[Erppcatoolkit-support] EP Toolkit 2.64 released

[Joseph D. <jd...@me...>]

EP Toolkit version 2.64 has been released:
https://sourceforge.net/projects/erppcatoolkit/ <https://sourceforge.net/projects/erppcatoolkit/>

The most serious bugfix was for the PCA cross-validation option (which I’m renaming "cross-verification" to avoid confusion with a common method for determining number of factors to retain).  It was not producing correct results.
In general, there has been continued improvements to the artifact correction routine, including addition of De Vos's BSS-CCA EMG correction algorithm.  Support for BrainVision files has also been improved, including an option for encoding subject and trial info in the trigger channel.

1) Fixed crash that could occur in the artifact correction function when there are bad time points that exceed the saturation threshold and the dataset is single-trial.  
2) Artifact correction median corrects data prior to saturation check to ensure channels with merely high offsets are not treated as bad data.  
3) Fixed autotemplate missing blinks when sampling rate not 250Hz.  
4) Loosened blink autotemplate slope criteria to better recognize slow blinks.  
5) Artifact correction interpolates bad EOG channels instead of aborting when too many are bad.  
6) Added EMG correction option to Preprocessing function using Maartens De Vos's BSS-CCA code by his kind permission.  
7) Saccade Potential correction works with sampling rates other than 250 Hz.  
8) Fixed crash when trying to save artifact correction summary figure due to Matlab changing their graphics objects AGAIN in 2017b.  
9) Segment function user interface now providing correct default sample range for sampling rates other than 250 Hz.  
10) Segment function user interface now provides only unique set of values for integer trials specs.  
11) Fixed crash when averaging files that have different trial specs.  
12) For artifact correction in general, detrend and baseline correct are now applied per trial for single-trial and average data rather than applied as if it were a continuous dataset (blink correction routine already treated baseline correction this way).  
13) In preprocessing function, eliminated x tick labels to address problem with subplots in summary artifact figure getting squeezed by formatting problem on PCs.  
14) Adding rank check prior to ICA in blink correction.  
15) When reading a file, if none of the channel names match but there are the same number of non-ref EEG channels, will not only use the CED channel names, will assume the ref channels are implicit and add them.  
16) Eliminated restrictions on location of CED files when reading a data file.  
17) Added support for impedance info in BrainVision files.  
18) Added -precedes- crit to the Segment function.  
19) In Segment function, fixed crash when changing value of a -precedes- or -follows- crit to an event name that has no keys.  
20) In Segment function, made fixes to -follows- function for relations other than = and ~= 
21) In Segment function, added support for TS-EPoffset field, which adjusts the epoch time points according to the contents of the field.  
22) Fixed crash when conducting PCA on data where there are trial spec fields that are identical across all the trials/cells and are characters.  
23) Fixed PCA cross-validation option giving wrong results!  
24) Added support for reading EP header information encoded in BrainVision's trigger channel.  
25) Fixed not recognizing ECG channels in mff files.

Joe


--------------------------------------------------------------------------------

Joseph Dien, PhD
Senior Research Scientist
Human Development and Quantitative Methodology Department
University of Maryland, College Park
http://joedien.com <http://joedien.com/>