jdmonaco / spikemaps

Adaptive kernel-based spatial maps for neural spike-trajectory data.

Note: spikemaps has not yet been refactored as an independent package. You should have no expectation that it will run or import correctly in its current state.

I hope to ameliorate this as soon as I have time; of course, PRs welcome.

 

spikemaps

The spikemaps package supports the creation of adaptive kernel-based maps from neurobehavioral datasets containing spikes and (x,y)-position trajectories within a 2D environment (e.g., average firing-rate maps for place-cell recordings with head-position tracking).

Origin

This code was used to generate all of the spatial map images presented in this paper:

• Monaco JD, De Guzman RM, Blair HT, and Zhang K. (2019). Spatial synchronization codes from coupled rate-phase neurons. PLOS Computational Biology, 15(1), e1006741. doi: 10.1371/journal.pcbi.1006741

The complete code archive for the paper is available on figshare (doi: 10.6084/m9.figshare.6072317.v1) and the dataset is archived on OSF (doi: 10.17605/osf.io/psbcw). The spikemaps package is based on the spc.tools subpackage in that code archive.

Dependencies

Note: This section will be updated as the packaging and dependencies are fixed.

The nearest-neighbor modeling depends on scikit-learn algorithms, which can be installed into an Anaconda environment as:

conda install scikit-learn

Similarly, for numpy, matplotlib, and pillow.

Todo

• Fix dependencies for other packages of mine (e.g., remove or add as submodules)
• Update the setup.py to ensure correct installation, etc.
• Improve function and class APIs to enhance usabililty and convenience
• Code style and formatting consistency (e.g., flake8 validation)