21.11 Velocity-dependent stability of gait for patientswith balance impairments can be explained by biomechanical stabilization

Chapter 21. Modeling approach of posture and gait I~

Sensorial input effects in postural control

G. Dietrich 1, M. Gilles 2, A.M. Wing 3. 1LAMA, University of Paris

5, eWorking Life Laboratory, INRS, Nancy, France," 3Behavioural Brain Sciences' Centre, Birmingham, UK Introduction: Since 1992 [1,2] effects of haptic input from the upper

$139

[2] Chang CC, Brown DR, Bloswick DS, Hsiang SM. Biomechanical simulation of manual lifting using spacetime optimization. J Biomech 2001; 34:527 532.

~

Nonlinear behaviours in human posture

R Fourcade 1, B.G. Bardy 1'2, S. Roudeix 1. 1Center for Research

limb were shown on posture and balance. In 2002 [3] a conceptual model was presented for studying the contribution of each leg to sideways stability of a four-link biped. The aim of this study is to analyze the effects of varying the height of the light touch contact point on sway during normal stance and to evaluate the contribution of biomechanical constraints by using an extended version of the previous model. Methods: Sway was measured during four conditions: i)Light touch support and hand raised up at shoulder level, ii) Light touch support and hand down at elbow height, iii) No light touch support and hand raised at shoulder level, and iv) No light touch support and hand down at elbow height. Recorded data were compared with simulations using a biomechanical model. Results: As all previous studies, we observed cross-correlation between postural sway and hand movement during the control situation. In the simulation similar cross-correlation was observed with maximum at a lag that depended on the stiffness ratio between the two controllers responsible for postural sway and arm position control. Discussion and conclusion: On the basis of these findings we suggest the hypothesis that the use of light touch in postural stabilisation uses two independent simple PD or PID controllers to achieve a functional "coordination unit". Coupling between controllers is achieved through sensorial information.

We present a mechanical model capturing the phase transitions observed in human posture [1], consisted of a double-inverted pendulum. The human body is modelled with two indeformable segments, one representing the head-trunk system rotating around the hips, the other representing the thigh-leg system rotating around the ankles. The oscillatory movement of the body is created by a torque, of constant amplitude, acting at the ankles. In order to keep constant the amplitude of the upper segment, a Van der Pol non-linear oscillator between the head and the vertical axis has been added. Simulations of the model for different movement frequencies (transition simulations) reveal two stable oscillation modes: an "in phase" mode for low frequencies (ankle-hip relative phase close to 0°), and an "anti-phase" mode for high frequencies (relative phase close to 180°). Simulations also show an hysteresis effect, with the between-modes frequency transition being higher when frequency is increased and lower when frequency is decreased. This mechanical model emphasizes the non-linear behaviour of human posture, and reveals that the postural system exhibits typical signatures are self-organized systems.

References

References

[1] Lackner JR. J. Vest Res 1992; 2:307 322. [2] Jeka JJ and Lackner JR. Exp. Brain Res 1994; 103:267 276. [3] Dietrich G, Wing A, Gilles M, Nimmo-Smith I. J. Appl. Biomech. 2002; 18:99 109.

[1] Bardy B.G., Oullier O., Bootsma R.J., Stoffregen T.A. Dynamics of human postural transitions. J Exp Psychol Hum Percept 2002; 28(3): 499 514.

I•

Balance control during an arm raising movement in bipedal stance: which biomechanical factor is controlled?

M. Ferry, L. Martin. Laboratoire Sport et Performance Motrice;

Universitd Grenoble 1- UFR-STAPS, France Introduction: With computational resources available today, large-

scale models of the body and techniques simulating human behavior [1,2] can provide data to support hypotheses on how musculoskeletal systems interact to produce movements. Methods: we performed an analysis of kinematics and dynamical aspects of arm raising movements by combining experimental work, biomechanical model of the body and dynamic optimization. We then performed an analysis based on computer simulations. Since keeping the CoP and the projection of the CoM inside the support area is essential for equilibrium, we modeled an arm raising movement where displacement of one or the other variable is limited. Results: The comparison between experimental and simulated joint kinematics showed that the minimum torque change model yielded realistic trajectories. The results show that: a) the choice of the regulated variable influences the strategy adopted by the system, and b) the system was not able to regulate the CoM for very fast movements without compromising its balance. Discussion and conclusion: The system is able to maintain balance while raising the arm by only controlling the CoP. This may be done mainly by using hip mechanisms and controlling net ankle torque.

References [1] Flash T, Hogan N. The co-ordination of arm movements: an experimentally confirmed mathematical model. J Neurosci 1985; 5:1688 1703.

in Sport Sciences, University of Paris-Sud XL Orsay, 2Institut Universitaire de France, France

I•

Velocity-dependent stability of gait for patients with balance impairments can be explained by biomechanical stabilization

W. Ilg, H. Golla, M.A. Giese. Laboratory for Action Representation

and Learning~Department of Cognitive Neurology, Hertie Institute for Clinical Brain Research, University of Tiibingen, Germany Introduction: Recent studies report velocity-dependent stability of walking and running movements for patients with balance impairments caused by vestibular neuritis [1] or cerebellar stroke. This phenomenon has been explained by circuits in spinal cord, which take over control for higher velocities and inhibit vestibular and cerebellar inputs. However, this explanation leaves open why running could be less dependent from balance control than slow walking. We provide an alternative explanation that is based on the dynamical stability of bipedal locomotion. Methods: We present data from a cerebellar patient performing tandem walk. Kinematic analysis reveals less missteps, and significantly decreased lateral sway for higher walking speeds. In order to investigate possible biomechanical influences we studied the velocitydependent stability of a 3D passive walker, a simplified bipedal model that can walk down a slope without external control [2]. Results: We examined the robustness of this model against lateral disturbances as a function of velocity, providing no lateral control. The amount of lateral perturbations tolerated without loss of stability is substantially larger for higher walking speeds. Discussion and Conclusion: These results suggest that biomechanics provides an explanation for the increased stability in gait for higher velocities. This explains why running could be less dependent from vestibular information and therefore might be controlled mainly by circuits in the spinal cord. This hypothesis is consistent with a study [3] that indicates increased demands for postural control in slow walking.

Chapter 21. Modeling approach of posture and gait

$140

References [1] Brandt T, Strupp M, Benson J. Lancet 1999; 354(9180): 746. [2] Kuo AD. International Journal of Robotic Research 1999; 18: 917 930. [3] Den Otter AR, Geurts AC, Mulder T, Duysens J. Gait & Posture 2004; 19(3): 270 8.

[

•

]

A

throw-and-catch pattern in postural sway does not exclude continuous feedback control

Conclusion: A simple spring-based model simulated knee function during gait and can be used for orthosis design.

References [1] Blaya J. and Herr H. (2004) IEEE Trans Neut. Syst. Rehab. Eng.

I•

Postural sway of older female twins while doing math task

S. Pajala 1, P. Era 1, M. Koskenvuo 2, J. Kaprio 3, T. Rantanen 1. C. Maurer 1, R.J. Peterka 2. 1Neurozentrum, Breisacher St~ 64, 79106

Freiburg, Germany," 2Neurological Sciences' Institute, Oregon Health & Science University, Portland, Oregon, USA When one tries to balance a stick at one's fingertip, a possible strategy is to apply small impulses in a given direction, with each being followed by an impulse in the opposite direction to counteract the evoked stick movement. The convenience of such a balance strategy is that the stick is not stabilized exactly at a certain position, but instead is "falling freely" between one impulse and the subsequent one, thereby reducing the control action to the series of discrete impulses. Loram and Lakie [1] hypothesized that humans use such a balancing strategy to maintain upright stance and called this a "throw-and-catch" strategy. Experimental data analyzed using newly developed techniques seemed to support the hypothesis that humans use a time-discrete balance strategy rather than a continuous control strategy. Using computer simulations, we demonstrate that a simple continuous feedback model (which includes internal noise and a 160ms feedback time delay) produces sway patterns that closely resemble a throw-and-catch pattern. Moreover, when we applied the analysis techniques of Loram and Lakie to our simulated data, our results were nearly identical to those of their subjects. The parameters we used in our model, however, markedly deviated from those calculated using their methods. We conclude, that a throw-and-catch pattern in postural sway does not exclude a simple sensory feedback control of human upright stance.

References [1] Loram & Lakie. J Physiol 2002; 540.3, 1111 1124.

[•3]

Simulation of knee function during gait with an orthosis by means of two springs of different stifnesses

J.C. Moreno 1, EJ. Brunetti 1, A. Cullell 1, A. Forner-Cordero 1'2,

J.L. Pons 1. 1C.S.LC, Inst.Automdtica Industrial, Madrid, Spain,"

2Motor Control Lab., KU Leuven, Belgium Introduction: Some orthotic systems are based on providing a certain

fixed stiffness at the knee joint. However, the knee has different requirements during swing or stance phases of gait. The knee has to flex during initial swing and extend on time for foot contact. During stance, it should provide support for body weight. Methods: Two springs were used to simulate the knee function during gait with an orthosis. During the initial swing phase a lowstiffness spring allows for knee flexion due to inertia. At mid-swing, this spring recovers length to assist the knee extension required before foot contact and, at its maximal elongation switches to a high stiffness spring. This mechanism was simulated on a computer and tested on healthy subjects. Results: The knee function was approximated by two linear springs acting at different phases of the gait cycle. It simulated knee function during gait given the inertial parameters of the leg and a limited range of gait speeds and step lengths. It was found that the ratio between both springs must be of 10 with a minimal low stiffness of 0.07 N/mm.kg. Discussion: The knee moment and knee joint angle can be locally approximated by a linear relation as given by the two springs. However, these stifihesses must be modulated depending on gait speed and, of course, inertial parameters of the leg [1].

1Finnish Centre for Interdisciplinary Gerontology, University of JyvSskyl& 2Department of Public Health, Helsinki Turku, Department of Public Health, University of Helsinla, Finland Introduction: Control of posture while performing another task

impairs with increasing age [1] and may expose older people to falls. The aim of the present study was to elucidate whether genetic effects determine individual differences in the ability to perform simultaneously a balance task and cognitive task among 63 to 76 year-old women. Methods: Among 90 monozygotic (MZ) pairs and 105 dizygotic (DZ) pairs, postural sway was measured on a force platform (Good Balance, Metitur Ldt.), first while standing with eyes open and feet slightly apart. Test was then repeated while counting backwards in threes (DT math). Medio-lateral (ml) sway velocity (mm/s), antero posterior (a p) sway velocity (mm/s) and velocity moment (mmZ/s) were used in the analyses. The effect of another task on postural sway is indicated as dual-task cost (DTC, %). Genetic contribution was assessed by comparing within MZ and DZ twin pair correlations (rmz and rdz). Results: On average, math task DTC was 36% (SD 24%) for ml velocity, 31% (SD 22%) for ap velocity and 56% (48%) for velocity moment. Correlations for DT math~est were rmz 0.37 and rdz 0.15 for ml velocity, rmz 0.40 and rdz 0.35 for ap velocity, and rmz 0.29 and rdz 0.20 for velocity moment. Discussion and conclusions: MZ and DZ correlations suggests that genetic influences account for 11 43% of the variance in DT math test. The genetic influences account for 35% of the variance in postural sway [2] and cognitive processes are also under substantial highly genetically controlled [3]. Therefore, it is plausible that genes have an influence on the individual differences also when another task is performed simultaneously with postural balance task.

References [1] Teasdale N et al. On the cognitive penetrability of posture control. Exp Aging Res. 1993; 19:1 13. [2] Pajala S et al. Contribution of genetic and environmental effects to postural balance in older female twins. J Appl Physiol 2004; 196:308 315. [3] Swan G & Carmelli D. Evidence for genetic mediation of executive control: a study of aging male twins. J Gerontol B Psychol Sci Soc Sci 2002; 57:133 143.

I•

Balance stability boundary for a feet-in-place strategy: the effect of time to peak ankle torque confirmed by a biomechanical model and an experimental validation

M. Simoneau, R Corbeil. Universitd Laval, PEPS Kinesiology,

Quebec G1K 7P4, Canada Introduction: Recent studies have suggested that the ability to maintain a stable posture depends not only on the magnitude of the restoring torque, but also on the time to generate this torque. The objective of this study was to build a biomechanical that predicts the balance stability boundary which includes time to peak ankle torque and to experimentally validate the model. Methods: Model validation: Dynamic equation of angular motion of an inverted pendulum and mechanical constraints of the foot