[Erppcatoolkit-support] ERP PCA Toolkit 2.11 release

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

I just posted version 2.11.  Aside from the usual bug fixes and minor enhancements, I went through the tutorial file and gave it some updating and polishing.  The exercise also turned up a number of bugs which I squashed.

A notable addition is relevant to NetStation users.  When opening EGIS files (only), NetStation is able to make use of montage information that is added to the file.  There are some incompatibilities in the montages between NetStation 4.3 and earlier versions for Hydrocel nets, with 4.3 unable to open montages from earlier versions and vice versa.  Because of this problem, I've added an option under the Files preferences to not add the montage to EGIS files (montage information is not used for other types of files).  In this case, NetStation will present you with a set of options the first time you open the files and add the information itself.

1) Fixed crash due to changes in 2.10 when loading factor file with CED information with Read function.  2) Fixed loss of some factor information (factor types and combined factor waveforms) when loading EP format factor file using Read function.  3) Fixed bug in code for follow-up ANOVAs that could cause crashes or the wrong factors to be used (but still correctly labeled).  4) Changed bad channel field to negative numbers for uncorrected bad channels to distinguish them from corrected bad channels.  5) Added "repChan" to QC information in the Edit Pane so that one can examine the proportion of replaced bad channels separately from the proportion of uncorrected bad channels.  6) Modified averaging function to accommodate uncorrected bad channels.  7) Counts of total blink trials and movement trials for averaged waveforms now computed by averaging function include only good trials.  8) Averaging function now handles possibility of no good trials for an average.  9) When importing data using the Read function, epochs that are entirely flat across all channels are marked as bad.  10) AvgNum and SubNum fields now calculated when computing grand averages.  11) Fixed crash in Read function when there are two REF channels (as in M1-M2 mean mastoid channels).  12) Can now handle displaying waves from multiple datasets where some of the channels are implicit in one but not the other dataset (as in Neuroscan missing channels).  13) BadChans numbers of QC subpane will now use number of subjects to calculate proportion for grand averages.  14) Fixed bug where during preprocessing, neighboring channels for determing whether a channel is not correlating with its neighbors was sometimes not chosen correctly, which could lead to too many channels being dubbed globally bad.  15) No longer treating shorted channels as being bad channels as this was proving too conservative a criterion.  Instead it now just provides a warning message.  16) Preprocessing can now be applied to subject average files with multiple subjects.  17) Analysis fields no longer optional parts of EP file format.  18) In preprocessing, turning off bad channel preferences affects only the relevant bad channel criteria rather than turning off all bad channel detection.  19) Now has option to use file's default reference channels (as indicated by the CED file) in the preprocess pane.  20) Fixed crash when loading in an EP file with ced information.  21) Fixed crash when loading in a .set file that does not have ced information included in its header.  22) Fixed crash when subject adds are being stripped from data with no subject specs (as in Neuroscan files).  23) Fixed crash when adding subject spec via Edit Pane.  24) When importing EP file format data, correctly checks for data type even for "factors" and "grand_average".  25) When running ANOVAs with between subject factors, fixed the addition of grand averages corresponding to the levels of the factors to the dataset.  26) Eliminated chantype field.  27) If there are two reference channels (as in mean mastoids), then no longer require that they have a -1 correlation as one may just be bad. 28) If there are two reference channels (as in mean mastoids), then they are still marked as bad channels if they are flat.  29) When writing out EGIS files, cell names and experiment names are terminated so that they are not padded out with spaces when read by some programs. 30) Fixed error message when reading factor data with combined (CMB) factor in EGIS format.  31) Now reads the nsweeps field of Neuroscan AVG files to fill in the avgNum and subNum fields.  32) Export file dialogs in Edit function now correctly indicate that the file with be of type .txt. 33) Made autoPCA option of Window function more memory efficient so it wouldn't run out of memory. 34) Fixed crash in ANOVA function when level names of between group factors were of different lengths. 35) Now zero-padding factor names (e.g., "003") to maximum number of digits so that they sort properly. 36) Close files after running batch of ANOVAs to avoid "too many files open" error. 37) When running ANOVAs with between subject factors using the output of the autoPCA option, fixed the addition of grand averages corresponding to the levels of the factors to the dataset. 38) Corrected the peak channel and peak time point identification of factors made by the autoPCA option of the window function. 39) Fixed bug that prevented one from fully deleting the contents of edit fields, such as "baseline" on the preprocessing pane. 40) Fixed bug that prevented one from changing the contents of the edit fields on the postprocessing pane. 41) Workaround for Matlab bug that appears to prevent buttons from being colored when pressed on Macs using version 2009b applied to windowing function's PCA guided channel group option too.  42) Fixed crash when viewing data with regional channels.  43) Eliminated updating of baseline field when baseline correction during postprocessing.  The part of the segment used for baseline correction may not necessarily correspond to the prestimulus period. 44) Fixed crash when using Edit function's Factor subpane to examine factors that include a combined factor add. 45) Fixed inability to set minimum variance setting for autoPCA preference to anything other than zero. 46) Implicit reference channels will no longer be marked as bad when reading them in. 47) Added option to turn off adding montage information to EGIS file format files due to incompatibility issues with some versions of NetStation. 48) Fixed crash when saving EGIS file format file with a "sex" subject spec field.

Cheers!

Joe

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

Joseph Dien,
Senior Research Scientist
University of Maryland 
7005 52nd Avenue
College Park , MD 20742-0025

E-mail: jd...@ma...
Phone: 301-226-8848
Fax: 301-226-8811
http://homepage.mac.com/jdien07/