Protocol

Mouse audiovisual spatial wheel task

Head-fixed mice localize visual and auditory stimuli presented across lateral screens and speakers, reporting left or right by turning a wheel. Visual-only, auditory-only, congruent audiovisual, and conflict audiovisual trials allow modality-specific and multisensory decision variables to be separated.

Task structure

Stimulus, choice, feedback, and training as declared in the protocol record.

Stimulus

evidence_schedule
Trials sample visual contrast and auditory location, including unisensory, congruent multisensory, and conflict multisensory conditions.
evidence_type
mixed
modalities
visual, auditory, multisensory
notes
—
units
percent contrast, degrees azimuth
variables
visual contrast, auditory azimuth, modality condition, cue congruence

Choice

action_mapping
Wheel movement reports the inferred side of the stimulus location.
alternatives
left, right
choice_type
2afc
notes
—
response_modalities
wheel

Feedback

feedback_type
mixed
notes
Correctness on conflict trials is analysis-specific and should be encoded explicitly.
penalty
Timeout or error feedback for incorrect choices, source-dependent.
reward
Water reward for correct choices on defined unisensory or congruent trials.

Training

notes
—
stages
visual localization, auditory localization, audiovisual integration

Trial timing

  1. inter_trial_interval
    1.5-2.5 s

    Delay between trials.

  2. wheel_quiescence
    100-250 ms

    Mouse must hold the wheel still before stimulus onset.

  3. open_loop_stimulus
    500 ms

    Stimuli are presented while wheel movement does not move the stimulus.

  4. closed_loop_response
    up to 1 s

    Stimulus position is yoked to wheel movement.

Vertical slices

Expected analyses

  • modality-split psychometric curves
  • additive audiovisual choice model
  • conflict-trial weighting
  • inactivation effects by cortical target
  • reaction time by stimulus condition

Interpretive claims

  • multisensory integration
    Confidence: high · ref.coen-2023-audiovisual-decisions

    Evidence for integration depends on modeling modality-specific cue values and conflict trials, not only pooled left/right accuracy.

Open questions

  • Verify a full local source export against the default adapter once BehaviorData.zip is downloaded.
  • Add a paper-matched additive audiovisual GLM and inactivation-effect analysis.