Objective. In silico experiments use mathematical models that capture as much as possible from the properties of the biological system under investigation. Our aim was to test the publicly available CA1 pyramidal cell models using the same simulation tasks, to compare them, and provide a systematic overview of their properties in order to improve the usefulness of these models as a tool for in silico experiments.
Methods. Parameters describing the morphology of the cells and the implemented biophysical mechanisms were collected from the ModelDB database of SenseLab Project. This data was analyzed in correlation with the purpose for which each particular model was developed. Multicompartmental simulations were run using the Neuron modeling platform. The properties of the action potentials generated in response to current injection, the firing pattern and the dendritic back-propagation were analyzed.
Results. The studied models were optimized to explore different physiological and pathological properties of the CA1 pyramidal cells. We could identify four broad classes of models focusing on: (i) initiation of the action potential, firing pattern and spike timing, (ii) dendritic back-propagation, (iii) dendritic integration of synaptic inputs and (iv) neuronal network activity. Despite the large variation of the active conductances implemented in the models, the properties of the individual action potentials were quite similar, but even the most complex models could not reproduce all studied biological phenomena.
Conclusions. At the moment the “perfect” pyramidal cell model is not yet available. Our work, hopefully, will help finding the best model for each scientific question under investigation.