A mathematical model of adaptive behavior in quadruped locomotion

Locomotion involves repetitive movements and is often executed unconsciously and automatically. In order to achieve smooth locomotion, the coordination of the rhythms of all physical parts is important. Neurophysiological studies have revealed that basic rhythms are produced in the spinal network called, the central pattern generator (CPG), where some neural oscillators interact to self-organize coordinated rhythms. We present a model of the adaptation of locomotion patterns to a variable environment, and attempt to elucidate how the dynamics of locomotion pattern generation are adjusted by the environmental changes. Recent experimental results indicate that decerebrate cats have the ability to learn new gait patterns in a changed environment. In those experiments, a decerebrate cat was set on a treadmill consisting of three moving belts. This treadmill provides a periodic perturbation to each limb through variation of the speed of each belt. When the belt for the left forelimb is quickened, the decerebrate cat initially loses interlimb coordination and stability, but gradually recovers them and finally walks with a new gait. Based on the above biological facts, we propose a CPG model whose rhythmic pattern adapts to periodic perturbation from the variable environment. First, we design the oscillator interactions to generate a desired rhythmic pattern. In our model, oscillator interactions are regarded as the forces that generate the desired motion pattern. If the desired pattern has already been realized, then the interactions are equal to zero. However, this rhythmic pattern is not reproducible when there is an environmental change. Also, if we do not adjust the rhythmic dynamics, the oscillator interactions will not be zero. Therefore, in our adaptation rule, we adjust the memorized rhythmic pattern so as to minimize the oscillator interactions. This rule can describe the adaptive behavior of decerebrate cats well. Finally, we propose a mathematical framework of an adaptation in rhythmic motion. Our framework consists of three types of dynamics: environmental, rhythmic motion, and adaptation dynamics. We conclude that the time scale of adaptation dynamics should be much larger than that of rhythmic motion dynamics, and the repetition of rhythmic motions in a stable environment is important for the convergence of adaptation. Received: 10 July 1997 / Accepted in revised form: 13 March 1998     

A mathematical model of the human thermal system

This paper describes a mathematical model developed to simulate the physical characteristics of the human thermal system in the transient state. Physiological parameters, such as local metabolic heat generation rates, local blood flow rates, and rates of sweating, must be specified as input data. Automatic computation of these parameters will be built into the model at a later date when it is used to study thermal regulation in the human. Finite-difference techniques have been used to solve the heat conduction equation on a Control Data Corporation 1604 computer. Since numerical techniques were used, it was possible to include many more factors in this model than in previous ones. The body was divided into 15 geometric regions, which were the head, the thorax, the abdomen, and the proximal, medial, and distal segments of the arms and legs. Axial gradients in a given segment were neglected. In each segment, the large arteries and veins were approximated by an arterial pool and a venous pool which were distributed radially throughout the segment. Accumulation of heat in the blood of the large arteries and veins, and heat transfer from the large arteries and veins to the surrounding tissue were taken into account. The venous streams were collected together at the heart before flowing into the capillaries of the lungs. Each of the segments was subdivided into 15 radial sections, thereby allowing considerable freedom in the assignment of physical properties such as thermal conductivity and rate of blood flow to the capillaries. The program has been carefully checked for errors, and it is now being used to analyze some problems of current interest. This study was supported by the office of the Surgeon General, U.S. Army, under contract no. DA 49-193-MD-2005.     

A mathematical model of the gate control theory of pain

The first test which any theory of pain must pass is that it must be able to explain the phenomena observed in acute pain in humans. This criterion is used to test the major theory of pain at present, the gate control theory of Melzack & Wall (1965, 1982). The theory is explicit enough to be cast in mathematical terms, and the mathematical model is shown to explain the observations considered. It also points up a common misconception on the consequences of the theory, and thus demolishes an argument which has been used against it. A hypothesis of the origin of rhythmic pain is then made, and consequent testable predictions given. This is the first time that the gate control theory has been used to explain any quality of pain. It has important consequences for the treatment of such pain. Finally, the applicability of the gate control theory as an explanation for chronic pain is discussed.     

A complex mathematical model of the human menstrual cycle

Despite the fact that more than 100 million women worldwide use birth control pills and that half of the world's population is concerned, the menstrual cycle has so far received comparatively little attention in the field of mathematical modeling. The term menstrual cycle comprises the processes of the control system in the female body that, under healthy circumstances, lead to ovulation at regular intervals, thus making reproduction possible. If this is not the case or ovulation is not desired, the question arises how this control system can be influenced, for example, by hormonal treatments. In order to be able to cover a vast range of external manipulations, the mathematical model must comprise the main components where the processes belonging to the menstrual cycle occur, as well as their interrelations. A system of differential equations serves as the mathematical model, describing the dynamics of hormones, enzymes, receptors, and follicular phases. Since the processes take place in different parts of the body and influence each other with a certain delay, passing over to delay differential equations is deemed a reasonable step. The pulsatile release of the gonadotropin-releasing hormone (GnRH) is controlled by a complex neural network. We choose to model the pulse time points of this GnRH pulse generator by a stochastic process. Focus in this paper is on the model development. This rather elaborate mathematical model is the basis for a detailed analysis and could be helpful for possible drug design.     

Modelling of direct motor program learning in fast human arm motions

We propose and simulate a new paradigm for organization of motor control in fast and accurate human arm motions. We call the paradigm direct motor program learning since the control programs are learned directly without knowing or learning the dynamics of a controlled system.The idea is to approximate the dependence of the motor control programs on the vector of the task parameters rather than to use a model of the system dynamics. We apply iterative learning control and scattered data multivariate approximation techniques to achieve the goal. The advantage of the paradigm is that the control complexity depends neither on the order nor on the nonlinearity of the system dynamics.We simulate the direct motor program learning paradigm in the task of point-to-point control of fast planar human arm motions. Simulation takes into account nonlinear arm dynamics, muscle force dynamics, delay in low-level reflex feedback, time dependence of the feedback gains and coactivation of antagonist muscles. Despite highly nonlinear time-variant dynamics of the controlled system, reasonably good motion precision is obtained over a wide range of the task parameters (initial and final positions of the arm). The simulation results demonstrate that the paradigm is indeed viable and could be considered as a possible explanation for the organization of motor control of fast motions.     

Impedance characteristics of a neuromusculoskeletal model of the human arm II. Movement control

The modulation of neuromusculoskeletal impedance during movements is analysed using a motor control model of the human arm. The motor control system combines feedback and feedforward control and both control modes are determined in one optimization process. In the model, the stiffness varies at the double movement frequency for 2-Hz oscillatory elbow movements and has high values at the movement reversals. During goal-directed two-degrees-of-freedom arm movements, the stiffness is decreased during the movement and may be increased in the initial and final phases, depending on the movement velocity. The stiffness has a considerable curl during the movement, as was also observed in experimental data. The dynamic stiffness patterns of the model can be explained basically by the α−γ coactivation scheme where feedback gains covary with motor control signals. In addition to the modulation of the gain factors, it is argued that the variation of the intrinsic stiffness has a considerable effect on movement control, especially during fast movements. Received: 14 October 1997 / Accepted in revised form: 18 May 1999     

Impedance characteristics of a neuromusculoskeletal model of the human arm I. Posture control

The mechanical impedance of neuromusculoskeletal models of the human arm is studied in this paper. The model analysis provides a better understanding of the contributions of possible intrinsic and reflexive components of arm impedance, makes clear the limitations of second-order mass-viscosity-stiffness models and reveals possible task effects on the impedance. The musculoskeletal model describes planar movements of the upper arm and forearm, which are moved by six lumped muscles with nonlinear dynamics. The motor control system is represented by a neural network which combines feedforward and feedback control. It is optimized for the control of movements or for posture control in the presence of external forces. The achieved impedance characteristics depend on the conditions during the learning process. In particular, the impedance is adapted in a suitable way to the frequency content and direction of external forces acting on the hand during an isometric task. The impedance characteristics of a model, which is optimized for movement control, are similar to experimental data in the literature. The achieved stiffness is, to a large extent, reflexively determined whereas the approximated viscosity is primarily due to intrinsic attributes. It is argued that usually applied Hill-type muscle models do not properly represent intrinsic muscle stiffness. Received: 14 October 1997 / Accepted in revised form: 18 May 1999     

A mathematical model of human thymidine kinase 2 activity

The mitochondrial enzyme thymidine kinase 2 (TK2) phosphorylates deoxythymidine (dT) and deoxycytidine (dC) to form dTMP and dCMP, which in cells rapidly become the negative-feedback end-products dTTP and dCTP. TK2 kinetic activity exhibits Hill coefficients of ～0.5 (apparent negative cooperativity) for dT and ～1 for dC. We present a mathematical model of TK2 activity that is applicable if TK2 exists as two monomer forms in equilibrium.     

A model of the cerebellum in adaptive control of saccadic gain

 We review data showing that the cerebellum is required for adaptation of saccadic gain to repeated presentations of dual-step visual targets and thus, presumably, for providing adaptive corrections for the brainstem saccade generator in response to any error created by the open-loop saccadic system. We model the adaptability of the system in terms of plasticity of synapses from parallel fibers to Purkinje cells in cerebellar cortex, stressing the integration of cerebellar cortex and nuclei in microzones as the units for correction of motor pattern generators. We propose a model of the inferior olive as an error detector, and use a ‘window of eligibility’ to insure that error signals that elicit a corrective movement are used to adjust the original movement, not the secondary movement. In a companion paper we simulate this large, realistic network of neural-like units to study the complex spatiotemporal behavior of neuronal subpopulations implicated in the control and adaptation of saccades. Received: 25 November 1994/Accepted in revised form: 6 February 1996     

A mathematical model of compartmentalized neurotransmitter metabolism in the human brain

After administration of enriched [1-13C]glucose, the rate of 13C label incorporation into glutamate C4, C3, and C2, glutamine C4, C3, and C2, and aspartate C2 and C3 was simultaneously measured in six normal subjects by 13C NMR at 4 Tesla in 45-ml volumes encompassing the visual cortex. The resulting eight time courses were simultaneously fitted to a mathematical model. The rate of (neuronal) tricarboxylic acid cycle flux (V(PDH)), 0.57 +/- 0.06 micromol. g(-1). min(-1), was comparable to the exchange rate between (mitochondrial) 2-oxoglutarate and (cytosolic) glutamate (Vx), 0.57 +/- 0.19 micromol. g(-1). min(-1)), which may reflect to a large extent malate-aspartate shuttle activity. At rest, oxidative glucose consumption [CMR(Glc(ox))] was 0.41 +/- 0.03 miccromol. g(-1). min(-1), and (glial) pyruvate carboxylation (VPC) was 0.09 +/- 0.02 micromol. g(-1). min(-1). The flux through glutamine synthetase (Vsyn) was 0.26 +/- 0.06 micromol. g(-1). min(-1). A fraction of Vsyn was attributed to be from (neuronal) glutamate, and the corresponding rate of apparent glutamatergic neurotransmission (VNT) was 0.17 +/- 0.05 micromol. g(-1). min(-1). The ratio [VNT/CMR(Glcox)] was 0.41 +/- 0.14 and thus clearly different from a 1:1 stoichiometry, consistent with a significant fraction (approximately 90%) of ATP generated in astrocytes being oxidative. The study underlines the importance of assumptions made in modeling 13C labeling data in brain.     

Active muscle response using feedback control of a finite element human arm model

Mathematical human body models (HBMs) are important research tools that are used to study the human response in car crash situations. Development of automotive safety systems requires the implementation of active muscle response in HBM, as novel safety systems also interact with vehicle occupants in the pre-crash phase. In this study, active muscle response was implemented using feedback control of a nonlinear muscle model in the right upper extremity of a finite element (FE) HBM. Hill-type line muscle elements were added, and the active and passive properties were assessed. Volunteer tests with low impact loading resulting in elbow flexion motions were performed. Simulations of posture maintenance in a gravity field and the volunteer tests were successfully conducted. It was concluded that feedback control of a nonlinear musculoskeletal model can be used to obtain posture maintenance and human-like reflexive responses in an FE HBM.     

A model of force and impedance in human arm movements

This paper describes a simple computational model of joint torque and impedance in human arm movements that can be used to simulate three-dimensional movements of the (redundant) arm or leg and to design the control of robots and human-machine interfaces. This model, based on recent physiological findings, assumes that (1) the central nervous system learns the force and impedance to perform a task successfully in a given stable or unstable dynamic environment and (2) stiffness is linearly related to the magnitude of the joint torque and increased to compensate for environment instability. Comparison with existing data shows that this simple model is able to predict impedance geometry well.     

A mathematical model for cell size control in fission yeast

Experimental investigations of cell size control in fission yeast Schizosaccharomyces pombe have illustrated that the cell cycle features ‘sizer’ and ‘timer’ phases which are distinguished by a growth rate changing point. Based on current biological knowledge of fission yeast size control, we propose here a model of ordinary differential equations (ODEs) for a possible explanation of the facts and control mechanism which is coupled with the cell cycle. Simulation results of the ODE model are demonstrated to agree with experimental data for the wild type and the cdc2-33 mutant. We show that the coupling of cell growth to cell division by translational control may account for observed properties of size control in fission yeast. As the translational control in the expression of cycle proteins Cdc13 and Cdc25 constructs positive feedback loops, the dynamical activities of the key components undergoes a rapid rising after a preliminary stage of slow increase. The coupling of this dynamical behavior to the elongation of the cell naturally gives rise to a rate change point and to ‘sizer’ and ‘timer’ phases, which characterize the cell cycle of fission yeast.     

Constraints for joint angle control of the human arm

The targeting movements of a human arm were examined when restricted to a horizontal plane. The three joints at shoulder, elbow, and wrist are allowed to move. Thus, the system is redundant and needs constraints. A model calculation using a simple form of constraint is found to describe the experimental results: a cost function is applied to each joint. The constraint consists in minimizing the sum of the costs of all three joints. The cost functions might be interpreted as to describing the energy cost necessary to move the joint and/or represent a mechanism which avoids singularities.     

A spring model and equivalent neural network for arm posture control

A model is introduced for the motor program which controls final position. The first part of the model relates the biomechanical properties of the muscles to the EMG activities of the extensor and flexor muscles and thereby generates quatitative predictions for the relationships between the EMGs, final position, external forces, muscle stiffness, and muscle tension. To the extent that comparable data exist, the model is shown to give correct quantitative predictions. When only qualitative comparisons can be made, the model is consistent with the data in the literature. The model is complete and can be tested quantitatively in detail in the future. An equivalent circuit for the neural network that innervates the muscles is given. It is shown to have the advantages of making the programming of final position simple to either compute or lookup in a table. In addition, new situations, such as adapting to a force, or an unusual viewing angle, lead to very simple changes in the basic program in terms of the equivalent circuit.     

Three-dimensional mathematical model of the human spine structure

A method is presented for constructing discrete-parameter type three-dimensional mathematical models and governing equations of motion of the spine structure. The anatomic structure is represented by any combination of rigid bodies, springs, and dashpots in space. These are positioned, orientated, and connected in a manner to represent the true mechanical function of the structure. The rigid bodies are of any shape and have 6 degrees-of-freedom, allowing three-dimensional motion. The springs and dashpots may have up to twenty-one stiffness and damping coefficients respectively to precisely represent the three-dimensional coupled behavior. The method is straightforward and simple to apply. The governing equations are in the matrix form and are easily generated and solved by computer techniques.     

Optimal posture control of a musculo-skeletal arm model

 In this paper maximal performance posture control of the human arm is investigated by means of model simulations. Recent experiments (F.C.T. van der Helm, submitted, 2000) have shown that the reflexive feedback during postural control varies with the bandwidth of the applied force disturbances. This paper focusses on the influence of the frequency content of force disturbances on the reflexive feedback gains by means of optimization. The arm is modelled by a non-linear musculo-skeletal model with two degrees of freedom and six muscles. To facilitate the optimization of the model parameters, the arm model is linearized. A performance criterion is minimized for stochastic force disturbances in a two-step procedure: (1) optimization of static muscle activations using an additional energy criterion to obtain a unique and energy-efficient solution; and (2) optimization of reflex gains using an additional control effort criterion to obtain a unique solution. The optimization reveals that for the given task and posture, the shoulder muscles have the largest contribution, whereas the bi-articular muscles have a relatively small contribution to the behaviour. The dynamics at the endpoint level are estimated so that a comparison can be made with the experiments. Compared to the experiments, the intrinsic damping of the model is relatively large (about 150%), whereas the intrinsic stiffness is relatively small (about 60%). These differences can be attributed to unmodelled mechanical effects of cross-bridges in Hill-type muscle models. The optimized reflex gains show remarkable similarities with the values found in the experiments, implying that humans can adjust their reflexive feedback gains in an optimal way, weighting the performance and energy. The approach in this paper could be useful in the study of various posture tasks, for example in the prediction of the relation between the control parameters of various musculo-skeletal models and different experimental variables. Received: 24 January 2000 / Accepted in revised form: 7 July 2000     

A mathematical model of hind-limb control in cats when walking backward

The “walking backward” mode was achieved within a single model of cat hind-limb locomotion with the balance maintenance only due to a change in the controlling actions (in addition to the “forward walking” mode). The skeletal part of the model contains the spine, pelvis, and two limbs consisting of the thigh, shin, and foot. The hip joint and spine mount in the thoracic region have three degrees of freedom; the knee and ankle joints have one degree of freedom. The pelvis is rigidly connected to the spine. Control is performed by model muscles (flexors and extensors of the thigh, shin, and foot). The muscle activation is performed by the effects that are typical for motoneurons that control the muscles. The feet in the support phase touch the treadmill, which moves at a constant speed. The model qualitatively reproduces multiple characteristics of feline movements during forward and backward walking (supporting its validity).