Adding CS+/CS- and Reward Delay to task

[rachelstephlee]

Rachel and Kenta would like to modify the standard foraging task a bit as shown below, based on our "data-driven-question-driven demand". The idea is having intermediate reinforcers (CS+/CS-).

• Essentially, a secondary reinforcer for the reward
• 100% deterministic sound CS+/CS- to let mouse know reward (or no-reward is coming)
• To start, 1s fixed timing between CS+/CS- and reward. We will implement the delay as a distribution, try with fix delay and introduce variability as we get more information

Motivation

When kenta recorded phasic DA in dynamic foraging, we could not get a clear RPE signal until we timelocked to ‘last lick’

• It’s difficult to disentangle ‘lick response’ and ‘reward response’
• We will not be able to clearly tell how learning is happening wrt to NMs. Especially if NMs already correlate with movement.

Recording designs:

• FIP DA x3 — VTA-DA-SomaCa / NAc-DA-AxonCa / NAc-DA release
• FIP DA in NAc + NE in Ctx
• Ephys DA soma / NAc MSNs

Plan

• Need to gather more information with @micahwoodard , @alexpiet , @XX-Yin, @bruno-f-cruz to figure out how we can implement this (with minimal disruption to other experiments, with as much aligned efforts wrt to refactoring code here)
• Rachel will scope/drive this project with feedback from @hagikent and @ZhixiaoSu
• Once we have buy-in, present at Tuesday morning behavior meeting.
• Then, Kenta will pilot this in test boxes in 428 so this will not affect 446, 447 experiments.

Expected change in code base

1. Python GUI: very minor, two new delay parameters.

2. Bonsai workflow: minor~intermediate (but not sure how to make it compatible to the current foraging task, -KH)
   Change in task logic will be something like:

In addition, some minor back-end updates to play two additional tones.

---

More detailed information here: https://alleninstitute-my.sharepoint.com/:p:/g/personal/rachel_lee_alleninstitute_org/Edi4EJV9YptFpQRpq3d59DwBo0GvP02s38wC5CFsXWbPmQ?e=en4zNa