Cricket VR

[glopesdev]

CricketVR 3.0

Beacons task

• Beacons task stage 0 #45
  • Linear track implemented by constraining VR movement to one axis
  • Forward movement only when facing direction of the corridor
  • Decide on visuals for linear track (gratings, textures, etc)
• Beacons task stage 1 - wider linear track
• Beacons task stage 2 - 2D world
• Track progression of training stages
• Logging of beacons task events
• Allow looming at any point during beacons task - at least in 2D environment

Hardware improvements

• Lick detection #55
  • Synchronization
  • Triggerable anytime anywhere
  • Acquisition hardware (can this be a Harp device, maybe the existing Behavior board?)
• Reward delivery #54
  • Same as above, ideally running on same hardware
• Bearing-mounted camera
  • is working somewhat, need to repeat benchmarking to check camera syncing
• Floor projectors #46
  • View mapping and synchronizing with monitors

Ideas to consider

• Ambient lighting to improve feeling of going into shelter (low priority)
  • Requires custom lighting system hardware / control

VR environment improvements

• Increase size of virtual environment (3x3 or 4x4m)
• Replace walls with a short wall or moat visible from afar
• Flashing stimulus when walls are too close - should be optional feature
• VR movement lower when running along walls - should be optional feature

Looming improvements

• Allow for repeats of looming stimuli #47
• Add timing parameters including duration, hang and inter-stimulus interval #48
  • Allow dynamically changing the parameters
• Allow specifying looming starting position based on current position relative to the shelter or head orientation #49
• Allow pairing looming with audio stimuli (review priority?)
  • Requires audio hardware (harp soundcard vs system soundcard, consider ephys)

Shelter appearance

• Make shelter darker inside and make it faster to reach peak darkness #50
• Decrease height of shelter and make it look more like a patch or burrow #51
  • One idea is to make it look like a hole on a slope so it looms as it approaches (self-looming)
  • Make it accessible from all sides vs directional
• Allow shelter to be made invisible while keeping all its logic

Cricket trajectories

• Update basic arm behaviour
  • A > B: if point A is close to mouse, move away from mouse first then go to target
  • Limit how close arm can get when mouse is facing away from the arm
• Allow option to move arm away from the mouse
• Increase number of possible triggers for initiating cricket hunt
  • Specific spot in arena
  • Manual trigger by experimenter
  • Trigger when it starts and trigger when it ends

Calibration and outstanding issues

• Benchmarking and retesting gain conversion from ball to VR world
• Function for re-homing arm controller if arm movements are dysfunctional
• Change arm controller homing position
  • This may be tricky since there are singularities and specific angular calculations to ensure arm stability
• Test how to change arm length and microcontroller parameters

Tracking head-orientation

• Readout orientation using image processing instead of magnetic encoder #52
  • Review requirements for this (mechanical constraints with lick port?)
• Benchmark closed-loop latency from orientation readout to arm movement #53

Next phase

Cricket trajectories

• Increase number of possible perturbations
  • Most important to allow option to move arm away from the mouse
• Model more complex cricket movement trajectories
  • Replay specific movement file
  • Define start and end of hunting

Hardware improvements