DeJa Vu Hypothesis and Experiment

Hypothesis

There’s significant work indicating that human action planning is simulated before action is taken in the world - called "unconscious hallucinations" by . For example, shooting a basketball is simulated prior to actually shooting it in a simulation internal to the body. This means that the simulation “runs” in some number of time steps ahead of actual environment input.

My hypothesis is that Deja Vu is an “ordering problem” between these two systems such that the current input is “behind” the “simulation” such that it feels like “now” is being pulled from memories rather than current experience

Methodology:

Using a controlled virtual reality (VR) environment that can precisely manipulate the temporal relationship between motor simulations (internal forward models of planned actions) and corresponding sensory feedback. By introducing subtle and systematically varied time shifts between participants’ predicted outcomes of their own motor actions and the actual sensory feedback, track the occurrence of subjective reports of déjà vu. If participants more frequently report déjà vu under conditions where the predicted sensory input (internal simulation) leads the actual input by a controlled margin, this supports the hypothesis that déjà vu is linked to an “ordering problem” between these two systems.

40 neurologically healthy adults with normal vision.

• VR system: A high-fidelity headset and controllers for recording motor actions.
• Motion tracking: High-speed cameras or built-in controller tracking to detect participant gestures in real-time.

• Motor task: Participants interact with virtual objects (e.g., picking up a ball and “shooting” it toward a target).
• Internal simulation phase: Before each action, instruct participants to mentally rehearse the action. Neural correlates of motor planning can be monitored using EEG or fNIRS placed on motor and premotor cortex regions. This data will confirm that internal simulations occur before action execution.
• Temporal manipulation: After the participant initiates the real action (e.g., moving their hand forward), introduce a systematic and subtle delay or advancement in the visual and auditory feedback of the corresponding motion in the virtual environment. For example, as the participant reaches for the ball, the VR display might show their arm movement 50–150 ms earlier or later than their actual movement.

• Baseline Condition (No Shift): Perfect synchronization between motor action and visual feedback.
• Lead Condition: VR display of action appears slightly before the participant’s real motion. For example, the hand in VR reaches the ball ~100 ms before the participant’s real hand movement is executed.
• Lag Condition: VR display of action lags behind the participant’s actual movement by a similar 100 ms.
• Vary these timings across trials (e.g., 0 ms, ±50 ms, ±100 ms, ±150 ms) to assess the threshold at which subjective déjà vu reports increase.

• Subjective reporting: After each trial, prompt participants to immediately rate their sense of familiarity or déjà vu on a Likert scale (e.g., 1 = no familiarity, 5 = strong déjà vu).
• Neural correlates: Use EEG or MEG to measure predictive motor signals (beta suppression in motor cortex, readiness potentials) and compare their timing with the perceived onset of the corresponding visual feedback.
• Behavioral metrics: Track reaction times, error rates in task performance, and changes in gaze patterns.

• Compare mean déjà vu ratings between baseline, lead, and lag conditions.
• Perform repeated-measures ANOVAs to test whether temporal misalignment significantly influences déjà vu ratings.
• Correlate EEG indices of motor simulation strength (readiness potentials, event-related desynchronization in motor areas) with subjective déjà vu scores to determine if stronger predictive signals combined with time shifts yield higher déjà vu reports.

If déjà vu ratings significantly increase in conditions where the visual feedback leads the participant’s actual movement (i.e., the simulation’s expected outcome is “seen” before the real input should arrive), this supports the hypothesis that déjà vu is tied to an internal timing and ordering mismatch. Conversely, if no significant effect is found, or if lag conditions produce higher déjà vu, my hypothesis may need refinement.

• Decoding the contents and strength of imagery before volitional engagement