What’s new in NEST 3.4¶

This page contains a summary of important breaking and non-breaking changes from NEST 3.3 to NEST 3.4. In addition to the release notes on GitHub, this page also contains transition information that helps you to update your simulation scripts when you come from an older version of NEST.

If you transition from an earlier version, please see our extensive transition guide from NEST 2.x to 3.0 and the list of updates for previous releases in the 3.x series.

Documentation restructuring and new theme¶

NEST documentation has a new theme! We did a major overhaul of the layout and structure of the documentation. The changes aim to improve findability and access of content. With a more modern layout, our wide range of docs can be discovered more easily.

The table of contents is simplified and the content is grouped based on topics (neurons, synapses etc) rather than type of documentation (e.g., ‘guides’).

We would be highly interested in any feedback you might have on the new look-and-feel either on our mailing list or as an issue on GitHub

Query spatially structured networks from target neuron perspective¶

PyNEST now provides functions GetSourceNodes(), GetSourcePositions(), and PlotSources() which allow you to query or plot the source neurons of a given target neuron.

Extent and center for spatial layers with freely placed neurons¶

Spatial layers in NEST can be created by specifying node positions in the call to Create() using spatial distributions from nest.spatial.

When using spatial.free, the layer’s extent will be determined automatically based on the positions of the lower-leftmost and upper-rightmost nodes in the layer, if omitted. While earlier versions of NEST added a hard-coded padding, NEST 3.4 will only use the node positions.

Likewise, the center of a layer is now automatically computed as the midpoint between the lower-leftmost and the upper-rightmost nodes.

When creating a layer with only a single node, the extent still has to be specified explicitly.

Disconnect with `SynapseCollection`¶

It is now possible to disconnect nodes using a SynapseCollection as argument to either disconnect() or the member function disconnect() of the SynapseCollection.

Removal of deprecated models¶

The models iaf_psc_alpha_canon and pp_pop_psc_delta have long been deprecated and were now removed from NEST. In case you depend on them, you will find similar functionality in the replacement models iaf_psc_alpha_ps and iaf_psc_alpha_ps, respectively.
Model spike_dilutor is now deprecated and can only be used in single-threaded mode. To implement connections which transmit spikes with fixed probability, use bernoulli_synapse instead.

Changed port of NEST Server¶

To avoid conflicts with other services, the default port for NEST Server has been changed from 5000 to 52025.