Hi Swann, Yes, I'm running the analysis on a 64 bit system. The matlab error message is: Warning: Variable 'SPM' cannot be saved to a MAT-file whose version is older than 7.3. To save this variable, use the -v7.3 switch. Skipping... > In spm_spm at 883 In spm_run_fmri_est at 33 In matlabbatch/private/cfg_run_cm at 29 In cfg_util>local_runcj at 1510 In cfg_util at 830 In spm_jobman at 217 In Labnic_1st_analysis_FACS>ana at 220 In Labnic_1st_analysis_FACS at 86 Failed 'Model estimation' Reference to a cleared variable SPM. In file "/usr/local/SPM/spm8/spm_spm.m" (v4191), function "spm_spm" at line 889. In file "/usr/local/SPM/spm8/config/spm_run_fmri_est.m" (v4022), function "spm_run_fmri_est" at line 33. The following modules did not run: Failed: Model estimation Best, Anne From: [log in to unmask] Date: Tue, 13 Nov 2012 15:00:11 +0100 Subject: Re: [SPM] 1st level analysis fails on large data set To: [log in to unmask] CC: [log in to unmask] what is your matlab error estimation of large matrices (eg experiments with more than ~5000 volumes) fails on 32 bit systems. Have you tu run tried a 64 bits system ? 2012/11/13 Anne Schobert <[log in to unmask]> ...I should add that I'm using spm8 and Matlab version 7.5.0! I hope somebody has some insight! Anne -- Swann Pichon, PhD Swiss Center for Affective Sciences (CISA) - University of Geneva 7, Rue des battoirs - 1211 GENEVA, Switzerland Phone/Fax: +41 22 37 99 823 / +41 22 37 99 844 GSM ch/fr: +41 77 46 84 910 / +33 6 26 43 83 61 http://www.affective-sciences.org/user/swannpichon