No, I mean the other way around:
Let's assume, I have use Z=3.5 to threshold my statistical map, and now
I would like to know what the corresponding p-Level of this threshold is
(corrected and uncorrected).
> Just use the 'ptoz' command without the -g switch:
>
> Z=`ptoz 0.05`; # gives you the uncorrected z-threshold for p=0.05
>
> Auer, Tibor M.D. Ph.D.
> Biomedizinische NMR Forschungs GmbH
> am Max-Planck Institut für Biophysikalische Chemie
> Am Fassberg 11
> 37077 Göttingen
> Germany
> Phone/Work: +49-(0)551-201-1725
> Phone/Home: +49-(0)551-387-0076
> Mobile: +49-(0)176-8012-7921
> Mail: [log in to unmask]
>
>
>
> 2010.09.14. 16:05 keltezéssel, Lucas Eggert írta:
>> Perfect!
>>
>> Thank you very much!
>>
>> Last question regarding these kinds of conversions:
>>
>> How do I determine the corrected and the uncorrected p-value for a
>> given Z-value?
>>
>> Thank you again!
>>
>> All the best,
>> -Lucas
>>> Dear Lucas,
>>>
>>> No. ptoz is an FSL-command. Its syntax is (for a p=0.05)
>>>
>>> ptoz 0.05 -g <number of resels>
>>>
>>> Number of resels can be calculated based on the 'smoothest' command
>>> or (even better) based on the information you can find in the
>>> "*.feat/stats/smoothness" file:
>>>
>>> number of resels = VOLUME / RESELS
>>>
>>> Then, the 'ptoz' command will give you the z-threshold for p = 0.05
>>> (or any) corrected for FWE
>>> In the next step, you just have to threshold your zstat with the
>>> given z-value:
>>>
>>> After you have stored the VOLUME and RESEL information in the
>>> respective variable: e.g.:
>>>
>>> VOLUME=47487
>>> RESEL=2.34
>>>
>>> NRES=`echo "scale=4; $VOLUME / $RESEL" | bc`; # you need 'bc',
>>> because this is a floating point operation
>>> Z=`ptoz 0.05 -g $NRES`;
>>> fslmaths zstat1 -thr $Z $thresholded_zstat1
>>>
>>> I hope it helps,
>>>
>>> Auer, Tibor M.D. Ph.D.
>>> Biomedizinische NMR Forschungs GmbH
>>> am Max-Planck Institut für Biophysikalische Chemie
>>> Am Fassberg 11
>>> 37077 Göttingen
>>> Germany
>>> Phone/Work: +49-(0)551-201-1725
>>> Phone/Home: +49-(0)551-387-0076
>>> Mobile: +49-(0)176-8012-7921
>>> Mail: [log in to unmask]
>>>
>>>
>>>
>>> 2010.09.14. 13:36 keltezéssel, Lucas Eggert írta:
>>>> Dear Tibor,
>>>>> Dear Lucas,
>>>>>
>>>>> AFAIK, you cannot convert a z-map to FWE-p-map; but if your goal
>>>>> is to threshold your z-map with FWE, then you can calculate the
>>>>> z(p=0.05 FWE) threshold with ptoz command with using the -g switch.
>>>> That is exactly what I want to do: Determine the Z-value that
>>>> represents a p-value of e.g. 0.05 (FWE-corrected).
>>>>
>>>> However, I do not understand what you mean with '-g switch'. Do you
>>>> mean something like
>>>>
>>>> fslmaths zstat1.nii.gz -g -ptoz output.nii.gz?
>>>>
>>>> Thank you very much!
>>>> -Lucas
>>>>
>>>>>
>>>>> Bests,
>>>>>
>>>>> Auer, Tibor M.D. Ph.D.
>>>>> Biomedizinische NMR Forschungs GmbH
>>>>> am Max-Planck Institut für Biophysikalische Chemie
>>>>> Am Fassberg 11
>>>>> 37077 Göttingen
>>>>> Germany
>>>>> Phone/Work: +49-(0)551-201-1725
>>>>> Phone/Home: +49-(0)551-387-0076
>>>>> Mobile: +49-(0)176-8012-7921
>>>>> Mail: [log in to unmask]
>>>>>
>>>>>
>>>>>
>>>>> 2010.09.14. 12:32 keltezéssel, Lucas Eggert írta:
>>>>>> Dear Experts!
>>>>>>
>>>>>> How can I determine the FWE-corrected p-Value for a given Z-value?
>>>>>>
>>>>>> I tried
>>>>>>
>>>>>> fslmaths zstat1.nii.gz -cpval fwe-corrected.nii.gz
>>>>>>
>>>>>> ... but that results in an image with only values of 1 in every
>>>>>> voxel.
>>>>>>
>>>>>> Best wishes,
>>>>>> -Lucas
>>>>>>
>>>>>
>>>>>
>>>>
>>>>
>>>
>>>
>>
>>
>
>
--
Lucas Eggert, M.Sc.
Institute of Cognitive Science
University of Osnabrueck
Albrechtstrasse 28
D-49076 Osnabrueck
Germany
Phone: +49-541-969-44-28
Website: http://www.cogsci.uni-osnabrueck.de/~leggert/
|