Attention modulates both serial dependence and central tendency bias in spatial judgments
Poster Presentation 56.439: Tuesday, May 19, 2026, 2:45 – 6:45 pm, Pavilion
Session: Perceptual Organization: Neural mechanisms, models
Schedule of Events | Search Abstracts | Symposia | Talk Sessions | Poster Sessions
Saija M. Niemi1 (), Toni P. Saarela1, Maria Olkkonen1; 1University of Helsinki
Visual judgments are biased toward stimuli from preceding trials (serial dependence) and toward the overall stimulus mean (central tendency bias). Whether these two sequential effects reflect a shared mechanism remains unclear. We investigated whether they are similarly affected by attention in spatial judgments by manipulating task and stimulus distributions. In two separate sessions, sixteen participants estimated the width or the hue of colored ellipses in alternating trials using the method of adjustment. Each trial began with a cue to attend to either width or hue, followed by a stimulus for 500 ms and a 1250 ms delay. Participants then adjusted a probe to match the cued feature. In width-attended trials in both sessions, ellipse widths were uniformly distributed between 4.5 and 7.5 degrees of visual angle. Depending on the session, the widths in hue-attended trials were randomized from different distributions. If attention gates stimulus integration into the internal reference underlying central tendency bias, the bias should always be towards the mean of the width-attended stimuli. If attention has no effect on central tendency bias, it should be towards the mean of all widths in a given session and thus be different between the sessions. A linear mixed-effects model was used to quantify biases in width judgements from the cumulative mean of stimuli (central tendency bias) and four preceding stimuli (serial dependence) and their interactions with attention. Serial dependence in width judgments was stronger when width was attended in the previous trial compared to when hue was attended. Independently of serial dependence, we also found a central tendency bias which was toward the mean of the width-attended stimuli in both sessions. To conclude, serial dependence and central tendency bias have independent effects on judgments but both are modulated by attentional top-down processes.
Acknowledgements: This work was supported by funding from the University of Helsinki Doctoral School to SMN.