (Happak et al. 1994, 1997), MU endplates were typically identified in two clusters (Fig. 7). The finding in a few subjects of DAO endplates in only a single cluster might be attributable to either an absence of a second endplate cluster (as a sign of interindividual variability) or, more likely, peaks of the smaller MUs belonging to an additional cluster were "covered" by the peaks of the larger MUs (Fig. 6) and were therefore not detected by our peak-detection algorithm.
The DLI, which is innervated from several motor nerve branches of the marginal mandibular branch, showed extreme eccentric endplate locations near its lower border, with a distribution of endplates in the mediolateral dimension in two (or in some subjects three) clusters (Fig. 8). Position of the MU endplates near the origin of this muscle at the lower mandibular border might be connected with the phylogenetic migration of the fibers from the neck to the area near the corners of the mouth and the interconnection with the platysma, respectively (Braus and Elze 1954). The predominance of either lateral or medial endplate location in the DLI muscle varied from subject to subject. Most DLI MUAPs had amplitude territories that spanned the region of the lower lip. This confirms histological findings of serial sagittal sections of human lower lip in which groups of vertical and skew muscle fibers with an origin inferior to the lower lip (i.e., in the central DLI muscle area) were found in between the horizontal OOI fibers (Blair 1986; Blair and Smith 1986). An interesting additional finding during our experiments was that some subjects were able to contract the DLI muscle either in its lateral or medial region. This ability reflects the endplate zone distribution on this muscle and can be interpreted as compartmentalization, i.e., within-muscle differences of anatomical and physiological parameters that partition the motoneuron pool of a single muscle into independently controlled subsets (Binder and Stuart 1980). The possibility of such heterogenic contraction was also observed in masticatory muscles (Blanksma and van Eijden 1995; Phanachet et al. 2001), which—similar to the facial muscles—stem from the branchial arches.
In the MEN muscle, a predominant MU endplate cluster was usually found in a location slightly lateral to the facial midline (vertically) between a quarter and half the distance from the lower lip vermilion's inferior border to the lower bony margin of the mandible. One or two additional, often less clearly delimited endplate zones were often located in the muscle's inferior portion. These results cannot be compared with other data because, to our knowledge, no histological study has been carried out on the MEN muscle. The finding of relatively dispersed MEN endplate locations in the combined data of all subjects might be connected with the high interindividual variability in soft-tissue thickness and contour in the lower lip and chin area (Kennedy and Abbs 1979) because the measured sEMG signals are a "projection" onto the surface of the electrode grids (which follows the uneven skin contours; see Fig. 2A). The difficulties in determination of the superior MEN muscle fiber orientation and the low amplitudes of the MEN MUAPs above the endplates, respectively, are explained by the three-dimensional fiber course and the filtering properties of the soft tissues: the MEN muscle has its origin on the anterior border of the alveolar process of the mandible and is therefore covered by the thick soft-tissue mass of the chin and the lower lip in its superior part.
The orbicularis oris muscle consists of four (upper/lower/right/left) quadrants, each of them anatomically subdivided in a marginal portion located within the lip vermilion, and a peripheral portion around the lips. The marginal and peripheral subcomponents of the investigated inferior orbicularis oris muscle (OOI) could not be separately examined in our study because of the impossibility of positioning sEMG electrode grids directly over the lip vermilion and the incapability of most subjects (even trained trumpeters) to contract the subcomponents separately. Anatomical and histological studies have shown that the OOI is innervated from its lateral side by two terminal nerve branches and has MU endplates spread over the muscle in multiple clusters. In agreement with these data from former studies, we observed dispersed MU endplate locations in the investigated inferior part of this muscle in most of the subjects, with a main concentration of endplates, however, in the center of the right or left lower orbicularis oris quadrants (i.e., approximately half the distance between the corner of the mouth and the facial midline). We were surprised to find MUs with only relatively small territories (when compared with the whole muscle's length) because in the anatomical literature horizontal OOI fibers are described to run from the modiolus to the facial midline without interruption (Braus and Elze 1954). This issue requires further electrophysiological and histological investigations. On the other hand, the small territories and distinct muscle fiber orientations might reflect the muscle's complicated fiber architecture. As already mentioned in the context of overlapping electrical territories of OOI and DLI MUs, serial sagittal sections of human lower lip revealed the presence of fiber groups with distinct orientation (i.e., horizontal, tangential, and vertical fibers) in the region "traditionally described as OOI." The relative masses of these fiber groups showed considerable variation as a function of medial–lateral and superior–inferior position (Blair 1986). Electrophysiological studies suggest that these distinct fiber groups in the OOI belong to MUs that are independently controlled and have differential mechanical functions (Abbs et al. 1984; McClean and Smith 1982). From all these electrophysiological and anatomical/histological data it can be concluded that MU architecture of the muscle tissue in the area of the lower lip is highly developed and very complex. Dispersed location of horizontal OOI MUs along the lips as well as the presence of intrinsic nonhorizontal OOI MUs and DLI MUs in this area endow the lips with the ability of highly differential variation of tension and 3-D shape. Such fine and differential motor control is required such as in production of speech sounds, drinking, mastication, and other oral functions.
An additional, frequent finding in averaged OOI MUAPs was the presence of two widely (i.e., 
10 mm) separated, propagating zones of high amplitude. Possible explanations for this phenomenon are either branching of the motor nerve in its distal part (Happak et al. 1994, 1997) or synchronization of the firings of distinct MUs. Synchronization of facial MUs has been described in previous studies (Folkins et al. 1988; Wohlert 1996a) and seems to be one of the components of differential neural regulation of lip muscle activity on a premotoneuron level depending on the motor task (Wohlert and Goffman 1994).
To conclude, in this electrophysiological study, the position and distribution of MU endplates and principal muscle fiber orientations of facial muscles have been mapped and quantitatively determined. Such fundamental data add substantially to the basic neurophysiologic and anatomical knowledge of the complex facial muscle system. The results obtained can also be regarded as a basis for improving both the accuracy and the reproducibility of conventional facial sEMG measurements. Facial sEMG has proved to be a valuable, noninvasive tool in many research and clinical applications in psychophysiology (Jancke et al. 1996; Lundqvist 1995; Root and Stephens 2003; Sloan et al. 2002), speech physiology and pathophysiology (Kelly et al. 1995; Leanderson et al. 1971; Ruark and Moore 1997; Wohlert 1996a,b; Wohlert and Hammen 2000), as well as in general dentistry and orthodontics (Lapatki et al. 2002; Tosello et al. 1998, 1999; Yamaguchi et al. 2000). Electrode placement in these studies has been orientated only toward the estimated positions and fiber directions of the examined muscles, with the intention in most of the studies to position the sensors on the center or "belly" of the examined muscle. Our results prove that this electrode position is suboptimal at least in some muscles, such as in the DAO muscle, because of the central location of most endplates (Fig. 7), which means that relevant factors for electrode placement (Freriks et al. 1999) had to be ignored in previous studies because of the lack of quantitative data on these variables. In our study, these gaps in fundamental anatomical and neurophysiological knowledge have been bridged to a large extent and, consequently, foundations have been laid for objective guidelines for facial EMG electrode placements. The presented results may also be relevant from a clinical perspective. The information about MU topography obtained in a group of healthy individuals is useful as normal data in the observation of regeneration and reinnervation of MUs after peripheral nerve injuries or muscle transplantation, and in the study of the topographical aspects and the characteristic alterations on the MU level of neuromuscular diseases affecting the facial musculature [e.g., facioscapulohumeral dystrophy (FSHD) and Möbius syndrome]. Finally, maxillofacial surgical reconstruction of facial musculature (e.g., in patients with cleft lip or tumors in the face) and therapeutic interventions with botulinum toxin (with respect to the definition of optimal locations for injections) may also benefit from this fundamental data on the facial musculature and myoneural junctions in healthy individuals.
|
|
GRANTS |
|---|
|
|
|
ACKNOWLEDGMENTS |
|---|
|
|
|
FOOTNOTES |
|---|
Address for reprint requests and other correspondence: B. Lapatki, Department of Orthodontics, School of Dental Medicine, University of Freiburg im Breisgau, Hugstetter Str. 55, D-79106 Freiburg i.Br., Germany (E-mail: bernd.lapatki{at}uniklinik-freiburg.de)
|
|
REFERENCES |
|---|
|
Aquilonius SM, Askmark H, Gillberg PG, Nandedkar S, Olsson Y, and Stålberg E. Topographical localization of motor endplates in cryosections of whole human muscles. Muscle Nerve 7: 287–293, 1984.[CrossRef][Web of Science][Medline]
Binder MD and Stuart DG. Motor unit–muscle receptor interactions: design features of the neuromuscular control system. In: Progress in Clinical Neurophysiology. Spinal and Supraspinal Mechanisms of Voluntary Motor contol and Locomotion, edited by Desmedt JD. Basel: Karger, 1980, p. 72–98.
Blair C. Interdigitating muscle fibers throughout orbicularis oris inferior: preliminary observations. J Speech Hear Res 29: 266–269, 1986.[Medline]
Blair C. Firing and contractile properties of human lower lip motor units during sustained isometric contractions. Exp Neurol 99: 269–280, 1988.[CrossRef][Web of Science][Medline]
Blair C and Smith A. EMG recording in human lip muscles: can single muscles be isolated? J Speech Hear Res 29: 256–266, 1986.[Medline]
Blanksma NG and van Eijden TM. Electromyographic heterogeneity in the human temporalis and masseter muscles during static biting, open/close excursions, and chewing. J Dent Res 74: 1318–1327, 1995.
Blok JH, van Dijk JP, Drost G, Zwarts MJ, and Stegeman DF. A high-density multichannel surface electromyography system for the characterization of single motor units. Rev Sci Instrum 73: 1887–1897, 2002.[CrossRef][Web of Science]
Braus H and Elze C. Anatomie des Menschen. Ein Lehrbuch für Studierende und 
rzte. Bewegungsapparat. Berlin: Springer-Verlag, vol. 1, 1954.
Christensen E. Topography of terminal motor innervation in striated muscles from stillborn infants. Am J Phys Med 38: 65–78, 1959.[Medline]
Cole KJ, Konopacki RA, and Abbs JH. A miniature electrode for surface electromyography during speech. J Acoust Soc Am 74: 1362–1366, 1983.[CrossRef][Web of Science][Medline]
Desmedt JE. Methods of studying neuromuscular function in humans: isometric myogram, excitation electromyogram, and topography of terminal innervation. Acta Neurol Psychiatr Belg 58: 977–1017, 1958.[Medline]
Dumitru D. Physiologic basis of potentials recorded in electromyography. Muscle Nerve 23: 1667–1685, 2000.[CrossRef][Web of Science][Medline]
Dumitru D and King JC. Motor unit action potential duration and muscle length. Muscle Nerve 22: 1188–1195, 1999.[CrossRef][Web of Science][Medline]
Everitt B, Landau S, and Leese M. Cluster Analysis. London: Arnold, 2001.
Farina D, Fortunato E, and Merletti R. Noninvasive estimation of motor unit conduction velocity distribution using linear electrode arrays. IEEE Trans Biomed Eng 47: 380–388, 2000.[CrossRef][Web of Science][Medline]
Folkins JW, Linville RN, Garrett JD, and Brown CK. Interactions in the labial musculature during speech. J Speech Hear Res 31: 253–264, 1988.[Medline]
Freriks B, Hermens HJ, Disselhorst-Klug C, and Rau G. The recommendations for sensors and sensor placement procedures for surface electromyography. In: European Recommendations for Surface Electromyography: Results of the SENIAM Project, edited by Hermens J. Enschede, The Netherlands: Roessingh Research and Development, 1999, p. 15–25.
Goffman L and Smith A. Motor unit territories in the human perioral musculature. J Speech Hear Res 37: 975–984, 1994.[Medline]
Happak W, Burggasser G, Liu J, Gruber H, and Freilinger G. Anatomy and histology of the mimic muscles and the supplying facial nerve. Eur Arch Otorhinolaryngol Suppl 7: S85–S86, 1994.
Happak W, Liu J, Burggasser G, Flowers A, Gruber H, and Freilinger G. Human facial muscles: dimensions, motor endplate distribution, and presence of muscle fibers with multiple motor endplates. Anat Rec 249: 276–284, 1997.[CrossRef][Medline]
Hilfiker P and Meyer M. Normal and myopathic propagation of surface motor unit action potentials. Electroencephalogr Clin Neurophysiol 57: 21–31, 1984.[CrossRef][Web of Science][Medline]
Hopf HC, Muller-Forell W, and Hopf NJ. Localization of emotional and volitional facial paresis. Neurology 42: 1918–1923, 1992.
Huppertz HJ, Disselhorst-Klug C, Silny J, Rau G, and Heimann G. Diagnostic yield of noninvasive high spatial resolution electromyography in neuromuscular diseases. Muscle Nerve 20: 1360–1370, 1997.[CrossRef][Web of Science][Medline]
Jancke L, Vogt J, Musial F, Lutz K, and Kalveram KT. Facial EMG responses to auditory stimuli. Int J Psychophysiol 22: 85–96, 1996.[CrossRef][Web of Science][Medline]
Kamen G and De Luca CJ. Firing rate interactions among human orbicularis oris motor units. Int J Neurosci 64: 167–175, 1992.[Web of Science][Medline]
Kelly EM, Smith A, and Goffman L. Orofacial muscle activity of children who stutter: a preliminary study. J Speech Hear Res 38: 1025–1036, 1995.[Medline]
Kennedy JG and Abbs JH. Anatomic studies of the perioral motor system: foundations for studies in speech physiology. In: Speech and Language, edited by Lass N. New York: Academic Press, 1979.
Kleine BU, Blok JH, Oostenveld R, Praamstra P, and Stegeman DF. Magnetic stimulation-induced modulations of motor unit firings extracted from multi-channel surface EMG. Muscle Nerve 23: 1005–1015, 2000.[CrossRef][Web of Science][Medline]
Lansing RW, Solomon NP, Kossev AR, and Andersen AB. Recording single motor unit activity of human nasal muscles with surface electrodes: applications for respiration and speech. Electroencephalogr Clin Neurophysiol 81: 167–175, 1991.[CrossRef][Web of Science][Medline]
Lapatki BG, Mager AS, Schulte-Moenting J, and Jonas IE. The importance of the level of the lip line and resting lip pressure in Class II, Division 2 malocclusion. J Dent Res 81: 323–328, 2002.
Lapatki BG, Stegeman DF, and Jonas IE. A surface EMG electrode for the simultaneous observation of multiple facial muscles. J Neurosci Methods 123: 117–128, 2003.[CrossRef][Web of Science][Medline]
Lapatki BG, Van Dijk JP, Jonas IE, Zwarts MJ, and Stegeman DF. A thin, flexible multielectrode grid for high-density surface EMG. J Appl Physiol 96: 327–336, 2004.
Lateva ZC and McGill KC. Estimating motor-unit architectural properties by analyzing motor-unit action potential morphology. Clin Neurophysiol 112: 127–135, 2001.[CrossRef][Web of Science][Medline]
Lateva ZC, McGill KC, and Johanson ME. Electrophysiological evidence of adult human skeletal muscle fibres with multiple endplates and polyneuronal innervation. J Physiol 544: 549–565, 2002.
Lateva ZC, McGill KC, and Johanson ME. Increased jitter and blocking in normal muscles due to doubly innervated muscle fibers. Muscle Nerve 28: 423–431, 2003.[CrossRef][Web of Science][Medline]
Leanderson R, Persson A, and Ohman S. Electromyographic studies of facial muscle activity in speech. Acta Otolaryngol 72: 361–369, 1971.[Medline]
Lundqvist LO. Facial EMG reactions to facial expressions: a case of facial emotional contagion? Scand J Psychol 36: 130–141, 1995.[Web of Science][Medline]
Masuda T, Miyano H, and Sadoyama T. The propagation of motor unit action potential and the location of neuromuscular junction investigated by surface electrode arrays. Electroencephalogr Clin Neurophysiol 55: 594–600, 1983.[CrossRef][Web of Science][Medline]
Masuda T, Miyano H, and Sadoyama T. A surface electrode array for detecting action potential trains of single motor units. Electroencephalogr Clin Neurophysiol 60: 435–443, 1985.[CrossRef][Web of Science][Medline]
Masuda T and Sadoyama T. Topographical map of innervation zones within single motor units measured with a grid surface electrode. IEEE Trans Biomed Eng 35: 623–628, 1988.[CrossRef][Web of Science][Medline]
Mateika JH, Essif EG, Dellorusso C, and Fregosi RF. Contractile properties of human nasal dilator motor units. J Neurophysiol 79: 371–378, 1998.
May T and Schaitkin BM. The Facial Nerve. New York: Thieme, 2000.
McClean MD. Lip muscle EMG responses to oral pressure stimulation. J Speech Hear Res 34: 248–251, 1991.[Medline]
McClean MD and Smith A. The reflex responses of single motor units in human lower lip muscles to mechanical stimulation. Brain Res 251: 65–75, 1982.[CrossRef][Web of Science][Medline]
Metting van Rijn AC, Peper A, and Grimbergen CA. High-quality recording of bioelectric events. Part 1. Interference reduction, theory and practice. Med Biol Eng Comput 28: 389–397, 1990.[CrossRef][Web of Science][Medline]
Morecraft RJ, Louie JL, Herrick JL, and Stilwell-Morecraft KS. Cortical innervation of the facial nucleus in the non-human primate: a new interpretation of the effects of stroke and related subtotal brain trauma on the muscles of facial expression. Brain 124: 176–208, 2001.
Nairn RI. The circumoral musculature: structure and function. Br Dent J 138: 49–56, 1975.[CrossRef][Web of Science][Medline]
Perie S, St Guily JL, Callard P, and Sebille A. Innervation of adult human laryngeal muscle fibers. J Neurol Sci 149: 81–86, 1997.[CrossRef][Web of Science][Medline]
Phanachet I, Whittle T, Wanigaratne K, and Murray GM. Functional properties of single motor units in inferior head of human lateral pterygoid muscle: task relations and thresholds. J Neurophysiol 86: 2204–2218, 2001.
Prutchi D. A high-resolution large array (HRLA) surface EMG system. Med Eng Phys 17: 442–454, 1995.[CrossRef][Web of Science][Medline]
Rau G and Disselhorst-Klug C. Principles of high-spatial-resolution surface EMG (HSR-EMG): single motor unit detection and application in the diagnosis of neuromuscular disorders. J Electromyogr Kinesiol 7: 233–239, 1997.[CrossRef][Web of Science][Medline]
Root AA and Stephens JA. Organization of the central control of muscles of facial expression in man. J Physiol 549: 289–298, 2003.
Rossi G. From the pattern of human vocal muscle fibre innervation to functional remarks. Acta Otolaryngol Suppl 473: 1–10, 1990.[Medline]
Ruark JL and Moore CA. Coordination of lip muscle activity by 2-year-old children during speech and nonspeech tasks. J Speech Lang Hear Res 40: 1373–1385, 1997.
Salmons S. Muscle. In: Gray's Anatomy. The Anatomical Basis of Medicine and Surgery (38th ed.), edited by Williams PL. New York: Churchill Livingstone, 1995.
Sloan DM, Bradley MM, Dimoulas E, and Lang PJ. Looking at facial expressions: dysphoria and facial EMG. Biol Psychol 60: 79–90, 2002.[CrossRef][Web of Science][Medline]
Smith A, Zimmermann GN, and Abbas PJ. Recruitment patterns of motor units in speech production. J Speech Hear Res 24: 567–576, 1981.[Medline]
Snobl D, Binaghi LE, and Zenker W. Microarchitecture and innervation of the human latissimus dorsi muscle. J Reconstr Microsurg 14: 171–177, 1998.[Web of Science][Medline]
Sobotta J. Atlas der Anatomie des Menschen. Munich: Verlag Urban & Schwarzenberg, 2000.
Stegeman DF, Dumitru D, King JC, and Roeleveld K. Near- and far-fields: source characteristics and the conducting medium in neurophysiology. J Clin Neurophysiol 14: 429–442, 1997.[CrossRef][Web of Science][Medline]
Stegeman DF, Houtman CJ, Lapatki BG, and Zwarts MJ. Multichannel surface EMG. In: Handbook of Clinical Neurophysiology, edited by Stålberg E. Amsterdam: Elsevier, 2004.
Tosello DO, Vitti M, and Berzin F. EMG activity of the orbicularis oris and mentalis muscles in children with malocclusion, incompetent lips and atypical swallowing—part I. J Oral Rehabil 25: 838–846, 1998.[CrossRef][Web of Science][Medline]
Tosello DO, Vitti M, and Berzin F. EMG activity of the orbicularis oris and mentalis muscles in children with malocclusion, incompetent lips and atypical swallowing—part II. J Oral Rehabil 26: 644–649, 1999.[CrossRef][Web of Science][Medline]
Valls-Sole J, Tolosa ES, and Pujol M. Myokymic discharges and enhanced facial nerve reflex responses after recovery from idiopathic facial palsy. Muscle Nerve 15: 37–42, 1992.[CrossRef][Web of Science][Medline]
Wohlert AB. Perioral muscle activity in young and older adults during speech and nonspeech tasks. J Speech Hear Res 39: 761–770, 1996a.[Medline]
Wohlert AB. Reflex responses of lip muscles in young and older women. J Speech Hear Res 39: 578–589, 1996b.[Medline]
Wohlert AB and Goffman L. Human perioral muscle activation patterns. J Speech Hear Res 37: 1032–1040, 1994.[Medline]
Wohlert AB and Hammen VL. Lip muscle activity related to speech rate and loudness. J Speech Lang Hear Res 43: 1229–1239, 2000.
Wood SM, Jarratt JA, Barker AT, and Brown BH. Surface electromyography using electrode arrays: a study of motor neuron disease. Muscle Nerve 24: 223–230, 2001.[CrossRef][Web of Science][Medline]
Yamada M, Kumagai K, and Uchiyama A. The distribution and propagation pattern of motor unit action potentials studied by multi-channel surface EMG. Electroencephalogr Clin Neurophysiol 67: 395–401, 1987.[CrossRef][Web of Science][Medline]
Yamaguchi K, Morimoto Y, Nanda RS, Ghosh J, and Tanne K. Morphological differences in individuals with lip competence and incompetence based on electromyographic diagnosis. J Oral Rehabil 27: 893–901, 2000.[CrossRef][Web of Science][Medline]
This article has been cited by other articles:
|
|
 |
|
  D. Farina, F. Negro, M. Gazzoni, and R. M. Enoka Detecting the Unique Representation of Motor-Unit Action Potentials in the Surface Electromyogram J Neurophysiol, September 1, 2008; 100(3): 1223 - 1233. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Visit Other APS Journals Online |
TOP
ABSTRACT
INTRODUCTION
