Bug fix

Incorporates a number of bug fixes made since the last major release.

Nuisance regressors

Added the ability to specify nuisance regessors in multiple regression. These regressors will be included in the model, but not subjected to TFCE, thereby reducing run time.

Progress and Parallelism

Command line feedback has been added for convenience. This includes a progress bar that runs during the permutation process so that users can observe the speed of execution live.

Additionally, parallel processing is now supported for all analyses via the command line or GUI. A new parameter "numworkers" specifies the number of parallel Matlab workers assigned to the permutation process. This has the potential to substantially accelerate the process on multi-core workstations and clusters. Acceleration appears almost linear (as a function of workers) on a quad-core workstation.

Multiple regression

Added support for TFCE-corrected multiple linear regression (at the random effects level) to the package. Produces corrected maps for each predictor in model. Between-subjects ANOVA and ANOCVA now possible through this approach (with appropriate variable coding). Also removed dependency on Mathworks stats package (randsample).

Initial release

Initial public release, featuring t-tests, correlation and ANOVAs

Bugfix'd build

Fixed one-tailed correlation bug

Also added one-tailed correlation demo

v0.1.1-beta

v0.1.1-beta

Feature complete initial version of the package.
Initial testing completed, FWE appropriate.
Wider testing and validation still needed.