Furthermore, as an in vitro model of a brain circuit, our device cannot fully model the complexity of the in vivo physiology, due at least to several challenges: the axonal projections that make up the circuit are largely unmyelinated; the devices do not contain all types of non-neuronal cells (e.g. endothelial cells and microglia); and we are not able to model all possible connections of the pathway. We can choose, however, to highlight different regions of a given brain circuit by seeding other subtypes into the central chamber. Lastly, in cases where the focus is on a three-way circuit, the regions of the device of greatest interest are the locations where all three neuron types are closely apposed. The current four-side chamber design may limit the number of central-chamber neurons available for patch clamp analysis for a three-way circuit. Additional side compartments (e.g. eight instead of four), with neurons plated in an alternating fashion will greatly increase the area of interaction between three distinct subtypes.
