Swimming and swim control in the medicinal leech: Computational, analytical, and imaging approaches

I examined three related questions: (1) How does the segmental leech swim central pattern generator (CPG) generate an oscillation? (2) How can a two-cell reciprocal inhibition network generate an oscillation by the mechanism of synaptic depression? (3) How does the leech swim-terminating neuron Tr2 stop swimming? The first was addressed by constructing a computational model of the leech swim CPG, based on the available information about the physiology. The main result of this work was that models employing strictly passive neurons (no active conductances) can generate realistic oscillations, but are unrealistically sensitive to small changes to the synaptic strengths and other model parameters. Therefore, the leech swim CPG probably employs some non-passive mechanism to generate swim oscillations. The most likely candidates, based on the physiology, are membrane relaxation, post-inhibitory rebound (PIR) or synaptic depression. The second question above was addressed by constructing an analytically tractable model of a two-cell circuit that oscillates due to synaptic depression. We used this model to make predictions about various features (e.g. frequency, amplitude) of a two-cell oscillation that is governed by synaptic depression. The third question was addressed by imaging fluorescence resonance energy transfer (FRET) voltage-sensitive dyes (VSDs) to efficiently search for neurons postsynaptic to cell Tr2. We identified three previously unknown synaptic targets of Tr2. Two of these targets stop swimming when stimulated during swimming, suggesting that they are each part of a neuronal pathway whereby Tr2 stops swimming.