I'm planning to do cluster-based permutation statistics with the Threshold-free cluster enhancement method which is available in fieldtrip (attached an image of the code). Is there any way I can do this in Brainstorm?
I need one more help. I have to do statistical comparisons (eg cluster-based permutation) across more than 2 groups (3 in one and 5 in other). The process 2 tab allows only comparison between 2 groups. Please let me know how I can perform this in Brainstorm.
I'm planning to do cluster-based permutation statistics with the Threshold-free cluster enhancement method which is available in fieldtrip (attached an image of the code). Is there any way I can do this in Brainstorm?
I need one more help. I have to do statistical comparisons (eg cluster-based permutation) across more than 2 groups (3 in one and 5 in other). The process 2 tab allows only comparison between 2 groups.
At the moment, Brainstorm can only run tests between two sets of inputs.
For more elaborated designs, you need to export the values (with processes like Extract values, or direct access to the database) you want to test and run your analysis with custom scripts in Matlab.
Thank you for the quick update. I will check this out and let you know.
About the multiple sets of inputs, can you elaborate a bit more about what you refer here. Are you telling me extract the values and run the test separately, for example, using fieldtrip or any other statistical program directly?
Exactly. Brainstorm does not offer at the moment the interface to compute interactions more elaborated than a single factor with two levels. But this can be done with Matlab or other programs - there are various solutions available to export data from the Brainstorm database in a structured way, across all the subjects at once.
I have one more doubt. I have epoched data of 25 subjects with approximately 240 trials of three conditions (no of trials not equal after trial rejection). I'm planning to do ft_timelockstatistics between 2 of the 3 conditions across subjects. To do this should I use the subjectwise trial averages of the epochs or the epochs themselves of all subjects(as in the tutorials)?
I tried this TFCE option that you added in the ft_timelockstatistics and it's working properly. But I'm not getting the options to visualize the significant clusters from the result of the statistics. Is there any way I can do this?
When calling ft_timelockstatistics with correctm='cluster', FieldTrip returns a structure with the fields posclusters and negclusters, that are then saved in Brainstorm structure StatClusters.
However, these fields are not returned when calling ft_timelockstatistics with correctm='tfce'.
You can explore this by putting a breakpoint at line 491 of function process_ft_timelockstatistics.m (= click on the line number). Execute the process, and when the debugger stops, you can explore the contents of the variable and execute the code step by step.
I'm not sure how this works and don't really have time to investigate this at the moment.
If you have time to explore this further and find a way to get clusters descriptions from the FieldTrip function with the TFCE option, that would like to visualize in Brainstorm, please let me know.