chipmunk

Freely moving auditory and visual discrimination task for mice (adapted from Raposo et al. 2012)

Introduction

This repo contains code, setup description, assembly instructions and parts design to build a freely moving behavior setup.

The task is a visual and/or auditory accumulation of evidence task. The stimuli is stochastic ticks and/or flashes.

Description of the setup

partslist here

Here goes a description in words.

3d parts for building the setup are in the custom_parts/3d folder.

The walls are laser cut from pololu and the drawings are in the custom_parts/enclosure

The visual and auditory stimuli printed circuit boards are here and can be ordered from SeeedStudio.

Software instalation

Bpod gen 2

Install Bpod from the repository here. This was tested with commit 329bd9e.

chipmunk task

Clone the repository to Bpod Local/Protocols. If you make changes; create a branch and remember to commit changes regularly.

Calibration

Stimuli

The visual stimuli LEDs response curve should have a plot here.

The auditory speaker response curves were measured with X microphone and a picture should be here.

This needs to be done for every rig? Stored somewhere? Where?

Water reward size

Use BPOD to do this. An explanation should be here.

Behavior monitoring

Cameras wiring for sync is done like this. (Needs diagram.)

Use cameras that have GPIO connections like the FLIR Chamaeleon 3. On the Chamaeleon 3 USB you can use the Purple wire (GPIO 2). Connect to BPod to signal trial start and/or end. To sync multiple cameras use the Green wire and set one of the cameras as master (It is smart to use the same sampling rate).

Cameras should be set like the following picture. This gives the tracking results with DeepLabCut below (figure).

Training strategy

Training has 3 phases:

1. Habituation - description of what happens in this phase. How long it lasts. When it ends.

2. Training phase - during these phase mice do the task but only with easy stimuli. This takes around 4 weeks.

3. Testing/experiment phase - mice are now considered experts and difficult stimuli are introduced. Mice perform X above chance.

Figure of performance, accross days. Code to produce this figure in the notebooks folder

TODO

Arena size is 20x20cm. An additional chamber can attached to hold an observer.

1. walls and floor - laser cut drawings there is an additional 3d part to connect the walls - use transparent red acrylic #2423 from Pololu, 6mm. Where can we find small hinges?
2. test the stimuli and put the results in a figure here somewhere. With a description of how to do this.
3. Do we use a sound card?? What the latency? Fixed? How fixed?
4. 3d parts for the ports 2 - adapt the design so that head gear on a mouse is not does not touch the port during licking. Can we do this just with a spout and capacitive sensing?
5. What's in the current port PCB, where the drawing?
6. try different training strategies. Find the best strategy. What the best way to get multimodal enhancement? What is the best loudness and visual saliency?

Most of these can be done in parallel, some can't. Version 1 can be rough for most of these (we can use code and audio from MudSkipper2 for now.)