Passive models of neurons in the deep cerebellar nuclei: the effect of reconstruction errors

The goal of this study was to determine how the .t of passive parameters in a compartmental model varies depending on the precise morphological reconstruction of the neuron. We performed whole-cell recordings of deep cerebellar nucleus neurons in brain slices, reconstructed the neuronal morphologies and converted them into detailed compartmental models. A genetic algorithm was used to .nd the best .t of speci.c capacitance CM, membrane resistance RM and axial resistivity RA of the model with recordings from the same cell. We then introduced morphological alterations that represented the likely consequence of shrinkage artefacts and reconstruction errors. We found that the optimal .ts of passive parameters change as much as 173% with such morphological alterations. In addition, dendrites cut during slicing could a$ect the value of RM, but not CM or RA. c © 2004 Elsevier B.V. All rights reserved.