NEURON 7.7.2 is now available for download. It is the recommended version for all new projects.
New features since NEURON 7.6 include:
- Reaction-diffusion enhancements
- Reimplemented reaction-diffusion simulation support, providing ~10x faster simulations with JIT compiled reactions
- faster 3D voxelization
- rxd.Parameters can now be used as fixed concentrations
- support for specifying electrophysiology models (see, e.g. the Hodgkin-Huxley model implemented using the rxd framework)
- extracellular simulation support enabled by default
- SBML export.
- See the tutorials for examples.
- Python-graphics aware .plot methods for RangeVarPlot and PlotShape.
- Mechanism introspection using new .code() and .file() methods in MechanismType.
- Updated help documentation; all examples should work for both Python 2 and 3.
- Ontology tagging with REPRESENTS keyword for NMODL and represents keyword for rxd.Species.
- Can now initialize MPI with h.nrnmpi_init(); i.e. no need for special flags or mpi4py.
- neuron.units submodule to simplify model specification.
- All platforms can now use nrnivmodl to compile NMODL files; previously Windows could not.
- Python-API enhancements
- constructor now accepts a Python iterable of sections
- methods accept a section argument; no need for sec=
- can directly set values; e.g. vec = 42
- vector.record now returns the vector, allowing one line recording specification; e.g. v = h.Vector().record(soma(0.5)._ref_v)
- More compact syntax for h.distance and h.RangeVarPlot
- Simpler NMODL pointer setting; instead of using h.setpointer, can use, e.g. seg._ref_POINTER_suffix = ptr
- new Section methods: wholetree, subtree that return Python lists of sections
- new Segment method: volume
- More meaningful return data types for CVode methods