The reconstruction of the SSCx involves the acquisition and organization of data from anatomical and electrophysiological cortical slices from rodent brains. Sparse data has been collected from our own laboratory and from published sources worldwide, both of which describes the structural and functional organization of the SSCx at various anatomical levels.

This ranges from individual neurons to synaptic connections and network activity in microcircuits. The data provides constraints, rules, and the principles to build computational models at specific levels of detail.

The SSCx workflow extracts all relevant information from sparse data and exploits interdependencies to build detailed and dense models of individual cells, synaptic connections and microcircuits. Rules and principles of organization are identified from sparse biological data sets and missing information is extrapolated to fill knowledge gaps, which enable a dense data-driven digital reconstruction of the entire SSCx region.

Digital reconstructions are built based on experimental datasets obtained from the developing SSCx in two-week old rats and integrate knowledge on the molecular and cellular composition of neurons, and the anatomy and physiology of synapses and microcircuits. Individually reconstructed neuron morphologies and their corresponding electrophysiological properties are assembled into atlas-based geometries of neocortical tissue to derive their synaptic connectivity.

Circuit reconstructions are based on a standardized workflow enabled by software tools and supercomputing infrastructure. The parameterization of the tissue model is based on biological data: directly, where available, generalized from data obtained in other similar systems; or, where unavailable, predicted from multi-constraints imposed by sparse data.

Validations are a crucial part of the data-driven modeling workflow. Through validation, we reduce the risk that errors may lead to major inaccuracies in the reconstruction or in simulations of emergent behavior.

Successful validations not only enable the systematic exploration of the emergent properties of the model, but also establish predictions for future in vitro experiments, or may call into question existing experimental data.

Failure in the validation process may indicate errors in experimental data, which enables us to identify future refinements.

Rigorous validation of a metric at one level of detail therefore also prevents error amplification to the next level, and triggers specific experimental refinements.

The Blue Brain Project Validation step in our workflow provides a scaffold that enables the integration of available experimental data, identifies missing experimental data, and facilitates the iterative refinement of constituent models.

The digital reconstruction of the SSCx provides an array of predictions across the many levels of organization in this brain region.

These predictions provide insights to link underlying structure with function. In addition, predictions are also a means to validate the component models of the SSCx model and identify missing data that could guide targeted experiments. In particular, we provide predictions on the propagation of activity across the different sub-regions of the SSCx.