Extracting trial-by-trial regressors for model-based EEG analysis

[Hanmo Yin]

Dear hBayesDM developers and users

Thank you so much for developing such a powerful and Encouraging package!

This is my first time doing computational neuroscience research, please bear with me asking the most basic questions.  I have collected behavioral and EEG data on the delay discounting task, now I'm having trouble in extracting the  trial-by-trial regressors. Currently the package does not provide regressor extraction function for this model, could anyone please tell me how I can caculate the value of the regressors? 

By the way, could any researchers with relevant experience tell me how to combine the regressors obtained by the model with EEG/ERP data, because the paper only describes the application in fMRI research (Ahn, W. Y., Haines, N., & Zhang, L., 2017).

Thanks!

Best regards,

Yin Hanmo

[wooyou...@gmail.com]

Hi Yin,

Sorry for the delayed reply. Model-based regressors are available only for models of certain tasks, which unfortunately do not include the delay discounting task. However, you can check how we generate trial-by-trial model-based regressors from some examples. 

For example, you can see the "generated quantities" block of "gng_m1" model here (starting line 81): https://github.com/CCS-Lab/hBayesDM/blob/534907d9cbe7101b3109feeec808cdbd380a93f2/commons/stan_files/gng_m1.stan#L81

Like this example, you can either directly change your Stan code, or you can just program your own R or Python codes for this. As long as you can extract such model-based regressors, I believe you can use them for any neuroimaging data analysis, such as fMRI or EEG. Hope this helps!

Best,

Young