cmind.utils.fsf_utils

FSL .fsf file creation/modification utilities

create_fsf create_con create_ev write_fsf fsf_substitute

Functions

create_con([mode]) Creates a contrast structure for specification of FEAT model
create_ev() ev=create_ev;
create_fsf([vers]) Populates FEAT fsf structure with default values
fsf_substitute(FSF_file, valdict)
write_fsf(fsf[, output_name])
Parameters:
cmind.utils.fsf_utils.create_con(mode='orig')[source]

Creates a contrast structure for specification of FEAT model

Parameters:

mode : {‘orig’,’real’}

contrast mode

Returns:

con : dict

dictionary with a default ev structure

cmind.utils.fsf_utils.create_ev()[source]

ev=create_ev; Creates a regressor for use in FEAT model

Returns:

ev : dict

dictionary with a default ev structure

cmind.utils.fsf_utils.create_fsf(vers=6.0)[source]

Populates FEAT fsf structure with default values

Parameters:

vers : float

Feat version to create the .fsf file for

Returns:

fsf : dict

fsf structure

See also

write_fsf

Notes

04/22/05 PJ Modified by GRL to update to feat v6.0

cmind.utils.fsf_utils.fsf_substitute(FSF_file, valdict)[source]
cmind.utils.fsf_utils.write_fsf(fsf, output_name=None)[source]
Parameters:

fsf : dict

fsf dictionary to be written

output_name : str

string for fsf file output. If unspecified it will be fsf[‘fsldir’]/design.fsf

Returns:

output_name : string

filename of the fsf file that was generated

Notes

write_fsf(fsf);

Writes a FEAT structure file (.fsf) for use with FSL’s FEAT analysis package

fsf is structure with all of the parameters necessary to write out the .fsf file. The default field values are set in create_fsf.m

NOTE: This is still a work in progress. Fortunately, the FEAT GUI will read a partially completed setup file. In effect, it should be possible to create a template for any given analysis and then fill a few fields with pointers to the subject specific files.

# based on a Matlab script by Petr Janata # writes the .fsf file based on information in the structure modification by Gregory Lee for CMIND project