Neuroscience drug discovery is challenged by the brain\'s structural and cell-type complexity, which is difficult to model in cellular systems compatible with high-throughput screening methods. Calcium oscillation assays, that harness neurons\' intrinsic capability to develop functional neural networks in cell culture, are currently the closest cellular models with a relevant functional endpoint to model human neuronal circuitry in a dish. Here we further develop this useful assay towards scalable drug discovery applications. We show the importance of defined neuron-to-astrocyte ratios for optimal cellular distribution and surface adherence in HTS-compatible cell culture vessels and how the cell type ratios determine network firing patterns. We identify DAPT, a molecule previously shown to promote neuronal maturation and synapse formation, as a negative regulator of astrocyte viability. Addition of GABA-ergic inhibitory neurons increases the network spike frequency while reducing network spike amplitudes. We develop a pixel-based analysis for plate reader data to access local field activity in an automated and scalable calcium imaging environment. Using this method, we identify a desynchronization of excitatory neuronal activity by GABA neurons that echoes GABA action in vivo, and dysregulation thereof in pathological conditions.