Hi Aaron, The S900 Group Average subjects included all subjects released with 100% of rfMRI timepoints collected in the 4 rfMRI scans. This included 184 subjects with the data reconstructed with the r177 reconstruction algorithm and 636 subjects reconstructed with the r227 reconstruction algorithm.
Best, Jenn Jennifer Elam, Ph.D. Scientific Outreach, Human Connectome Project Washington University School of Medicine Department of Neuroscience, Box 8108 660 South Euclid Avenue St. Louis, MO 63110 314-362-9387<tel:314-362-9387> e...@wustl.edu<mailto:e...@wustl.edu> www.humanconnectome.org<http://www.humanconnectome.org/> ________________________________ From: hcp-users-boun...@humanconnectome.org <hcp-users-boun...@humanconnectome.org> on behalf of Aaron C <aaroncr...@outlook.com> Sent: Friday, January 27, 2017 4:25:52 PM To: hcp-users@humanconnectome.org Subject: Re: [HCP-Users] HCP-Users Digest, Vol 50, Issue 40 Dear HCP experts, I have a question about the different reconstruction algorithms of the rfMRI data Jennifer mentioned in her previous email. Did the S900 group-average dense connectome (820 subjects) mix the subjects with different rfMRI reconstruction algorithms? Thank you. ________________________________ From: hcp-users-boun...@humanconnectome.org <hcp-users-boun...@humanconnectome.org> on behalf of hcp-users-requ...@humanconnectome.org <hcp-users-requ...@humanconnectome.org> Sent: Friday, January 27, 2017 1:00:01 PM To: hcp-users@humanconnectome.org Subject: HCP-Users Digest, Vol 50, Issue 40 Send HCP-Users mailing list submissions to hcp-users@humanconnectome.org To subscribe or unsubscribe via the World Wide Web, visit http://lists.humanconnectome.org/mailman/listinfo/hcp-users or, via email, send a message with subject or body 'help' to hcp-users-requ...@humanconnectome.org You can reach the person managing the list at hcp-users-ow...@humanconnectome.org When replying, please edit your Subject line so it is more specific than "Re: Contents of HCP-Users digest..." Today's Topics: 1. Re: Unrelated subjects (Elam, Jennifer) 2. Very large z values for task contrasts in S900_ALLTASKS_level3_zstat file: what does this mean in terms of statistical significance? (Xavier Guell Paradis) 3. Re: Very large z values for task contrasts in S900_ALLTASKS_level3_zstat file: what does this mean in terms of statistical significance? (Glasser, Matthew) 4. Re: Very large z values for task contrasts in S900_ALLTASKS_level3_zstat file: what does this mean in terms of statistical significance? (David Van Essen) 5. Re: Very large z values for task contrasts in S900_ALLTASKS_level3_zstat file: what does this mean in terms of statistical significance? (Xavier Guell Paradis) 6. Re: Very large z values for task contrasts in S900_ALLTASKS_level3_zstat file: what does this mean in terms of statistical significance? (Glasser, Matthew) 7. Re: Very large z values for task contrasts in S900_ALLTASKS_level3_zstat file: what does this mean in terms of statistical significance? (Xavier Guell Paradis) ---------------------------------------------------------------------- Message: 1 Date: Thu, 26 Jan 2017 17:57:43 +0000 From: "Elam, Jennifer" <e...@wustl.edu> Subject: Re: [HCP-Users] Unrelated subjects To: "hcp-users@humanconnectome.org" <hcp-users@humanconnectome.org> Message-ID: <dm2pr0201mb0560fd78b36b157862b2aa9fcc...@dm2pr0201mb0560.namprd02.prod.outlook.com> Content-Type: text/plain; charset="windows-1252" It looks like the message I sent with the attached CSV of unrelated S900 subjects with complete rfMRI and physio data did not go through to the list (likely blocked due to the attachment). The original message is pasted below. If you would like that file please email me directly and I will send it on. Best, Jenn Hi Siobhan, Here is a csv spreadsheet we pulled together for another user that has 339 unrelated subjects from the S900 release with a T1, a T2, complete rfMRI, and physiological data. If you end up using the fMRI data for these subjects, please note that two different image reconstruction algorithms were used on this data (r177 early in the project, r227 later) and data from the different reconstructions shouldn't be mixed in an analysis. The structural data should be unaffected by this and all the diffusion data was reconstructed using the r227 algorithm. Best, Jenn Jennifer Elam, Ph.D. Scientific Outreach, Human Connectome Project Washington University School of Medicine Department of Neuroscience, Box 8108 660 South Euclid Avenue St. Louis, MO 63110 314-362-9387<tel:314-362-9387> e...@wustl.edu<mailto:e...@wustl.edu> www.humanconnectome.org<http://www.humanconnectome.org/<http://www.humanconnectome.org<http://www.humanconnectome.org/>> ________________________________ From: Elam, Jennifer Sent: Thursday, January 26, 2017 10:30 AM To: Ewert, Siobhan Geraldine; hcp-users@humanconnectome.org Subject: Re: Unrelated subjects I forgot to mention for everyone's benefit that you can find the reconstruction version for 3T dMRI and fMRI in the dMRI_3T_ReconVrs and fMRI_3T_ReconVrs columns in the CSV (these are columns W and X when viewed in Excel) of unrelated subjects attached to the previous message. These are also something you can filter for in ConnectomeDB to make sure you are using subjects with compatible data. Best, Jenn Jennifer Elam, Ph.D. Scientific Outreach, Human Connectome Project Washington University School of Medicine Department of Neuroscience, Box 8108 660 South Euclid Avenue St. Louis, MO 63110 314-362-9387<tel:314-362-9387> e...@wustl.edu<mailto:e...@wustl.edu> www.humanconnectome.org<http://www.humanconnectome.org/<http://www.humanconnectome.org<http://www.humanconnectome.org/>> ________________________________ From: Elam, Jennifer Sent: Thursday, January 26, 2017 10:12:08 AM To: Ewert, Siobhan Geraldine; hcp-users@humanconnectome.org Subject: Re: Unrelated subjects Hi Siobhan, Here is a csv spreadsheet we pulled together for another user that has 339 unrelated subjects from the S900 release with a T1, a T2, complete rfMRI, and physiological data. If you end up using the fMRI data for these subjects, please note that two different image reconstruction algorithms were used on this data (r177 early in the project, r227 later) and data from the different reconstructions shouldn't be mixed in an analysis. The structural data should be unaffected by this and all the diffusion data was reconstructed using the r227 algorithm. Best, Jenn Jennifer Elam, Ph.D. Scientific Outreach, Human Connectome Project Washington University School of Medicine Department of Neuroscience, Box 8108 660 South Euclid Avenue St. Louis, MO 63110 314-362-9387<tel:314-362-9387> e...@wustl.edu<mailto:e...@wustl.edu> www.humanconnectome.org<http://www.humanconnectome.org/<http://www.humanconnectome.org<http://www.humanconnectome.org/>> ________________________________ From: hcp-users-boun...@humanconnectome.org <hcp-users-boun...@humanconnectome.org> on behalf of Ewert, Siobhan Geraldine <sew...@mgh.harvard.edu> Sent: Thursday, January 26, 2017 9:36 AM To: hcp-users@humanconnectome.org Subject: [HCP-Users] Unrelated subjects Dear all, I?m looking for a large number of unrelated subjects with anonymised anatomical (T1 & T2) 3T brain scans from the connectome dataset. I found the option ?100 unrelated subjects? but would need a higher number than 100 (appr. 200-300). Is it possible to access more than 100 unrelated subjects without having to apply for restricted data? Your help is greatly appreciated. Thank you for your great work! Best, Siobhan 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. _______________________________________________ HCP-Users mailing list HCP-Users@humanconnectome.org http://lists.humanconnectome.org/mailman/listinfo/hcp-users -------------- next part -------------- An HTML attachment was scrubbed... URL: http://lists.humanconnectome.org/pipermail/hcp-users/attachments/20170126/0425752f/attachment-0001.html ------------------------------ Message: 2 Date: Thu, 26 Jan 2017 21:46:40 +0000 From: Xavier Guell Paradis <xavie...@mit.edu> Subject: [HCP-Users] Very large z values for task contrasts in S900_ALLTASKS_level3_zstat file: what does this mean in terms of statistical significance? To: "hcp-users@humanconnectome.org" <hcp-users@humanconnectome.org> Message-ID: <c754ed01015254479875671a9bf5527d2973b...@oc11expo32.exchange.mit.edu> Content-Type: text/plain; charset="iso-8859-1" Dear HCP team, I have seen that the zstat values for tasks contrasts are very large in the HCP_S900_787_tfMRI_ALLTASKS_level3_zstat1_hp200_s2_MSMAll.dscalar.nii file, to the point that one can observe areas of activation in task contrasts by setting very high z value thresholds (e.g., a z threshold of +14). I think (please correct me if I'm wrong) that the z values of the S900 file are very large because the group is very large, therefore the standard deviation is very small (given that there will be less variability in a group if one takes a very large group of people rather than a small group of people), and if the standard deviation is very small then even small differences from the mean will lead to very large z values. I was wondering what implication does this have in terms of statistical significance. A z value of 14 or larger would correspond to an extremely small p value, i.e. it would be extremely unlikely to observe by chance a measure which is 14 times the standard deviation away from the mean. Would it therefore be correct to assume that the areas that we can observe in the S900 tfMRI_ALLTASKS task contrasts with a very high zstat threshold (e.g., 14) are statistically significant, without having to worry about multiple comparisons or family structure? Thank you very much, Xavier. -------------- next part -------------- An HTML attachment was scrubbed... URL: http://lists.humanconnectome.org/pipermail/hcp-users/attachments/20170126/ff6ebf38/attachment-0001.html ------------------------------ Message: 3 Date: Thu, 26 Jan 2017 22:33:01 +0000 From: "Glasser, Matthew" <glass...@wustl.edu> Subject: Re: [HCP-Users] Very large z values for task contrasts in S900_ALLTASKS_level3_zstat file: what does this mean in terms of statistical significance? To: Xavier Guell Paradis <xavie...@mit.edu>, "hcp-users@humanconnectome.org" <hcp-users@humanconnectome.org> Message-ID: <d4afd044.135910%glass...@wustl.edu> Content-Type: text/plain; charset="iso-8859-1" Standard error scales with sample size, standard deviation does not. Things like Z, t, and p all also scale with sample size and are measures of statistical significance via various transformations. Thus for a large group of subjects, Z and t will be very high and p will be very low. Z, t and p are thus all not biologically interpretable, as their values also depend on the amount and quality of the data. In the limit with infinite amounts of data, the entire brain will be significant for any task, but wether a region is statistically significant tells us little about its importance functionally. Measures like appropriately scaled GLM regression betas, %BOLD change, or Cohen's d are biologically interpretable measures of effect size because their values should not change as sample size and data amount go up (rather the uncertainty on their estimates goes down). Regions with a large effect size in a task are likely important to that task (and will probably also meet crite ria for statistical significance given a reasonable amount of data). A common problem in neuroimaging studies is showing thresholded statistical significance maps rather than effect size maps (ideally unthresholded with an indication of which portions meet tests of statistical significance), and in general focusing on statistically significant blobs rather than the effect size in identifiable brain areas (which should often show stepwise changes in activity at their borders). Peace, Matt. From: <hcp-users-boun...@humanconnectome.org<mailto:hcp-users-boun...@humanconnectome.org>> on behalf of Xavier Guell Paradis <xavie...@mit.edu<mailto:xavie...@mit.edu>> Date: Thursday, January 26, 2017 at 3:46 PM To: "hcp-users@humanconnectome.org<mailto:hcp-users@humanconnectome.org>" <hcp-users@humanconnectome.org<mailto:hcp-users@humanconnectome.org>> Subject: [HCP-Users] Very large z values for task contrasts in S900_ALLTASKS_level3_zstat file: what does this mean in terms of statistical significance? Dear HCP team, I have seen that the zstat values for tasks contrasts are very large in the HCP_S900_787_tfMRI_ALLTASKS_level3_zstat1_hp200_s2_MSMAll.dscalar.nii file, to the point that one can observe areas of activation in task contrasts by setting very high z value thresholds (e.g., a z threshold of +14). I think (please correct me if I'm wrong) that the z values of the S900 file are very large because the group is very large, therefore the standard deviation is very small (given that there will be less variability in a group if one takes a very large group of people rather than a small group of people), and if the standard deviation is very small then even small differences from the mean will lead to very large z values. I was wondering what implication does this have in terms of statistical significance. A z value of 14 or larger would correspond to an extremely small p value, i.e. it would be extremely unlikely to observe by chance a measure which is 14 times the standard deviation away from the mean. Would it therefore be correct to assume that the areas that we can observe in the S900 tfMRI_ALLTASKS task contrasts with a very high zstat threshold (e.g., 14) are statistically significant, without having to worry about multiple comparisons or family structure? Thank you very much, Xavier. _______________________________________________ HCP-Users mailing list HCP-Users@humanconnectome.org<mailto:HCP-Users@humanconnectome.org> http://lists.humanconnectome.org/mailman/listinfo/hcp-users ________________________________ The materials in this message are private and may contain Protected Healthcare Information or other information of a sensitive nature. If you are not the intended recipient, be advised that any unauthorized use, disclosure, copying or the taking of any action in reliance on the contents of this information is strictly prohibited. If you have received this email in error, please immediately notify the sender via telephone or return mail. -------------- next part -------------- An HTML attachment was scrubbed... URL: http://lists.humanconnectome.org/pipermail/hcp-users/attachments/20170126/5c140efa/attachment-0001.html ------------------------------ Message: 4 Date: Thu, 26 Jan 2017 17:39:44 -0600 From: David Van Essen <vanes...@wustl.edu> Subject: Re: [HCP-Users] Very large z values for task contrasts in S900_ALLTASKS_level3_zstat file: what does this mean in terms of statistical significance? To: Xavier Guell Paradis <xavie...@mit.edu> Cc: "hcp-users@humanconnectome.org" <hcp-users@humanconnectome.org> Message-ID: <7a36516d-7846-4283-982f-eeded7571...@wustl.edu> Content-Type: text/plain; charset="utf-8" Xavier et al., An important secondary factor is the improved intersubject alignment provided by ?MSMAll? alignment (based on areal features) over folding-based alignment (MSMSulc or fsaverage). Having better alignment of functionally defined areas increases effect sizes (and z-stats as well), making the type of analysis Matt recommends more robust whatever the number of subjects in a particular analysis. These and a variety of other recommendations for "HCP-style? analysis are discussed more fully in Glasser et al., (Nature Neuroscience, 2016; https://www.ncbi.nlm.nih.gov/pubmed/27571196). David > On Jan 26, 2017, at 4:33 PM, Glasser, Matthew <glass...@wustl.edu> wrote: > > Standard error scales with sample size, standard deviation does not. Things > like Z, t, and p all also scale with sample size and are measures of > statistical significance via various transformations. Thus for a large group > of subjects, Z and t will be very high and p will be very low. Z, t and p > are thus all not biologically interpretable, as their values also depend on > the amount and quality of the data. In the limit with infinite amounts of > data, the entire brain will be significant for any task, but wether a region > is statistically significant tells us little about its importance > functionally. Measures like appropriately scaled GLM regression betas, %BOLD > change, or Cohen?s d are biologically interpretable measures of effect size > because their values should not change as sample size and data amount go up > (rather the uncertainty on their estimates goes down). Regions with a large > effect size in a task are likely important to that task (and will probably > also meet cri teria for statistical significance given a reasonable amount of data). > > A common problem in neuroimaging studies is showing thresholded statistical > significance maps rather than effect size maps (ideally unthresholded with an > indication of which portions meet tests of statistical significance), and in > general focusing on statistically significant blobs rather than the effect > size in identifiable brain areas (which should often show stepwise changes in > activity at their borders). > > Peace, > > Matt. > > From: <hcp-users-boun...@humanconnectome.org > <mailto:hcp-users-boun...@humanconnectome.org>> on behalf of Xavier Guell > Paradis <xavie...@mit.edu <mailto:xavie...@mit.edu>> > Date: Thursday, January 26, 2017 at 3:46 PM > To: "hcp-users@humanconnectome.org <mailto:hcp-users@humanconnectome.org>" > <hcp-users@humanconnectome.org <mailto:hcp-users@humanconnectome.org>> > Subject: [HCP-Users] Very large z values for task contrasts in > S900_ALLTASKS_level3_zstat file: what does this mean in terms of statistical > significance? > > Dear HCP team, > I have seen that the zstat values for tasks contrasts are very large in the > HCP_S900_787_tfMRI_ALLTASKS_level3_zstat1_hp200_s2_MSMAll.dscalar.nii file, > to the point that one can observe areas of activation in task contrasts by > setting very high z value thresholds (e.g., a z threshold of +14). > I think (please correct me if I'm wrong) that the z values of the S900 file > are very large because the group is very large, therefore the standard > deviation is very small (given that there will be less variability in a group > if one takes a very large group of people rather than a small group of > people), and if the standard deviation is very small then even small > differences from the mean will lead to very large z values. > > I was wondering what implication does this have in terms of statistical > significance. A z value of 14 or larger would correspond to an extremely > small p value, i.e. it would be extremely unlikely to observe by chance a > measure which is 14 times the standard deviation away from the mean. Would it > therefore be correct to assume that the areas that we can observe in the S900 > tfMRI_ALLTASKS task contrasts with a very high zstat threshold (e.g., 14) are > statistically significant, without having to worry about multiple comparisons > or family structure? > > Thank you very much, > Xavier. > _______________________________________________ > HCP-Users mailing list > HCP-Users@humanconnectome.org <mailto:HCP-Users@humanconnectome.org> > http://lists.humanconnectome.org/mailman/listinfo/hcp-users > <http://lists.humanconnectome.org/mailman/listinfo/hcp-users> > > The materials in this message are private and may contain Protected > Healthcare Information or other information of a sensitive nature. If you are > not the intended recipient, be advised that any unauthorized use, disclosure, > copying or the taking of any action in reliance on the contents of this > information is strictly prohibited. If you have received this email in error, > please immediately notify the sender via telephone or return mail. > _______________________________________________ > HCP-Users mailing list > HCP-Users@humanconnectome.org <mailto:HCP-Users@humanconnectome.org> > http://lists.humanconnectome.org/mailman/listinfo/hcp-users > <http://lists.humanconnectome.org/mailman/listinfo/hcp-users> -------------- next part -------------- An HTML attachment was scrubbed... URL: http://lists.humanconnectome.org/pipermail/hcp-users/attachments/20170126/cd30d47f/attachment-0001.html ------------------------------ Message: 5 Date: Thu, 26 Jan 2017 23:41:54 +0000 From: Xavier Guell Paradis <xavie...@mit.edu> Subject: Re: [HCP-Users] Very large z values for task contrasts in S900_ALLTASKS_level3_zstat file: what does this mean in terms of statistical significance? To: "Glasser, Matthew" <glass...@wustl.edu>, "hcp-users@humanconnectome.org" <hcp-users@humanconnectome.org> Message-ID: <c754ed01015254479875671a9bf5527d2973b...@oc11expo32.exchange.mit.edu> Content-Type: text/plain; charset="windows-1252" Dear Matt, Thank you very much for your very helpful reply. I will have to investigate this topic more, but is there any approach you would suggest to obtain effect size maps from the S900 group HCP data? I was wondering if the zstat data of the S900 group task contrasts could be converted to effect size values without having to go back to the individual subjects. Thank you very much, Xavier. ________________________________ From: Glasser, Matthew [glass...@wustl.edu] Sent: Thursday, January 26, 2017 5:33 PM To: Xavier Guell Paradis; hcp-users@humanconnectome.org Subject: Re: [HCP-Users] Very large z values for task contrasts in S900_ALLTASKS_level3_zstat file: what does this mean in terms of statistical significance? Standard error scales with sample size, standard deviation does not. Things like Z, t, and p all also scale with sample size and are measures of statistical significance via various transformations. Thus for a large group of subjects, Z and t will be very high and p will be very low. Z, t and p are thus all not biologically interpretable, as their values also depend on the amount and quality of the data. In the limit with infinite amounts of data, the entire brain will be significant for any task, but wether a region is statistically significant tells us little about its importance functionally. Measures like appropriately scaled GLM regression betas, %BOLD change, or Cohen?s d are biologically interpretable measures of effect size because their values should not change as sample size and data amount go up (rather the uncertainty on their estimates goes down). Regions with a large effect size in a task are likely important to that task (and will probably also meet crite ria for statistical significance given a reasonable amount of data). A common problem in neuroimaging studies is showing thresholded statistical significance maps rather than effect size maps (ideally unthresholded with an indication of which portions meet tests of statistical significance), and in general focusing on statistically significant blobs rather than the effect size in identifiable brain areas (which should often show stepwise changes in activity at their borders). Peace, Matt. From: <hcp-users-boun...@humanconnectome.org<mailto:hcp-users-boun...@humanconnectome.org>> on behalf of Xavier Guell Paradis <xavie...@mit.edu<mailto:xavie...@mit.edu>> Date: Thursday, January 26, 2017 at 3:46 PM To: "hcp-users@humanconnectome.org<mailto:hcp-users@humanconnectome.org>" <hcp-users@humanconnectome.org<mailto:hcp-users@humanconnectome.org>> Subject: [HCP-Users] Very large z values for task contrasts in S900_ALLTASKS_level3_zstat file: what does this mean in terms of statistical significance? Dear HCP team, I have seen that the zstat values for tasks contrasts are very large in the HCP_S900_787_tfMRI_ALLTASKS_level3_zstat1_hp200_s2_MSMAll.dscalar.nii file, to the point that one can observe areas of activation in task contrasts by setting very high z value thresholds (e.g., a z threshold of +14). I think (please correct me if I'm wrong) that the z values of the S900 file are very large because the group is very large, therefore the standard deviation is very small (given that there will be less variability in a group if one takes a very large group of people rather than a small group of people), and if the standard deviation is very small then even small differences from the mean will lead to very large z values. I was wondering what implication does this have in terms of statistical significance. A z value of 14 or larger would correspond to an extremely small p value, i.e. it would be extremely unlikely to observe by chance a measure which is 14 times the standard deviation away from the mean. Would it therefore be correct to assume that the areas that we can observe in the S900 tfMRI_ALLTASKS task contrasts with a very high zstat threshold (e.g., 14) are statistically significant, without having to worry about multiple comparisons or family structure? Thank you very much, Xavier. _______________________________________________ HCP-Users mailing list HCP-Users@humanconnectome.org<mailto:HCP-Users@humanconnectome.org> http://lists.humanconnectome.org/mailman/listinfo/hcp-users ________________________________ The materials in this message are private and may contain Protected Healthcare Information or other information of a sensitive nature. If you are not the intended recipient, be advised that any unauthorized use, disclosure, copying or the taking of any action in reliance on the contents of this information is strictly prohibited. If you have received this email in error, please immediately notify the sender via telephone or return mail. -------------- next part -------------- An HTML attachment was scrubbed... URL: http://lists.humanconnectome.org/pipermail/hcp-users/attachments/20170126/fa3047f1/attachment-0001.html ------------------------------ Message: 6 Date: Thu, 26 Jan 2017 23:53:05 +0000 From: "Glasser, Matthew" <glass...@wustl.edu> Subject: Re: [HCP-Users] Very large z values for task contrasts in S900_ALLTASKS_level3_zstat file: what does this mean in terms of statistical significance? To: Xavier Guell Paradis <xavie...@mit.edu>, "hcp-users@humanconnectome.org" <hcp-users@humanconnectome.org> Message-ID: <d4afe7a9.1359a3%glass...@wustl.edu> Content-Type: text/plain; charset="windows-1252" The files called cope1 or beta are an effect size measure, however the released versions are not optimally scaled (because of a non-optimal intensity bias field correction). We plan to correct this in the future. Peace, Matt. From: Xavier Guell Paradis <xavie...@mit.edu<mailto:xavie...@mit.edu>> Date: Thursday, January 26, 2017 at 5:41 PM To: Matt Glasser <glass...@wustl.edu<mailto:glass...@wustl.edu>>, "hcp-users@humanconnectome.org<mailto:hcp-users@humanconnectome.org>" <hcp-users@humanconnectome.org<mailto:hcp-users@humanconnectome.org>> Subject: RE: [HCP-Users] Very large z values for task contrasts in S900_ALLTASKS_level3_zstat file: what does this mean in terms of statistical significance? Dear Matt, Thank you very much for your very helpful reply. I will have to investigate this topic more, but is there any approach you would suggest to obtain effect size maps from the S900 group HCP data? I was wondering if the zstat data of the S900 group task contrasts could be converted to effect size values without having to go back to the individual subjects. Thank you very much, Xavier. ________________________________ From: Glasser, Matthew [glass...@wustl.edu<mailto:glass...@wustl.edu>] Sent: Thursday, January 26, 2017 5:33 PM To: Xavier Guell Paradis; hcp-users@humanconnectome.org<mailto:hcp-users@humanconnectome.org> Subject: Re: [HCP-Users] Very large z values for task contrasts in S900_ALLTASKS_level3_zstat file: what does this mean in terms of statistical significance? Standard error scales with sample size, standard deviation does not. Things like Z, t, and p all also scale with sample size and are measures of statistical significance via various transformations. Thus for a large group of subjects, Z and t will be very high and p will be very low. Z, t and p are thus all not biologically interpretable, as their values also depend on the amount and quality of the data. In the limit with infinite amounts of data, the entire brain will be significant for any task, but wether a region is statistically significant tells us little about its importance functionally. Measures like appropriately scaled GLM regression betas, %BOLD change, or Cohen?s d are biologically interpretable measures of effect size because their values should not change as sample size and data amount go up (rather the uncertainty on their estimates goes down). Regions with a large effect size in a task are likely important to that task (and will probably also meet crite ria for statistical significance given a reasonable amount of data). A common problem in neuroimaging studies is showing thresholded statistical significance maps rather than effect size maps (ideally unthresholded with an indication of which portions meet tests of statistical significance), and in general focusing on statistically significant blobs rather than the effect size in identifiable brain areas (which should often show stepwise changes in activity at their borders). Peace, Matt. From: <hcp-users-boun...@humanconnectome.org<mailto:hcp-users-boun...@humanconnectome.org>> on behalf of Xavier Guell Paradis <xavie...@mit.edu<mailto:xavie...@mit.edu>> Date: Thursday, January 26, 2017 at 3:46 PM To: "hcp-users@humanconnectome.org<mailto:hcp-users@humanconnectome.org>" <hcp-users@humanconnectome.org<mailto:hcp-users@humanconnectome.org>> Subject: [HCP-Users] Very large z values for task contrasts in S900_ALLTASKS_level3_zstat file: what does this mean in terms of statistical significance? Dear HCP team, I have seen that the zstat values for tasks contrasts are very large in the HCP_S900_787_tfMRI_ALLTASKS_level3_zstat1_hp200_s2_MSMAll.dscalar.nii file, to the point that one can observe areas of activation in task contrasts by setting very high z value thresholds (e.g., a z threshold of +14). I think (please correct me if I'm wrong) that the z values of the S900 file are very large because the group is very large, therefore the standard deviation is very small (given that there will be less variability in a group if one takes a very large group of people rather than a small group of people), and if the standard deviation is very small then even small differences from the mean will lead to very large z values. I was wondering what implication does this have in terms of statistical significance. A z value of 14 or larger would correspond to an extremely small p value, i.e. it would be extremely unlikely to observe by chance a measure which is 14 times the standard deviation away from the mean. Would it therefore be correct to assume that the areas that we can observe in the S900 tfMRI_ALLTASKS task contrasts with a very high zstat threshold (e.g., 14) are statistically significant, without having to worry about multiple comparisons or family structure? Thank you very much, Xavier. _______________________________________________ HCP-Users mailing list HCP-Users@humanconnectome.org<mailto:HCP-Users@humanconnectome.org> http://lists.humanconnectome.org/mailman/listinfo/hcp-users ________________________________ The materials in this message are private and may contain Protected Healthcare Information or other information of a sensitive nature. If you are not the intended recipient, be advised that any unauthorized use, disclosure, copying or the taking of any action in reliance on the contents of this information is strictly prohibited. If you have received this email in error, please immediately notify the sender via telephone or return mail. ________________________________ The materials in this message are private and may contain Protected Healthcare Information or other information of a sensitive nature. If you are not the intended recipient, be advised that any unauthorized use, disclosure, copying or the taking of any action in reliance on the contents of this information is strictly prohibited. If you have received this email in error, please immediately notify the sender via telephone or return mail. -------------- next part -------------- An HTML attachment was scrubbed... URL: http://lists.humanconnectome.org/pipermail/hcp-users/attachments/20170126/904699cb/attachment-0001.html ------------------------------ Message: 7 Date: Fri, 27 Jan 2017 16:04:03 +0000 From: Xavier Guell Paradis <xavie...@mit.edu> Subject: Re: [HCP-Users] Very large z values for task contrasts in S900_ALLTASKS_level3_zstat file: what does this mean in terms of statistical significance? To: "Glasser, Matthew" <glass...@wustl.edu>, "hcp-users@humanconnectome.org" <hcp-users@humanconnectome.org> Message-ID: <c754ed01015254479875671a9bf5527d2973b...@oc11expo32.exchange.mit.edu> Content-Type: text/plain; charset="windows-1252" Dear Matt, Thank you again for your reply. I have been able to find cope1 files for single subject task contrasts (e.g. cope1 file for working memory contrasts of subject 996782), but not for the S900 group (e.g. I have not been able to find a cope1 file for the S900 group for working memory contrasts). I was wondering: a) Is there any task contrast effect size map available for the S900 group? (even if they are not optimally scaled) b) If not, would it be possible to generate a task contrast effect size map by using available S900 group data (e.g. the task contrasts zstat maps of the S900 group), or would it be necessary to go back to the data of each individual subject? c) If it is necessary to go back to the data of each individual subject, which approach would you suggest to combine all cope1 files of each subject of the S900 group into one effect size map of all subjects? Would something like normalizing the cope1 file of each subject (using wb_command as written below) and then averaging all normalized cope1 files work? Or would something as simple as averaging all cope1 files work? wb_command -cifti-reduce <input> MEAN mean.dtseries.nii wb_command -cifti-reduce <input> STDEV stdev.dtseries.nii wb_command -cifti-math '(x - mean) / stdev' <output> -fixnan 0 -var x <input> -var mean mean.dtseries.nii -select 1 1 -repeat -var stdev stdev.dtseries.nii -select 1 1 -repeat Thank you very much, Xavier. ________________________________ From: Glasser, Matthew [glass...@wustl.edu] Sent: Thursday, January 26, 2017 6:53 PM To: Xavier Guell Paradis; hcp-users@humanconnectome.org Subject: Re: [HCP-Users] Very large z values for task contrasts in S900_ALLTASKS_level3_zstat file: what does this mean in terms of statistical significance? The files called cope1 or beta are an effect size measure, however the released versions are not optimally scaled (because of a non-optimal intensity bias field correction). We plan to correct this in the future. Peace, Matt. From: Xavier Guell Paradis <xavie...@mit.edu<mailto:xavie...@mit.edu>> Date: Thursday, January 26, 2017 at 5:41 PM To: Matt Glasser <glass...@wustl.edu<mailto:glass...@wustl.edu>>, "hcp-users@humanconnectome.org<mailto:hcp-users@humanconnectome.org>" <hcp-users@humanconnectome.org<mailto:hcp-users@humanconnectome.org>> Subject: RE: [HCP-Users] Very large z values for task contrasts in S900_ALLTASKS_level3_zstat file: what does this mean in terms of statistical significance? Dear Matt, Thank you very much for your very helpful reply. I will have to investigate this topic more, but is there any approach you would suggest to obtain effect size maps from the S900 group HCP data? I was wondering if the zstat data of the S900 group task contrasts could be converted to effect size values without having to go back to the individual subjects. Thank you very much, Xavier. ________________________________ From: Glasser, Matthew [glass...@wustl.edu<mailto:glass...@wustl.edu>] Sent: Thursday, January 26, 2017 5:33 PM To: Xavier Guell Paradis; hcp-users@humanconnectome.org<mailto:hcp-users@humanconnectome.org> Subject: Re: [HCP-Users] Very large z values for task contrasts in S900_ALLTASKS_level3_zstat file: what does this mean in terms of statistical significance? Standard error scales with sample size, standard deviation does not. Things like Z, t, and p all also scale with sample size and are measures of statistical significance via various transformations. Thus for a large group of subjects, Z and t will be very high and p will be very low. Z, t and p are thus all not biologically interpretable, as their values also depend on the amount and quality of the data. In the limit with infinite amounts of data, the entire brain will be significant for any task, but wether a region is statistically significant tells us little about its importance functionally. Measures like appropriately scaled GLM regression betas, %BOLD change, or Cohen?s d are biologically interpretable measures of effect size because their values should not change as sample size and data amount go up (rather the uncertainty on their estimates goes down). Regions with a large effect size in a task are likely important to that task (and will probably also meet crite ria for statistical significance given a reasonable amount of data). A common problem in neuroimaging studies is showing thresholded statistical significance maps rather than effect size maps (ideally unthresholded with an indication of which portions meet tests of statistical significance), and in general focusing on statistically significant blobs rather than the effect size in identifiable brain areas (which should often show stepwise changes in activity at their borders). Peace, Matt. From: <hcp-users-boun...@humanconnectome.org<mailto:hcp-users-boun...@humanconnectome.org>> on behalf of Xavier Guell Paradis <xavie...@mit.edu<mailto:xavie...@mit.edu>> Date: Thursday, January 26, 2017 at 3:46 PM To: "hcp-users@humanconnectome.org<mailto:hcp-users@humanconnectome.org>" <hcp-users@humanconnectome.org<mailto:hcp-users@humanconnectome.org>> Subject: [HCP-Users] Very large z values for task contrasts in S900_ALLTASKS_level3_zstat file: what does this mean in terms of statistical significance? Dear HCP team, I have seen that the zstat values for tasks contrasts are very large in the HCP_S900_787_tfMRI_ALLTASKS_level3_zstat1_hp200_s2_MSMAll.dscalar.nii file, to the point that one can observe areas of activation in task contrasts by setting very high z value thresholds (e.g., a z threshold of +14). I think (please correct me if I'm wrong) that the z values of the S900 file are very large because the group is very large, therefore the standard deviation is very small (given that there will be less variability in a group if one takes a very large group of people rather than a small group of people), and if the standard deviation is very small then even small differences from the mean will lead to very large z values. I was wondering what implication does this have in terms of statistical significance. A z value of 14 or larger would correspond to an extremely small p value, i.e. it would be extremely unlikely to observe by chance a measure which is 14 times the standard deviation away from the mean. Would it therefore be correct to assume that the areas that we can observe in the S900 tfMRI_ALLTASKS task contrasts with a very high zstat threshold (e.g., 14) are statistically significant, without having to worry about multiple comparisons or family structure? Thank you very much, Xavier. _______________________________________________ HCP-Users mailing list HCP-Users@humanconnectome.org<mailto:HCP-Users@humanconnectome.org> http://lists.humanconnectome.org/mailman/listinfo/hcp-users ________________________________ The materials in this message are private and may contain Protected Healthcare Information or other information of a sensitive nature. If you are not the intended recipient, be advised that any unauthorized use, disclosure, copying or the taking of any action in reliance on the contents of this information is strictly prohibited. If you have received this email in error, please immediately notify the sender via telephone or return mail. ________________________________ The materials in this message are private and may contain Protected Healthcare Information or other information of a sensitive nature. If you are not the intended recipient, be advised that any unauthorized use, disclosure, copying or the taking of any action in reliance on the contents of this information is strictly prohibited. If you have received this email in error, please immediately notify the sender via telephone or return mail. -------------- next part -------------- An HTML attachment was scrubbed... URL: http://lists.humanconnectome.org/pipermail/hcp-users/attachments/20170127/c50f76f1/attachment-0001.html ------------------------------ _______________________________________________ HCP-Users mailing list HCP-Users@humanconnectome.org http://lists.humanconnectome.org/mailman/listinfo/hcp-users End of HCP-Users Digest, Vol 50, Issue 40 ***************************************** _______________________________________________ HCP-Users mailing list HCP-Users@humanconnectome.org http://lists.humanconnectome.org/mailman/listinfo/hcp-users _______________________________________________ HCP-Users mailing list HCP-Users@humanconnectome.org http://lists.humanconnectome.org/mailman/listinfo/hcp-users
