Recent work argues that similar network performance can result from highly variable sets of network parameters, raising the question of whether neuromodulation can be reliable across individuals with networks with different sets of synaptic strengths and intrinsic membrane conductances. To address this question, we used the dynamic clamp to construct 2-cell reciprocally inhibitory networks from gastric mill (GM) neurons of the crab stomatogastric ganglion. When the strength of the artificial inhibitory synapses (gsyn) and the conductance of an artificial Ih (gh) were varied with the dynamic clamp, a variety of network behaviors resulted, including regions of stable alternating bursting. Maps of network output as a function of gsyn and gh were constructed in normal saline and again in the presence of serotonin or oxotremorine. Both serotonin and oxotremorine depolarize and excite isolated individual GM neurons, but by different cellular mechanisms. Serotonin and oxotremorine each increased the size of the parameter regions that supported alternating bursting, and, on average, increased burst frequency. Nonetheless, in both cases some parameter sets within the sample space deviated from the mean population response and decreased in frequency. These data provide insight into why pharmacological treatments that work in most individuals can generate anomalous actions in a few individuals, and they have implications for understanding the evolution of nervous systems. dynamic clamp ͉ evolution ͉ serotonin ͉ stomatogastric ganglion ͉ oscillator H ow different are the brains in a population of healthy individuals of the same species? Both modeling and experimental studies demonstrate that similar circuit function can result from a large variation in the synaptic and intrinsic conductances that contribute to circuit function (1-11). This finding suggests that each individual could, during development, build a brain that differs significantly from all others and poses the question of whether circuits with diverse underlying structure, yet similar behavior, can respond reliably (12, 13) to neuromodulatory substances that alter brain states (14-21). Therefore, we used the dynamic clamp (22, 23) to construct 2-cell circuits (24) from neurons of the crab stomatogastric ganglion (STG). The strength of the reciprocal inhibitory synapses (g syn ) and the conductance of an artificial I h (g h ) were varied (24), and the resulting activity patterns were recorded in control saline and in response to 2 neuromodulatory substances, serotonin and the muscarinic agonist, oxotremorine.In these hybrid experiments, the intrinsic properties of these 2 biological neurons vary between each other and across preparations while 2 of the important parameters for network performance are under experimental control. This method ensures that the variability intrinsic to biological systems is retained, while allowing access to some of the parameters controlling network performance.
Results and DiscussionWe used gastric mill (GM) neurons of the STG of t...