Abstract: Spiking microcircuit models simulate neurons, action potentials, and synaptic transmission to reproduce healthy or pathological brain dynamics. These detailed models offer potential for personalized treatments of brain disorders, but identifying parameters that reproduce patient data in biologically detailed models remains challenging. We apply neural posterior estimation, a modern simulation-based inference method, to efficiently estimate parameter distributions from observed dynamics. With these parameter distributions, we identified neuronal network mechanisms that can compensate for simulated epileptogenic perturbations. This allowed us to create a ranking of compensatory mechanisms and show that interneuron loss, excitatory synaptic sprouting, and intrinsic hyperexcitability require different compensatory mechanisms to maintain excitability. We believe that modern simulation-based inference methods, such as neural posterior estimation, can turn microcircuit simulators into useful clinical tools.