ROI Toolbox Docs

Documentation: ROI HRFestimate (Roi_HRFestimate.m)
----------------------------
Mevan Ranasinghe
12/01/04, and update 04/22/05.


The ROI-HRFestimate function estimates the HRF function in an ROI. The program
provides an estimate of the rest state so that activation sizes can be 
measured, and it provides a standard deviation of the sampled data to 
estimate one-sigma error bars. 
 
Note SPM can calculate variations in HRF (using canonical HRF with time and 
dispersion derivatives), but the resulting possible HRF shapes are limited. 
The HRF Estimation function can find ANY possible HRF shape.

The full flexibility of HRF estimation comes at a cost of careful interpretation. 
HRF estimation can be a hard problem that is numerically unstable, so the 
program can be tricky to use correctly. With care, it gives nice results.
An entire manual for it in MSWord format is located in spm2/devel/help2/ROI_HRFEstimateDoc.

The program is derived from AFNI 3dDeconvolve, and reuses many parts of
Jeff Cooper's program for percent signal change in SPM99.

The Version 1.0.1 upgrade in April 2005 allows block design experiments, and allows
long time frames to be estimated. The program is designed to work on single session
experiments- multiple session experiments will not work properly.



