Each file contains the instantaneous power time courses at different recording sites during an ictal event. Recordings comprise the ictal epoch along with 60 seconds of pre- and post-ictal activity shown for comparison. Each seizure is represented in 5 different frequency bands.
The filenames have the following structure:
> "P1sz{**seizure-id**}_pw{**frequency-band**}"
The parameter **seizure-id** ranges from 1 to 9 and the parameter **frequency-band** takes one of the following values:
- **DTh**, delta-theta range (3–8 Hz).
- **AB**, power in the alpha-beta range (8–20 Hz).
- **G**, gamma range (20–70 Hz).
- **HG**, high-gamma range (70–165 Hz).
- **Broad**, broadband (3–165 Hz).
The files can be opened using the Epylib python package (https://github.com/mvilavidal/Epylib):
import epylib as epy
sz=1
bd='Broad'
pw=epy.loadHDF5('power_hdf5/P1sz'+str(sz)+'_pw'+bd+'.hdf5')
The function `epy.loadHDF5` returns a `MetaArray` object. `MetaArray` is a subclass of the numpy `ndarray` class with extended metadata functionality.
The variable `pw` is a 2-D array, wherein each row contains the power time course of one channel in the specified frequency range. Additionally, the metadata is stored as attributes:
pw.channels # number of channels
pw.samples # number of time points
pw.sampling # sampling rate
pw.span # time span of the whole recording in seconds
pw.subject # subject id
pw.seizure # seizure id
pw.filename # filename
pw.channellabels # channel labels
