Error message when choosing option 8

[ab698]

Hi Chris and the team,
I'm getting the following error when running option 8. I've just
downloaded the latest version of the code onto one of our other
servers with a bit more resources but now I'm getting the following
message running matlab 7.0.4.352 (R14) Service Pack 2

Please provide filename prefix for the ICV (inverse coefficient of
variation) voxel map:
COMT-N81
Where would you like to place the file? Please input target path
(ending WITHOUT "/"):
.

Beginning behavioral-fit bootstrap resampling...

Bootstrap Sample #1 out of 1000.
??? Error using ==> unique
Unknown flag.

Error in ==> bootstrap_resample_f at 35
[class first_instances standardized_label] = unique(label,
'first');

Error in ==> behav_fit_bootstrap_module_f at 142
subj_sample = bootstrap_resample_f(nOfSubj, behav);

Error in ==> PCA_suite_f at 367
vars.behav_fit_icv_map_noZeroes ...

>>

[Christian Habeck]

Oops. You ran this option with exactly these settings before? - I am a
little suspicious about the hyphen in your filename. I'll look into it
with Tony.

--
Sent from my mobile device

[annib]

Hi Chris,
you're up early !

Hyphen hasn't been a problem before ever however this is matlab 7
(wbic server) and I was running it previously on matlab 7.3 (bcni
server) so may be there is some mismatch there. I think you say
something about previous versions of matlab on your download page.

Anyway I've had a word with my SA and he says as long as I'm not
accessing 4GB of memory in one go I should be fine. Is there any risk
that with 81 images I am accessing 4GB in one go?

thanks
Anna

[Christian Habeck]

Hmm,

not exactly sure, but that could be a problem.... In terms of memory management this program might be quite greedy and 81 images is a quite a bit, but then I would have expected a different error message.

Sorry my suggestions are quite empirical:  maybe you could mask this images much more conservatively and reduce the number of active voxels, or maybe only analyze a subset of images? If the error still shows up maybe we can be confident it's NOT a memory issue?

Chris

-- 
Christian Habeck
ch...@columbia.edu
http://www.cumc.columbia.edu/dept/sergievsky/cnd/habeck.html