Dear John, Elliott,

Thanks. Just a comment on two of the points made  - 

1. I don't think the AAL mask is trying to be "too precise", rather the contrary: I've just measured the AAL hippocampi and they weigh in at a hefty ~7,500mm^3. Ok, they're in Colin Holmes / MNI space which is about 35% (Ashburner et al. 1997) to 46% (Hammers, Allom et al. 2003) bigger than average, but still they're big - as partly explained by the protocol (Tzourio-Mazoyer et al. 2002). For comparison, our protocol (Niemann K et al. 2000) yields volumes of about 2100mm^3 in native space and 3300mm^3 in MNI space (Hammers, Allom et al. 2003); they are the volumes that went into the maxprobmap in MNI space recently discussed here. There are reviews/overviews of protocols and differences (e.g. Konrad C et al. 2009).

If a "GM only" mask bothers you, you can always get rid of the "holes" with simple morphological operations - e.g. by smoothing say with a 4mm kernel and then thresholding back. Obviously if you loose many voxels, your activation may not be in the hippocampus at all - which might also be interesting.

2. "moreover, it's not left-right symmetric": Well, why would it be? The physiological R>L asymmetry is well known (e.g. Pedraza O et al. 2004), certainly depends a bit on protocols, but should probably be around 5-12%; personally I use its detection as a rough indicator of the quality/plausability of a delineation protocol. In the case of AAL, it's present but a bit small at 1.5% (remember this is an atlas based on a _single_subject_). Ignoring the physiological asymmetry is in my view dangerous, especially when clinical decisions are at stake, as for example in epilepsy surgery (see Hammers et al. NI 2007 for a discussion with example).
The only instance when hippocampi would be expected to be symmetrical is when only _one_hemisphere_ is available and then mirrored, as in the Talairach atlas (and derivatives). But of course that carries no information about physiological asymmetry.

Hope this helps,

Best wishes,


Alexander Hammers, MD PhD

Chair in Functional Neuroimaging
Neurodis Foundation
Postal Address:
CERMEP – Imagerie du Vivant
Hôpital Neurologique Pierre Wertheimer
59 Boulevard Pinel, 69003 Lyon, France

Telephone +33-(0)4-72 68 86 34
Fax +33-(0)4-72 68 86 10
Email [log in to unmask];[log in to unmask]
Other affiliations:

Visiting Reader; Honorary Consultant Neurologist
Division of Neuroscience and Mental Health, Faculty of Medicine
Imperial College London, UK
Honorary Reader in Neurology; Honorary Consultant Neurologist
Department of Clinical and Experimental Epilepsy
National Hospital for Neurology and Neurosurgery/ Institute of Neurology, University College London, UK

On 21 Feb 2011, at 13:50, John Fredy wrote:

Thanks Alexander, Michael and Elliot for the reply, I will see the AAL atlas and the  maxprobmap more deeply. I am in the beggining in fmri in the hippocampus for 
epilepsy patients, so, I will have time to explore more and more in working with roi analysis.

Best wishes from Antioquia

On Sun, Feb 20, 2011 at 10:35 PM, G Elliott Wimmer <[log in to unmask]> wrote:

Regarding the AAL template (which is most likely what you have already seen in the WFU pickatlas), I too have noticed that the AAL hippocampus definition tries to only cover grey matter. This exclusion leads to gaps and holes in the mask that I think you are referring to.

Given the anatomical variability between subjects in the hippocampus, it appears to me that the AAL mask attempts to be far too precise; moreover, it is not left-right symmetric.  In the past, I have used AFNI's 'draw dataset' function to correct these problems with the AAL mask and used the resulting modified mask to define ROIs.

Alternatively, I've seen probabilistic anatomical masks that I think were used in FSL, and that may be worth checking out.

Best Regards,

G Elliott Wimmer

Dept. of Psychology
Columbia University