In trainingChoiceWorld, when the stimulus reaches the center, it does not actually trigger a reward yet - the wheel needs to be moved slightly further to trigger the valve opening. It's almost like the 'real' gain of the wheel is not matching the 'virtual' gain, i.e. the mouse needs to move the wheel by a larger amount than what the stimulus position on the screen suggests.
Conversely, when the stiumulus has reached the center and the wheel is moved in the opposite direction, the stimulus stays locked in the center until the 'error' feedback is triggered.
Spotted on _iblrig_cortexlab_mesoscope running trainingChoiceWorld (phase 0, default adaptive reward & adaptive gain) in iblrig v8.24.3. The gain is set to high (this is the 3rd day of training for this mouse).
In trainingChoiceWorld, when the stimulus reaches the center, it does not actually trigger a reward yet - the wheel needs to be moved slightly further to trigger the valve opening. It's almost like the 'real' gain of the wheel is not matching the 'virtual' gain, i.e. the mouse needs to move the wheel by a larger amount than what the stimulus position on the screen suggests.
Conversely, when the stiumulus has reached the center and the wheel is moved in the opposite direction, the stimulus stays locked in the center until the 'error' feedback is triggered.
Spotted on _iblrig_cortexlab_mesoscope running trainingChoiceWorld (phase 0, default adaptive reward & adaptive gain) in iblrig v8.24.3. The gain is set to high (this is the 3rd day of training for this mouse).