Dear all, the current module in FreeSurfer works with 1mm T1 data, by relying on strong shape priors. So, at this point, feeding the algorithm data from a 3T scanner or a 1.5T scanner is pretty much the same. Joshua, it is indeed inaccurate to say that the method relies on a generated hippocampal surface, but you are definitely right regarding the SLRM: it is not modeled at this point (we have a new version that models it coming out hopefully soon!). Cheers, /Eugenio
Juan Eugenio Iglesias Postdoctoral researcher BCBL www.jeiglesias.com www.bcbl.eu Legal disclaimer/Aviso legal/Lege-oharra: www.bcbl.eu/legal-disclaimer ----- Original Message ----- From: "Joshua Lee" <[email protected]> To: "Freesurfer support list" <[email protected]> Sent: Friday, April 25, 2014 2:10:29 AM Subject: Re: [Freesurfer] Hippocampal subfields on 1.5 Tesla Hi Alan, Typically subfields segmentation requires hi-resolution data (e.g. 0.4 x 0.4 mm in-plane resolution). The thickness of a CA subfield typically range between 0.5-1.00 mm, but 1.5 T data does not achieve sub-millimeter resolutions. Further, subfield segmentation typically requires high-contrast data to discern the internal boundaries formed by the stratum radiatum/stratum lacunosum-moleculare (SLRM). I doubt that images produced on a 1.5 T magnet can achieve the necessary contrast. Last, and please someone correct me if what I say is inaccurate, but doesn't the Van Leemput method use statistical priors to apply label probabilities in reference to a generated hippocampal surface? This would imply that the method assigns label probabilities without reference to a subject's SLRM intensity information. For volumetry, I am somewhat skeptical that a method that only relies on a generated surface would be sensitive to group x subfield interactions; especially double dissociations in which ove rall volume/shape of the hippocampus may be similar across groups. That the that was generated from potentially low resolution, low contrast data cannot help the matter. Some may disagree about this though and I'd be interested in hearing what other people think about the matter. In general, I am quite optimistic about automated methods to segment the subfields. Joshua - Joshua K. Lee Doctoral Candidate Department of Psychology & Center for Mind and Brain University of California, Davis On Thu, Apr 24, 2014 at 12:24 PM, Alan Francis < [email protected] > wrote: Hi Bruce and FreeSurfers: I have received a manuscript to review for possible publication. The authors have used the subfields algorithm on 1.5T scans and obtained a parcellation with values. They have drawn some major conclusions on the basis of the findings. My understanding is that this method can only be done on 3T. Is the 1.5T results valid? Please advice. thanks, Alan Francis _______________________________________________ Freesurfer mailing list [email protected] https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer The information in this e-mail is intended only for the person to whom it is addressed. If you believe this e-mail was sent to you in error and the e-mail contains patient information, please contact the Partners Compliance HelpLine at http://www.partners.org/complianceline . If the e-mail was sent to you in error but does not contain patient information, please contact the sender and properly dispose of the e-mail. _______________________________________________ Freesurfer mailing list [email protected] https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer The information in this e-mail is intended only for the person to whom it is addressed. If you believe this e-mail was sent to you in error and the e-mail contains patient information, please contact the Partners Compliance HelpLine at http://www.partners.org/complianceline . If the e-mail was sent to you in error but does not contain patient information, please contact the sender and properly dispose of the e-mail. _______________________________________________ Freesurfer mailing list [email protected] https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer
