Thanks, that helped me a lot!
Best wishes,
-Lucas
> Sorry!
> Uncorrected is easy: there is a `ztop` command:
>
> ztop 3.5 --> 0.000233: this is an uncorrected p.
>
> For corrected however...
> I would simply try a "manual iterative" approach. :-) Of course, you
> can script it.
>
> 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. 16:15 keltezéssel, Lucas Eggert írta:
>> 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/
|