Tactile length contraction as Bayesian inference

Area of Science:

• Neuroscience
• Psychophysics
• Computational Neuroscience

Background:

• Perception relies on interpreting neural activity, which is inherently uncertain due to neural response stochasticity.
• Bayesian inference optimizes perception by integrating sensory input with prior expectations, but can cause illusions when expectations are violated.
• Length contraction illusions, a consistent underestimation of distances between rapidly presented stimuli, occur across tactile, auditory, and visual modalities.

Purpose of the Study:

• To test the prediction that tactile length contraction illusions intensify with decreased stimulus localization precision.
• To validate a Bayesian observer model proposing that these illusions arise from a low-velocity prior expectation.

Main Methods:

• A tactile psychophysical study involving 20 human participants.
• Participants compared a fixed reference distance (two taps, 1-s separation) with an adjustable comparison distance (taps with temporal separation t ≤ 1 s).
• Stimulus localization precision was manipulated by varying tap intensity (weaker taps reduce precision).

Main Results:

• Significant length contraction was observed, with perceived distances decreasing as temporal separation decreased.
• Weaker tactile stimuli, indicating reduced localization precision, significantly enhanced the magnitude of the length contraction illusion.
• These findings align with the predictions of the Bayesian observer model.

Conclusions:

• The results support the hypothesis that tactile length contraction illusions are a consequence of Bayesian inference with a low-velocity prior.
• The study reinforces the view that spatiotemporal perception is a probabilistic estimation process, integrating sensory evidence with prior expectations.
• This work provides empirical evidence for the role of Bayesian inference in explaining perceptual illusions.