Version 0.16 (2018-05-01)

Changelog

• Add possibility to pass dict of floats as argument to mne.make_ad_hoc_cov() by Nathalie Gayraud

• Add support for metadata in mne.Epochs by Chris Holdgraf, Alex Gramfort, Jona Sassenhagen, and Eric Larson

• Add support for plotting a dense head in mne.viz.plot_alignment() by Eric Larson

• Allow plotting in user-created mayavi Scene in mne.viz.plot_alignment() by Daniel McCloy

• Reduce memory consumption and do not require data to be loaded in mne.Epochs.apply_baseline() by Eric Larson

• Add option render_bem to mne.Report.parse_folder() by Eric Larson

• Add to mne.viz.plot_alignment() plotting of coordinate frame axes via show_axes and terrain-style interaction via interaction, by Eric Larson

• Add option initial_event to mne.find_events() by Clemens Brunner

• Left and right arrow keys now scroll by 25% of the visible data, whereas Shift+left/right scroll by a whole page in mne.io.Raw.plot() by Clemens Brunner

• Add support for gantry tilt angle determination from Elekta FIF file header by Chris Bailey

• Add possibility to concatenate mne.Annotations objects with + or += operators by Clemens Brunner

• Add support for MaxShield raw files in mne.Report by Eric Larson

• Add ability to plot whitened data in mne.io.Raw.plot(), mne.Epochs.plot(), mne.Evoked.plot(), and mne.Evoked.plot_topo() by Eric Larson

• Workaround for reading EGI MFF files with physiological signals that also present a bug from the EGI system in mne.io.read_raw_egi() by Federico Raimondo

• Add support for reading subject height and weight in info['subject_info'] by Eric Larson

• Improve online filtering of raw data when plotting with mne.io.Raw.plot() to filter in segments in accordance with the default skip_by_annotation=('edge', 'bad_acq_skip') of mne.io.Raw.filter() to avoid edge ringing by Eric Larson

• Add support for multiple head position files, plotting of sensors, and control of plotting color and axes in mne.viz.plot_head_positions() by Eric Larson

• Add ability to read and write Annotations separate from mne.io.Raw instances via Annotations.save() and read_annotations() by Eric Larson

• Add option to unset a montage by passing None to mne.io.Raw.set_montage() by Clemens Brunner

• Add sensor denoising via mne.preprocessing.oversampled_temporal_projection() by Eric Larson

• Add mne.io.pick.get_channel_types which returns all available channel types in MNE by Clemens Brunner

• Use standard PCA instead of randomized PCA whitening prior to ICA to increase reproducibility by Clemens Brunner

• Plot sEEG electrodes in mne.viz.plot_alignment() by Alex Gramfort

• Add support for any data type like sEEG or ECoG in covariance related functions (estimation, whitening and plotting) by Alex Gramfort and Eric Larson

• Add function mne.io.read_annotations_eeglab to allow loading annotations from EEGLAB files, by Alex Gramfort

• mne.io.Raw.set_montage() now accepts a string as its montage argument; this will set a builtin montage, by Clemens Brunner

• Add 4D BTi phantom dataset mne.datasets.phantom_4dbti.data_path(), by Alex Gramfort

• Changed the background color to grey in mne.viz.plot_alignment() to make helmet more visible, by Alex Gramfort

• Add mne.io.Raw.reorder_channels(), mne.Evoked.reorder_channels(), etc. to reorder channels, by Eric Larson

• Add to mne coreg and mne.gui.coregistration() by Eric Larson:

  • Improved visibility of points inside the head

  • Projection of EEG electrodes

  • Orientation of extra points toward the surface

  • Scaling points by distance to the head surface

  • Display of HPI points

  • ICP fitting with convergence criteria

  • Faster display updates

  • Scaling of mri/*.mgz files

  • Scaling of mri/trainsforms/talairach.xfm files for conversion to MNI space

• Add ability to exclude components interactively by clicking on their labels in mne.preprocessing.ICA.plot_components() by Mikołaj Magnuski

• Add reader for manual annotations of raw data produced by Brainstorm by Anne-Sophie Dubarry

• Add eLORETA noise normalization for minimum-norm solvers by Eric Larson

• Tighter duality gap computation in mne.inverse_sparse.tf_mxne_optim and new parametrization with alpha and l1_ratio instead of alpha_space and alpha_time by Mathurin Massias and Daniel Strohmeier

• Add dgap_freq parameter in mne.inverse_sparse.mxne_optim solvers to control the frequency of duality gap computation by Mathurin Massias and Daniel Strohmeier

• Add support for reading Eximia files by Eric Larson and Federico Raimondo

• Add the Picard algorithm to perform ICA for mne.preprocessing.ICA, by Pierre Ablin and Alex Gramfort

• Add ability to supply a mask to the plot in mne.viz.plot_evoked_image() by Jona Sassenhagen

• Add connectivity=False to cluster-based statistical functions to perform non-clustering stats by Eric Larson

• Add mne.time_frequency.csd_morlet() and mne.time_frequency.csd_array_morlet() to estimate cross-spectral density using Morlet wavelets, by Marijn van Vliet

• Add multidictionary time-frequency support to mne.inverse_sparse.tf_mixed_norm() by Mathurin Massias and Daniel Strohmeier

• Add new DICS implementation as mne.beamformer.make_dics(), mne.beamformer.apply_dics(), mne.beamformer.apply_dics_csd() and mne.beamformer.apply_dics_epochs(), by Marijn van Vliet and Susanna Aro

Bug

• Fix bug in EEG interpolation code to do nothing if there is no channel to interpolate by Mainak Jas

• Fix bug in mne.preprocessing.peak_finder to output datatype consistently and added input check for empty vectors by Tommy Clausner

• Fix bug in mne.io.read_raw_brainvision() to use the correct conversion for filters from time constant to frequency by Stefan Appelhoff

• Fix bug with events when saving split files using mne.Epochs.save() by Eric Larson

• Fix bug in mne.decoding.SlidingEstimator and mne.decoding.GeneralizingEstimator to allow mne.decoding.cross_val_multiscore() to automatically detect whether the base_estimator is a classifier and use a StratifiedKFold instead of a KFold when cv is not specified, by Jean-Rémi King

• Fix bug in mne.set_eeg_reference() to remove an average reference projector when setting the reference to [] (i.e. do not change the existing reference) by Clemens Brunner

• Fix bug in threshold-free cluster enhancement parameter validation (mne.stats.permutation_cluster_1samp_test() and mne.stats.permutation_cluster_test()) by Clemens Brunner

• Fix bug in mne.io.Raw.plot() to correctly display event types when annotations are present by Clemens Brunner

• Fix bug in mne.stats.spatio_temporal_cluster_test() default value for threshold is now calculated based on the array sizes in X, by Eric Larson

• Fix bug in mne.simulation.simulate_raw() with use_cps=True where CPS was not actually used by Eric Larson

• Fix bug in mne.simulation.simulate_raw() where 1- and 3-layer BEMs were not properly transformed using trans by Eric Larson

• Fix bug in mne.viz.plot_alignment() where the head surface file -head.fif was not used even though present by Chris Bailey

• Fix bug when writing compressed sparse column matrices (e.g., Maxwell filtering cross-talk matrices) by Marijn van Vliet and Eric Larson

• Fix bug in mne.io.Raw.plot_psd() to correctly deal with reject_by_annotation=False by Clemens Brunner

• Fix bug in mne.make_fixed_length_events() when hitting corner case problems rounding to sample numbers by Eric Larson

• Fix bug in mne.Epochs when passing events as list with event_id=None by Alex Gramfort

• Fix bug in mne.Report.add_figs_to_section when passing numpy.ndarray by Eric Larson

• Fix bug in CSS class setting in mne.Report BEM section by Eric Larson

• Fix bug in Annotations where annotations that extend to the end of a recording were not extended properly by Eric Larson

• Fix bug in mne.io.Raw.filter() to properly raw data with acquisition skips in separate segments by Eric Larson

• Fix bug in mne.preprocessing.maxwell_filter() where homogeneous fields were not removed for CTF systems by Eric Larson

• Fix computation of average quaternions in mne.preprocessing.maxwell_filter() by Eric Larson

• Fix bug in writing raw.annotations where empty annotations could not be written to disk, by Eric Larson

• Fix support for writing FIF files with acquisition skips by using empty buffers rather than writing zeros by Eric Larson

• Fix bug in the mne make_scalp_surfaces command where --force (to bypass topology check failures) was ignored by Eric Larson

• Fix bug in mne.preprocessing.maxwell_filter() when providing origin in 'meg' coordinate frame for recordings with a MEG to head transform (i.e., non empty-room recordings) by Eric Larson

• Fix bug in mne.viz.plot_cov() that ignored colorbar argument by Nathalie Gayraud

• Fix bug when picking CTF channels that could cause data saved to disk to be unreadable by Eric Larson

• Fix bug when reading event latencies (in samples) from eeglab files didn’t translate indices to 0-based python indexing by Mikołaj Magnuski

• Fix consistency between mne.Epochs and mne.stats.linear_regression_raw() in converting between samples and times (mne.stats.linear_regression_raw() now rounds, instead of truncating) by Phillip Alday

• Fix bug in mne coreg where sphere surfaces were scaled by Eric Larson

• Fix bug in mne.Evoked.plot_topomap() when using proj='interactive' mode by Eric Larson

• Fix bug when passing show_sensors=1 to mne.viz.plot_compare_evokeds() resulted in sensors legend placed in lower right of the figure (position 4 in matplotlib), not upper right by Mikołaj Magnuski

• Fix handling of annotations when cropping and concatenating raw data by Alex Gramfort and Eric Larson

• Fix bug in mne.preprocessing.create_ecg_epochs() where keep_ecg=False was ignored by Eric Larson

• Fix bug in mne.io.Raw.plot_psd() when picks is not None and picks spans more than one channel type by Eric Larson

• Fix bug in mne.make_forward_solution when passing data with compensation channels (e.g. CTF) that contain bad channels by Alex Gramfort

• Fix bug in mne.SourceEstimate.get_peak() and mne.VolSourceEstimate.get_peak() when there is only a single time point by Marijn van Vliet

• Fix bug in mne.io.read_raw_edf() when reading BDF files stimulus channels are now not scaled anymore by Clemens Brunner

API

• Channels with unknown locations are now assigned position [np.nan, np.nan, np.nan] instead of [0., 0., 0.], by Eric Larson

• Removed unused image_mask argument from mne.viz.plot_topomap() by Eric Larson

• Unknown measurement dates are now stored as info['meas_date'] = None rather than using the current date. None is also now used when anonymizing data and when determining the machine ID for writing files, by Mainak Jas and Eric Larson

• mne.Evoked.plot() will now append the number of epochs averaged for the evoked data in the first plot title, by Eric Larson

• Changed the line width in mne.viz.plot_bem() from 2.0 to 1.0 for better visibility of underlying structures, by Eric Larson

• Changed the behavior of mne.io.Raw.pick_channels() and similar methods to be consistent with mne.pick_channels() to treat channel list as a set (ignoring order) – if reordering is necessary use inst.reorder_channels, by Eric Larson

• Changed the labeling of some plotting functions to use more standard capitalization and units, e.g. “Time (s)” instead of “time [sec]” by Eric Larson

• mne.time_frequency.csd_epochs has been refactored into mne.time_frequency.csd_fourier() and mne.time_frequency.csd_multitaper(), by Marijn van Vliet

• mne.time_frequency.csd_array has been refactored into mne.time_frequency.csd_array_fourier() and mne.time_frequency.csd_array_multitaper(), by Marijn van Vliet

• Added clean_names=False parameter to mne.io.read_raw_ctf() for control over cleaning of main channel names and compensation channel names from CTF suffixes by Oleh Kozynets

• The functions lcmv, lcmv_epochs, and lcmv_raw are now deprecated in favor of mne.beamformer.make_lcmv() and mne.beamformer.apply_lcmv(), mne.beamformer.apply_lcmv_epochs(), and mne.beamformer.apply_lcmv_raw(), by Britta Westner

• The functions mne.beamformer.dics, mne.beamformer.dics_epochs and mne.beamformer.dics_source_power are now deprecated in favor of mne.beamformer.make_dics(), mne.beamformer.apply_dics(), and mne.beamformer.apply_dics_csd(), by Marijn van Vliet

Authors

People who contributed to this release (in alphabetical order):

• Alejandro Weinstein

• Alexandre Gramfort

• Annalisa Pascarella

• Anne-Sophie Dubarry

• Britta Westner

• Chris Bailey

• Chris Holdgraf

• Christian Brodbeck

• Claire Braboszcz

• Clemens Brunner

• Daniel McCloy

• Denis A. Engemann

• Desislava Petkova

• Dominik Krzemiński

• Eric Larson

• Erik Hornberger

• Fede Raimondo

• Henrich Kolkhorst

• Jean-Rémi King

• Jen Evans

• Joan Massich

• Jon Houck

• Jona Sassenhagen

• Juergen Dammers

• Jussi Nurminen

• Kambiz Tavabi

• Katrin Leinweber

• Kostiantyn Maksymenko

• Larry Eisenman

• Luke Bloy

• Mainak Jas

• Marijn van Vliet

• Mathurin Massias

• Mikolaj Magnuski

• Nathalie Gayraud

• Oleh Kozynets

• Phillip Alday

• Pierre Ablin

• Stefan Appelhoff

• Stefan Repplinger

• Tommy Clausner

• Yaroslav Halchenko