Manuel Hulliger

Quantitative analysis of human movement synergies: constructive pattern analysis for gait.

To record three-dimensional coordinates of the joints from normal human subjects during locomotion, we used a digital motion analysis system (ELITE). Recordings were obtained under several different conditions, which included normal walking and stepping over obstacles. Principal component analysis was used to analyze coordinate data after conversion of the data to segmental angles. This technique gave a stable summary of the redundancy in gait kinematic data in the form of reduced variables (principal components). By modeling the shapes of the phase plots of reduced variables (distortion analysis) and using a limited number of model parameters, good resolution was obtained between subtly different conditions. Hence, it was possible to accurately resolve small distributed changes in gait patterns within subjects. These methods seem particularly suited to longitudinal studies in which relevant movement features are not known a priori. Assumptions and neurophysiological applications are discussed.

Reciprocal Ia inhibition in patients with spinal spasticity

Reciprocal Ia inhibition from ankle flexors to extensors was studied in five patients with spasticity due to incomplete traumatic spinal cord lesions. Nine healthy subjects were tested as controls. Excitability of the soleus motoneuron pool was estimated by H-reflex testing in the resting state. Ia inhibition was activated by conditioning stimuli to the peroneal nerve. Ia inhibition was detected in all patients tested, the amount of inhibition ranging from 8% to more than 50% of the test H-reflex size. In the control subjects only weak Ia inhibitory effects were present. These findings indicate increased excitability of the Ia inhibitory pathway to ankle extensor motoneurons in patients with spasticity due to spinal cord injury.

Antigenic compartmentation of the cat cerebellar cortex

Despite the apparent uniformity in cellular composition of the mammalian cerebellar cortex, a complex topography is revealed by several expression patterns. Zebrin II, a polypeptide antigen identified as aldolase C, is one such marker which, in several species of mammals, is restricted to a subset of Purkinje cells that are clustered together to form a symmetrical and reproducible array of zones and stripes. In rodents the cerebellar cortex is divided into four transverse zones--anterior, central, posterior, and nodular. Each transverse zone is further subdivided mediolaterally into an array of parasagittal stripes. The similar zone and stripe organization partitions the hemispheres. Based upon a novel whole mount immunohistochemical staining procedure, we have now identified homologous zones and stripes in the feline cerebellum. In the cat cerebellum the somata of most Purkinje cells express zebrin II but parasagittal stripes may still be delineated owing to the alternating high and low zebrin II expression levels in the dendritic arbors. As in rodents, the cat cerebellum consists of four transverse zones with each zone subdivided into a unique combination of zebrin II parasagittal stripes, suggesting that a common architecture underlies the organization of the mammalian cerebellum.

A comparative analysis of the encapsulated end-organs of mammalian skeletal muscles and of their sensory nerve endings

The encapsulated sensory endings of mammalian skeletal muscles are all mechanoreceptors. At the most basic functional level they serve as length sensors (muscle spindle primary and secondary endings), tension sensors (tendon organs), and pressure or vibration sensors (lamellated corpuscles). At a higher functional level, the differing roles of individual muscles in, for example, postural adjustment and locomotion might be expected to be reflected in characteristic complements of the various end‐organs, their sensory endings and afferent nerve fibres. This has previously been demonstrated with regard to the number of muscle‐spindle capsules; however, information on the other types of end‐organ, as well as the complements of primary and secondary endings of the spindles themselves, is sporadic and inconclusive regarding their comparative provision in different muscles. Our general conclusion that muscle‐specific variability in the provision of encapsulated sensory endings does exist demonstrates the necessity for the acquisition of more data of this type if we are to understand the underlying adaptive relationships between motor control and the structure and function of skeletal muscle. The present quantitative and comparative analysis of encapsulated muscle afferents is based on teased, silver‐impregnated preparations. We begin with a statistical analysis of the number and distribution of muscle‐spindle afferents in hind‐limb muscles of the cat, particularly tenuissimus. We show that: (i) taking account of the necessity for at least one primary ending to be present, muscles differ significantly in the mean number of additional afferents per spindle capsule; (ii) the frequency of occurrence of spindles with different sensory complements is consistent with a stochastic, rather than deterministic, developmental process; and (iii) notwithstanding the previous finding, there is a differential distribution of spindles intramuscularly such that the more complex ones tend to be located closer to the main divisions of the nerve. Next, based on a sample of tendon organs from several hind‐foot muscles of the cat, we demonstrate the existence in at least a large proportion of tendon organs of a structural substrate to account for multiple spike‐initiation sites and pacemaker switching, namely the distribution of sensory terminals supplied by the different first‐order branches of the Ib afferent to separate, parallel, tendinous compartments of individual tendon organs. We then show that the numbers of spindles, tendon organs and paciniform corpuscles vary independently in a sample of (mainly) hind‐foot muscles of the cat. Grouping muscles by anatomical region in the cat indicated the existence of a gradual proximo‐distal decline in the overall average size of the afferent complement of muscle spindles from axial through hind limb to intrinsic foot muscles, but with considerable muscle‐specific variability. Finally, we present some comparative data on muscle‐spindle afferent complements of rat, rabbit and guinea pig, one particularly notable feature being the high incidence of multiple primary endings in the rat.

A method for the study of the effects of combining multiple pseudorandom fusimotor stimulation on the responses of muscle-spindle primary-ending afferents

We describe a new method of investigation of the integrative action of fusimotor inputs in mammalian muscle spindles by stimulation of multiple fusimotor axons using independent pseudorandom pulse trains, each of low mean rate with pseudorandomly distributed stimulus intervals. Technically it was feasible only because of the development of (1) a novel, highly efficient approach to functional isolation of fusimotor efferents in ventral-root filaments, which we have called the isodyne strategy; (2) a real-time, microprocessor-based stimulus artefact cancellation device (SACAD); and (3) a highly adjustable, multi-branch stimulation electrode array. The general approach of using multiple, independent, pseudorandom stimulation of several input channels has wider applications in controlled-activation paradigms.

Correlated histological and physiological observations on a case of common sensory output and motor input of the bag(1) fibre and a chain fibre in a cat tenuissimus spindle

In muscle spindles of the cat, independent control of dynamic and static components of the response of the primary sensory ending to stretch is provided by separate motor inputs to the various kinds of intrafusal muscle fibre: dynamic axons (γ or β) to the bag1 fibres and static axons to the bag2 (typically γ only) and chain (γ or β) fibres. Nonlinear summation of separately evoked effects during combined stimulation of dynamic and static motor axons appears to be due to mutual resetting by antidromic invasion of separate encoding sites, leading to partial occlusion of the momentarily lesser response by the greater. The encoding sites are thought to be located within the primary endings preterminal branches which from first‐order level are normally segregated to the bag1 fibre and to the bag2 and chain fibres. Here we describe the analysis of a special case that arose in a histophysiological study which had shown that the degree of occlusion was related to the minimum number of nodes between the putative encoding sites. Three‐dimensional reconstruction of the primary ending revealed that the terminals of one chain fibre were derived entirely from the first‐order branch that supplied the bag1 fibre, including one terminal that was shared directly with the bag1 (sensory cross‐terminal). The other first‐order branch supplied the bag2 and remaining chain fibres as normal. The degree of occlusion seen during simultaneous stimulation of a dynamic β axon and a static γ axon indicated that the encoding sites were separated by both first‐order branches. Schematic reconstruction of the motor innervation revealed that the static γ axon was most unlikely to have supplied the chain fibre which shared sensory terminals with the bag1, but that these fibres also shared a motor input with histological characteristics of β type. Ramp‐frequency stimulation of the dynamic β axon at constant length evoked a driving effect which persisted after fatiguing the extrafusal component and was therefore explicable on the basis of the observed pattern of motor innervation, though the identity of the axon could not be conclusively proved. Individually, instances of shared sensory terminals and motor input of bag1 and chain fibres are rare in the cat; their combination in a single spindle with correlated physiology is described here for the first time. The observation is considered in relation to the importance of dynamic and static segregation in motor control, since it may imply that there is a lower limit to the degree of segregation that the developmental programme can provide.

Direct dynamics simulation of FES-assisted locomotion

Using functional electrical stimulation (FES), muscles of spinal-cord injured patients can be activated by externally generated electrical currents in order to restore function. As for gait, the question arises when during the gait cycle and two what extent individual muscles should be stimulated. Computer simulation provides the designer with a tool to evaluate the performance of different muscle stimulation patterns without the need to test patients at every stage of system development. The goals of this paper are: first, to identify, using computer simulation, multi-channel stimulation patterns that are capable of reproducing normal gait kinematics for a full gait cycle, without relying on sensory feedback (open-loop control); second, to briefly assess the stability of the gait obtained. A two-dimensional musculo-skeletal model was developed, based on mathematical representations of muscle properties (including force-length and force velocity characteristics and muscle activation dynamics). A visco-elastic model, including non-linear heel-pad properties, was used to describe the foot-ground interaction. A seven segment skeletal model was actuated by 8 major muscle groups in each leg. Rectangular muscle stimulation patterns were defined by 3 parameters: onset, termination and level of stimulation. Thus, the minimization of the differences between simulated and measured normal gait kinematics was a 24 (3 by 8) parameter optimization problem. Although a good agrement was found between simulated and measured kinematics (rms difference equals 6.5 degrees), stable cyclic locomotion was not achieved. At this point it is concluded that muscle properties do not provide sufficient stability to permit cyclic locomotion with sixteen channels of muscle stimulation, and that incorporation of sensory feedback control will be necessary to achieve this goal.

Modelling of Chaotic and Regular Ia Afferent Discharge During Fusimotor Stimulation

Mammalian Ia afferents have a tendency to fire action potentials phase-locked to periodic stimuli, when certain static gamma axons are stimulated at constant frequency. This driving is thought to be mediated by chain fibres, which, in contrast to other intrafusal fibres, have fast contraction properties, comparable to type II extrafusal fibres. Intrafusal contractions are mechanically transmitted to the sensory ending, which transduces the movement into a receptor current. The ensuing oscillatory changes in receptor potential which accompany these rapid chain fibre contractions then cause the generation of one spike during each contraction cycle. If stimulation frequencies are high, or the bag2 fibre is coactivated, the afferent may fire at subharmonic frequencies, giving 1:2 or 1:3 driving. Also, very irregular spike intervals can be observed, which gives the impression of an erratic firing response. Changes in the initial muscle length often cause transitions between subharmonic, 1:1 driving, and irregular firing (Boyd, Murphy & Mann, 1985; Banks, 1991).

Pacemaker Competition and the Role of Preterminal-Branch Tree Architecture: A Combined Morphological, Physiological and Modelling Study

Peripheral sensory endings often consist of groups of unmyelinated nerve terminals, specialised for transduction, that ultimately converge on the single afferent axon through a system of myelinated preterminal branches. At least in the mammalian muscle spindle the sensory terminals are in continuity only through the preterminal branches (Banks, 1986), so each may be supposed to have an associated spike-initiation (encoding) site, potentially able to act as a separate pacemaker. Interaction in such a system is expected to be highly competitive (Eagles & Purple, 1974) so that the final output of the afferent is in general a non-linear function of the activities of the separate encoders.

Domestic cat walking parallels human constrained optimization: optimization strategies and the comparison of normal and sensory deficient individuals.

To evaluate how fundamental gait parameters used in walking (stride length, frequency, speed) are selected by cats we compared stride characteristics selected when walking on a solid surface to those selected when they were constrained to specific stride lengths using a pedestal walkway. Humans spontaneously select substantially different stride length-stride frequency-speed relationships in walking when each of these parameters is constrained, as in walking to a metronome beat (frequency constrained), evenly spaced floor markers (stride length constrained) or on a treadmill (speed constrained). In humans such adjustments largely provide energetic economy under the prescribed walking conditions. Cats show a similar shift in gait parameter selection between conditions as observed in humans. This suggests that cats (and by extension, quadrupedal mammals) also select gait parameters to optimize walking cost-effectiveness. Cats with a profound peripheral sensory deficit (from pyridoxine overdose) appeared to parallel the optimization seen in healthy cats, but without the same level of precision. Recent studies in humans suggest that gait optimization may proceed in two stages - a fast perception-based stage that provides the initial gait selection strategy which is then fine-tuned by feedback. The sensory deficit cats appeared unable to accomplish the feedback-dependent aspect of this process.