Go/NoGo model modification (eliminating parameter "pi" in gng_m4)

[Azadeh Nazemoroaya]

Hi hbayesdm-users,

I would like to test sth on gng example data by eliminating pavlovian parameter "pi" in gng_m4 model. Is there an easy way to do this? Or should I go through the contribution process to add this as a new model to the package in github.

Many thanks for your help!

Best,

Azadeh

[Azadeh Nazemoroaya]

Could anyone let me to know where these gng_example data are form? from Marc Guitart-Masip 2013 paper?

[wooyou...@gmail.com]

Hi Azadeh, 

Sorry for the long delay!

Just to double check - so, you want to test a model with separate sensitivity to reward and punishment, but with no Pavlovian bias, right? Then, you need to modify its Stan code and run it outside the hBayesDM. You don't have to add your modified model to hBayesDM. 

If it's okay to use the same sensitivity to reward and punishment, you can just use gng_m2: https://github.com/CCS-Lab/hBayesDM/blob/develop/commons/stan_files/gng_m2.stan. 

The "example" data are NOT from the Guitart-Masip study. I believe they are generated from a simulation code. 

Best,

Young