ISSN 0439-755X
CN 11-1911/B

Acta Psychologica Sinica ›› 2026, Vol. 58 ›› Issue (7): 1312-1324.doi: 10.3724/SP.J.1041.2026.1312

• Reports of Empirical Studies • Previous Articles     Next Articles

Serial dependence for same-category and cross-category stimuli in social attention

WANG Da, YANG Zhihao, MEI Gaoxing   

  1. Department of Psychology, School of Education Science, Jiangsu Normal University, Xuzhou 221116, China
  • Received:2025-08-22 Published:2026-07-25 Online:2026-05-15

Abstract: Serial dependence refers to a phenomenon in which current perceptual judgments are biased toward the recent past. This effect plays an elemental role in maintaining perceptual continuity in changing and noisy visual environments. Serial dependence is widely found not only for low-level stimulus features (e.g., grating direction), but also operates within high-level social cues (e.g., eye gaze direction) in social attention. Social attention refers to the ability to detect others' focus of attention and infer their intentions through social cues such as eye gaze. Although different social cues (e.g., eye gaze direction and biological motion direction) vary in their perceptual features, they can convey the same information about others' intentions and effectively recruit social attention. However, whether the serial dependence effect could occur between different categories of social cues remains largely unknown. To this end, the present study used an inducer paradigm in four experiments to investigate whether the same-category and cross-category serial dependence could exist in social attention.
Experiment 1a (27 participants) examined whether the serial dependence effect could exist between social cues of the same category (i.e., eye gaze direction). In each trial, an inducer stimulus (extremely leftward or rightward gaze direction), a reference stimulus (straight gaze), and a probe stimulus (randomly selected from seven levels of eye gaze directions ranging from -40% to +40% with a 13.3% step) were sequentially presented. Participants were then asked to judge which stimulus, the reference or probe, was perceived more leftward or rightward. Using the same task as Experiment 1a, Experiment 1b (24 participants) investigated whether an emotional valence (happy/angry) could modulate the same-category serial dependence in social attention. In this experiment, the inducer, reference, and probe stimuli could each be either happy or angry. Crucially, within a given trial, the emotional valence was held constant across all three stimuli. Experiment 2a (27 participants) investigated whether a serial dependence effect between different categories of social cues could exist. The same task as that in Experiment 1a was used except that the inducer stimulus was replaced with a biological motion stimulus (leftward or rightward). Experiment 2b (27 participants) used a non-social cue (i.e., arrow) as the inducer stimulus to further examine whether the cross-category serial dependence effect could exist between social and non-social cues.
The results of Experiment 1a revealed a significant attractive serial dependence between eye gaze directions. Specifically, individuals' judgments of the current reference stimulus were biased toward the gaze direction of the previously perceived inducer stimulus. Consistent with previous studies, this finding thereby supported the existence of a serial dependence effect in the perception of eye gaze direction. Moreover, the serial dependence effect was modulated by facial emotional valence of eye gaze (Experiment 1b), with stronger serial dependence observed under the angry condition compared to the happy condition. Notably, the results of Experiment 2a showed that the direction of biological motion also induced an attractive serial dependence in judgments of gaze direction, indicating that serial dependence occurred across different categories of social cues. This experiment provided the first evidence for the cross-category serial dependence in social attention. However, although arrow direction (a non-social cue) shares similar directional properties with eye gaze, the serial dependence effect did not emerge when the arrow was used as the inducer stimulus (Experiment 2b).
Taken together, our current findings suggested that serial dependence effects in social attention existed not only between same-category social cues, but also between different- category social cues. However, the effect did not appear between social and non-social cues. These results indicate that the serial dependence effect in social attention arises from higher-order social information rather than low-level directional cues.

Key words: serial dependence, social attention, eye gaze, biological motion