Statistical Parametric Mapping (SPM) in 6 min

[@biomechstu]

Thank you for watching. You can find a full playlist of SPM tutorials here: kzbin.info/aero/PLI7VLEjUJidBBtyWFcZYF7f8t8IgMT9cg Please note that I do reply to questions/comments, but all previous responses have been lost in transitioning the channel over to my new email address.

[@javadtaei4674]

I need hellp. I have two independent groups that ran in three diffrent condition. Which code is two way repeated measure anova. And how should I use it?

[@biomechstu]

Hi@@javadtaei4674 Please see my reply to your other comment with the link to my ANOVA video. Hopefully that helps. Best wishes.

[@StevenHughes0]

GREAT video! I used Python to run SPM for my first analysis and it took me 2 months to learn how to understand the code and get it to work through trial and error. With a little bit of instruction it is actually quite easy so this is amazingly useful

[@catherinetucker]

Thank you for this, Stuart. It is very helpful.

[@TheWuTangClanForever]

Hi Stuart! Thank you so much for this video. Is there any chance you still have these original variableA/B excel files? I'd like to use them for a class demonstration on SPM and think they are the perfect size and simple enough to use.

[@user-ib3zp5sj6z]

I can not believe that someone uploaded what I exactly finding!! But I want to compare 3 Groups of Patients. Can I use one-way or two-way ANOVA with SPM??

[@tifahan936]

Do you know where could download test data code of SPM12?

[@user-us9ph1hw4c]

Dear Stuart, may I ask how to get the specific P value when P>0.05 at a certain stage during SPM analysis?

[@biomechstu]

Hi. I think you've also emailed me, so I'll reply there.

[@raqueljcosta]

but if the sample is small (you have 11 subjects and i have 8), shouldn't we use a non parametric test like Wilcoxon?

[@biomechstu]

There is usually very little difference in results, but this paper provides more detail: doi.org/10.1016/j.jbiomech.2015.02.051

[@raqueljcosta]

@@biomechstu thank you!

[@javadtaei4674]

Hello, I want to use two way repeated measure anova which code should I use????

[@biomechstu]

Hi. If you haven't already seen it, hopefully this video will help: kzbin.info/www/bejne/eqGQfH5orJpji6c

[@kasperjepsen7066]

Hi Stuart. Could you explain how i can use the SPM code on a dataset that isn't 10x101? I have datasets which are 1x371, but how do i adjust the code so it works? Thanks in advance

[@kasperjepsen7066]

@@biomechstu Thank you very much, this fixed the problem. Just by curiosity, is there a way to plot the SPM chart combined with the mean and SD plot in Matlab (one chart - not a subplot)? I have seen in a few scientific studies implementing SPM, that you can show the areas of significans as a sort of bar plot between the x axis values and the chart itself.

[@ababababababababa9317]

hello Stuart, is there any wat to calculate SPM for 2D analysis? I found SPM12, however its not looking for easy for me, do u have any suggestion?How can ı figure it out ? I need to solve this :/ Thank you in advance

[@ababababababababa9317]

@@biomechstu Thank you Stuart, just started to watch

[@jiphelsloot4138]

Dear Stuart, thanks for the helpfull clear video. We are trying to run a similar script with two variables but we get the following error: SPM{t} (0D) inference z: 18.1209 df: [1 354] alpha: 0.0500 zstar: 1.9667 h0reject: 1 p: 0 Unrecognized method, property, or field 'plot' for class 'spm1d.stats.spm.SPM0Di'. Error in script_spm (line 6) spmi.plot() Is there a way to contact you for any help? Looking forwards hearing from you. Jip and Eva

[@biomechstu]

Hi both. You can happily contact me at [e-mail] (if so, can you send the code) although being honest you would be better to use the forum on the SPM1D website and contact Todd Pataky.

[@nitzanbreitman]

Hey@@biomechstu , I am getting the same problem and didn't find a solution in the forum. So I'm asking here for future users, did you figure out the problem and found the solution? Is there a minimum for the number of rows required to compare? In your example you are working with 11x101. Thank you for the great videos!

[@biomechstu]

@@nitzanbreitman Hi. I have just managed to run my most recent SPM analysis (a two-way repeated measures ANOVA) with only 3 participants after deleting the rest of the data. So that suggests that the number of participants is unlikely to be the issue. Besides, the output above looks like the actual statistical testing worked okay but it is the plot that failed to run. You may be able to plot it manually using something such as: plot(spmi.z,'k') % this plots the t-values as a black line yline(spmi.zstar,'--k') % this plots the significance threshold as a dashed black line

[@nitzanbreitman]

@@biomechstu ​ Yes I get the same result for 3 participants. Furthermore, when I ran it for 2 participants (2x101), Matlab couldn't finish the process at all (I let it run for 5 min). I guess we can infer that the minimum here is 3x101. Thanks again

[@biomechstu]

@@nitzanbreitman It's also worth noting that with that few participants, you are unlikely to have the statistical power to reliably detect any effect sizes of interest.

[@zininmed98]

my red lines for significance are at t=6.9 way above what yours looks like. Can anyone explain what I did wrong?

[@zininmed98]

@@biomechstu thank you for your fast reply. Embarrassingly, yes I have realized now my sample size is extremely small (n=5), your video on G*Power was super helpful in finding what sample size would be more useful. Do you have any recommendations for using MatLabs more efficiently when there are several variables? I am investigating gait kinematics and kinetics (with respect to leg length discrepancy) and have noticed I have had to clear spm and spmi in my "workspace" or my plot will combine data for two variables. I have tried replacing "SPM" in the code with the variable name so that multiple analyses can be plotted together without the coding interfering with each other, however, maybe it is best to replace what is in the brackets each time? I hope I have made sense. Certainly, looking forward to similar videos in the future!

[@melodysmith3728]

Hi! I have downloaded the spm1d files just like it's done here and tried running my own analysis using the exact same lines of code showed in this video. However, I get the following error: Undefined variable "spm1d" or class "spm1d.stats.ttest_paired". Can anyone help me understand what I'm doing wrong?

[@biomechstu]

Hi. Did you copy the files out into the main folder? And also run your code from there?

[@melodysmith3728]

@@biomechstu yes, I did. I just found out that I hadn't added the folder correctly onto the matlab path, I have fixed it and it runs perfectly now. Thank you very much for your tutorials and for your quick reply!

[@biomechstu]

Great news!