Dear FieldTrip Support Team, I hope this message finds you well. I a

hi Guillermo, <a class="user-mention notranslate" data-hovercard-type="user" data-hove

Dear <a class="user-mention notranslate" data-hovercard-type="user" data-hovercard-url

Inconsistency between stat.mask and p-values in ft_freqstatistics,about fieldtrip/fieldtrip

schoffelen commented on June 11, 2024

hi Guillermo, @GuilleSGCUPO I am very sorry, but again I want to refer you to the e-mail discussion list. If the answer that was provided by one of the people on the list was not to your satisfaction, it will not work to re-post the same question here.
Rather, I suggest to reply to the discussion list's e-mail (I think that the reply was given by our colleague Mikkel), and perhaps think a little bit about additional details that you could provide, in order to increase the probability that the readership of the list can actually help you. Your initial question was apparently not formulated well enough to inspire the readers of the list.

Note that at minimum you should at least lift a tip of the veil as to what the observed discrepancy between mask and p-values is, rather than just stating that it's not according to your expectation.

Also, you may want to critically evaluate the pasted code that you used for ft_freqstatistics. If it's your intention to try and obtain an analytically derived p-value, cfg.method = 'montecarlo' is inappropriate.

Also, you may want to look into the code of ft_statistics_montecarlo (or perhaps better: ft_statistics_analytic) to find the line of code that converts a p-value into a mask. This will give you some clues as to whether indeed the FieldTrip implementation is wrong, or whether your expectations are wrong.

from fieldtrip.

GuilleSGCUPO commented on June 11, 2024

Hi Jan, Thank you for your prompt response, and no worries at all. Following my discussion with Mikkel, he suggested that I raise an issue on Github as he was unable to provide assistance with this statistical matter. Consequently, I plan to send an email to the discussion list with additional details on the topic. Despite trying different correction methods--Monte Carlo, Analytic, and even no correction--I consistently observe the same discrepancy between stat.mask and stat.prob. Of particular note is the observation I made regarding stat.cirange, which seems to contribute to this discrepancy. Despite reviewing various resources on Github related to statistics, I couldn't find a comprehensive explanation for how the mask variable is derived. Once again, thank you for your assistance. I'll proceed with submitting the issue to the discussion list. Best regards, Guille El 2024-03-25 16:01, Jan-Mathijs Schoffelen escribió:

hi Guillermo, @GuilleSGCUPO [1] I am very sorry, but again I want to refer you to the e-mail discussion list. If the answer that was provided by one of the people on the list was not to your satisfaction, it will not work to re-post the same question here. Rather, I suggest to reply to the discussion list's e-mail (I think that the reply was given by our colleague Mikkel), and perhaps think a little bit about additional details that you could provide, in order to increase the probability that the readership of the list can actually help you. Your initial question was apparently not formulated well enough to inspire the readers of the list. Note that at minimum you should at least lift a tip of the veil as to what the observed discrepancy between mask and p-values is, rather than just stating that it's not according to your expectation. Also, you may want to critically evaluate the pasted code that you used for ft_freqstatistics. If it's your intention to try and obtain an analytically derived p-value, cfg.method = 'montecarlo' is inappropriate. Also, you may want to look into the code of ft_statistics_montecarlo (or perhaps better: ft_statistics_analytic) to find the line of code that converts a p-value into a mask. This will give you some clues as to whether indeed the FieldTrip implementation is wrong, or whether your expectations are wrong. -- Reply to this email directly, view it on GitHub [2], or unsubscribe [3]. You are receiving this because you were mentioned.Message ID: ***@***.***>

Links: ------ [1] https://urldefense.com/v3/__https://github.com/GuilleSGCUPO__;!!D9dNQwwGXtA!TAn73-Syc2C6tn8RgStl7Dc0ATOQ5dhEMOmN0rAPIgYiEbQbh3-NQVPKUOQTtlD9gZ4MrVGTdJW4p1-r95-XPg$ [2] https://urldefense.com/v3/__https://github.com/fieldtrip/fieldtrip/issues/2402*issuecomment-2018209593__;Iw!!D9dNQwwGXtA!TAn73-Syc2C6tn8RgStl7Dc0ATOQ5dhEMOmN0rAPIgYiEbQbh3-NQVPKUOQTtlD9gZ4MrVGTdJW4p19aPeYwFw$ [3] https://urldefense.com/v3/__https://github.com/notifications/unsubscribe-auth/AZ7RUMATGBW7MJYUKUBUSHTY2A35VAVCNFSM6AAAAABFHEXH46VHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDAMJYGIYDSNJZGM__;!!D9dNQwwGXtA!TAn73-Syc2C6tn8RgStl7Dc0ATOQ5dhEMOmN0rAPIgYiEbQbh3-NQVPKUOQTtlD9gZ4MrVGTdJW4p18FZrHvXQ$ --=_96fbe284015548d0c323ac38823380d2 Content-Type: multipart/related; boundary="=_1f2a49c9b6cdd3aeea2a69054d878ab1"

--=_1f2a49c9b6cdd3aeea2a69054d878ab1 Content-Transfer-Encoding: quoted-printable Content-Type: text/html; charset=UTF-8 <html><head><meta http-equiv=3D"Content-Type" content=3D"text/html; charset= =3DUTF-8" /></head><body style=3D'font-size: 10pt; font-family: Verdana,Gen= eva,sans-serif'> Hi Jan, Thank you for your prompt response, and no worries at all. Following my = discussion with Mikkel, he suggested that I raise an issue on Github as he = was unable to provide assistance with this statistical matter. Consequently= , I plan to send an email to the discussion list with additional details on= the topic. Despite trying different correction methods—Monte Carlo, Analytic,= and even no correction—I consistently observe the same discrepancy b= etween stat.mask and stat.prob. Of particular note is the observation I mad= e regarding stat.cirange, which seems to contribute to this discrepancy. De= spite reviewing various resources on Github related to statistics, I couldn= 't find a comprehensive explanation for how the mask variable is derived.</= p> Once again, thank you for your assistance. I'll proceed with submitting = the issue to the discussion list. Best regards, Guille El 2024-03-25 16:01, Jan-Mathijs Schoffelen escribi&o= acute;: <blockquote type=3D"cite" style=3D"padding: 0 0.4em; border-left: #1010ff 2= px solid; margin: 0"> <div id=3D"replybody1"> hi Guillermo, <a class=3D"v1user-mention v1notranslate" hre= f=3D"https://urldefense.com/v3/__https://github.com/GuilleSGCUPO__;!!D9dNQw= wGXtA!TAn73-Syc2C6tn8RgStl7Dc0ATOQ5dhEMOmN0rAPIgYiEbQbh3-NQVPKUOQTtlD9gZ4Mr= VGTdJW4p1-r95-XPg$" target=3D"_blank" rel=3D"noopener ***@***.***= CUPO</a> I am very sorry, but again I want to refer you to the e-mail discu= ssion list. If the answer that was provided by one of the people on the lis= t was not to your satisfaction, it will not work to re-post the same questi= on here. Rather, I suggest to reply to the discussion list's e-mail (I= think that the reply was given by our colleague Mikkel), and perhaps think= a little bit about additional details that you could provide, in order to = increase the probability that the readership of the list can actually help = you. Your initial question was apparently not formulated well enough to ins= pire the readers of the list. Note that at minimum you should at least lift a tip of the = veil as to what the observed discrepancy between mask and p-values is, rath= er than just stating that it's not according to your expectation. Also, you may want to critically evaluate the pasted code t= hat you used for ft_freqstatistics. If it's your intention to try and obtai= n an analytically derived p-value, cfg.method =3D 'montecarlo' is inappropr= iate. Also, you may want to look into the code of ft_statistics_m= ontecarlo (or perhaps better: ft_statistics_analytic) to find the line of c= ode that converts a p-value into a mask. This will give you some clues as t= o whether indeed the FieldTrip implementation is wrong, or whether your exp= ectations are wrong. <p style=3D"font-size: small; -webkit-text-size-adjust: none; color: #666;"=

— Reply to this email directly, <a href=3D"https://urldefense.c=

om/v3/__#2402*issuecomment-201= 8209593__;Iw!!D9dNQwwGXtA!TAn73-Syc2C6tn8RgStl7Dc0ATOQ5dhEMOmN0rAPIgYiEbQbh= 3-NQVPKUOQTtlD9gZ4MrVGTdJW4p19aPeYwFw$" target=3D"_blank" rel=3D"noopener n= oreferrer">view it on GitHub</a>, or <a href=3D"https://urldefense.com/v3/_= _https://github.com/notifications/unsubscribe-auth/AZ7RUMATGBW7MJYUKUBUSHTY= 2A35VAVCNFSM6AAAAABFHEXH46VHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDAMJYGIY= DSNJZGM__;!!D9dNQwwGXtA!TAn73-Syc2C6tn8RgStl7Dc0ATOQ5dhEMOmN0rAPIgYiEbQbh3-= NQVPKUOQTtlD9gZ4MrVGTdJW4p18FZrHvXQ$" target=3D"_blank" rel=3D"noopener nor= eferrer">unsubscribe</a>. You are receiving this because you were ment= ioned.<img ***@***.***" width=3D"1" = height=3D"1" />Message ID: <fiel= dtrip/fieldtrip/issues/2402/2018209593@github</sp= an>.com> </div> </blockquote> </body></html>

…

--=_1f2a49c9b6cdd3aeea2a69054d878ab1 Content-Transfer-Encoding: base64 Content-ID: ***@***.***> Content-Type: image/gif; name=blocked.gif Content-Disposition: inline; filename=blocked.gif; size=118 R0lGODlhZAAyAIAAAPrOzgAAACH5BAAAAAAALAAAAABkADIAAAJNhI+py+0Po5y02ouz3rz7D4bi SJbmiabqyrbuC8fyTNf2jef6zvf+DwwKh8Si8YhMKpfMpvMJjUqn1Kr1is1qt9yu9wsOi8fksvls KwAAOw== --=_1f2a49c9b6cdd3aeea2a69054d878ab1--

--=_96fbe284015548d0c323ac38823380d2--

from fieldtrip.

GuilleSGCUPO commented on June 11, 2024

Sorry, I meant using Montecarlo or Analytic for the method, with "fdr" for correction or not correction at all, still having the same discrepancy in the mask results. El 2024-03-25 17:13, Guillermo Sanchez-Garrido Campos escribió:

Hi Jan, Thank you for your prompt response, and no worries at all. Following my discussion with Mikkel, he suggested that I raise an issue on Github as he was unable to provide assistance with this statistical matter. Consequently, I plan to send an email to the discussion list with additional details on the topic. Despite trying different correction methods--Monte Carlo, Analytic, and even no correction--I consistently observe the same discrepancy between stat.mask and stat.prob. Of particular note is the observation I made regarding stat.cirange, which seems to contribute to this discrepancy. Despite reviewing various resources on Github related to statistics, I couldn't find a comprehensive explanation for how the mask variable is derived. Once again, thank you for your assistance. I'll proceed with submitting the issue to the discussion list. Best regards, Guille El 2024-03-25 16:01, Jan-Mathijs Schoffelen escribió: > hi Guillermo, @GuilleSGCUPO [1] I am very sorry, but again I want to > refer you to the e-mail discussion list. If the answer that was > provided by one of the people on the list was not to your > satisfaction, it will not work to re-post the same question here. > Rather, I suggest to reply to the discussion list's e-mail (I think > that the reply was given by our colleague Mikkel), and perhaps think a > little bit about additional details that you could provide, in order > to increase the probability that the readership of the list can > actually help you. Your initial question was apparently not formulated > well enough to inspire the readers of the list. > > Note that at minimum you should at least lift a tip of the veil as to > what the observed discrepancy between mask and p-values is, rather > than just stating that it's not according to your expectation. > > Also, you may want to critically evaluate the pasted code that you > used for ft_freqstatistics. If it's your intention to try and obtain > an analytically derived p-value, cfg.method = 'montecarlo' is > inappropriate. > > Also, you may want to look into the code of ft_statistics_montecarlo > (or perhaps better: ft_statistics_analytic) to find the line of code > that converts a p-value into a mask. This will give you some clues as > to whether indeed the FieldTrip implementation is wrong, or whether > your expectations are wrong. > > -- > Reply to this email directly, view it on GitHub [2], or unsubscribe > [3]. > You are receiving this because you were mentioned.Message ID: > ***@***.***>

Links: ------ [1] https://urldefense.com/v3/__https://github.com/GuilleSGCUPO__;!!D9dNQwwGXtA!TAn73-Syc2C6tn8RgStl7Dc0ATOQ5dhEMOmN0rAPIgYiEbQbh3-NQVPKUOQTtlD9gZ4MrVGTdJW4p1-r95-XPg$ [2] https://urldefense.com/v3/__https://github.com/fieldtrip/fieldtrip/issues/2402*issuecomment-2018209593__;Iw!!D9dNQwwGXtA!TAn73-Syc2C6tn8RgStl7Dc0ATOQ5dhEMOmN0rAPIgYiEbQbh3-NQVPKUOQTtlD9gZ4MrVGTdJW4p19aPeYwFw$ [3] https://urldefense.com/v3/__https://github.com/notifications/unsubscribe-auth/AZ7RUMATGBW7MJYUKUBUSHTY2A35VAVCNFSM6AAAAABFHEXH46VHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDAMJYGIYDSNJZGM__;!!D9dNQwwGXtA!TAn73-Syc2C6tn8RgStl7Dc0ATOQ5dhEMOmN0rAPIgYiEbQbh3-NQVPKUOQTtlD9gZ4MrVGTdJW4p18FZrHvXQ$ --=_b53b019be0a466101944bdd49bc3135a Content-Type: multipart/related; boundary="=_f99b96bf23872daa9a0c266f741f67a8"

--=_f99b96bf23872daa9a0c266f741f67a8 Content-Transfer-Encoding: quoted-printable Content-Type: text/html; charset=UTF-8 <html><head><meta http-equiv=3D"Content-Type" content=3D"text/html; charset= =3DUTF-8" /></head><body style=3D'font-size: 10pt; font-family: Verdana,Gen= eva,sans-serif'> Sorry, I meant using Montecarlo or Analytic for the method, with "fdr" f= or correction or not correction at all, still having the same discrepancy i= n the mask results. <div id=3D"signature"></div> El 2024-03-25 17:13, Guillermo Sanchez-Garrido Campos= escribió: <blockquote type=3D"cite" style=3D"padding: 0 0.4em; border-left: #1010ff 2= px solid; margin: 0"> <div id=3D"replybody1"> <div style=3D"font-size: 10pt; font-family: Verdana,Geneva,sans-serif;"> Hi Jan, Thank you for your prompt response, and no worries at all. Following my = discussion with Mikkel, he suggested that I raise an issue on Github as he = was unable to provide assistance with this statistical matter. Consequently= , I plan to send an email to the discussion list with additional details on= the topic. Despite trying different correction methods—Monte Carlo, Analytic,= and even no correction—I consistently observe the same discrepancy b= etween stat.mask and stat.prob. Of particular note is the observation I mad= e regarding stat.cirange, which seems to contribute to this discrepancy. De= spite reviewing various resources on Github related to statistics, I couldn= 't find a comprehensive explanation for how the mask variable is derived.</= p> Once again, thank you for your assistance. I'll proceed with submitting = the issue to the discussion list. Best regards, Guille El 2024-03-25 16:01, Jan-Mathijs Schoffelen escribi= ó: <blockquote style=3D"padding: 0 0.4em; border-left: #1010ff 2px solid; marg= in: 0;"> <div id=3D"v1replybody1"> hi Guillermo, <a class=3D"v1v1user-mention v1v1notranslate"= href=3D"https://urldefense.com/v3/__https://github.com/GuilleSGCUPO__;!!D9= dNQwwGXtA!TAn73-Syc2C6tn8RgStl7Dc0ATOQ5dhEMOmN0rAPIgYiEbQbh3-NQVPKUOQTtlD9g= Z4MrVGTdJW4p1-r95-XPg$" target=3D"_blank" rel=3D"noopener ***@***.***= leSGCUPO</a> I am very sorry, but again I want to refer you to the e-mail d= iscussion list. If the answer that was provided by one of the people on the= list was not to your satisfaction, it will not work to re-post the same qu= estion here. Rather, I suggest to reply to the discussion list's e-mai= l (I think that the reply was given by our colleague Mikkel), and perhaps t= hink a little bit about additional details that you could provide, in order= to increase the probability that the readership of the list can actually h= elp you. Your initial question was apparently not formulated well enough to= inspire the readers of the list. Note that at minimum you should at least lift a tip of the = veil as to what the observed discrepancy between mask and p-values is, rath= er than just stating that it's not according to your expectation. Also, you may want to critically evaluate the pasted code t= hat you used for ft_freqstatistics. If it's your intention to try and obtai= n an analytically derived p-value, cfg.method =3D 'montecarlo' is inappropr= iate. Also, you may want to look into the code of ft_statistics_m= ontecarlo (or perhaps better: ft_statistics_analytic) to find the line of c= ode that converts a p-value into a mask. This will give you some clues as t= o whether indeed the FieldTrip implementation is wrong, or whether your exp= ectations are wrong. <p style=3D"font-size: small; -webkit-text-size-adjust: none; color: #666;"=

— Reply to this email directly, <a href=3D"https://urldefense.c=

om/v3/__#2402*issuecomment-201= 8209593__;Iw!!D9dNQwwGXtA!TAn73-Syc2C6tn8RgStl7Dc0ATOQ5dhEMOmN0rAPIgYiEbQbh= 3-NQVPKUOQTtlD9gZ4MrVGTdJW4p19aPeYwFw$" target=3D"_blank" rel=3D"noopener n= oreferrer">view it on GitHub</a>, or <a href=3D"https://urldefense.com/v3/_= _https://github.com/notifications/unsubscribe-auth/AZ7RUMATGBW7MJYUKUBUSHTY= 2A35VAVCNFSM6AAAAABFHEXH46VHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDAMJYGIY= DSNJZGM__;!!D9dNQwwGXtA!TAn73-Syc2C6tn8RgStl7Dc0ATOQ5dhEMOmN0rAPIgYiEbQbh3-= NQVPKUOQTtlD9gZ4MrVGTdJW4p18FZrHvXQ$" target=3D"_blank" rel=3D"noopener nor= eferrer">unsubscribe</a>. You are receiving this because you were ment= ioned.<img ***@***.***" width=3D"1" = height=3D"1" />Message ID: <fiel= dtrip/fieldtrip/issues/2402/2018209593@github</sp= an>.com> </div> </blockquote> </div> </div> </blockquote> </body></html>

…

--=_f99b96bf23872daa9a0c266f741f67a8 Content-Transfer-Encoding: base64 Content-ID: ***@***.***> Content-Type: image/gif; name=blocked.gif Content-Disposition: inline; filename=blocked.gif; size=118 R0lGODlhZAAyAIAAAPrOzgAAACH5BAAAAAAALAAAAABkADIAAAJNhI+py+0Po5y02ouz3rz7D4bi SJbmiabqyrbuC8fyTNf2jef6zvf+DwwKh8Si8YhMKpfMpvMJjUqn1Kr1is1qt9yu9wsOi8fksvls KwAAOw== --=_f99b96bf23872daa9a0c266f741f67a8--

--=_b53b019be0a466101944bdd49bc3135a--

from fieldtrip.

robertoostenveld commented on June 11, 2024

Dear @GuilleSGCUPO

I'd appreciate if you would stop referring to us as the "support team". We are researchers, probably just like you. Furthermore, please stop spamming this github issue with unformatted raw html content.

I am not sure how you have followed up Mikkel's suggestion on the mailing list, but please refer to https://doi.org/10.1371/journal.pcbi.1002202.

If you are still not sure what the suggestion of Mikkel refers to, you may want to read up (and carry out) the tutorial on https://www.fieldtriptoolbox.org/tutorial/eventrelatedstatistics/ an follow the links that appear at the bottom of the tutorial

If you still believe there to be a bug in the code that is to be reported here as issue, please see https://www.fieldtriptoolbox.org/development/issues/#provide-informative-details-about-your-issue for guidance.

best regards,
Robert

from fieldtrip.

GuilleSGCUPO commented on June 11, 2024

Dear Robert,

Thank you for your email and your constructive feedback. I want to assure you that at no point did I intend to belittle the valuable work you do as researchers by referring to the team as the "support team." I fully acknowledge the importance and significant impact your work has on advancing scientific knowledge.

Regarding the HTML format you mentioned, I realize there must have been an error when sending emails through the email client, resulting in the posting of unformatted content. This was unintentional, and I apologize for any inconvenience it may have caused.

Furthermore, I appreciate the link you provided for the FieldTrip statistics tutorial. I am confident that it will prove to be a valuable resource, and I look forward to utilizing it to enhance my understanding and skills in this area.

Once again, thank you.

Best regards,
Guille

from fieldtrip.

Inconsistency between stat.mask and p-values in ft_freqstatistics about fieldtrip HOT 5 CLOSED

Comments (5)

Related Issues (20)

Recommend Projects

React

Vue.js

Typescript

TensorFlow

Django

Laravel

D3

Recommend Topics

javascript

web

server

Machine learning

Visualization

Game

Recommend Org

Facebook

Microsoft

Google

Alibaba

D3

Tencent