Abstract
We report a novel illusion in the perception of sequential random texture displays. In 3 experiments, participants consistently perceive alternating vertical apparent motion when random textures were shown sequentially. In Experiment 1, 18 participants viewed 6-frame sequences of 140x140 pixel square arrays presented at 2.5 Hz. The first frame was a random array of black and white pixels; the 5 subsequent frames were shifted by 4 pixels in some cardinal direction, with additional pixels added to the receding edges. Half of the displays showed alternating apparent motion (Up-Down-Up-Down-Up or Right-Left-Right-Left-Right) and half showed same-direction motion (Up-Up-Up-Up-Up or Right-Right-Right-Right-Right). Participants were asked to indicate if they detected apparent motion, and if so, which type. Across trials, varying amounts of noise were added to the frames, ranging from 0-noise (100% coherence) to 1-noise (0% coherence). Participants' performance at 0-noise was nearly perfect (95.5%) and dropped sharply as noise was added. Interestingly, strong biases emerged at noise levels above 0.5. Specifically, at 1-noise participants reported perceiving coherent motion on 41% of trials; of those, 56% were perceived as alternating vertical, and 25% as alternating horizontal, indicating statistically reliable alternating (p=0.002) and vertical (p=0.01) biases. In Experiments 2 and 3, we measured hysteresis by adding or subtracting noise dynamically. In Experiment 2, 27 participants saw sequences starting at 0-noise in which noise was incrementally added, and in Experiment 3, 35 participants saw sequences at 1-noise in which noise was incrementally removed, revealing one of the 4 types of underlying apparent motion. In both experiments, participants consistently perceived alternating vertical motion even at 1-noise. Together our data indicate 3 biases in the perception of sequential random texture displays: (1) perception of coherent apparent motion when there is none; (2) perception of alternating (over same-direction) motion; and (3) perception of vertical (over horizontal) motion.
Meeting abstract presented at VSS 2014