bci2kreader's Introduction

BCI2kReader

Python 3 and Python 2 compatible BCI2000 .dat file reader.

Reader for BCI2000 (https://www.neurotechcenter.org/research/bci2000/dissemination) .dat files. This project is a wrapper using the reader developed for the BcPy2000 project (http://bci2000.org/downloads/BCPy2000/BCPy2000.html)

to install the current release package use:

pip install BCI2kReader

This package is still under development, use with caution!

USAGE:

from BCI2kReader import BCI2kReader as b2k

with b2k.BCI2kReader('yourbci2000testfile.dat') as test: #opens a stream to the dat file
     test.samplingrate # sampling rate

# you can use the reader for random access using read(), seek()

    my_signals=test.signals #reads the whole file and stores it in a numpy matrix channels (channels,datapoints)
    my_states=test.states #reads all states as a dictionary ..
    my_signals, my_states=test.readall() #
    my_signals, my_states=test.read(-1) # reads from current position until end
    my_states['Running'] # access to the Running state

# By default calling test.signals, test.states and test.readall() caches all data in the object, this default behaviour can be changed by either calling the constructor with usecache=false or by calling .usecache(False). The cache can be cleared by calling .purge(), which should be called after you set usecache to false to free the memory.

# the reader object also supports direct slicing

    signalslice, stateslice = test[0:100] #returns the first 100 items,
    # this works with cached and non cached mode and does not alter the current position of the file pointer

# Slicing with states

    test.states[:100] # returns all states within the slice frame 
    # test.states['Running'][:,:100] is equivalent to test.states[:100]['Running']
    signals, states=test[test.States['StimulusCode'] == 1] # binary masks work in cached and uncached mode

Recommend Projects