3dMVM issue with many response values

AFNI Message Board
1

Dear AFNI experts,

I’m trying to analyze data from 17 subject using 3dMVM. Brain activity was estimated using TENT basis functions (8 basis functions). I have 2 additional within subjects factors: Condition (Neutral, Negative), and Delay (Short, Medium, Long). This means 17 * 8 * 2 * 3 = 816 response values.

When I run the analysis, 3dMVM correctly reports the contingency tables and succesfully reads input files.
Unfortunately, then the program seems to take an infinite time to perform the analysis. I waited for 8 hours but it stood still, without reporting the analysis of the first slice.

May this be related to the quite large number of response values? Is there something wrong in my code? (see below)

Useful informations: I run AFNI on “virtual” Ubuntu (through VirtualBox). The virtual machine has 4 processors (of 8 total in my PC), and 8 GB RAM (of 12). So it is not a particularly slow machine, and was in fact able to perform many analysis using AFNI.

These are the first lines of the script:
-3dMVM
-prefix MVM_results
-jobs 4
-wsVars “ConditionDelayTime”
-mask …/groupmask/NoiseMask+tlrc
-SS_type 3
-num_glt 3
-gltLabel 1 ‘Negative’ -gltCode 1 ‘Condition : 1Negative’
-gltLabel 2 ‘Neutral’ -gltCode 2 'Condition : 1Neutral’
-gltLabel 3 ‘Neg-Neu’ -gltCode 3 ‘Condition : 1Negative -1Neutral’
-dataTable
Subj Condition Delay Time InputFile
s1 Neutral Short t1 …/subjF1/Decon/bucket+tlrc’[1]’
s1 Neutral Short t2 …/subjF1/Decon/bucket+tlrc’[3]’
…

Thanks in advance,
Simone

2

Simone, could you paste the whole message on your terminal when you run the 3dMVM script?

3

Hi Simone,

The problem may lie in using " vs '. At least, that’s the exact problem I had today :slight_smile: I’m working on a MAC and had to change from using ’ to ". Also, you are using both in your command.

-wsVars “ConditionDelayTime”
vs
-gltLabel 1 ‘Negative’ -gltCode 1 ‘Condition : 1*Negative’ \

Hoping this does the trick,

Nic

4

Loading required package: lme4
Loading required package: Matrix
Loading required package: reshape2
Loading required package: lsmeans
Loading required package: estimability

Welcome to afex. Important changes in the current version:

  • Functions for ANOVAs have been renamed to: aov_car(), aov_ez(), and aov_4().
  • ANOVA functions return an object of class ‘afex_aov’ as default, see: ?aov_car
  • ‘afex_aov’ objects can be passed to lsmeans for contrasts and follow-up tests.
  • Reset previous (faster) behavior via: afex_options(return_aov=‘nice’)
  • Many more arguments can now be set globally via options, see: afex_options()

Loading required package: car

++++++++++++++++++++++++++++++++++++++++++++++++++++
***** Summary information of data structure *****
17 subjects : s1 s10 s11 s12 s13 s14 s15 s16 s17 s2 s3 s4 s5 s6 s7 s8 s9
816 response values
2 levels for factor Condition : Negative Neutral
3 levels for factor Delay : Long Medium Short
8 levels for factor Time : t1 t2 t3 t4 t5 t6 t7 t8
3 post hoc tests

Contingency tables of subject distributions among the categorical variables:

, , Time = t1

      Delay

Condition Long Medium Short
Negative 17 17 17
Neutral 17 17 17

, , Time = t2

      Delay

Condition Long Medium Short
Negative 17 17 17
Neutral 17 17 17

, , Time = t3

      Delay

Condition Long Medium Short
Negative 17 17 17
Neutral 17 17 17

, , Time = t4

      Delay

Condition Long Medium Short
Negative 17 17 17
Neutral 17 17 17

, , Time = t5

      Delay

Condition Long Medium Short
Negative 17 17 17
Neutral 17 17 17

, , Time = t6

      Delay

Condition Long Medium Short
Negative 17 17 17
Neutral 17 17 17

, , Time = t7

      Delay

Condition Long Medium Short
Negative 17 17 17
Neutral 17 17 17

, , Time = t8

      Delay

Condition Long Medium Short
Negative 17 17 17
Neutral 17 17 17

Tabulation of subjects against each of the categorical variables:

lop$nSubj vs Condition:
     
      Negative Neutral
  s1        24      24
  s10       24      24
  s11       24      24
  s12       24      24
  s13       24      24
  s14       24      24
  s15       24      24
  s16       24      24
  s17       24      24
  s2        24      24
  s3        24      24
  s4        24      24
  s5        24      24
  s6        24      24
  s7        24      24
  s8        24      24
  s9        24      24

lop$nSubj vs Delay:

  Long Medium Short

s1 16 16 16
s10 16 16 16
s11 16 16 16
s12 16 16 16
s13 16 16 16
s14 16 16 16
s15 16 16 16
s16 16 16 16
s17 16 16 16
s2 16 16 16
s3 16 16 16
s4 16 16 16
s5 16 16 16
s6 16 16 16
s7 16 16 16
s8 16 16 16
s9 16 16 16

lop$nSubj vs Time:
     
      t1 t2 t3 t4 t5 t6 t7 t8
  s1   6  6  6  6  6  6  6  6
  s10  6  6  6  6  6  6  6  6
  s11  6  6  6  6  6  6  6  6
  s12  6  6  6  6  6  6  6  6
  s13  6  6  6  6  6  6  6  6
  s14  6  6  6  6  6  6  6  6
  s15  6  6  6  6  6  6  6  6
  s16  6  6  6  6  6  6  6  6
  s17  6  6  6  6  6  6  6  6
  s2   6  6  6  6  6  6  6  6
  s3   6  6  6  6  6  6  6  6
  s4   6  6  6  6  6  6  6  6
  s5   6  6  6  6  6  6  6  6
  s6   6  6  6  6  6  6  6  6
  s7   6  6  6  6  6  6  6  6
  s8   6  6  6  6  6  6  6  6
  s9   6  6  6  6  6  6  6  6

***** End of data structure information *****
++++++++++++++++++++++++++++++++++++++++++++++++++++


Reading input files now...

Reading input files: Done!

If the program hangs here for more than, for example, half an hour,
kill the process because the model specification or the GLT coding
is likely inappropriate.

[1] "Great, test run passed at voxel (22, 32, 25)!"
[1] "Start to compute 50 slices along Z axis. You can monitor the progress"
[1] "and estimate the total run time as shown below."
[1] "08/31/17 14:29:14.027"
Loading required package: snow
Package snow loaded successfully!

----------------------------

Then the program stucks here for eternity...

Simone
5

Thank you for the answer,

anyway, to me it is unclear what you are suggesting. Can you please specify?

Simone

6

Simone, I fail to see anything obviously wrong. Kill the current job, and rerun the script. If the problem recurs, I’ll have to ask you to upload the data so that I can take a close look.