search

AllenNeuralDynamics / dynamic-foraging-task

Bonsai/Harp workflow for Dynamic Foraging with Python GUI for visualization and control
MIT License
5 stars 4 forks source link

Documentation about latent rules of reward probability #533

Open KanghoonJ opened 3 months ago

KanghoonJ commented 3 months ago

I wonder where we have a document about latent rules of reward probability such as 1) minimum reward probabilities should not be assigned to both sides simultaneously and 2) add four trials to both sides due to bias. The rules seem to be latent in the code. But it would be great to make the rules visible in a document and add rationales for each rule. Any thoughts? @alexpiet @hanhou @XX-Yin @ZhixiaoSu @JeremiahYCohen

hanhou commented 3 months ago

Since this PR, the rules are exactly the same as in Grossman 2022:

image

But yes, we should add this to the repo's document for clarity.

ZhixiaoSu commented 3 months ago

Yes, as Han said. I think the only few differences are:

  1. There's no 'CSminus' all cues are 'CSplus''
  2. The session doesn't pause for 30s when mouse stops responding for 3 trials.