BdPy

Python library for brain decoding

Requirements

Python 2 (mainly developed with Python 2.7.12)
- Python 3 support is upcoming.
numpy
scipy
scikit-learn
h5py

Optional requirements

'dataform' module
- pandas
'mri' module
- nipy

Installation

Run the following command:

$ pip install git+https://github.com/KamitaniLab/bdpy.git

$ git clone git@github.com:KamitaniLab/bdpy.git
$ cd bdpy/
$ python setup.py install

Packages

bdata: BdPy data format
dataform: Third-party data format (e.g., Pandas)
distcomp: Distributed computation
fig: Making figures
ml: Machine learning
mri: MRI
preproc: Preprocessing
stats: Statistics
util: Miscellaneous utilities

BdPy data format

BdPy data format (or BrainDecoderToolbox2 data format; bdata) consists of two variables: dataset and metadata. 'dataset' stores brain activity data (e.g., voxel signal value for fMRI data), targets (e.g., ID of stimuli for vision experiments), and additional information specifying experimental design (e.g., run and block numbers for fMRI experiments). Each row corresponds to a single 'sample', and each column representes either single feature (voxel), target, or experiment design information. 'metadata' contains data describing meta-information for each column in dataset.

Data API usage

Import module and initialization.

from bdpy import BData

# Create a BData instance
dat = BData()

An instance of the class 'BData' contains dataset and metadata as instance variables. dataset is a M x N numpy array (M: the number of samples, N: the number of all features (voxels), labels, design information, etc).

Load data

# Load BData from 'data_file.h5'
dat.load('data_file.h5')

Show data

# Show 'key' and 'description' of the betaData
dat.show_meatadata()

# Get 'value' of the metadata specified by the 'key'
voxel_x = dat.get_metadata('voxel_x', where='VoxelData')

Data extraction

# Get an array of voxel data in V1 (shape = (M, [no. of voxels in V1]))
v1_data = dat.select('ROI_V1')

# Another way to select data
v1_data = dat.select('ROI_V1 = 1')

# Select data in several ROIs
data_v1v2 = dat.select('ROI_V1 = 1 | ROI_V2 = 1')

# Get labels ('Label_A') in the dataset
label_a  = dat.select('Label_A')

Data creation

# Add new data
dat.add(numpy.random.rand(dat.dataset.shape[0]), 'random_data')

# Set description of metadata
dat.set_metadatadescription('random_data', 'Random data (just for test)')

# Save data
dat.save('output_file.h5')  # File format is selected automatically by extension. .mat, .h5,and .npy are supported.

For developers

Please send your pull requests to dev branch, not to master.

Contributors

Shuntaro C. Aoki (DNI, ATR)
Misato Tanaka (DNI, ATR)

miyosuda/bdpy