Releases: pennmem/cmlreaders
Releases · pennmem/cmlreaders
Version 0.10.0
v0.10.0 Version 0.10.0
Version 0.9.10
- Now compatible with mne > 0.20.0, which previously caused EEG loading failure
- Fixed a bug that caused automatic baseline correction of scalp data without allowing user control. Reader no longer applies baseline correction. User can always do so with mne.baseline.rescale function (or a user-defined method)
- Baseline correction bug also prevented .load_eeg() from being called on scalp data with a rel_start>=0
Version 0.9.9
- Added checks to prevent using a reader object to load data from other subjects and sessions, preventing an error where data could be silently coerced into incorrect formats
- Added support for loading from subjects where the bipolar montage includes a set of contacts referenced to each other, (e.g. LA1-LA2 and LA2-LA-1)
v0.9.8
Bug fixes:
- EEGContainer now correctly sets the time dimension of EpochsArrays created when using EEGContainer.to_mne(). Previously, the time dimension always started from 0. Now, it will be set based on what rel_start is. For example, if the epochs begin at -200 ms, the time dimension will now correctly start from -0.2.
v0.9.7 - Scalp EEG Support
New features:
- Added ability to load continuous and epoched EEG data from scalp lab (protocol "ltp") sessions using the standard load() and load_eeg() functions. Includes support for EGI's .raw and .mff formats and BioSemi's .bdf format.
- Added keyword argument "clean" to load() and load_eeg(), providing a choice to load either cleaned (.fif) or raw data for ltp sessions. Currently no support for r1 sessions.
v0.9.4
New features:
- Added
stim_params
accessor to unpack the stim_params field
Bug fixes:
- LocationSearch events are 'task_events' by default, similar to PS
- EEGReader now uses the reader's subject, experiment, session when determining full EEG filename whenever possible
v0.9.3
Version 0.9.2
- Ignore events with negative
eegoffset
values when loading EEG data (#192).
v0.9.1
v0.9.0
New features:
- Added initial support for caching some data types (#143)
- Added a new tutorial to the documentation (#151)
Improvements:
- Improved the reading EEG metadata for resumed sessions (#139)
- Taught
CMLReader.load_events
how to handle string arguments in addition to
lists (#150) - Ramulator HDF5 reader now handles missing channels without crashing (#158)
- Updated
EEGReader
to userel_start
as the start time given to
EEGContainer
(#167) - Allowed the use of
contacts
data for thescheme
keyword argument in
CMLReader.load_eeg
(#169) - Made
get_data_index
a static method ofCMLReader
to simplify imports
(#170)
Bug fixes: