Qoala-T / QC

Qoala-T is a supervised-learning tool for quality control of FreeSurfer segmented MRI data
Other
72 stars 15 forks source link

Freesurfer 7.1 ? #30

Open jlhanson5 opened 4 years ago

jlhanson5 commented 4 years ago

Hello Qoala-T repo,

My research group was interested in deploying Qoala-T in a decent size sample of subjects (~2K) that had completed Freesurfer 7.1. I wondered how well Qoala-T performs with that new FS version? Anything to watch out for, etc.? I thought I would get some thoughts before we dug too deeply into things. Any thoughts are deeply appreciated!

Thanks much, Jamie.

eduardklap commented 4 years ago

Hi Jamie,

Thanks for posting, sounds great! We are currently working on it, planning to test how well it works with the new FS version. Until we have done that thoroughly I would recommend to use the subset-based approach and not the model-based approach, such that the the training and validation sets are run using the same FS version. Given the size of your sample, training on a subset would be the best strategy anyway.

Will post our findings in the repo and in this issue as well.

Best, Eduard

vims1 commented 2 years ago

Hi,

Would love to use your software, however, does the reply to Jamie entail that the .txt-files from newer fs is not compatible with the current Qoala-T? I am receiving an error message when uploading - might the fs version (7.1) be underlying the issue?

larawierenga commented 2 years ago

HI there,

Yes there is a newer version for fs 710: https://github.com/Qoala-T/QC/blob/master/Scripts/Stats2Table/Stats2Table_fs7.R We compared this to the fs v600 outcomes, and showed that results are similar, but do note that sensitivity is a little lower and shows larger variation. You can find more on the results in the read me file here https://github.com/Qoala-T/QC#validation-of-qoala-t-tool-in-freesurfer-v710

Best, Lara

On 17 Nov 2021, at 19:06, vims1 @.**@.>> wrote:

Hi,

Would love to use your software, however, does the reply to Jamie entail that the .txt-files from newer fs is not compatible with the current Qoala-T? I am receiving an error message when uploading - might the fs version (7.1) be underlying the issue?

— You are receiving this because you commented. Reply to this email directly, view it on GitHubhttps://eur03.safelinks.protection.outlook.com/?url=https%3A%2F%2Fgithub.com%2FQoala-T%2FQC%2Fissues%2F30%23issuecomment-971828253&data=04%7C01%7Cl.m.wierenga%40fsw.leidenuniv.nl%7C75f5624ff1e4477c1e5e08d9a9f4f926%7Cca2a7f76dbd74ec091086b3d524fb7c8%7C0%7C0%7C637727691923125136%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C3000&sdata=40TSkFm7S6Z89m0ne64G8jQoMCwT4J482bpwGvHBDes%3D&reserved=0, or unsubscribehttps://eur03.safelinks.protection.outlook.com/?url=https%3A%2F%2Fgithub.com%2Fnotifications%2Funsubscribe-auth%2FAITGLADOXOD44MP5MWURK4TUMPVJ5ANCNFSM4NZSXJBQ&data=04%7C01%7Cl.m.wierenga%40fsw.leidenuniv.nl%7C75f5624ff1e4477c1e5e08d9a9f4f926%7Cca2a7f76dbd74ec091086b3d524fb7c8%7C0%7C0%7C637727691923135126%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C3000&sdata=0PhKvVwNCjZHn0pz9djhhgpjHmIwoQzZd%2BDKRCqZfiE%3D&reserved=0. Triage notifications on the go with GitHub Mobile for iOShttps://eur03.safelinks.protection.outlook.com/?url=https%3A%2F%2Fapps.apple.com%2Fapp%2Fapple-store%2Fid1477376905%3Fct%3Dnotification-email%26mt%3D8%26pt%3D524675&data=04%7C01%7Cl.m.wierenga%40fsw.leidenuniv.nl%7C75f5624ff1e4477c1e5e08d9a9f4f926%7Cca2a7f76dbd74ec091086b3d524fb7c8%7C0%7C0%7C637727691923145123%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C3000&sdata=XYHAhbmbp1%2F5HNnvBm8UsXMqDuUBohiGhZNzcHVtdQg%3D&reserved=0 or Androidhttps://eur03.safelinks.protection.outlook.com/?url=https%3A%2F%2Fplay.google.com%2Fstore%2Fapps%2Fdetails%3Fid%3Dcom.github.android%26referrer%3Dutm_campaign%253Dnotification-email%2526utm_medium%253Demail%2526utm_source%253Dgithub&data=04%7C01%7Cl.m.wierenga%40fsw.leidenuniv.nl%7C75f5624ff1e4477c1e5e08d9a9f4f926%7Cca2a7f76dbd74ec091086b3d524fb7c8%7C0%7C0%7C637727691923145123%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C3000&sdata=lXkU%2Fvsh6QEhbaeA6lC7C6XM81D38VPA8PPZItwePqE%3D&reserved=0.

ealyoubi commented 1 year ago

Hello, I would like to try the Qoala-T tool on my dataset, but I am wondering if the tool is compatible with data processed in Freesufer (V7.2) ?

Thanks in advance Esraa

larawierenga commented 1 year ago

HI Esraa,

Thank you for your interest in our tool. Unfortunately there is not an update version for FS 7.2 at the moment. We are working on funding to to do so. But for now we would recommend to use vs 710 for this tool.

All the best, Lara

On 8 Oct 2022, at 22:13, Esraa Alyoubi @.**@.>> wrote:

Hello, I would like to try the Qoala-T tool on my dataset, but I am wondering if the tool is compatible with data processed in Freesufer (V7.2) ?

Thanks in advance Esraa

— Reply to this email directly, view it on GitHubhttps://eur03.safelinks.protection.outlook.com/?url=https%3A%2F%2Fgithub.com%2FQoala-T%2FQC%2Fissues%2F30%23issuecomment-1272390358&data=05%7C01%7Cl.m.wierenga%40fsw.leidenuniv.nl%7C9fb300554cd6478fd78108daa9699337%7Cca2a7f76dbd74ec091086b3d524fb7c8%7C0%7C0%7C638008568145228603%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C3000%7C%7C%7C&sdata=UTqtLCylvlL%2BOqOLJcXjWxd4CE%2F2ZutJLl%2BV4Htok9g%3D&reserved=0, or unsubscribehttps://eur03.safelinks.protection.outlook.com/?url=https%3A%2F%2Fgithub.com%2Fnotifications%2Funsubscribe-auth%2FAITGLAFPVCF75XIGKMNZ2RLWCHIWZANCNFSM4NZSXJBQ&data=05%7C01%7Cl.m.wierenga%40fsw.leidenuniv.nl%7C9fb300554cd6478fd78108daa9699337%7Cca2a7f76dbd74ec091086b3d524fb7c8%7C0%7C0%7C638008568145228603%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C3000%7C%7C%7C&sdata=iU88TIKstgWC8bGFV2EhTwlO%2FUEkGtsQ%2F45DZgeoc14%3D&reserved=0. You are receiving this because you commented.Message ID: @.***>

Adali35 commented 1 year ago

Hello, we are a group of students working on analyzing MRI data with Freesurfer and we were planning to use Qoala-T to do QC of our data but during the process we encountered some problems:

First of all, can we use subset-based approach for outputs of Freesurfer version 7.3.2 or generally any other version of Freesurfer which is higher than 7.1? If not are you supposed to validate the architecture for every single new version of Freesurfer, perhaps the variation between close versions such as 7.1 and 7.3 not that much.

Secondly, there are many manually rated recon-all outputs from Freesurfer that is version 7 or higher, can we use these data from other studies to create a bigger training data for the subset-based approach? Or do we have to use the data that come from the same study to create a training set. We could not make sense of the latter but since there is not any trained model for versions that are higher than version 7 we could not be sure either.

We really appreciate the work and thank you for your effort.

Sincerely, Yıldırım