Nounou is a JVM-based interface for loading neurophysiological data, written mainly in Scala. It is a full rewrite of a prior version in Java (Takagaki, 2011), and old functions are gradually being translated to Scala and refactored in as needed.
- to provide an adapter to dynamically load neurophysiology data into MATLAB, Mathematica, Scala REPL, and Java (and Python).
- Other file readers are designed to read whole data files into memory at once. This can become quite problematic for large/long files.
- Nounou is designed around data streams, where the data can be loaded as-needed on demand from disk/network.
- Nounou can handle channel layout data within the main data structure to facilitate geometric analyses. This was originally for voltage-sensitive dye/intrinsic imaging, but is applicable for ECoG arrays and multishank electrodes too.
- to provide building blocks for computationally intensive neurophysiology analysis algorithms, such as flow-analysis [1].
- (perhaps at some point) to provide a rudimentary JavaFX-based graphical interface for browsing neurophysiology data
- Nounou is in no way intended as an "uber package" to serve as a one-step analysis tool. Instead, it is intended for interactive use with a REPL/notebook such as Mathematica/Matlab/iPython Notebook and for incorporation into your custom programs as a citable, Git-versioned library. Nounou focuses only on (1) loading data in a coherent way, and (2) performing very basic but calculation-intensive analyses.
Many of Nounou's routines are based on the numerical processing library breeze.
Why Scala?
- JVM Benefits
- Since Scala is JVM-based, it is easily accessed from MATLAB/Mathematica for further advanced analyses and graphing.
- The JVM ensures unmatched cross-platform compatiblity.
- Takes advantage of the latest processor-specific optimizations in the Oracle JVM. This can make tightly-written Java-based programs faster than non-optimized generic C code in certain circumstances.
- Scala Benefits
- Fast compiled routines can be used even from a REPL.
- Scala makes parallel/distributed programming simple and relatively safe
- The object-oriented nature of Scala is key to keep track of complex data structures
- The functional programming capabilities of Scala make parallelization and streaming safer/much easier
- Scala is both theoretically well concieved and also practical. It is cool.
- Why not Python: speed issues with uncompiled Python and there is already a relatively large Python project available (neo) (A Python bridge should be relatively simple to make at some point)
see wiki
The license is as stated.
Academically, there is no citable publication as of yet. Please cite the following, which uses a prior version of the code:
[1] Takagaki, et. al. (2011) Flow detection of propagating waves with temporospatial correlation of activity. J Neurosci Meth. 200(2):207-218
Please contribute if you share the above goals! Significant contributions may also contribute to a potential future publication (~2016 timeframe)