Long-term unit recording from somatosensory neurons in the spinal ganglia of the freely walking cat.

A new technique has been developed for stable, long-term recording from groups of individual primary afferent neurons in the freely walking cat. A number of fine, flexible wires are inserted into dorsal root ganglia via a small laminotomy in the lumbar spine. The cut end of each wire can record stable and separable action potentials from one to three dorsal root ganglion neurons; each unit has typically held for 1 to 4 days. A broad range of myelinated somatosensory afferent (conduction velocities of 30 to 120 meters per second) have been studies during locomotion. Most cutaneous and proprioceptive afferent studied have been sensitive monitors of complex combinations of step-cycle components, and their firing patterns would often have been difficult to predict from existing information.

Discharge patterns of hindlimb motoneurons during normal cat locomotion.

Long-term recording from single lumbar motoneurons of intact cats revealed activation patterns fundamentally different from those seen in decerebrate preparations. In intact cats, motoneuron bursts showed marked rate modulation without initial doublets. Each unit's frequencygram generally resembled the envelope of the gross electromyogram simultaneously recorded from the corresponding muscle. Average and peak discharge rates increased for faster gaits. These findings suggest that, in cat locomotion, rate modulation is a more important contributor to force regulation than was previously thought.

Computationally efficient models of neuromuscular recruitment and mechanics.

We have improved the stability and computational efficiency of a physiologically realistic, virtual muscle (VM 3.*) model (Cheng et al 2000 J. Neurosci. Methods 101 117-30) by a simpler structure of lumped fiber types and a novel recruitment algorithm. In the new version (VM 4.0), the mathematical equations are reformulated into state-space representation and structured into a CMEX S-function in SIMULINK. A continuous recruitment scheme approximates the discrete recruitment of slow and fast motor units under physiological conditions. This makes it possible to predict force output during smooth recruitment and derecruitment without having to simulate explicitly a large number of independently recruited units. We removed the intermediate state variable, effective length (Leff), which had been introduced to model the delayed length dependency of the activation-frequency relationship, but which had little effect and could introduce instability under physiological conditions of use. Both of these changes greatly reduce the number of state variables with little loss of accuracy compared to the original VM. The performance of VM 4.0 was validated by comparison with VM 3.1.5 for both single-muscle force production and a multi-joint task. The improved VM 4.0 model is more suitable for the analysis of neural control of movements and for design of prosthetic systems to restore lost or impaired motor functions. VM 4.0 is available via the internet and includes options to use the original VM model, which remains useful for detailed simulations of single motor unit behavior.

Neural prosthetic interfaces with the nervous system.

Once damaged, the adult mammalian nervous system is capable of little functional regeneration. Thus, clinical disorders such as deafness, blindness and paralysis have been treated primarily by substitution rather than correction (e.g. teletype for telephone, braille for print, wheelchair instead of walking). However, recent advances in the technology of miniature electronic implants and in the basic understanding of sensory and motor functions have made it possible to build neural prosthetic devices that work by exchanging information directly between computing devices and neurons. Such interfaces are already permitting thousands of otherwise deaf patients to hear sounds directly; some have appreciable speech comprehension without visual cues. There is active research on restoring many types of sensory, motor and autonomic function. Two interesting synergies have emerged. The first stems from the recognition that the biophysical processes involved in stimulating and recording from neurons are universal; thus, technical advances have broad implications in many clinical areas. The second stems from the unique opportunity that such prostheses present to conduct neurophysiological and psychophysical studies directly in conscious human subjects.

Distribution of motoneurons supplying cat sartorius and tensor fasciae latae, demonstrated by retrograde multiple-labelling methods.

Sartorius (SART) and tensor fasciae latae (TFL) in the cat hindlimb are functionally heterogeneous muscles with regions that differ in their skeletal actions and electromyographic recruitment during normal activity. The topographical organization of motoneurons supplying different regions of SART or TFL has been investigated by exposing cut nerve branches supplying different peripheral territories to a combination of retrograde tracers, including Fast Blue (FB), Fluorogold (FG), and horseradish peroxidase (HRP). Motoneurons supplying medial, central, and anterior regions of SART were intermixed extensively throughout a single columnar nucleus located in the ventrolateral part of segments L4 and L5. With this column, motoneurons supplying medial SART tended to lie more rostrally than those supplying anterior regions, but the gradient was modest and showed some cat-to-cat variation. Two major branches entered anterior SART at different proximodistal levels. When these two branches were exposed to different tracers, most motoneurons contained a single tracer; only a few double-labelled cells were apparent. The labelling suggests that anterior SART may contain two separate, in-series divisions of motor units. In TFL, motoneurons supplying nerve branches to posterior, central, and anterior parts of the muscle were intermingled indiscriminately in a single ventrolateral cell column in L6 and rostral L7. These results suggest that topographical organization in lumbar motor nuclei does not always reflect the highly ordered biomechanical and functional specialization evident in the peripheral organization of the muscles themselves.

Cuff electrodes for chronic stimulation and recording of peripheral nerve activity.

A comparative study of 5 different designs of nerve cuff electrodes was undertaken to determine their relative merits for stimulating and recording whole-nerve activity over extended periods of chronic implantation on large and small peripheral nerves in 8 cats. Four of the designs represent novel fabrication strategies, including 2 based on flexible, thin-film substrates and 2 based on dip-coating silicone elastomer on a cylindrical mandrel. Various advantages and shortcomings of these materials and designs are discussed in the context of the biophysical factors that influence these electrophysiological interfaces, particularly the problem of recording microvolt-level neurograms in the presence of millivolt-level electromyograms from adjacent muscles in freely behaving subjects. The most effective design was one in which a thin sheath of silicone rubber was wrapped around and intra-operatively sealed to a longitudinally slit, tripolar cuff made by dip-coating silicone over stranded stainless steel leads that were prepositioned on a mandrel using polyvinyl alcohol as a temporary adhesive. When properly installed, these electrodes had stable impedances, recruitment thresholds and relatively interference-free recording properties for the duration of this study (up to 9 weeks).

Toward the ultimate metal microelectrode.

The performance of metal microelectrodes for stimulating and recording neuronal action potentials depends on precise control of their geometrical, electrical and mechanical properties. We describe a combination of materials whose properties approach fundamental physical limitations on achievable performance and reproducible fabrication techniques that provide probes with very small dimensions. Pure iridium wire is electrolytically sharpened, vapor-coated with Parylene-C insulation and the tip exposed using an automatically steerable UV laser. Electrochemical activation of the iridium increases the capacitance of the metal-electrolyte interface so that the overall impedance in the relevant frequency band (100-10,000 Hz) is dominated by the access resistance of the surrounding tissues.

A multiple-contact EMG recording array for mapping single muscle unit territories.

The glycogen-depletion technique has become a well-established method for determining histologically the cross-sectional distribution of a single muscle unit. A major drawback of this method is its low yield of one depleted unit per experiment. Furthermore, this technique is particularly unsuited for determining the longitudinal distribution of single muscle units in long, broad muscles because of the formidable serial sectioning job that would be required. Our alternative, electrophysiological method utilizes a multiple-contact, two-dimensional EMG recording array to map efficiently the cross-sectional and longitudinal distributions of numerous single muscle units in anatomically diverse muscles. Additionally, architectural information on muscle fiber lengths, end-plate locations, motor subunit (MSU) arrangements and muscle conduction velocities can be determined.

Virtual muscle: a computational approach to understanding the effects of muscle properties on motor control.

This paper describes a computational approach to modeling the complex mechanical properties of muscles and tendons under physiological conditions of recruitment and kinematics. It is embodied as a software package for use with Matlab and Simulink that allows the creation of realistic musculotendon elements for use in motor control simulations. The software employs graphic user interfaces (GUI) and dynamic data exchange (DDE) to facilitate building custom muscle model blocks and linking them to kinetic analyses of complete musculoskeletal systems. It is scalable in complexity and accuracy. The model is based on recently published data on muscle and tendon properties measured in feline slow- and fast-twitch muscle, and incorporates a novel approach to simulating recruitment and frequency modulation of different fiber-types in mixed muscles. This software is distributed freely over the Internet at http://ami.usc.edu/mddf/virtualmuscle.

Microminiature molding techniques for cochlear electrode arrays.

We provide a general method for producing a variety of small, complex electrode arrays based on injection molds produced using computer-aided drafting and machining (CAD-CAM) procedures and a novel method for connecting to the very fine electrical leads associated with the individual contacts of such arrays. Cat-sized cochlear electrode arrays with up to eight contacts were built according to these methods and their electrical contacts were characterized in vitro by impedance spectroscopy and in vivo by monitoring impedance for over 1 year of intermittent stimulation in chronically instrumented animals.

Physiological characterization of motor unit properties in intact cats.

Single motor units were isolated in intact cats, by microstimulation through chronically implanted microwires in the L5 ventral roots. Motor unit axonal and mechanical properties were obtained by stimulus-triggered averaging the signals from an implanted femoral nerve recording cuff and patellar tendon force transducer. All unit types were sampled with this technique, and it was also possible to stimulate in isolation an axon whose ventral root spike was recorded during treadmill locomotion. A new technique was described, spike-triggered microstimulation, for verifying the identity of a stimulated and a recorded axon.

A simple technique for determination of footfall patterns of animals during treadmill locomotion.

This paper describes a direct on-line electronic method for determination of the quadrupedal footfall patterns of animals during locomotion. This technique involves the application of a 30 kHz AC carrier signal to a conductive treadmill belt. The time of contact of each paw on the belt is determined by recording the increased conduction of these carrier signals from the treadmill surface through the footpads to monopolar electrodes implanted in each limb. Since the carrier signal frequency is very high, filters may be employed to electronically separate this signal from electromyographic signals simultaneously obtained through the same electrodes. In comparison with the cinematographic methods typically used for analysis of footfall patterns, this technique is less time consuming, is just as accurate and also yields signals which can be used for on-line triggering applications.

Single unit conduction velocities from averaged nerve cuff electrode records in freely moving cats.

The conduction velocity of peripheral neurons recorded by wire microelectrodes implanted in intact, freely moving cats was determined on-line using the technique of spike-triggered averaging of nerve cuff electrode records described here. Axonal velocity was estimated from the conduction latency between two adjacent sets of tripolar recording electrodes inside a cuff, thereby avoiding uncertainties that could arise from differences in spike shape, variable conduction distance, or unknown stimulus utilization time. This method rendered conduction velocity values for individual afferent and efferent myelinated fibers ranging from 27 to 120 m/sec, estimated with an uncertainty of +/-5%. In addition, predictions from theoretical models relating extracellular potential amplitude, wavelength, and conduction velocity were confirmed experimentally for en passant records obtained from intact myelinated fibers.

Cross-correlation of EMG reveals widespread synchronization of motor units during some slow movements in intact cats.

It is commonly assumed that the motor units comprising a single mammalian muscle will be recruited asynchronously at subtetanic firing rates to produce smoothly modulated force output. However, electromyograms from certain neck muscles, recorded by implanted bipolar "patch" electrodes having large contacts, often exhibited a rhythmic clustering of spike activity whose patterns suggested that motor-unit firing was synchronized both within and across muscles. We have developed a computerized processing system that digitizes EMG activity and calculates auto- and cross-correlation products of selected segments. The presence or absence of synchronization caused by neural mechanisms can be identified and differentiated from that due to the rhythmicity of the behavior itself (e.g. shaking) or due to cross-talk, according to the shapes of the resultant correlograms. These methods have so far been applied to the study of hindlimb and neck muscle EMG during various natural motor behaviors, but they provide a general, quantitative tool for the study of an important aspect of motor control that may be overlooked by conventional sampling and smoothing techniques.

An intrinsic mechanism to stabilize posture--joint-angle-dependent moment arms of the feline ankle muscles.

The neuromuscular system can use sensory feedback to regulate motion, but these feedback loops involve relatively long delays (50-100 ms) and may produce undesirable oscillations. However, stabilizing changes in muscle force can also be provided intrinsically (i.e. without feedback) by 2 properties of the muscle itself, its force-length and force-velocity relationships. We have discovered another intrinsically stabilizing mechanism in the musculoskeletal architecture of the cat ankle joint. Many of its muscles have their predominant moment arms about the adduction/abduction axis, with smaller moment arms for inversion/eversion and about the principal axis of motion, dorsiflexion/extension. The magnitudes of the moment arms in ab/adduction and in/eversion depend strongly on joint angle, increasing for positions away from neutral that lengthen the muscles. Thus, co-activation of agonist-antagonist pairs, known to occur in these muscles, would provide immediate stabilizing changes in torque when the ankle is perturbed away from neutral position.

Why cats pace on the treadmill.

There have been many studies suggesting that locomotion on a treadmill tends to be different than locomotion at similar velocities overground, but no satisfactory mechanical or neural mechanisms to account for the differences have been identified. The most prominent difference is the tendency to adopt a pacing gait for both walking and trotting speeds, in which the legs on one side of the body move in phase as lateral couplets rather than the typical diagonal couplet pattern seen overground. Using conventional video analysis, we quantified the gait patterns of intact, adult cats walking at various speeds overground and in a motorized treadmill. We noted that cats paced most frequently when they were at the front end of the treadmill enclosure, and that this gait was associated with an extended stride length that permitted the animals to maintain a higher duty factor of support (mean number of feet on the ground). We propose that the animal extends its stride specifically to improve the duty factor in anticipation of sudden stops of the treadmill belt and that it converts abruptly from diagonal to lateral gait because the extended stride results in collisions between ipsilateral hind and front feet.

Mechanical properties of aponeurosis and tendon of the cat soleus muscle during whole-muscle isometric contractions.

Recent studies have suggested that the mechanical properties of aponeurosis are not similar to the properties of external tendon. In the present study, the lengths of aponeurosis, tendon, and muscle fascicles were recorded individually, using piezoelectric crystals attached to the surface of each structure during isometric contractions in the cat soleus muscle. We used a surgical microscope to observe the surface of the aponeurosis, which revealed a confounding effect on measures of aponeurosis length due to sliding of a thin layer of epimysium over the proximal aponeurosis. After correcting for this artifact, the stiffness computed for aponeurosis was similar to tendon, with both increasing from around 8 F0/Lc (F0 is maximum isometric force and Lc is tissue length) at 0.1 F0 to 30 F0/Lc at forces greater than 0.4 F0. At low force levels only (0.1 F0), aponeurotic stiffness increased somewhat as fascicle length increased. There was a gradient in the thickness of the aponeurosis along its length: its thickness was minimal at the proximal end and maximal at the distal end, where it converged to form the external tendon. This gradient in thickness appeared to match the gradient in tension transmitted along this structure. We conclude that the specific mechanical properties of aponeurosis are similar to those of tendon.

Architecture and consequent physiological properties of the semitendinosus muscle in domestic goats.

Morphological and physiological analyses confirm that the semitendinosus muscle of goats contains two separate compartments in series, each with distinct innervation. These compartments of the muscle are in turn composed of short fibers (approximately four fibers in series in the proximal compartment and seven to eight fibers in the distal compartment) which overlap each other for more than 30% of their length, with much of the overlapping portions consisting of slender tails that terminate at one-tenth of the midfiber diameter. Groups of fibers are associated into relatively narrow bands that run end-to-end in each compartment. The data suggest that the maximum length of muscle fibers may be limited; even the fibers of parallel-fibered muscles may not scale with the dimension of the animal.

Relationships between range of motion, lo, and passive force in five strap-like muscles of the feline hind limb.

The relationships between range of motion, optimal length for force production (lo), and passive force provide useful insights into the structure and function of muscles but are unknown for most individual muscles. We measured these values and examined their relationships in five strap-like muscles of the cat hind limb: caudofemoralis, semitendinosus, sartorius anterior, tenuissimus, and biceps femoris anterior. The range of motion relative to lo was found to vary significantly between different muscles and even between different specimens of the same muscle. The passive force-length (FL) curve was found to be correlated with both lo and lmax (maximal in situ muscle length) but was correlated more strongly with lmax. The mean passive force produced by these muscles at lmax was less than 7% of estimated maximal isometric force, suggesting that passive force may not be important in these muscles during normal activation patterns. The variance in passive FL curves between specimens of the same muscle was found to be significantly lower when length was scaled by lmax as opposed to lo. These results suggest that lmax may provide a more useful scaling factor for generic models of muscle. However, the passive length-tension properties of mammalian muscle appear to reflect a complex mix of structures at both the myofilament and connective tissue levels that may differ depending on muscle-fiber architecture and perhaps on the history of trophic influences on a particular specimen.

Distribution and innervation of short, interdigitated muscle fibers in parallel-fibered muscles of the cat hindlimb.

The cat hindlimb contains several long, biarticular strap muscles composed of parallel muscle fascicles that attach to short tendons. Three of these muscles--sartorius, tenuissimus, and semitendinosus--were studied by dissecting individual gold-stained fibers and determining the surface distribution of acetylcholinesterase-stained end-plate zones. In each muscle, fascicles were composed of muscle fibers that ran only part of the fascicle length and tapered to end as fine strands that interdigitated with other tapering fibers within the muscle mass. Most muscle fibers measured 2-3 cm in length. Fascicles of muscle fibers were crossed by short transverse bands of endplates (1 mm wide by 1-5 mm long) that were spaced at fairly regular intervals from the origin to the insertion of the muscle. The endplate pattern suggested that the fiber fascicles were organized into multiple longitudinal strips. In the sartorius, the temporospatial distribution of electromyographic (EMG) activity evoked by stimulating fine, longitudinal branches of the parent nerve confirmed that each strip was selectively innervated by a small subset of the motor axons. These axons appeared to distribute their endings throughout the entire length of the fascicles, providing for synchronous activation of their in-series fibers.