Drift-diffusion model of Go/No-Go task

[Nathan Benton]

Hello! I just recently found this group and have already found it helpful as a new hBayesDM user. I'm hoping to see if someone might be able to provide some guidance on a current issue that I am experiencing and maybe some advice on the best way to approach my project. 

I am attempting to fit a DDM on data from a Go/No-Go task. So on some trials, participants did not respond, resulting in censored data where I do not know when the exact decision not to respond was made. I did not see in the documentation for the package if there was any way to handle censored data such as this, like there is in 'brms'.

I have also tried fitting a single-choice model looking only at the Go trials, but I am getting some errors in trying to implement the 'choiceRT_ddm_single' function. It is showing that model fitting is complete, but with the following warnings:

************************************ **** Model fitting is complete! **** ************************************ Warning messages: 1: In mean.default(parVals[[x]]) : argument is not numeric or logical: returning NA 2: In mean.default(parVals[[x]]) : argument is not numeric or logical: returning NA 3: In mean.default(parVals[[x]]) : argument is not numeric or logical: returning NA 4: In mean.default(parVals[[x]]) : argument is not numeric or logical: returning NA

Subsequently, I am not able to extract any parameters or diagnostic plots. My dataset is properly formatted as far as I can tell. Any insight into what might be causing this issue would be greatly appreciated!!!

Thanks,

Nathan

[Jeongyeon Shin]

Hi Nathan,

Unfortunately, our current package doesn’t support the option for handling censored data. As long as you don’t have that many trials with such responses, I suggest you just remove those trials before fitting the data with the hBayesDM package.Additionally, choiceRT_ddm_single model in the package is designed for multiple choice, not single choice as you noticed (see this link: https://github.com/CCS-Lab/hBayesDM/blob/develop/commons/models/choiceRT_ddm_single.yml)Feel free to ask if you have more questions!

Best regards,
Jeongyeon

2024년 6월 30일 일요일 오후 12시 29분 0초 UTC+9에 natha...@gmail.com님이 작성: