Index TableOfContents
Name
sphsmooth-sess - smoothes hemodynamic averages on the surface
Synopsis
sphsmooth-sess ??
Arguments
Positional Arguments
none
Required Flagged Arguments
??
Optional Flagged Arguments
-smoothsteps <n>  | 
  number of smoothing steps  | 
-analysis <analysisname>  | 
  session-level functional analysis name  | 
-insphdir <dirname>  | 
  input spherical space directory name <sph>  | 
-outsphdir <dirname>  | 
  output spherical space directory name  | 
-hemi <hemisphere>  | 
  <hemisphere> = lh or rh  | 
-sf <sessidfile>  | 
  ...  | 
-df <srchdirfile>  | 
  ...  | 
-s <sessid>  | 
  ...  | 
-d <srchdir>  | 
  ...  | 
-umask <umask>  | 
  set unix file permission mask  | 
-scriptonly  | 
  don't run, just generate a script  | 
-version  | 
  print version and exit  | 
Outputs
<??>  | 
  ??  | 
Description
["sphsmooth-sess"] will smooth an analysis on the surface. The results of the analysis must have been resampled onto the sphere using ["func2sph-sess"]. By default, the input directory will be sph (as created by ["func2sph-sess"]). The user must supply the output directory. To access the smoothed data in subsequent calls (eg, ["isxavg-re-sess"] or ["isxavg-fe-sess"]), add -spacedir followed by the name given to the output directory.
The amount of smoothing is controlled by the number of smoothing steps (-smoothsteps). A single smoothing step is when each vertex is averaged with its neighbors. This will be performed iteratively as many times as specified. This is ROUGHLY equivalent to a gassian kernel with StdDev of SQRT(nsmooth*2/pi). The FWHM = (2.3*StdDev)/1.24. The StdDev and FWHM are in mm. I emphasize that this is a rough equivalence; the actual StdDev will probably be a little more.
How much smoothing you do depends upon your data. Ideally, it should be the same as your activation size.
See http://www.nmr.mgh.harvard.edu/~greve/surf-smooth
Examples
Example 1
??
Example 2
??
Bugs
??
See Also
["func2sph-sess"]
Links
Methods Description
??
References
["References/Lastname###"]
Reporting Bugs
Report bugs to <analysis-bugs@nmr.mgh.harvard.edu>
