From jaston@bic.mni.mcgill.ca  Tue Sep 11 18:17:54 2001
Received: from localhost (jaston@localhost)
	by bottom.bic.mni.mcgill.ca (8.11.3/8.11.3) with ESMTP id f8BMHrp04280;
	Tue, 11 Sep 2001 18:17:53 -0400 (EDT)
Date: Tue, 11 Sep 2001 18:17:52 -0400
From: John ASTON <jaston@bic.mni.mcgill.ca>
To: Miriam Beauchamp <beauchmi@MAGELLAN.UMontreal.CA>
cc: <kate@bic.mni.mcgill.ca>
Subject: Re: Euler
In-Reply-To: <3.0.32.20010911173504.008e9160@magellan.umontreal.ca>
Message-ID: <Pine.SGI.4.33.0109111812020.3829-100000@bottom.bic.mni.mcgill.ca>
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII
X-Keywords:                 
X-UID: 159
Status: RO
Content-Length: 1589
Lines: 46


Hi

Fine - The value sounds more likely now.

I think that certainly the values from Worsley et al, are a good first
approximation, although they might not be exactly correct depending on
your and his definition. Just do a rough calculation by roughly counting
the number of voxels (to the nearest few) and then check that this is
roughly the same volume as in Worsley et al.

You seem to be using euler  correctly - that is the p-value you get out.
The thing to do is iterate (keep choosing) the thresholds so they get to
roughly p=0.05(or p=whatever you want).

The values in Worsley et al are for 20mm smoothing so you cant use the
thresholds directly you'll have to increase them a bit using euler to
check.

Hope this all works, let me know if you have any probs.

John

On Tue, 11 Sep 2001, Miriam Beauchamp wrote:

> Heeeelllllo,
>
> 	Just to confirm with you...the number of resels in the search volume is
> 544.68 and the search volume is 1494.59 cc with an FWHM of 14mm.
>
> 	Two questions:  First, Kate and I are just trying to figure out how best
> to calculate the volumes of our structures of interest (do you think we can
> use the values directly from Worsley et al., 96?).  And, once we get the
> Euler characteristic, how do we use this information?!  For example:
>
> Structure of interest = 207 cc (total frontal gyri)
> FWHM = 14 mm
> Resel volume = 207  / (1.4)^3 = 75.44
> t for P(M >= t) = 3.85 (from Worsley et al., 96)
> EC = 0.093 (from /usr/local/mni/bin/euler)
>
> So, does EC give us the p-value for this structure???
>
> Thanks a lot.  Miriam and Kate.
>
>