Classification of dynamics of a model of motor coordination and comparison with Parkinson's disease data

In our recent reports motor coordination of human lower limbs has been investigated during pedaling a special kind of ergometer which allows its left and right pedals to rotate independently. In particular, relative phase between left and right rotational-velocity waveforms of the pedals and their amplitude modulation have been analyzed for patients with Parkinson's disease (PD). Several patients showed peculiar interlimb coordination different from the regular anti-phase pattern of normal subjects. We have reported that these disordered patterns could be classified into four groups. Moreover, it has been demonstrated that a mathematical model could reproduce most of the disordered patterns. Such a model includes a schematization of the central pattern generator with two identical half-centers mutually coupled and two tonic control signals from higher motor centers, each of which inputs to one of the half-centers. Depending on the intensities of the tonic signals and on the differences between them, the model could generate a range of dynamics comparable to the clinically observed disordered patterns. In this paper, we explore the dynamics of the model by varying the intensities of the tonic signals in the model. Using the same method used for classifying the clinical data, the dynamics of the model are classified into several groups. The classified groups for the simulated data are compared with those for the clinical data to look at qualitative correspondence. Our systematic exploration of the model's dynamics in a wide range of the parameter space has revealed global organization of the bifurcations including Hopf bifurcations and cascades of period-doubling bifurcations among others, suggesting that the bifurcations, induced by instability of stable dynamics of the human motor control system, are responsible for the emergence of the disordered coordination in PD patients.