Three-dimensional (3D) Topography of the Motor Endplates of the Rat Gastrocnemius Muscle

 

Dimiter Prodanov1,2, MD, Marie-Anne Thil3, MSc, Enrico Marani1,2, PhD, Jean Delbeke3, MD, PhD, and Jan Holsheimer2, PhD

1Neuroregulation Group, Department of Neurosurgery, Leiden University Medical Center, Leiden, The Netherlands

2Biomedical Signals & Systems Group, Faculty of Electrical Engineering, Twente University, Enschede, The Netherlands

3Neural Rehabilitation Engineering Laboratory, Catholic University of Louvain, Brussels, Belgium

 

Published in Muscle and Nerve, 2005.

Copyright - pre-acceptance version © 2004 by the authors; published version © 2005 by John Wiley & Sons, Inc. Personal use of this material is permitted. However, permission to reprint or republish this material for advertising or promotional purposes or for creating new collective works for resale or redistribution to servers or lists, or to reuse any copyrighted component of this work in other works, must be obtained from the copyright holder.

To avoid copyright issues, here follows the pre-acceptance version of the manuscript.

 

Abstract

Spatial distribution of motor endplates affects the shape of the electrical activity recorded from muscle. To provide information for realistic models of action potential propagation within the rat medial gastrocnemius (MGM) and lateral gastrocnemius (MGL) muscles, we mapped the distribution of their motor endplates in three-dimensional (3D) space. The maps were assembled from histological cross-sections stained for acetylcholinesterase activity. Within MGL, the motor endplates formed three columns along its longitudinal axis and were concentrated in the middle of muscle as suggested by the distribution of their cross-sectional projection area. Within MGM, the motor endplates were arranged in a leaf-like body that shifted obliquely from proximal to distal and were concentrated in the distal third of muscle. Regions of maximal motor endplate concentration are considered most suitable for injections of neuroactive substances, such as neuronal tracers. The assembled 3D maps of the gastrocnemius muscles can be used as guides for such injections within the motor endplate zones.

 

Address for correspondence: Prof. Dr. Enrico Marani, PhD, Neuroregulation Group, Leiden University, PO Box 9604, NL-2300 RC Leiden, The Netherlands

Key words: 3D-map, EMG simulation, modeling of action potentials, motor endplate, regeneration, VRML

Abbreviation list: AChE - Acetylcholinesterase; EMG - electromyogram; HRP - horseradish peroxidase; MGL - gastrocnemius lateralis; MGM - gastrocnemius lateralis; TPA - Total acetylcholinesterase positive particle area; VRML - Virtual Reality Modeling Language

 

Introduction

The rat gastrocnemius muscle consists of two main heads that are considered to be distinct muscles – the tripennate gastrocnemius lateralis (MGL) and the unipennate gastrocnemius medialis (MGM). MGL comprises lateral, intermediate, and medial subheads3,4. MGL and MGM are innervated by separate motor branches arising from the tibial nerve14. Rostrally, the gastrocnemius muscle is attached to the femur; caudally, its tendon joins the tendon of the soleus muscle in order to form the common calcaneal tendon (Achilles tendon)14.

The rat gastrocnemius muscle is often used as a model in the electrophysiological and functional assessment of recovery during peripheral nerve regeneration20, 26, 32, 33, 34. A common approach for such assessment is to inject retrograde tracers into the muscle and count the number of retrogradely labeled motoneurons5, 25, 33. Water-soluble retrograde tracers such as horseradish peroxidase (HRP) and Fluoro-Gold are mainly taken up by the motor endplates18, 31. Since the motor endplates are not restricted to a single spot inside the muscle and uptake depends on the availability of the tracer near them, “blind” injections may result in underestimation of the number of innervating motoneurons in the spinal cord16, 22, 31. Therefore, topographic description of the motor endplates will allow placement of the tracer injections in a specific and reproducible manner, which in turn will increase the quantity and reproducibility of labeling in the spinal cord.

Variations of single anatomical or physiological parameters are difficult to study in the electromyogram (EMG) in vivo. Therefore, numerous studies addressing the manifestation of different anatomical and physiological features of normal and pathological muscles have been performed in simulated EMGs9, 23, 28, 29, 30, 36. Theoretically, features such as the duration, area, and amplitude of the motor unit action potentials are determined by the geometry of muscle fibers within different parts of the motor unit territory24. Moreover, the presence of even a single muscle fiber close to a concentric needle electrode could lead to the occurrence of serrated motor unit potentials36. Accordingly, precise anatomical descriptions of the three-dimensional (3D) geometry of the muscle fibers and motor endplates of different muscles may help to model their electrical activity more realistically. 

Morphology and topography of the motor endplates in various muscles are typically revealed at the light-microscopic level by acetylcholinesterase (AChE) histochemistry (see, for example, the classic works of Coërs and Woolf 6 and Koelle and Friedenwald19). In the present study, motor endplates were visualized by means of the Karnovsky-Roots thiocholine iodide method17. The locations of the motor endplates were reconstructed three-dimensionally using specialized software as previously described by Fiala and Harris12.

The present study aims to provide a guide for both modeling studies and experimental procedures requiring the precise localization of the endplate zones within the gastrocnemius muscle in the rat. We assembled maps of the locations of motor endplates and quantified the longitudinal distribution of their projection area within the rat gastrocnemius muscles. Three columns of motor endplates were present within MGL, while within MGM the motor endplates were arranged in a structure resembling a curved plant leaf.

Materials and Methods

Animals

The experiments were performed in accordance with international (EU Directive 86/609/EEC) and national laws governing the protection of animals used for experimental purposes (Leiden University Animal Care and Experimentation Commission, UDEC 02076A). Four female Wistar Albino-Glaxo rats, weighing on average 250 g, were deeply anesthetized with Nembutal (Ceva Sante Animale, Libourne, France) and then perfused transcardially with 250 ml 0.9% sodium chloride solution, followed by 250 ml modified Karnovsky fixative (1% paraformaldehyde, 1.25% glutaraldehyde, 0.1M phosphate buffer, pH 7.20) 11, and finally by 250 ml 0.1M phosphate-buffered saline (PBS), pH 7.2. Under a surgical microscope, the right gastrocnemic muscles were dissected free from the surrounding tissues and collected in the same fixative. In three cases, MGM and MGL were dissected separately.

AChE histochemistry

The motor endplates were visualized by means of AChE histochemistry, employing the Karnovsky-Roots modified thiocholine iodide method17. The tissue was stained twice, once as a whole mount (in toto) prior to sectioning and once after sectioning and mounting. 

The first AChE staining procedure was performed to visualize the superficial motor endplates. The muscles were incubated for 2 hours at 37 ºC in a reaction medium containing 0.17 mM thiocholine iodide (Sigma-Aldrich, St. Louis, Missouri), 0.53 mM sodium citrate, 0.34 mM copper sulfate, and 0.05 mM potassium ferricyanide (Sigma-Aldrich) dissolved in a 0.05M sodium maleate buffer (pH 6.0) (see also Marani21). Photographs of the entire muscles were taken to assist in the 3D-map reconstruction process (Fig. 1).

Subsequently the muscles were embedded in 13% gelatin. In order to orient and align future sections, the gelatin blocks were trimmed obliquely so that an asymmetric shape was formed in the section plane. The gelatin blocks were serially sectioned on a freezing microtome (Leica Jung 1205, Leica Microsystems, Bensheim, Germany). One of the muscles was sectioned parallel to its longitudinal axis, whereas the others were sectioned perpendicularly. The muscles were sectioned at a thickness of 120 mm and each fifth section was mounted and air-dried. The remaining sections were kept until the end of the second staining procedure.

During the second AChE staining, the sections on the slides were incubated in the reaction medium at 37 ºC for 3 to 5 hours under microscopic control until a brownish-red reaction product was formed. Finally, the sections were rinsed in distilled water, passed through ascending series of alcohols, cleared in xylene, and permanently cover-slipped using Fluoromount (Merck, Darmstadt, Germany).

Reconstruction procedure

Three specimens of each muscle were used for assembly of the 3D map. The reconstructions were derived from 15-17 sections evenly spaced at 1.2 mm from each other (i.e., every 10th section in the series). This allowed for replacement of a section with an adjacent one with a minimal displacement error in space (10%) if a tissue defect was introduced in the motor endplate zone during clearing. Each cross-section was photographed with an CyberShot 5Mpix digital camera (Olympus Imaging Corp., Tokyo, Japan) under identical lighting conditions.

The subsequent pre-processing and reconstruction steps were performed using the specialized software for 3D reconstructions sEM Align and IGL Trace (Boston University, Boston, MA; http://synapses.bu.edu/tools/) as previously described12. The digital images of the sections were aligned by performing translations and rotations, i.e., “rigid” geometric transformations only.

Three classes of objects were defined and represented by contours in the images – muscle borders, motor endplate clusters, and motor endplate zones (Coërs’s terminal innervation bands6). In every image, the muscle border was manually outlined and the motor endplates were traced by circles, each representing a single cluster. Each motor endplate zone was defined as the minimal enclosing convex outline around a macroscopically distinct cluster of motor endplates (Figs. 2A and 2C) and manually drawn. The tracings of the objects were compared to the original sections by microscopic examination and improved if necessary. The objects were further exported to a common format for 3D data (Virtual Reality Modeling Language; VRML, ISO/IEC 14772) and assembled as solid bodies by surface rendering with the Actify Spinfire software program (Actify Inc.,
San Francisco, CA). Hereafter this composite 3D representation is referred to as the 3D map. In the 3D map, the muscle tissue boundaries were represented by their planar contours and the motor endplate clusters by spheres situated on the section planes. The motor endplate zones were connected by surfaces so that solid bodies could be formed. The 3D map was interactively inspected by rotations and translations in 3D space. The resulting muscle surfaces were compared to photographs of the original muscles and the alignment and assembly procedures were repeated if necessary.

Image processing and morphometric analysis

All of the processing and measurement steps employed in the morphometric analysis were performed by means of the freeware image analysis software package ImageJ (NIH, Bethesda, MA; http://rsb.info.nih.gov/ij/)1. Based on their color histograms, the digital images of the cross-sections of the gastrocnemius muscle were thresholded. Depending on its red, green and blue intensities, a foreground label was interactively assigned to every pixel in the image, so that a maximal overlap between the AChE-positive particles and the foreground was obtained. In such a way, binary images were produced where each foreground pixel represented an AChE-positive spot in the original section. The individual particles in the binary images were then automatically identified and measured. The total AChE-positive particle area per section (TPA) was calculated as the arithmetic sum of the individual particle areas obtained from the automatic measurements in each section. The close spacing of the individual motor endplates inside the bands resulted in merging of particles in the derived binary images. Therefore, TPA was chosen over the number of particles as a better measure for the quantity of motor endplates. The TPA of each section was plotted against its position along the longitudinal axis of the series. Hereafter this plot is referred to as the proximodistal axial distribution.

The numerical values are reported as mean ± standard deviation. Data were compared by performing Student’s t-test. Probability levels less than 0.05 were considered significant.

Results

Macroscopic observations

In the living animal, the length of the muscle belly was on average 22 mm. For descriptive purposes, we distinguished a proximal end, distal end, dorsolateral aspect, and tibial aspect of MGL. In MGM, we distinguished proximal end, distal end, dorsomedial aspect, and lateral aspect. The lateral aspect of MGM overlapped the medial third of the dorsolateral surface of MGL. The parva saphenous vein and sural nerve crossed obliquely the dorsolateral surface of MGL. After their removal, a shallow sulcus could be recognized on the surface of the muscle.

Fig. 1: Superficial AChE staining of the gastrocnemius muscle

Photographs of right MGM (tibial aspect, A; dorsîmedial aspect, B) and right MGL (tibial aspect, C; dorsîlateral aspect, D). The innervation bands of MGM appear as punctuated arches (A and B; arrow heads); the innervation band of MGL is present only on its lateral subhead (C; arrow head). Muscle aponeurosis appears as a white translucent band beginning from the Achilles tendon (A). Top, proximal end; scale bar, 1 cm.

The in toto staining of MGM revealed punctated bands of motor endplates on both dorsomedial and tibial surfaces (arrow heads, Figs.1A and 1B). On the tibial surface, the innervation band was situated at 1 cm from the proximal end of the muscle (arrow head, Fig. 1A), whereas dorsîmedially it was situated 1.25 cm from the distal end of the muscle tendon (arrow head, Fig. 1B).

The in toto staining of MGL revealed a faint oblique band of motor endplates on the surface of its lateral subhead, situated near the rim of the tibial surface at 1.4 cm from the distal end of the muscle tendon (arrow head, Fig. 1C).

Microscopic anatomy of the motor endplates

In cross sections and longitudinal sections alike, the motor endplates formed clusters, typically 150-200 µm in diameter (Figs. 2 D and E). The clusters were arranged along imaginary straight lines (Figs. 2 A, B, and C), each one representing an individual motor endplate zone.

In most sections of MGL, there were three motor endplate zones separated by the muscle aponeuroses. The distance between motor endplates in the different motor endplate zones was substantially larger than between individual motor endplates (Fig. 2A). Conversely, in the cross sections of MGM the motor endplates were arranged along a single line (i.e., in one motor endplate zone; Fig. 2C), which in some sections had a dashed appearance. At the proximal end of the series, the motor endplate zone was situated peripherally near the dorsomedial surface of the muscle (Fig. 1A). Further in the series, the motor endplate zone shifted its position and through the center of the section it eventually reached the tibial surface (Fig. 1B).

Microscopic examination of the cross sections revealed that the motor endplate zones were larger in the middle of MGL and in the distal third of MGM. The terminal arborizations formed clusters of motor endplates of the “en plaque” type (Fig. 2 D,E) as previously described for mammals35. In the longitudinal sections, the motor endplates were situated at the middle of the muscle fiber (Fig. 2B) in most instances.

Fig 2: AChE staining of cross and longitudinal sections of the gastrocnemius muscle

A. Cross section of MGL, case A1066. Three bands of motor endplates are visible (arrow heads). Offset from the beginning of the series: 6.0 mm. Orientation: left, medial; right, lateral; T, tibial surface; DL, dorsolateral surface. Scale bar, 1 mm. B. Longitudinal section, case A1135. The complex pennate structure of the muscle can be seen. Note the presence of three aponeuroses, enhanced by black dash-dot lines and the N-like configuration of the motor endplate bands within MGL (arrow heads). The motor endplate clusters can be seen as dark dots of approximately equal size. Top, proximal pole; bottom, Achilles tendon; scale bar, 0.25 cm. C. Cross section of MGM, case A1066. One punctated band of motor endplates is visible across the muscle section (arrow head). Offset from the beginning of the series: 10.8 mm. Orientation: left, medial; right, lateral; T, tibial surface; DL, dorsolateral surface; DM, dorsomedial surface. Scale bar, 1 mm. D. Cross section of MGL. A terminal nerve branch (asterisks) approaches a cluster of motor endplates (arrow head); scale bar, 50 µm. E. Longitudinal section of MGL. A cluster of motor endplates of the “en plaque” type (arrow head) is situated on top of several transversely striated muscle fibers; scale bar, 50 µm.

3D map representation

Lateral, central, and medial columns of motor endplates could be discerned in the 3D map of MGL. The columns remained separated throughout their course in all of the animals. Comparison with the longitudinal sections of MGL demonstrated a single column in every muscle subhead. The central column was situated bellow the sulcus left from the saphenous vein (Fig. 3).

On a frontal projection, the medial and central columns formed a sharp angle pointing towards the Achilles tendon (Fig. 3C). In the same projection, the medial and lateral columns formed a sharp angle pointing towards the proximal pole of the muscle (Fig. 3C). The entire constellation resembled a broad “N”-shaped structure (Fig. 3A). On a coronal projection, the lateral and the central columns formed a sharp angle, pointing towards the dorsolateral surface; the medial and central columns formed a sharp angle pointing in the direction of the tibia (Fig. 3B).

Fig. 3: 3D map of MGL

Case A1066; Section planes are perpendicular to the longitudinal axis of muscle (z-axis). Sections are 1.2 mm apart. The X-axis points in a medial direction; the Y-axis, to the tibial side of the muscle; the Z- axis, to the Achilles tendon. Boundaries of the muscle tissue are represented by planar contours. The motor endplate clusters are represented by spheres. A. Isometric projection of the 3D map (dorsal view). Three columns of motor endplates, boundaries indicated by blue space filling, are present within the muscle. The map is rotated so that the N-shaped figure is best seen (compare to Fig. 1C). B. Coronal projection. Top, tibial surface; bottom, dorsolateral surface. C. Frontal projection from the side of the tibia. The angle between the lateral and the central column is 18°. The angle between the medial and the central column is 20°.

In the 3D map of MGM, motor endplates were arranged in a structure resembling a plant leaf (Fig. 4). The longitudinal axis of this arrangement formed a sharp angle with both the muscle longitudinal axis and the aponeurosis (Fig. 4A and 2B). The leaf-like arrangement extended from the tibial surface of the muscle cranially to the dorsomedial surface caudally. The surface innervation bands corresponded to the 3D bodies of motor endplates inside the muscles. For MGM, this relationship can be demonstrated in Figs. 1 A, B, and 4, where the motor endplates touch the dorsomedial surface of the muscle. Only in the lateral compartment of MGL, the motor endplates could be revealed on the surface.

Fig. 4: 3D map of MGM

Case C6221; Motor endplate clusters are represented by spheres. Boundaries of the muscle tissue are represented by planar contours. Section planes are perpendicular to the longitudinal axis of the muscle (z-axis). Sections are 1.2 mm apart. The X-axis points in a medial direction; the Y-axis, to the tibial side of the muscle and the Z- axis, to the Achilles tendon. A. Isometric projection of the 3D map (medial view). The motor endplates form a leaf-like structure. The 3D map is rotated so that the configuration of the motor endplates in space is best observed. B. Coronal projection. C. Frontal projection from the side of the tibia.

The VRML versions of the 3D maps are available for inspection at http://www.neuromorf.com/models.php.

Morphometric analysis of the motor endplates

TPA was calculated as a measure for the quantity of motor endplates in each section. The mean TPA for MGL (0.412 mm2± 0.263 mm2) was larger than for MGM (0.295 mm2± 0.178 mm2; p<0.05, t-test). In order to compare and collate the axial distributions in the different muscles, we had to account for the variability caused by trimming before and during sectioning. To do so, the MGL series were aligned to each other by the mid-belly point of muscle. Each mid-belly point was identified by the maximum of the cross-section projection area of the muscle in the series. The same alignment minimized the case to case “similarity measure” of the individual TPA distribution curves, defined as the negative of the sum of the squared differences between the corresponding individual data points. The mean axial TPA distribution is represented in Figure 5. The axial distribution of MGL showed a broad peak with a 3.7-fold increase of TPA in the center of the series compared to the median value (from L=7.2 to L=14.4 mm; where L is the distance from the proximal end of the series; Fig. 5A).

Fig 5. Total particle area per section

Data are presented as mean ± SD. Proximal end of the C6221 case series is taken for offset 0 mm. A. MGL, the TPA ranged from 0.0360 mm2 to 0.7817 mm2. The peak spanned from 7.2 mm to 14.4 mm. B. MGM, inter-animal variation was larger than for MGL. A broad peak is present from 10.8 to 14.4 mm.

For MGM the variation in data was larger. Nevertheless, a broad peak in the TPA could also be discerned starting from L=10.8 to L=14.4 mm from the proximal end of the series (Fig. 5B). The increase was 2.21-fold compared to the median value of the series.

Discussion

Relationship between the 3D map and muscle compartmentalization

MGL is divided in different physiological compartments that can contract separately, each supplied by a distinct motor branch 3,4,10. The N-shape arrangement of the motor endplate columns is in good accordance with the tri-pennate structure of MGL. Surprisingly, the functional compartmentalization dictated by the motor branch pattern corresponds only partly to columnar architecture. We can only relate the medial column to the previously described medial muscle compartment 4. Little is known about the course and the pattern of the motor branches of MGL in the rat. Previous studies suggest similarity with the anatomy of cat MGL where five primary muscle nerve branches were described10. Therefore, we can possibly relate the central column of motor endplates to two of the centrally situated primary branches, and the lateral column to the distal lateral primary branch10.

MGM is innervated by up to four primary nerve branches, of which the proximal and distal ones are the largest7. This pattern can be recognized in the 3D map of the motor endplate zone that resembled a curved plant leaf with its veins probably corresponding to the pattern of the nerve branches.

Relationship between the 3D-map and TPA distribution

Observed concentration of the motor endplates in the 3D map corresponded to the location of the maxima in TPA distribution (Figs. 5 and 6). It is unlikely that the observed pattern of TPA distribution is caused by regional variations in density or numbers of muscle fibers, as that would imply a 2- to 4-fold increase in either of these parameters (Fig. 5). Regional differences in 3D fiber geometry (orientation and alignment of the muscle fibers) are more likely to explain the observed maxima in the projection area. Differences in size of motor endplates may also contribute, because the fiber composition varies in MGM7, 15 and the motor endplate size is related to the fiber type6.

Applicability in modeling research

The shape, amplitude, and duration of recorded motor unit potentials are highly dependent on the distance between the electrode and the motor endplates8,36 and therefore on the spatial configuration of the endplate zones. Complex motor endplate distributions, such as reported in the present study, will result in an increase in variation of these parameters of motor unit potentials. Indeed, such variability in the maximum-to-minimum time interval of the biphasic potential and in the peak-to-peak amplitude of potentials was shown experimentally in recordings from single motor units from the rat MGM13. However, even the most realistic EMG modeling studies performed so far have assumed only simple motor endplate spatial distributions28, 29. Therefore, realistic modeling of EMG requires that the geometry of the motor endplate zones is incorporated into models.

Applicability in tracing studies

Three-dimensional reconstructions of different muscles (biceps brachii, tibialis anterior, sartorius2 and thyroarytenoid27) have been utilized primarily in the clinic as guides for Botulinum toxin injections. We propose similar approach to be used for the application of substances, such as retrograde tracers, in the gastrocnemius muscle of rat. In our experience, this muscle is best approached for injections ad dorsum after small incision of the skin along the midline of the calf. In order to provide a guide for such injections, the locations of the motor endplate zones inside the muscles are demonstrated in combination with the quantity of the motor endplates in one representative case (Fig 6).

Fig 6. Relationship between the 3D map and the TPA distribution

Case A1066. A, Dorsolateral view of the 3D map of MGL; B, Corresponding TPA, horizontal bars are aligned to the section planes. C, Lateral view of the 3D map of MGM; D, Corresponding TPA, with horizontal bars aligned to the section planes. Microscopic inspection showed that the notch in the TPA distribution of MGM (asterisk) was due to nerve branches present in the cross sections at that level.

In Figure 6, muscles are displayed as approximately seen in such a surgical approach. Based on the depicted relationship between the topography and the area of motor endplates, we recommend injections to be made as following:

  1. In MGM, in a rostral direction along the longitudinal axis of the muscle at a depth of 2-2.5 mm, preferably in the caudal third of the muscle;
  2. In MGL, in a rostral direction oblique to the longitudinal axis of the muscle at a depth of 4-6 mm. For the lateral column, in a lateral direction; for the central column, parallel to the course of the saphena parva vein and the sural nerve; and for the medial column, in a medial direction.

The volumetric representation of the anatomical objects by means of VRML models provides better understanding of the topographical anatomy of muscle and the relationship with its innervation. A major benefit of the presented 3D map is the user interactivity, which enables a thorough examination of the 3D arrangement of the motor endplates and their relation to the muscle surface. The 3D map will facilitate precise injections close to the motor endplates, such as those needed in retrograde tracing experiments. Secondly, the spatial distribution of the motor endplates in the 3D map can be incorporated into EMG simulation models of the gastrocnemius muscle so that it can increase our understanding of the influence of the motor endplate geometry on EMG both in normal and in pathological conditions.

Acknowledgements

This research work was funded by the Research and Training Network NeuralPRO (European Commission Human Potential project, shared cost contract No HPRN-CT-2000-00030 - Neural Prostheses). Part of the study was reported in preliminary form at the Dutch Annual Conference on Biomedical Engineering, October 2003, Papendal, The Netherlands.

 

Reference List

1. Abramoff MD, Magelhaes PJ, Ram SJ. Image processing with ImageJ. Biophotonics Int 2004; 11: 36-42.

2. Aquilonius SM, Askmark H, Gillberg PG, Nandedkar S, Olsson Y, Stalberg E. Topographical localization of motor endplates in cryosections of whole human muscles. Muscle Nerve 1984; 7: 287-293.

3. Bennett M, Ho S, Lavidis N. Competition between segmental nerves at end-plates in rat gastrocnemius muscle during loss of polyneuronal innervation. J Physiol (Lond) 1986; 381: 351-376.

4. Bennett MR, Ho S. The formation of topographical maps in developing rat gastrocnemius muscle during synapse elimination. J Physiol (Lond) 1988; 396: 471-496.

 5. Brushart TM, Gerber J, Kessens P, Chen YG, Royall RM. Contributions of pathway and neuron to preferential motor reinnervation. J Neurosci 1998; 18: 8674-8681.

 6. Coers C, Woolf AL. Normal histology of the intramuscular nerves and nerve endings. In: Coers C, Woolf AL, editors. The innervation of muscle; a biopsy study. Oxford: Blackwell Scientific Publications; 1959. p 12-41.

 7. De Ruiter CJ, de Haan A, Sargeant AJ. Physiological characteristics of two extreme muscle compartments in gastrocnemius medialis of the anaesthetized rat. Acta Physiol Scand 1995; 153: 313-324.

 8. Dekhuijzen AJ, van Koetsveld PA, Baan GC, Woittiez RD, Huijing PA. Motor endplate position of rat gastrocnemius muscle. Muscle Nerve 1986; 9: 642-647.

 9. Dimitrova NA, Disselhorst-Klug C, Trachterna A, Dimitrov GV, Rau G. Surface detected potentials of normal and reinnervated motor units: a simulation study for muscles consisted of short fibres. Electromyogr Clin Neurophysiol 2002; 42: 219-230.

10. English AW, Letbetter WD. Anatomy and innervation patterns of cat lateral gastrocnemius and plantaris muscles. Am J Anat 1982; 164: 67-77.

11. Feirabend HK, Kok P, Choufoer H, Ploeger S. Preservation of myelinated fibers for electron microscopy: a qualitative comparison of aldehyde fixation, microwave stabilisation and other procedures all completed by osmication. J Neurosci Methods 1994; 55: 137-153.

12. Fiala JC, Harris KM. Computer-based alignment and reconstruction of serial sections. Microsc Anal. 2002; 52: 5-7.

13. Grottel K, Celichowski J, Kowalski K. Spatial analysis of motor unit potentials of the rat medial gastrocnemius. Acta Physiol Pol 1986; 37: 219-227.

14. Hebel R, Stromberg MW. Myology. In: Hebel R, Stromberg MW, editors. Anatomy and embryology of the laboratory rat. Worthsee: BioMed Verlag; 1986. p 25-44.

15. Holewijn M, Plantinga P, Woittiez RD, Huijing PA. The number of sarcomeres and architecture of the M. gastrocnemius of the rat. Acta Morphol Neerl Scand 1984; 22: 257-263.

16. Janjua MZ, Leong SK. Labelling of cervical and thoracic spinal neurones following injection of horseradish peroxidase in the leg muscles of rats. Brain Res 1981; 223: 386-390.

17. Karnovsky MJ, Roots L. A "direct-coloring" thiocholine method for cholinesterase. J Histochem Cytochem 1964; 12: 219-221.

18. Kobbert C, Apps R, Bechmann I, Lanciego JL, Mey J, Thanos S. Current concepts in neuroanatomical tracing. Prog Neurobiol 2000; 62: 327-351.

19. Koelle GB, Friedenwald JS. A histochemical method for localizing cholinesterase activity. Proc Soc Exp Biol (N Y) 1949; 70: 617-622.

20. Liu S, Damhieu P, Devanze P, Said G, Heard JM, Tadie M. Efficient reinnervation of hindlimb muscles by thoracic motor neurons after nerve cross-anastomosis in rats. J Neurosurg 2003; 99: 879-885.

21. Marani E. Enzyme histochemistry. In: Lahue R, editor. Methods in neurobiology. New York: Plenum Press; 1981. p 481-581.

22. McHanwell S, Biscoe TJ. The localization of motoneurons supplying the hindlimb muscles of the mouse. Philos Trans R Soc Lond B Biol Sci 1981; 293: 477-508.

23. Miller-Larsson A. Motor unit action potential field - modelling results. Biol Cybern 1986; 53: 307-321.

24. Nandedkar SD, Sanders DB, Stalberg EV, Andreassen S. Simulation of concentric needle EMG motor unit action potentials. Muscle Nerve 1988; 11: 151-159.

25. Popratiloff AS, Neiss WF, Skouras E, Streppel M, Guntinas-Lichius O, Angelov DN. Evaluation of muscle re-innervation employing pre- and post-axotomy injections of fluorescent retrograde tracers. Brain Res Bull 2001; 54: 115-123.

26. Rafiuddin AM, Jayakumar R. Peripheral nerve regeneration in RGD peptide incorporated collagen tubes. Brain Res 2003; 993: 208-216.

27. Rosen M, Malmgren LT, Gacek RR. Three-dimensional computer reconstruction of the distribution of neuromuscular junctions in the thyroarytenoid muscle. Ann Otol Rhinol Laryngol 1983; 92: 424-429.

28. Stalberg E, Karlsson L. Simulation of EMG in pathological situations. Clin Neurophysiol 2001; 112: 869-878.

29. Stalberg E, Karlsson L. Simulation of the normal concentric needle electromyogram by using a muscle model. Clin Neurophysiol 2001; 112: 464-471.

30. Stalberg E, Karlsson L. The motor nerve simulator. Clin Neurophysiol 2001; 112: 2118-2132.

31. Swett JE, Wikholm RP, Blanks RH, Swett AL, Conley LC. Motoneurons of the rat sciatic nerve. Exp Neurol 1986; 93: 227-252.

32. Thomas CK, Sesodia S, Erb DE, Grumbles RM. Properties of medial gastrocnemius motor units and muscle fibers reinnervated by embryonic ventral spinal cord cells. Exp Neurol 2003; 180: 25-31.

33. Valero-Cabre A, Tsironis K, Skouras E, Navarro X, Neiss WF. Peripheral and spinal motor reorganization after nerve injury and repair. J Neurotrauma 2004; 21: 95-108.

34. Varejao AS, Melo-Pinto P, Meek MF, Filipe VM, Bulas-Cruz J. Methods for the experimental functional assessment of rat sciatic nerve regeneration. Neurol Res 2004; 26: 186-194.

35. Zacks SI. The motor endplate . In: Zacks SI, editor. The motor endplate. Philadelphia: W.B. Saunders; 1964. p 1-28.

36. Zalewska E, Hausmanowa-Petrusewicz I, Stalberg E. Modeling studies on irregular motor unit potentials. Clin Neurophysiol 2004; 115: 543-556.