Abstract
Serial dependence, how recent experiences bias our current estimations, has been described experimentally during delayed-estimation of many different visual features, with subjects tending to make estimates biased towards previous ones. It has been proposed that these attractive biases help perception stabilization in the face of correlated natural scene statistics as an adaptive mechanism, although this remains mostly theoretical. Color, which is strongly correlated in natural scenes, has never been studied with regard to its serial dependencies. Here, we found significant serial dependence in 7 out of 8 datasets with behavioral data of humans (total n=760) performing delayed-estimation of color with uncorrelated sequential stimuli. Moreover, serial dependence strength built up through the experimental session, suggesting metaplastic mechanisms operating at a slower time scale than previously proposed (e.g. short-term synaptic facilitation). Because, in contrast with natural scenes, stimuli were temporally uncorrelated, this build-up casts doubt on serial dependencies being an ongoing adaptation to the stable statistics of the environment.
Footnotes
In this revised version we added a dataset from CamCan, which substantially increased our total n=760. In this dataset we also found significant build-up of serial dependence in the course of an experimental session. Additionally, we show that serial dependence build-up was not associated with trends in performance dynamics as a result of subjects getting familiar with the task or tiredness.
