“ The leans” is a spatial illusion encountered by pilots flying without reliable visual references. It can occur when a pilot promptly returns to straight-and-level flight after unawarely having entered a banked turn. The return to straight-and-level flight is sensed by the semicircular canals and erroneously perceived as banking in the opposite direction. The purpose was to establish (1) the magnitude and interindividual variability of the leans, and (2) if centrifuge-based spatial-orientation (SO) training, improving roll-tilt assessment, would affect the illusion. In a fighter-jet-trainer aircraft, pilots ( n = 10) were sitting in darkness and setting a luminous line parallel with their imagined horizon, while being exposed to a gradual entering of a 15–20° banked turn, followed by a prompt return to straight-and-level flight. Measurements were performed before and after a 4-week period, during which five of the pilots undertook a 9 × 1 h SO-training regimen. During the pre-training trial, the pilots perceived on average 47% (range = 0–140%) of the roll-angular displacement encountered during the transition from the turn to level flight, i.e., they perceived being roll-tilted 10 ± 9° in the opposite direction. SO-training did not affect the leans magnitude ( p = .84). The results confirm anecdotal accounts of marked interindividual variability of the leans illusion, and suggest that such variability in the canal-induced illusion exists even without perceptual interference from visual input.
We investigated whether the pattern of illusion effects differs between length and position judgments in the Giovanelli illusion, the gravity lens illusion, and Yang's iris illusion. The results revealed that length judgments consistently produced assimilation across all three illusions, whereas position judgments produced assimilation in the Giovanelli and gravity lens illusions but contrast in the iris illusion. The common assimilative effects observed in length judgments may reflect similarities in the global configurations of these illusions. In contrast, the different effects observed in position judgments may arise from differences in the perceptual organization of local stimulus elements. These findings suggest that perceptual organization depends not only on stimulus configuration but also on the perceptual dimension being judged and that different reference structures may underlie length and position processing.
Active vision relies on different forms of eye movements that require the usage of different spatial coordinate systems for successful vision. This is often overlooked in visual psychophysics, when tasks are performed under fixation only, inherently confounding retinal and real-world coordinates. Here we investigated whether the psychophysical principle of luminance integration across time generalizes to smooth pursuit eye movements and in which spatial coordinate system integration operates. In Experiment 1, participants smoothly tracked a moving object, and two brief liminal luminance targets (LTs) were presented with a short interval between them in retinally congruent but incongruent real-world (screen) coordinates. Luminance of the successive LTs was integrated, as reflected in enhanced detectability and perceptual awareness ratings compared to a single LT and to two LTs with an interval exceeding 100 ms. In Experiment 2, we ruled out alternative perceptual accounts through apparent motion and through the percept of a single, spatially extended stimulus. In Experiment 3, we replicated luminance integration in retinal coordinates during smooth-pursuit eye movements but observed no robust evidence for luminance integration operating in real-world coordinates. Our findings confirm the implicit assumption that temporal luminance integration mainly relies on early (e.g., retinal) processes and illustrate the importance of testing alternatives, including active vision, for a full understanding of the principles of visual psychophysics.
Previous studies have revealed that inverted faces are perceived as physically larger than upright faces, called the inversion-induced size illusion—a phenomenon thought to arise from disrupted configural processing. It remains unclear whether this illusion generalizes across other face types (such as real child and cartoon child faces). Thirty Chinese university students completed a psychophysical size-comparison task using the method of constant stimuli under a simultaneous presentation paradigm with limited exposure duration. Experiment 1a replicated the classic inversion-induced size illusion with adult faces. Experiment 1b observed a comparable and significant illusion for real child faces, indicating that developmental differences in facial morphology do not abolish the effect. In contrast, no inversion-induced size illusion was observed for cartoon child faces. Direct comparisons revealed a larger magnitude of the size illusion effect for real child faces as compared to cartoon child faces. Together, these findings suggest that the inversion-induced size illusion is not a general property of all face-like stimuli. Instead, its occurrence appears to depend on specific stimulus characteristics under the present experimental conditions, including—but not necessarily limited to—the preservation of biologically plausible facial structure.
Perceptual illusions provide important clues about visual processing mechanisms. The filled-space illusion demonstrates how contextual information within a visual stimulus can distort perceived length, whereas polar anisotropy reflects systematic differences in perceptual performance across the visual field. This study investigated the relationship between the filled-space illusion and visual-field anisotropy by quantifying their combined effects on perceived length. In psychophysical experiments, stimuli consisted of two spatial intervals. The tilt angle of the test interval was fixed at one of the five orientations relative to the horizon, while the orientation of the reference interval varied from 0° to 360° relative to the test interval. In the first block of experimental trials, both intervals were empty, allowing us to assess visual-field polar anisotropy. In the second and third blocks, radially or tangentially oriented line segments were attached to the terminator of the reference interval to examine the combined effects of visual-field anisotropy and the filled-space illusion. The experimental data were analyzed using computational procedures based on a previously developed model of the hypothetical visual mechanism underlying the filled-space illusion. The results showed that the filled-space illusion component increased as the stimulus expanded and reached its maximum at the straight internal angle. The model calculations were consistent with the experimental data, suggesting that polar anisotropy may be related to vertical scale stretching and that its contribution may be distinguishable from the filled-space illusion component, although alternative explanations cannot be excluded.
