Hi Akira,

These look mostly fine. The age hasn't been completely mean-centered (that is, it has a non-zero mean that isn't negligible). Other than that, everything seems fine.

The nuisance absorbs some of the variance that otherwise would to into the residuals of the GLM. Since this variance can be explained by the model, and the residuals become smaller, it is expected that the result for C1 (average laterality) become more significant, even if age alone fails to reach significance (there is also the multiple testing correction that makes any relationship less clear w.r.t. significance of the variables).

I'd say as done it's fine, just please, make sure that the age column as a zero mean.

All the best,

Anderson

On 2 August 2015 at 21:21, Yogi, Akira <[log in to unmask]> wrote:

Hi Anderson,

I am sorry for sending wrong files.

Please see the attached.

Best,

Akria Yogi

差出人: FSL - FMRIB's Software Library [[log in to unmask]] が Anderson M. Winkler [[log in to unmask]] の代理で送信しました
送信日時: 2015年8月2日 1:31

宛先: [log in to unmask]
件名: Re: [FSL] FW: question about the interpretation of one-sample t test with covariate

Hi Akira,

These aren't design matrices I'm afraid. The files that would be helpful to look are the design.mat and design.con that were used as inputs to randomise. Thanks.

All the best,

Anderson

On 31 July 2015 at 19:45, Yogi, Akira <[log in to unmask]> wrote:

Hi Anderson,

Thank you for your suggestion.

Please see the attache files of matrices.

i am sure that all the values fed into FEAT were demeaned: I subtracted the mean value from each subjects' original values.

Thanks,

Akira Yogi

差出人: FSL - FMRIB's Software Library [[log in to unmask]] が Anderson M. Winkler [[log in to unmask]] の代理で送信しました
送信日時: 2015年7月31日 0:37
宛先: [log in to unmask]
件名: Re: [FSL] FW: question about the interpretation of one-sample t test with covariate

Hi Akira,

Are you completely sure that age was mean-centered before being entered in the model? Please, feel free to attach the design.mat and design.con so we could have a look.

All the best,

Anderson

On 30 July 2015 at 17:03, Akira Yogi <[log in to unmask]> wrote:

Thank you for replying Anderson,

I see.
Let me make sure that I can get the average mean laterality (adjusted for age) and positive and negative relationship with age if I set C1 as [1, 0], C2 as [0, 1], and C3 as [0, -1].

I processed FEAT as above and moved randomisation following the notes in FEATwiki. Though many tracts showed FA laterality before adjusting for age, only a few voxels still showed FA laterality after adjusting for age ("tstat1.nii.gz"). I thought it means that age may have a somewhat great effect on FA lateraliy. But there were no positive or negative correlation ("tstat2.nii.gz", "tstat3.nii.gz"). It's so weird.

Please tell me if you know my mistakes.

Best regards,
Akira Yogi

IMPORTANT WARNING: This email (and any attachments) is only intended for the use of the person or entity to which it is addressed, and may contain information that is privileged and confidential. You, the recipient, are obligated to maintain it in a safe, secure and confidential manner. Unauthorized redisclosure or failure to maintain confidentiality may subject you to federal and state penalties. If you are not the intended recipient, please immediately notify us by return email, and delete this message from your computer.

--

Anderson M. Winkler

FMRIB / Analysis Group

[ Blog | Twitter ]

IMPORTANT WARNING: This email (and any attachments) is only intended for the use of the person or entity to which it is addressed, and may contain information that is privileged and confidential. You, the recipient, are obligated to maintain it in a safe, secure and confidential manner. Unauthorized redisclosure or failure to maintain confidentiality may subject you to federal and state penalties. If you are not the intended recipient, please immediately notify us by return email, and delete this message from your computer.

Anderson M. Winkler

FMRIB / Analysis Group

[ Blog | Twitter ]