rev007 2015-06-17 task002 stimulus auditory features
  - Add extracted auditory features for task002 stimuli (incl. source code).
    Thanks to Michael Casey for this contribution.

rev006 2015-06-15 task002 bug fix
  - Strip all samples prior to the first trigger from task002 physio data to
    homogenize with task001.
  - Physio data for task002, sub005 runs 3 and up were mis-numbered and run008 
    was missing, due to a conversion error that did not take the aborted run
    into account. All runs are available and correctly numbered now.
  - Remove invalid log files for behavioral data of sub020, task002, runs005-8.
    Subject aborted after run004.
  - Add acquisition protocol for task002.
  - Add task002 experiment implementation and stimulus files.
  - Homogenize task002 stimulus file names with release experiment implementation.

rev005 2015-04-06 task002 data added; emotion annotation, fixes

  - New 7T fMRI data with music stimulation (task002).
  - New group and per-subject EPI template images based on all available 7T
    data.
  - Re-aligned task001 BOLD data using the new templates.
  - Updated, now frame-accurate annotation of movie scene boundaries.
  - Frame accurate camera shot annotation for the audio-visual movie (task003).
  - Multi-observer emotion annotation for the audio-visual and audio-movie
    stimulus (see Reich et al. F1000Research 2015).

rev004 2014-09-29 NIfTI header fixes

  - Task001 BOLD images reported a wrong volume repetition time
    of ~24s in the NIfTI header field pixdim[4] (MRIConvert issue).
    This release updates the images to the correct value of 2.0s.

rev003 2014-06-16 Bug fixes, aggregate data

  - Individual brain mask images from non-linear alignment had
    only the linear transformation component applied to them.
    Consequently the masks did not match the brain outline.
    This release contains the fixed brain mask images. No other
    data was modified.
  - Include aggregate data (mean, median, min, max, std) for
    more than 200 regions-of-interests taken from the Harvard-Oxford
    cortical and sub-cortical atlas, as well as the Juelich brain atlas
    shipped with FSL. See the scripts in 'scripts/parcellation_ts/' in
    gumpdata source code repository for details. Data is available
    in CSV and HDF5 format.

rev002 2014-02-20 Physiological data fixes and additions

  - physiological data for all participants in original
    sampling rate (physio_pristine.txt.gz) was added
  - physiological data for sub008 run005 was updated to
    strip leading data samples prior to the fixes MR
    trigger signal. Thanks to Christine Guo for the report.
  - trigger log for first MR trigger (only) was offset by
    one data sample (5-10ms)

rev001 2014-01-22 Initial release

