Abstract
Central pattern generators (CPGs) are neural circuits that are able to produce rhythmic outputs to drive critical behaviors such as respiration and locomotion. In this work I exploited the well-studied CPG networks of the stomatogastric ganglion (STG) of the crab, Cancer borealis. Although the STG is able to produce a stereotyped output pattern across animals, and also across species, direct observation of the properties of STG neurons has revealed wide animal-to-animal variability. Additionally, these networks are heavily modulated by a wide array of amines and peptides that can alter all aspects of circuit function. I investigated the interaction of variability and neuromodulation in this system by performing a set of long-term experiments on preparations of the stomatogastric nervous system (STNS) in the presence and absence of neuromodulatory inputs. Long-term preparations of the STNS with modulatory inputs intact continued to display activity patterns that resembled those observed immediately after dissection. The removal of modulatory inputs (decentralization) had profound and long-lasting effects on STG output that did not recover. In both the continuously active pyloric rhythm, and the episodic gastric rhythm, decentralization reduced the burst frequency of CPG neurons, and increased their variability. Decentralization altered phase relationships in both rhythms and disrupted coordination between the gastric and pyloric cycle periods. Interestingly, I observed large variability in the response of individual preparations to the loss of neuromodulatory inputs. To examine the potential sources of this variability, I applied the modulatory peptide proctolin at several time points following decentralization. I found that responses to proctolin were variable, but highly correlated at time points long following decentralization. Together, these data support the hypothesis that features of STG circuits are variable across animals, and that neuromodulation is able to compensate for these differences to produce stereotyped output. I also assessed the competence of the severed axons of modulatory projection neurons, finding that these axons still contain peptides and can influence STG activity many days after transection. These results comprise the most comprehensive description of the response of the STG to the loss of neuromodulatory input, a common and interesting experimental procedure in this system.