Abstract
Background: Spontaneous eye-blinking is known to serve an important physiological function, but recent evidence show that it is also linked with cognitive processes. Here we show that the eye-blinks, although rare, are similar to microsaccades in their modulation pattern in response to transient perceptual events, including the temporal dependencies on preceding events, which uncover stimulus predictability. Method: During fixation observers (n=8) viewed and silently counted sequences of 100 randomly-ordered stimuli of two types, presented at 1Hz repetition-rate. These included small patches of contrast (high/low), color (red/blue) and audio-visual stimuli (beep/circle), with an additional go/no-go CPT experiment (n=22 observers, 20min-long, 0.5Hz-rate) on target position. Eye-tracking data were used to compute the average latencies of the blinks (if present) and their rate modulation following stimulus onset. Results: The overall blink-rate varied across observers and conditions (1-80% of trials, 20-60% on-average). Blinking was inhibited around stimulus onset in all conditions, with a pattern generally similar to that observed with microsaccades, including higher-magnitude and shorter-latency for higher contrast. For the audio-visual and CPT experiments, repetition (e.g. sound after a sequence of sounds) decreased the inhibition-rebound latency and magnitude, while change (e.g. sound after a sequence of visual-patches) increased them. The magnitude of these effects changed systematically with the number of preceding items (10-20ms per item, ~100ms total). The contrast experiment showed only a blink-rate increase for high-contrast patches preceded by low-contrast sequences. Conclusion: Eye-blinks, like microsaccades, obey a general pattern of oculomotor inhibition in a lawful and precise manner. This inhibitory pattern depends on stimulus saliency and in some cases on the relation between the current and preceding events, with higher-rate and delayed blinking for "surprise", which we account for by a simple prediction-error model. Given the sparse and apparently imprecise pattern of blinking, reading the current findings might induce "blinking by surprise".
Meeting abstract presented at VSS 2015