Hi Bruce, What if she created an FSGD file, and instead of just having "patients" and "controls", she would have "patients1.5", "patients3", etc.? With over 400 subjects this should be feasible, i.e. the loss of d.f. is worth running the combined analysis.

Chris ________________________________________ From: freesurfer-bounces@nmr.mgh.harvard.edu [freesurfer-bounces@nmr.mgh.harvard.edu] on behalf of Bruce Fischl [fischl@nmr.mgh.harvard.edu] Sent: Monday, February 18, 2013 10:59 AM To: Jorge Jovicich Cc: Sinead Kelly; freesurfer@nmr.mgh.harvard.edu Subject: Re: [Freesurfer] Cortical thickness analysis on data acquired from multiple sites

Hi Sinead

I agree with Jorge - there is bound to be a substantial scanner effect. You might be better off keeping the data separate and treating the 3T as a confirmatory study.

cheers Bruce

On Mon, 18 Feb 2013, Jorge Jovicich wrote:

Dear Sinead,

we found global significant differences in thickness between 1.5T and 3T, in a group of subjects that was scanned at both scanners (http://www.ncbi.nlm.nih.gov/pubmed/16651008). I think that nothing stops you from doing the analysis, but maybe model in a field effect to asset it in your own data.

Cheers,

jorge

On 18/02/2013 16:32, Sinead Kelly wrote:

...

Dear members,

I would like to get your opinion on this issue - I have a dataset of over 400 subjects but under half of this data was acquired on a 1.5T scanner and the rest was acquired on a 3T scanner. Would it be acceptable to conduct cortical thickness analysis on the combined dataset? From reading the literature it seems that this is not a major problem but I just wanted to get some more thoughts on this.

Thank you for your help,

Sinead

-- Sinead Kelly Neuropsychiatric Genetics Group Trinity Centre St. James's Hospital Dublin 8

_______________________________________________ Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu 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.

Thank you all so much for your advice on this - The patient:control ratio is disproportionate in both datasets and I will also be splitting the subjects into genotype groups so I think I may consider running separate analyses on the two datasets to see if we find similar results.

Thanks again for your help!

Sinead

On 18 February 2013 16:54, Bruce Fischl fischl@nmr.mgh.harvard.edu wrote:

Hi Chris

you could do that but the increased variance may make it less. You also could have a disease interaction such as changing tissue properties having different effects at 1.5T and 3T. With that many subjects I would think the more powerful statement is if you find maps that are similar when splitting the data into 1.5T and 3T samples.

cheers Bruce

On Mon, 18 Feb 2013, Watson, Christopher wrote:

Hi Bruce,

...

What if she created an FSGD file, and instead of just having "patients" and "controls", she would have "patients1.5", "patients3", etc.? With over 400 subjects this should be feasible, i.e. the loss of d.f. is worth running the combined analysis.

Chris ______________________________**__________ From: freesurfer-bounces@nmr.mgh.**harvard.edufreesurfer-bounces@nmr.mgh.harvard.edu[ freesurfer-bounces@nmr.mgh.**harvard.edufreesurfer-bounces@nmr.mgh.harvard.edu] on behalf of Bruce Fischl [fischl@nmr.mgh.harvard.edu] Sent: Monday, February 18, 2013 10:59 AM To: Jorge Jovicich Cc: Sinead Kelly; freesurfer@nmr.mgh.harvard.edu Subject: Re: [Freesurfer] Cortical thickness analysis on data acquired from multiple sites

Hi Sinead

I agree with Jorge - there is bound to be a substantial scanner effect. You might be better off keeping the data separate and treating the 3T as a confirmatory study.

cheers Bruce

On Mon, 18 Feb 2013, Jorge Jovicich wrote:

Dear Sinead,

...

we found global significant differences in thickness between 1.5T and 3T, in a group of subjects that was scanned at both scanners (http://www.ncbi.nlm.nih.gov/**pubmed/16651008http://www.ncbi.nlm.nih.gov/pubmed/16651008). I think that nothing stops you from doing the analysis, but maybe model in a field effect to asset it in your own data.

Cheers,

jorge

On 18/02/2013 16:32, Sinead Kelly wrote:

...

Dear members,

I would like to get your opinion on this issue - I have a dataset of over 400 subjects but under half of this data was acquired on a 1.5T scanner and the rest was acquired on a 3T scanner. Would it be acceptable to conduct cortical thickness analysis on the combined dataset? From reading the literature it seems that this is not a major problem but I just wanted to get some more thoughts on this.

Thank you for your help,

Sinead

-- Sinead Kelly Neuropsychiatric Genetics Group Trinity Centre St. James's Hospital Dublin 8

______________________________**_________________

Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.**edu/mailman/listinfo/**freesurferhttps://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/**compliancelinehttp://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.