v0.14.rst

Version 0.14 (2017-03-24)

Changelog

• Add example of time-frequency decoding with CSP by `Laura Gwilliams`_
• Automatically create a legend in :func:`mne.viz.plot_evoked_topo` by `Jussi Nurminen`_
• Add I/O support for Artemis123 infant/toddler MEG data by `Luke Bloy`_
• Add filter plotting functions :func:`mne.viz.plot_filter` and :func:`mne.viz.plot_ideal_filter` as well as filter creation function :func:`mne.filter.create_filter` by `Eric Larson`_
• Add HCP-MMP1.0 parcellation dataset downloader by `Eric Larson`_
• Add option to project EEG electrodes onto the scalp in mne.viz.plot_trans by `Eric Larson`_
• Add option to plot individual sensors in :meth:`mne.io.Raw.plot_psd` by `Alex Gramfort`_ and `Eric Larson`_
• Add option to plot ECoG electrodes in mne.viz.plot_trans by `Eric Larson`_
• Add convenient default values to :meth:`mne.io.Raw.apply_hilbert` and :meth:`mne.io.Raw.apply_function` by `Denis Engemann`_
• Remove MNE-C requirement for :ref:`mne make_scalp_surfaces` by `Eric Larson`_
• Add support for FastTrack Polhemus .mat file outputs in hsp argument of mne.channels.read_dig_montage by `Eric Larson`_
• Add option to convert 3d electrode plots to a snapshot with 2d electrode positions with :func:`mne.viz.snapshot_brain_montage` by `Chris Holdgraf`_
• Add skull surface plotting option to mne.viz.plot_trans by `Jaakko Leppakangas`_
• Add minimum-phase filtering option in :meth:`mne.io.Raw.filter` by `Eric Larson`_
• Add support for reading ASCII BrainVision files in :func:`mne.io.read_raw_brainvision` by `Eric Larson`_
• Add method of ICA objects for retrieving the component maps :meth:`mne.preprocessing.ICA.get_components` by `Jona Sassenhagen`_
• Add option to plot events in :func:`mne.viz.plot_epochs` by `Jaakko Leppakangas`_
• Add dipole definitions for older phantom at Otaniemi in :func:`mne.dipole.get_phantom_dipoles` by `Eric Larson`_
• Add spatial colors option for :func:`mne.viz.plot_raw_psd` by `Jaakko Leppakangas`_
• Add functions like :func:`get_volume_labels_from_src` to handle mixed source spaces by `Annalisa Pascarella`_
• Add convenience function for opening MNE documentation :func:`open_docs` by `Eric Larson`_
• Add option in :meth:`mne.io.Raw.plot` to display the time axis relative to raw.first_samp by `Mainak Jas`_
• Add new :mod:`mne.datasets.visual_92_categories <mne.datasets.visual_92_categories.data_path>` dataset by `Jaakko Leppakangas`_
• Add option in :func:`mne.io.read_raw_edf` to allow channel exclusion by `Jaakko Leppakangas`_
• Allow integer event codes in :func:`mne.read_epochs_eeglab` by `Jaakko Leppakangas`_
• Add ability to match channel names in a case insensitive manner when applying a mne.channels.Montage by `Marijn van Vliet`_
• Add yscale keyword argument to :meth:`mne.time_frequency.AverageTFR.plot` that allows specifying whether to present the frequency axis in linear ('linear') or log ('log') scale. The default value is 'auto' which detects whether frequencies are log-spaced and sets yscale to log. Added by `Mikołaj Magnuski`_
• Add :ref:`Representational Similarity Analysis (RSA) <ex-rsa-noplot>` example on :mod:`mne.datasets.visual_92_categories.data_path` dataset by `Jaakko Leppakangas`_, `Jean-Rémi King`_ and `Alex Gramfort`_
• Add support for NeuroScan files with event type 3 in :func:`mne.io.read_raw_cnt` by `Marijn van Vliet`_
• Add interactive annotation mode to :meth:`mne.io.Raw.plot` (accessed by pressing 'a') by `Jaakko Leppakangas`_
• Add support for deleting all projectors or a list of indices in :meth:`mne.io.Raw.del_proj` by `Eric Larson`_
• Add source space plotting with :meth:`mne.SourceSpaces.plot` using mne.viz.plot_trans by `Eric Larson`_
• Add :func:`mne.decoding.get_coef` to retrieve and inverse the coefficients of a linear model - typically a spatial filter or pattern, by `Jean-Rémi King`_
• Add support for reading in EGI MFF digitization coordinate files in mne.channels.read_dig_montage by `Matt Boggess`_
• Add n_per_seg keyword argument to mne.time_frequency.psd_welch and :func:`mne.time_frequency.psd_array_welch` that allows to control segment length independently of n_fft and use zero-padding when n_fft > n_per_seg by `Mikołaj Magnuski`_
• Add annotation aware data getter :meth:`mne.io.Raw.get_data` by `Jaakko Leppakangas`_
• Add support of dipole location visualization with MRI slice overlay with matplotlib to :func:`mne.viz.plot_dipole_locations` via mode='orthoview' parameter by `Jaakko Leppakangas`_ and `Alex Gramfort`_
• Add plotting of head positions as a function of time in :func:`mne.viz.plot_head_positions` by `Eric Larson`_
• Add real_filter option to mne.beamformer.dics, mne.beamformer.dics_source_power, mne.beamformer.tf_dics and mne.beamformer.dics_epochs by `Eric Larson`_, `Alex Gramfort`_ and `Andrea Brovelli`_.
• Add a demo script showing how to use a custom inverse solver with MNE by `Alex Gramfort`_
• Functions :func:`mne.preprocessing.create_ecg_epochs`, :func:`mne.preprocessing.create_eog_epochs`, :func:`mne.compute_raw_covariance` and ICA methods :meth:`mne.preprocessing.ICA.score_sources`, :meth:`mne.preprocessing.ICA.find_bads_ecg`, :meth:`mne.preprocessing.ICA.find_bads_eog` are now annotation aware by `Jaakko Leppakangas`_
• Allow using spatial_colors for non-standard layouts by creating custom layouts from channel locations and add to_sphere keyword to :func:`mne.viz.plot_sensors` to allow plotting sensors that are not on the head surface by `Jaakko Leppakangas`_
• Concatenating raws with :func:`mne.concatenate_raws` now creates boundary annotations automatically by `Jaakko Leppakangas`_
• :func:`mne.viz.plot_projs_topomap` now supports plotting EEG topomaps by passing in :class:`mne.Info` by `Eric Larson`_

BUG

• Fix bug with DICS and LCMV (e.g., mne.beamformer.lcmv, mne.beamformer.dics) where regularization was done improperly. The default reg=0.01 has been changed to reg=0.05, by `Andrea Brovelli`_, `Alex Gramfort`_, and `Eric Larson`_
• Fix callback function call in mne.viz.topo._plot_topo_onpick by `Erkka Heinila`_
• Fix reading multi-file CTF recordings in :func:`mne.io.read_raw_ctf` by `Niklas Wilming`_
• Fix computation of AR coefficients across channels in :func:`mne.time_frequency.fit_iir_model_raw` by `Eric Larson`_
• Fix maxfilter channel names extra space bug in :func:`mne.preprocessing.maxwell_filter` by `Sheraz Khan`_
• :func:`mne.channels.find_layout` now leaves out the excluded channels by `Jaakko Leppakangas`_
• Array data constructors :class:`mne.io.RawArray` and :class:`EvokedArray` now make a copy of the info structure by `Jaakko Leppakangas`_
• Fix bug with finding layouts in :func:`mne.viz.plot_projs_topomap` by `Eric Larson`_
• Fix bug :func:`mne.io.anonymize_info` when Info does not contain 'file_id' or 'meas_id' fields by `Jean-Rémi King`_
• Fix colormap selection in :func:`mne.viz.plot_evoked_topomap` when using positive vmin with negative data by `Jaakko Leppakangas`_
• Fix channel name comparison in mne.channels.read_montage so that if ch_names is provided, the returned montage will have channel names in the same letter case by `Jaakko Leppakangas`_
• Fix :meth:`inst.set_montage(montage) <mne.io.Raw.set_montage>` to only set inst.info['dev_head_t'] if dev_head_t=True in mne.channels.read_dig_montage by `Eric Larson`_
• Fix handling of events in mne.realtime.RtEpochs when the triggers were split between two buffers resulting in missing and/or duplicate epochs by `Mainak Jas`_ and `Antti Rantala`_
• Fix bug with automatic decimation in :func:`mne.io.read_raw_kit` by `Keith Doelling`_
• Fix bug with :func:`setup_volume_source_space` where arguments subject and subjects_dir were ignored by `Jaakko Leppakangas`_
• Fix sanity check for incompatible threshold and tail values in clustering functions like :func:`mne.stats.spatio_temporal_cluster_1samp_test` by `Eric Larson`_
• Fix _bad_dropped not being set when loading eeglab epoched files via :func:`mne.read_epochs_eeglab` which resulted in :func:`len` not working by `Mikołaj Magnuski`_
• Fix a bug in :meth:`mne.time_frequency.AverageTFR.plot` when plotting without a colorbar by `Jaakko Leppakangas`_
• Fix _filenames attribute in creation of :class:`mne.io.RawArray` with :meth:`mne.preprocessing.ICA.get_sources` by `Paul Pasler`_
• Fix contour levels in :func:`mne.viz.plot_evoked_topomap` to be uniform across topomaps by `Jaakko Leppakangas`_
• Fix bug in :func:`mne.preprocessing.maxwell_filter` where fine calibration indices were mismatched leading to an AssertionError by `Eric Larson`_
• Fix bug in :func:`mne.preprocessing.fix_stim_artifact` where non-data channels were interpolated by `Eric Larson`_
• :class:`mne.decoding.Scaler` now scales each channel independently using data from all time points (epochs and times) instead of scaling all channels for each time point. It also now accepts parameter scalings to determine the data scaling method (default is None to use static channel-type-based scaling), by `Asish Panda`_, `Jean-Rémi King`_, and `Eric Larson`_
• Raise error if the cv parameter of mne.decoding.GeneralizationAcrossTime and mne.decoding.TimeDecoding is not a partition and the predict_mode is "cross-validation" by `Jean-Rémi King`_
• Fix bug in :func:`mne.io.read_raw_edf` when preload=False and channels have different sampling rates by `Jaakko Leppakangas`_
• Fix :func:`mne.read_labels_from_annot` to set label.values[:]=1 rather than 0 for consistency with the :class:`Label` class by `Jon Houck`_
• Fix plotting non-uniform freqs (for example log-spaced) in :meth:`mne.time_frequency.AverageTFR.plot` by `Mikołaj Magnuski`_
• Fix :func:`mne.minimum_norm.compute_source_psd` when used with pick_ori=None by `Annalisa Pascarella`_ and `Alex Gramfort`_
• Fix bug in :class:`mne.Annotations` where concatenating two raws where orig_time of the second run is None by `Jaakko Leppakangas`_
• Fix reading channel location from eeglab .set files when some of the channels do not provide this information. Previously all channel locations were ignored in such case, now they are read - unless a montage is provided by the user in which case only channel names are read from set file. By `Mikołaj Magnuski`_
• Fix reading eeglab .set files when .chanlocs structure does not contain X, Y or Z fields by `Mikołaj Magnuski`_
• Fix bug with :func:`mne.simulation.simulate_raw` when interp != 'zero' by `Eric Larson`_
• Fix :func:`mne.fit_dipole` to handle sphere model rank deficiency properly by `Alex Gramfort`_
• Raise error in :func:`mne.concatenate_epochs` when concatenated epochs have conflicting event_id by `Mikołaj Magnuski`_
• Fix handling of n_components=None in :class:`mne.preprocessing.ICA` by `Richard Höchenberger`_
• Fix reading of fiducials correctly from CTF data in :func:`mne.io.read_raw_ctf` by `Jaakko Leppakangas`_
• Fix :func:`mne.beamformer.rap_music` to return dipoles with amplitudes in Am instead of nAm by `Jaakko Leppakangas`_
• Fix computation of duality gap in mne.inverse_sparse.mxne_optim.dgap_l21 by `Mathurin Massias`_

API

• The filtering functions band_pass_filter, band_stop_filter, low_pass_filter, and high_pass_filter have been deprecated in favor of :func:`mne.filter.filter_data` by `Eric Larson`_
• :class:`EvokedArray` now has default value tmin=0. by `Jaakko Leppakangas`_
• The ch_type argument for mne.viz.plot_trans has been deprecated, use eeg_sensors and meg_sensors instead, by `Eric Larson`_
• The default tmax=60. in :meth:`mne.io.Raw.plot_psd` will change to tmax=np.inf in 0.15, by `Eric Larson`_
• Base classes :class:`mne.io.BaseRaw` and :class:`mne.BaseEpochs` are now public to allow easier typechecking, by `Daniel McCloy`_
• :func:`mne.io.read_raw_edf` now combines triggers from multiple tal channels to 'STI 014' by `Jaakko Leppakangas`_
• The measurement info :class:`Info` no longer contains a potentially misleading info['filename'] entry. Use class properties like :attr:`mne.io.Raw.filenames` or :attr:`mne.Epochs.filename` instead by `Eric Larson`_
• Default fiducial name change from 'nz' to 'nasion' in mne.channels.read_montage, so that it is the same for both mne.channels.Montage and :class: mne.channels.DigMontage by `Leonardo Barbosa`_
• MNE's additional files for the fsaverage head/brain model are now included in MNE-Python, and the now superfluous mne_root parameter to :func:`create_default_subject` has been deprecated by `Christian Brodbeck`_
• An overwrite=False default parameter has been added to :func:`write_source_spaces` to protect against accidental overwrites, by `Eric Larson`_
• The :class:`mne.decoding.LinearModel` class will no longer support plot_filters and plot_patterns, use :class:`mne.EvokedArray` with :func:`mne.decoding.get_coef` instead, by `Jean-Rémi King`_
• Made functions :func:`mne.time_frequency.tfr_array_multitaper`, :func:`mne.time_frequency.tfr_array_morlet`, :func:`mne.time_frequency.tfr_array_stockwell`, :func:`mne.time_frequency.psd_array_multitaper` and :func:`mne.time_frequency.psd_array_welch` public to allow computing TFRs and PSDs on numpy arrays by `Jaakko Leppakangas`_
• :meth:`mne.preprocessing.ICA.fit` now rejects data annotated bad by default. Turn off with reject_by_annotation=False, by `Jaakko Leppakangas`_
• :func:`mne.io.read_raw_egi` now names channels with pattern 'E<idx>'. This behavior can be changed with parameter channel_naming by `Jaakko Leppakangas`_
• the name parameter in :class:`mne.Epochs` is deprecated, by `Jaakko Leppakangas`_

Authors

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

• Alexander Rudiuk
• Alexandre Gramfort
• Annalisa Pascarella
• Antti Rantala
• Asish Panda
• Burkhard Maess
• Chris Holdgraf
• Christian Brodbeck
• Cristóbal Moënne-Loccoz
• Daniel McCloy
• Denis A. Engemann
• Eric Larson
• Erkka Heinila
• Hermann Sonntag
• Jaakko Leppakangas
• Jakub Kaczmarzyk
• Jean-Rémi King
• Jon Houck
• Jona Sassenhagen
• Jussi Nurminen
• Keith Doelling
• Leonardo S. Barbosa
• Lorenz Esch
• Lorenzo Alfine
• Luke Bloy
• Mainak Jas
• Marijn van Vliet
• Matt Boggess
• Matteo Visconti
• Mikolaj Magnuski
• Niklas Wilming
• Paul Pasler
• Richard Höchenberger
• Sheraz Khan
• Stefan Repplinger
• Teon Brooks
• Yaroslav Halchenko