Large-Scale Models of Spinal and Brainstem Locomotor Circuits
Royal Institute of Technology
The fictive swimming preparation of the lamprey spinal cord is used as an experimental model for vertebrate locomotion. It has allowed detailed studies of the cellular and synaptic properties of the spinal rhythm generating circuitry. This system has also been subject to a series of modeling studies. Previous sub-sampled simulation models have been able to explain a number of experimental findings and have generated several experimentally testable hypotheses. However, despite the use of biophysically detailed compartmental model neurons, individual simulated cells typically show a too high frequency and too regular spiking patterns. We have recently investigated to what extent this effect can be overcome by the use of a larger network and distribution of cell properties. Instead of just one cell of each type we study a full-scale model of a piece of lamprey spinal cord with about hundred cells per segment and a realistic density of synapses. Such a simulated spinal cord piece produces robust rhythmic activity within the physiological frequency range. Intracellular traces and the synthesized EMG look much like what is actually recorded in vitro. When activated asymmetrically, turning activity including a rebound ipsi-lateral burst can be qualitatively reproduced. In sufficiently long spinal cord pieces inter-segmental coordination patterns can be studied. We further compare with the bursting activity produced by a reduced connectionist type model where graded output units represent local populations of neurons. Such reduced models are currently subject to analysis with regard to their different activity modes. The swimming movements generated by the simulated spinal pattern generator have been studied by means of a neuro-mechanical model where the lamprey body and surrounding water are also included in the simulation.
This work illustrates that it may be useful to describe the same system by a range of models at different levels of abstraction, while making careful transformations between them.