with the -m flag the metrics was not identical, but differences were minimal
> Something isn't right here. You should not be getting drastically
> different DVARS values if all you've done is change the normalization
> factor of your input data, since the scaling is accounted for within
> 'fsl_motion_outliers'. You said you got identical results if you
> constructed an explicit mask and gave that as an input using the -m flag?
>
> --
> Michael Harms, Ph.D.
>
> -----------------------------------------------------------
> Conte Center for the Neuroscience of Mental Disorders
> Washington University School of Medicine
> Department of Psychiatry, Box 8134
> 660 South Euclid Ave. Tel: 314-747-6173
> St. Louis, MO 63110 Email: [log in to unmask]
>
>
>
>
> On 2/25/14 3:58 PM, "Angela Favaro" <[log in to unmask]> wrote:
>
>>yes, I am using the last version of FSL
>>
>>I have tried to change my pre-processing script and normalize to 1000
>>I think the DVARS now are more suitable:
>>
>>brainmed = 1128.405762 ; maskmean = 0.197097
>>Calculating outliers
>>Range of metric values: 0.059909 6.443539
>>Found 14 outliers over 4.9505675
>>Generating EVs
>>Found spikes at 176 177 191 192 193 200 203 225 226 227 243 244 248 249
>>
>>what do you think?
>>
>>I would need another hint.
>>Which is the best way to scrab my data?
>>I would like to scrab it before performing ICA and/or seed-based
>> analyses.
>>The papers describe a 'temporal masking' process, but what does it mean?
>>
>>THanks for any help!
>>
>>Angela
>>
>>
>>> Are you using the latest (FSL 5.0.6) version of 'fsl_motion_outliers'?
>>> Because I know that at least that version internally adjusts for the
>>> intensity scaling, so as to yield a DVARS scaling comparable to the
>>> mode
>>> 1000 scaling that JP uses.
>>>
>>> If your preprocessed data is normalized to 10000, then it shouldn't be
>>> reporting a 'brainmed' (median value within the brain mask) of only
>>>99.8.
>>>
>>> Try using the hidden --nocleanup option, and examine the intermediate
>>> outputs to see if you can see what is going on.
>>>
>>> cheers,
>>> -MH
>>>
>>> --
>>> Michael Harms, Ph.D.
>>>
>>> -----------------------------------------------------------
>>> Conte Center for the Neuroscience of Mental Disorders
>>> Washington University School of Medicine
>>> Department of Psychiatry, Box 8134
>>> 660 South Euclid Ave. Tel: 314-747-6173
>>> St. Louis, MO 63110 Email: [log in to unmask]
>>>
>>>
>>>
>>>
>>> On 2/25/14 12:49 PM, "Angela Favaro" <[log in to unmask]> wrote:
>>>
>>>>Can it be that the intensity normalization was done by a factor of
>>>>10000?
>>>>Could DVARS be influenced by this?
>>>>
>>>>thank you
>>>>Angela
>>>>
>>>>
>>>>> If you compute DVARS without motion correction, you are going to get
>>>>> a
>>>>> much higher range of values. While Jonathan may have done some
>>>>> analyses
>>>>> without motion *regressors* in the preprocessing, I'm pretty sure
>>>>> that
>>>>> motion *correction* is always included prior to DVARS computation.
>>>>>
>>>>> Depending on the temporal SNR of your data, a DVARS baseline of 30-40
>>>>> after motion correction is certainly reasonable.
>>>>>
>>>>> BTW: You probably want to be specifying a mask as part of your input
>>>>>to
>>>>> 'fsl_motion_outliers'.
>>>>>
>>>>> cheers,
>>>>> -MH
>>>>>
>>>>> --
>>>>> Michael Harms, Ph.D.
>>>>>
>>>>> -----------------------------------------------------------
>>>>> Conte Center for the Neuroscience of Mental Disorders
>>>>> Washington University School of Medicine
>>>>> Department of Psychiatry, Box 8134
>>>>> 660 South Euclid Ave. Tel: 314-747-6173
>>>>> St. Louis, MO 63110 Email: [log in to unmask]
>>>>>
>>>>>
>>>>>
>>>>>
>>>>> On 2/25/14 11:15 AM, "Angela Favaro" <[log in to unmask]> wrote:
>>>>>
>>>>>>Hi all,
>>>>>>I have some problemi in using the script fsl_motion_outliers and I
>>>>>>need
>>>>>>some clarification. I would like to 'scrab' my dataset before using
>>>>>> it
>>>>>>with both melodic and seed-based resting-state analyses.
>>>>>>
>>>>>>I tried to use the script with a subject with:
>>>>>>MCFLIRT Motion correction Mean displacements: absolute=0.08mm,
>>>>>>relative=0.04mm.
>>>>>>
>>>>>>The command I used is:
>>>>>>fsl_motion_outliers -i file -o file --dvars -nomoco -v
>>>>>>and I tried to run it in preprocessed data (as described in the paper
>>>>>> by
>>>>>>POwer 2012).
>>>>>>the output was:
>>>>>>brainmed = 99.815247 ; maskmean = 0.249414
>>>>>>Calculating outliers
>>>>>>Range of metric values: 99.421906 458.086700
>>>>>>Found 9 outliers over 377.0876780
>>>>>>Generating EVs
>>>>>>Found spikes at 89 172 175 192 212 214 215 220 223
>>>>>>
>>>>>>I do not understand why the values of DVARS are so high. With
>>>>>>unprocessed
>>>>>>data (and moco allowed) values are lower, but not as expected (I
>>>>>> think):
>>>>>>
>>>>>>brainmed = 382.000000 ; maskmean = 0.238453
>>>>>>Calculating outliers
>>>>>>Range of metric values: 26.647026 39.988392
>>>>>>Found 5 outliers over 37.0337425
>>>>>>Generating EVs
>>>>>>Found spikes at 134 192 203 226 239
>>>>>>
>>>>>>What am I doing wrong?
>>>>>>
>>>>>>Thank you!
>>>>>>
>>>>>>Angela
>>>>>
>>>>>
>>>>> ________________________________
>>>>> The materials in this message are private and may contain Protected
>>>>> Healthcare Information or other information of a sensitive nature. If
>>>>>you
>>>>> are not the intended recipient, be advised that any unauthorized use,
>>>>> disclosure, copying or the taking of any action in reliance on the
>>>>> contents of this information is strictly prohibited. If you have
>>>>>received
>>>>> this email in error, please immediately notify the sender via
>>>>>telephone
>>>>>or
>>>>> return mail.
>>>>>
>>>>>
>>>
>>>
>>> ________________________________
>>> The materials in this message are private and may contain Protected
>>> Healthcare Information or other information of a sensitive nature. If
>>>you
>>> are not the intended recipient, be advised that any unauthorized use,
>>> disclosure, copying or the taking of any action in reliance on the
>>> contents of this information is strictly prohibited. If you have
>>>received
>>> this email in error, please immediately notify the sender via telephone
>>>or
>>> return mail.
>>>
>>>
>
>
> ________________________________
> The materials in this message are private and may contain Protected
> Healthcare Information or other information of a sensitive nature. If you
> are not the intended recipient, be advised that any unauthorized use,
> disclosure, copying or the taking of any action in reliance on the
> contents of this information is strictly prohibited. If you have received
> this email in error, please immediately notify the sender via telephone or
> return mail.
>
>
|