Temporal Trend in Serial Dependence
Advised by Prof. Huihui Zhang and Prof. Huan Luo, PKU
Serial dependence, a systematic bias on our current estimations caused by immediate past experiences, is thought to be an integrative mechanism adaptive to the autocorrelated structure of the natural world, implying that it should get weaker when exposed to long sessions of uncorrelated stimuli. However, few studies have investigated this prediction directly and little is known about what aspects of serial dependence would be affected.
In this study, with two experiments and a reanalysis of a published dataset, we found a narrowing trend in the spatial tuning window of serial dependence, revealed by a decrease in serial dependence over time between stimuli with different positions but not between stimuli with the same position. We also found a modulation effect of color that such a trend was only found between stimuli with the same color but not between stimuli with different colors. These results provide evidence in support of the adaptive nature of serial dependence and offer hints for understanding the mechanism that controls the temporal trend of serial dependence.
