A tolerance window exists between tactile feedback and visual/auditory events, beyond which they're perceived as out of sync

What it is

For tactile feedback and an accompanying image or sound to be perceived as "the same event," they don't need to be aligned in time down to zero delay — the brain tolerates a range of temporal offset between them, and as long as the offset falls within that range, the two are still experienced as one synchronized event. This range is the temporal binding window. Beyond it, two signals that should have been paired are instead noticed as "out of sync," or even treated as two unrelated events.

Why it happens

This window exists, rather than the brain demanding strict zero-delay alignment, for two reasons. First, the neural conduction delay from physical stimulus to cortex differs across modalities — tactile and auditory signals generally reach cortex faster than visual signals — so demanding absolute synchrony would actually cause the brain to misjudge multisensory signals from one real-world event as out of sync, purely because of this inherent difference in conduction delay. Second, in the real world, light, sound, and mechanical vibration travel at different physical speeds, so the visual, auditory, and tactile signals produced by one real event (striking an object, say) were never going to arrive at the senses at exactly the same instant to begin with — through long-term statistical learning, the brain has internalized this naturally occurring, bounded temporal offset as "normal," and only flags a mismatch as abnormal, judging it out of sync, once it exceeds this learned tolerance.

Studying it

The standard paradigm for measuring this window is a simultaneity judgment or temporal order judgment task: the time gap between two modality-specific stimuli (the stimulus onset asynchrony, or SOA) is systematically varied, from one modality clearly leading to the other clearly leading, and subjects judge "whether the two occurred at the same time" or "which appeared first." Judgments across different SOAs are fit to a psychometric function, from which the window width and the point of subjective simultaneity (the actual SOA at which the proportion of "simultaneous" judgments peaks, often not zero) are extracted. This paradigm is reusable across modality pairings — audio-visual, tactile-visual, tactile-auditory.

Where it stops holding

Window width isn't a fixed constant: different modality pairings (touch with vision, touch with audition) have their own window widths to begin with; specialized training (a musician's experience with auditory-tactile synchrony, for instance) narrows the window and makes judgments stricter; higher cognitive load or divided attention widens the window and raises tolerance; and whether the two stimuli share a spatial location or a semantic relationship (feeling more like "two sides of the same event") also affects window width — the more spatially or semantically related the two stimuli are, the wider the tolerance window generally is.

Applying it

• When designing feedback that pairs touch with vision or audio — a click sound plus a confirmation vibration, an animation plus an emphasis buzz — keep the presentation delay between them within the known tolerance window, rather than assuming "roughly at the same time" is good enough. The specific window width should come from measurements for the target modality pairing, not borrowed from a different pairing's window (audio-visual, say).
• How to verify: run a simultaneity-judgment test using the actual target software/hardware pipeline (the full delay from triggering event to tactile feedback actually reaching the user), confirming the measured end-to-end delay falls within the range subjects report as "synchronized," rather than relying on the theoretical delay figure listed on a device's spec sheet.

Related

• Same group: A4.17.1 Touch resolves rapid pulse trains worse than hearing does, but better than visual flicker resolution · A4.17.3 Tactile delay is noticed more readily than visual delay, because touch is often bound to a self-generated action · A4.17.4 Multiple simultaneous touch points delivered too close together in time fuse into a single perceived event instead of several
• Nearby: A4.03 Vibrotactile frequency sensitivity band
• Search terms: temporal binding window · simultaneity judgment · point of subjective simultaneity · multisensory integration