Psychology includes many examples of models that make very different psychological assumptions but provide equally good fits to behavioral data. One way to progress is to add extra constraints to the models. The goal of CCN is to build and test models that are simultaneously consistent with the available behavioral and neuroscience data. So a successful CCN model should provide good fits to the available accuracy and response time data, but also to some available neuroscience data, which might include single-neuron spike trains, neural recordings from EEG, fMRI, PET, DTI/DSI, MEG, or optical imaging experiments, or behavioral results from neuropsychological patient, TMS, gene, or drug studies. Adding these extra constraints greatly facilitates the search for valid models. For example, CCN researchers independently modeling the same behavior are likely to converge on highly similar models.

Three important CCN Modeling Principles :

1. The Neuroscience Ideal: A CCN model should not make any assumptions that are known to contradict the current neuroscience literature. 

2. The Simplicity Heuristic: No extra neuroscientific detail should be added to the model unless there are data to test this component of the model or past research has shown that the neuroscientific detail is a major contributor to the explanation. 

3. The Set-in-Stone Ideal: Once set, the architecture of the network and the models of each individual unit should remain fixed throughout all applications.