Journal of Computational Neuroscience

, Volume 31, Issue 2, pp 385–400

Virtual NEURON: a strategy for merged biochemical and electrophysiological modeling

  • Sherry-Ann Brown
  • Ion I. Moraru
  • James C. Schaff
  • Leslie M. Loew

DOI: 10.1007/s10827-011-0317-0

Cite this article as:
Brown, SA., Moraru, I.I., Schaff, J.C. et al. J Comput Neurosci (2011) 31: 385. doi:10.1007/s10827-011-0317-0


Because of its highly branched dendrite, the Purkinje neuron requires significant computational resources if coupled electrical and biochemical activity are to be simulated. To address this challenge, we developed a scheme for reducing the geometric complexity; while preserving the essential features of activity in both the soma and a remote dendritic spine. We merged our previously published biochemical model of calcium dynamics and lipid signaling in the Purkinje neuron, developed in the Virtual Cell modeling and simulation environment, with an electrophysiological model based on a Purkinje neuron model available in NEURON. A novel reduction method was applied to the Purkinje neuron geometry to obtain a model with fewer compartments that is tractable in Virtual Cell. Most of the dendritic tree was subject to reduction, but we retained the neuron’s explicit electrical and geometric features along a specified path from spine to soma. Further, unlike previous simplification methods, the dendrites that branch off along the preserved explicit path are retained as reduced branches. We conserved axial resistivity and adjusted passive properties and active channel conductances for the reduction in surface area, and cytosolic calcium for the reduction in volume. Rallpacks are used to validate the reduction algorithm and show that it can be generalized to other complex neuronal geometries. For the Purkinje cell, we found that current injections at the soma were able to produce similar trains of action potentials and membrane potential propagation in the full and reduced models in NEURON; the reduced model produces identical spiking patterns in NEURON and Virtual Cell. Importantly, our reduced model can simulate communication between the soma and a distal spine; an alpha function applied at the spine to represent synaptic stimulation gave similar results in the full and reduced models for potential changes associated with both the spine and the soma. Finally, we combined phosphoinositol signaling and electrophysiology in the reduced model in Virtual Cell. Thus, a strategy has been developed to combine electrophysiology and biochemistry as a step toward merging neuronal and systems biology modeling.


Virtual Cell NEURON Model reduction Compartmental modeling Biochemical simulation Electrophysiology modeling 

Supplementary material

10827_2011_317_MOESM1_ESM.pdf (126 kb)
ESM 1(PDF 125 kb)
10827_2011_317_MOESM2_ESM.pdf (1.2 mb)
ESM 2(PDF 1267 kb)
10827_2011_317_MOESM3_ESM.hoc (0 kb)
ESM 3(HOC 0 kb)
10827_2011_317_MOESM4_ESM.hoc (0 kb)
ESM 4(HOC 0 kb)
10827_2011_317_MOESM5_ESM.hoc (94 kb)
ESM 5(HOC 93 kb)
10827_2011_317_MOESM6_ESM.hoc (4 kb)
ESM 6(HOC 4 kb) (2 kb)
ESM 7(SES 2 kb)
10827_2011_317_MOESM8_ESM.xls (316 kb)
ESM 8(XLS 316 kb)
10827_2011_317_MOESM9_ESM.vcml (2.1 mb)
ESM 9(VCML 2167 kb)

Copyright information

© Springer Science+Business Media, LLC 2011

Authors and Affiliations

  • Sherry-Ann Brown
    • 1
  • Ion I. Moraru
    • 1
  • James C. Schaff
    • 1
  • Leslie M. Loew
    • 1
  1. 1.Richard D. Berlin Center for Cell Analysis & ModelingUniversity of Connecticut Health CenterFarmingtonUSA

Personalised recommendations