要旨
Swimming in lampreys is generated by a Central Pattern Generator (CPG) in the spinal cord. This CPG produces neural and motor activity which oscillates between the two sides of the body and propagates along the spinal cord. Central control switches the CPG between different behavior regimes including forward and backward swimming of various speeds, and turning. These behaviors are characterized by oscillation frequencies, direction of the oscillation phase gradient along the body, and the balance between the left and right side of the oscillators. Previously, apart from computer simulations, the CPG network has only been modeled analytically as an abstract chain of coupled phase oscillators, restricting the applicability and the insights that can be gained. We present analytical results and simulations of a rate model of the CPG including all three types of relevant neurons known from experiments, thus capturing phase, amplitude, and generation of oscillations. Our model explains features obscured in a chain of phase oscillators, including oscillations in a neural circuit in which local neural populations do not oscillate spontaneously, and specific connection structures from different neural types. Our mathematical analysis provides significant insights which would be extremely difficult to obtain by simulations alone. In particular, we analyzed how the CPG behavior can be generated and controlled. These understandings were confirmed in simulations which generated forward swimming, backward swimming, turning, and switching of behavior between these regimes. We use the model to make experimental predictions about the structure and strength of the synaptic connections within and between body segments needed for the network to function as a controllable CPG.
