whatsnew.txt

绝对经典,老外制作的功能强大的matlab实现PLS_TOOBOX

Version 5.0705021  May 8, 2007
	1.	Command-line PLS help information is included in the
		Appendix of PLSgui User's Guide.

Version 5.0705021  May 2, 2007
	1.	Modified "load_ant.m" based on the changed .avr file
		information provided by ANT staff. So now PLSgui can
		load ANT's average binary data format again. Please
		see detail in:
		http://www.rotman-baycrest.on.ca/pls/FAQ.txt

Version 5.0704241  April 24, 2007
	1.	A rounding issue in the code caused multiple voxel
		extraction routine to fail in some situations. This
		problem is fixed now.

Version 5.0704231  April 23, 2007
	1.	Stop supporting ANT average binary data format.
	2.	Start supporting users' own scalp electrode system.
	3.	Updated http://www.rotman-baycrest.on.ca/pls/FAQ.txt
		and http://www.rotman-baycrest.on.ca/pls/UserGuide.htm

Version 5.0704201  April 20, 2007
	1.	Fixed the display of CTF150 location for ERP module.
	2.	Jordan Poppenk found that the voxel value can be
		negative, if a brain mask is used. So the previous
		restriction is remvoed from FAQ file.
	3.	"load_nii" will automatically load SPM affine matrix
		M in .mat file, if it detects that the file name is
		the same as the Analyze file.

Version 5.0704131  April 13, 2007
	1.	Modified load_acq.m.

Version 5.0704091  April 9, 2007
	1.	Modified normalize.m and add load_acq.m.

Version 5.0703272  March 27, 2007
	1.	Fixed a bug for the wrong scaling dispay in the
		Subject and Average Amplitude window in the ERP
		module, which was introduced from on March 9
		from version 5.0703091.

Version 5.0703221  March 22, 2007
	1.	Fixed a bug in saving data to a text file.

Version 5.0703191  March 19, 2007
	1.	Fixed a bug in exporting cluster report all location.

Version 5.0703151  March 15, 2007
	1.	Fixed a bug in bootstrap test for Multiblock PLS.

Version 5.0703091  March 9, 2007
	1.	The display of "Design LV and Design Scores" window
		in other module follows the way that is already in
		place for E.R.fMRI. i.e. all the groups will display
		together in the same window.
	2.	The subject name in all MRI module is now using the
		datamat prefix name defined in its session file.
	3.	"Scale by Singular Value" button is added in all
		result display window. It is the same button called
		"Rescale" previously only for PET and STRUCT.
	4.	The "Voxel Intensity Response" window will get the
		same axis for all the groups. However, if you first
		click a new group that exceeds the min/max of currnet
		value, the axis will still adjust to the new min/max.
		If you click back, the axis will stay unchanged.
	5.	Added a "Create Batch File" button in the "Run PLS
		Analysis" window. So people can using this created
		plain text file (batch file) to feed "batch_plsgui"
		program. We may also do some test with an external
		program that is compiled in C language, and it will
		also use the same batch text file to read MATLAB
		files (Session / Datamat file) and write a MATLAB
		file (Result file). but it is not available yet.
	6.	Updated PLS User's Guide.

Version 5.0703011  March 1, 2007
	1.	Prompt user to delete some subjects in STRUCT session
		information window when there is too much data to
		process.

Version 5.0702261  February 26, 2007
	1.	Mofified User's Guide.
	2.	Fixed false warning when there is no match subject files.

Version 5.0702221  February 22, 2007
	1.	Used different way to check low variability of behavior
		data for bootstrap test.

Version 5.0702212  February 21, 2007
	1.	"num_LowVariability_behav_boots" field is added to the
		"boot_result" struct for the bootstrap test containing
		behavior data. In the case that behavior data does not 
		have enough variance the following warning message will
		also be displayed in the command window: "For at least
		one behavior measure, the minimum unique values of
		resampled behavdata doesnot exceed 50% of its total".
	2.	Fixed display problem for the plot with special character
		in subject name or condition name. The program now can
		accept special character (e.g. underscore "_").
	3.	Removed "Merge Conditions" button for Structure PLS.

Version 5.0702161  February 16, 2007
	1.	Fixed a bug in bootstrap test in handling bad behavior
		data.
	2.	MATLAB 7 has a little difference in handling "uiputfile"
		in that it will include the wildcard string PLS program
		provided. This problem has been solved.
	3.	Made a workaround for single condition Mean-Centering PLS.
	4.	Structure MRI module is implemented.

Version 5.0702061  February 6, 2007
	1.	Fixed a bug in permutation test of Non-Rotated PLS for
		MRI modules.
	2.	Fixed a dialog box error message asking about default
		button.
	3.	In datamat correlation window for behavior PLS result,
		the "Save ... to IMG file" will now apply the bootstrap
		mask that filters out the unreliable voxels.

Version 5.0701151  January 15, 2007
	1.	Fixed a bug in detecting group contrast rank deficiency.

Version 5.0701101  January 10, 2007
	1.	When exporting all locations of cluster report, values
		(BrainLV or BSRatio) for the corresponding voxels are
		also exported into "*_MODULEcluster_num.txt". The first
		column of this file is still Cluster ID#; the second
		column of this file is now the corresponding BrainLV
		or BSRatio value; the last 3 columns are still XYZ
		locations.

Version 5.0701032  January 3, 2007
	1.	Integrated single precision math operation feature of
		MATLAB 7 (R14) above into the PET & MRI modules of the
		program. If MATLAB is running on 64-bit system, a dialog
		box will be popped out to ask user for the system's 
		processor. If the answer is "Don't know" or "Intel 
		processor", double precision will be used according to 
		MATLAB Bug Report ID 268001. (If you are using RRI Linux
		cluster, "telesto" is the only 64-bit system that has
		non-Intel processor, which supports single precision math
		operation).
	2.	Removed unnecessary rounding (ceil) for the bootstrap
		test of behavior PLS analysis, which caused inaccuracy
		when number of bootstrap is small. "zero_brain_se" is
		also added in "boot_result", which represents the index
		of those brain_se with zero standard error.

Version 5.0612081  December 8, 2006
	1.	Neuroscan version 4.3 also has 32-bit acquisition
		for its continuous data files. By default, it collects
		16-bit integer data with the original SynAmp (SynAmp1).
		With SynAmp2, it will collect 32-bit integer data. 
		According to Neuroscan tech support, "this file format 
		must be determined prior to reading the file, i.e. there
		is no flag within the file itself to identify a file as 
		32-bit". In order to read 32-bit data format, optional 
		input "fileformat" of "load_neuroscan.m" function must
		be specified to 'cnt32'. For example, you can load data
		using: ns = load_neuroscan('myfile.cnt', 'cnt32'); Type
		"help load_neuroscan" for the usage of "fileformat".
		Because now the continuous data file can be very large,
		you can specify cnt_start_time and cnt_end_time. Their
		units are "second". For example, if you only want to
		extract data from 0 second to 1.998 second, you can type:
		ns = load_neuroscan('myfile.cnt', 'cnt', '', 0, 1.998);
		The actual rounded start_time and end_time will be 
		displayed in ns.start_time and ns.end_time.

Version 5.0612051  December 5, 2006
	1.	Added "behavior_name" field for the "analysis batch
		file" processing program. Demo "analysis batch file"
		are: batch_pet_analysis.txt, batch_erp_analysis.txt,
		batch_bfm_analysis.txt, and batch_fmri_analysis.txt.

Version 5.0611021  November 2, 2006
	1.	Fixed a bug when brain mask only contains 1 slice.

Version 5.0609251  September 25, 2006
	1.	Both EGI 128-Channel and EGI 256-Channel are provided
		along with BESA and CTF-150.
	2.	Script "erp_single_folder" is included in PLSgui for
		a special reason, and we do not recommand normal user
		to use this script.
	3.	Added "mat_into_hdr.m" script to convert SPM ANALYZE
		Reoriented image file into NIfTI format and integrate
		its matlab file into its header file.

Version 5.0609151  September 15, 2006
	1.	We decided to use XYZmm, voxel location with unit 
		in millimeter, for the entries of Multiple Voxel 
		Extraction" voxel location text file. In earlier
		version, we were using XYZ, absolute voxel location,
		for the entries of voxel location file. In order
		to convert your absolute voxel location file to 
		voxel location with unit in millimeter, we provided
		a handy script called "xyz2xyzmm.m".
	2.	Both "Peak Locations" and "All Locations" that are
		exported from "File" menu in "Cluster Report" window
		are also changed to voxel location with unit in 
		millimeter. This way, people perform "Seed PLS" do not
		have to to any extra conversion.
	3.	"Posthoc Data" button in "PLS Analysis" window is
		now removed.
	4.	Since new BESA file can have more than 1 header lines,
		ERP code is modified to accommodate this change. That
		means, ERP module can load BESA file with no header,
		or with any amount of headerlines, as well as binary
		files like NeuroScan, ANT, or EGI.

Version 4.0609071  September 7, 2006
	1.	A bug was introduced during last change (Version
		4.0608291), in which the neighborhood voxels were
		not exported without averaging together. Since last
		change only applied on MRI "Multiple Voxel Extraction",
		only MRI module was affected. It is now fixed.
	2.	In Matlab 2006a, the bootstrap test sometimes giving
		you a feeling of "frozen". This is caused by the 
		bootstrap progress bar sometimes conflict with the 
		permutation progress bar. I have also fixed this.

Version 4.0608291  August 29, 2006
	1.	Modified "Multiple Voxel Extraction" part of the
		program for MRI module, so it will consume less 
		memory. The result remains the same as the earlier
		version.

Version 4.0608111  August 11, 2006
	1.	The message of "Duplicate bootstrap order" was not
		popped out, and it is now fixed.

Version 4.0608091  August 9, 2006
	1.	Implemented the workaround for "persistent" variable
		that is provided by Matlab Tech Support, which can 
		prevent from using global variable as a replacement.
	2.	When exporting all locations of cluster report, a new
		file that is named "*_MODULEcluster_num.txt" will be
		created at the same time as "*_MODULEcluster_all.txt"
		file. In this new file, a column with cluster # is added
		in front of XYZ locations to indicate which cluster each
		exported voxel belongs to.

Version 4.0608021  August 2, 2006
	1.	Provided a workaround for "persistent" variable issue
		in Matlab R2006a (a Matlab bug).

Version 4.0607111  July 11, 2006
	1.	Fixed a bug when moving existing session file while the
		sub-folders also needs to be changed.

Version 4.0606231  June 23, 2006
	1.	"expand_nii_scan.m" is modified so the expanded image
		volume scan files will be appended with a 4-digit 
		number starting from 0000.

Version 4.0606201  June 20, 2006
	1.	The original display routine assumed the result data
		have both positive and negative values, and it crashed
		otherwise. Now the problem is fixed, and it can display
		all positive (or all negative) data.

Version 4.0606091  June 9, 2006
	1.	Cluster sizes were over counted during combination. It
		is fixed now.

Version 4.0606061  June 6, 2006
	1.	Program "m2hdr.m" is created to integrate affine matrix
		generated from old SPM into NIfTI *.hdr file.  
	2.	There was a bug in converting some permuted images to
		RAS orientation. It has been fixed now.

Version 4.0606011  June 1, 2006
	1.	There is a bug (crash) in fMRI non-rotate, if just
		doing bootstrap test. It is fixed.

Version 4.0605261  May 26, 2006
	1.	Modified cluster report "Export All Location" menu,
		and now the voxel is in the order of clusters. If you
		have two clusters, one with 30 voxels, and the other 
		one with 20 voxels, the first 30 lines correspond to
		cluster one and the second one correspond to cluster
		two. If you have saved "*cluster.mat" from previous
		version, you have to re-generate it before exporting
		all locations.

Version 4.0605111  May 11, 2006
	1.	When clicking "Save Current Display to IMG file", or 
		"Save BLV/BSR result to IMG file" under "File" menu,
		those img/hdr files can be saved in a directory that
		is different from result file directory.

Version 4.0604261  April 26, 2006
	1.	When editing existing fMRI session file, existing
		conditions were protected to prevent being modified.
		Now, the protection is removed.

Version 4.0604201  April 20, 2006
	1.	Size of some image data is incorrectly larger than 
		what specified in "hdr.dime.dim" field. So it caused 
		problem to load image data. Image data is now always
		loaded according to the size that is specified in 
		its header.
	2.	When editing existing fMRI session file, the "DELETE"
		buttons in "Edit Condition" window is now enabled.

Version 4.0603211  March 21, 2006
	1.	Tools for NIFTI image now supports both COMPLEX64 and
		COMPLEX128 data type.

Version 4.0603192  March 19, 2006
	1.	"UserGuide.htm" file is included in the package to
		replace the old out-of-dated "*_hlp.txt" files. The
		entire PLSgui User's Guide is still under processing.

Version 4.0603062  March 6, 2006
	1.	Modified "Export Data" for fMRI "Response Function Plot"
		and "Temporal Brain Scores Plot". The "st_evt_list" in
		the exported file is always sorted as condition-within-
		subjects, so as the "data" variable.
	2.	Added "data_mean" variable to the above exported MATLAB
		data file. Rows of "data_mean" stand for conditions, and
		subjects for each condition are averaged together.

Version 4.0602241  February 24, 2006
	1.	When loading some NIFTI data, there is an ordering 
		problem in "hdr.dime.pixdim" and "hdr.hist.originator"
		fields. It is now fixed.

Version 4.0602231  February 23, 2006
	1.	Modified NIFTI package to tell if the loaded NIFTI data
		is converted to RAS system. (see detail in FAQ.txt)

Version 4.0602072  February 7, 2006
	1.	Removed unused "Group Analysis" check box from E.R.fMRI
		and Blocked fMRI analysis window.
	2.	Made fields in PET and ERP session window look even,
		because behavior data fields were removed from session
		window on version 4.0512201.
	3.	Moved "Load Behavior Data" to a button in PLS option
		frame, which looks better.

Version 4.0601271  January 27, 2006
	1.	"Load Onsets from a text file for this run" and "Save
		Onsets to a text file for this run" submenu is added
		under "Edit" menu in both E.R fMRI & Blocked fMRI "Run
		Information" window.
	2.	When "DONE" button in above window is clicked, "Data
		Directory", and "Data Files" for all the runs will be
		checked, to make sure they are all filled.
	3.	Check session information further before creating MRI
		datamat.
	4.	When adding more subjects for PET and ERP module, the
		program will not always start from the working directory.
		Instead, it will remember parent directory of the first 
		subject, and start to look for more subject directories
		from there.

Version 4.0601251  January 25, 2006
	1.	Change "Save Background Image" label in result window
		to "Save brain region to IMG file".
	2.	When display activation map, changed inaccurate display