Overview

Our domain adaptive model-based inference framework, dastNeuro aims to infer various neural properties from diverse experimental data. Our framework takes a set of experimental data and a model hypothesis as the input, see Fig. 2 in (Sheng et al., 2022) for details, which are defined by users.

In this tutorial, we'll focus on two applications as in (Sheng et al., 2022). The first is to infer biophysical properties of neurons, and the second is to infer monosynaptic connectivity in a network. These two applications cover two different categories of applications in our paper. For neuronal inference, we aim to fully replicate the experimental input-output pattern by the model. The associated code is under branch BiophysicsInference, and please refer to the tutorial Inferring Neuronal Properties. For monosynaptic connectivity inference, we aim to use features extracted from synthesized data to guide the inference on experimental data. The associated code is under branch ConnectivityInference, and please refer to the tutorial Inferring Monosynaptic Connectivity from Cross-correlograms.