Density is a precursor in the perception of visual numerosity: Evidence from adaptation.

To dissociate density-based and number-based explanations of the aftereffects affecting visually perceived numerosity, a preregistered psychophysical experiment was conducted with 60 participants.The experiment was designed to address a recent report claiming that numerosity aftereffects could not be explained by density adaptation.Each participant was adapted to either high-or low-density adapters that were either larger or the same size as the standard test patch.During separate measurement blocks, the perceived density, numerosity, and size of the test patch were each assessed using comparison stimuli that varied along all three dimensions, so that the intended test dimension was dissociable from the other two.The primary results showed that numerosity aftereffects were fully accounted for by a mixture of size and density adaptation, whereas density aftereffects were solely dependent on adapter density and unaffected by adapter size or number.Exploratory analyses found that large aftereffects in numerosity were correlated with measured aftereffects of patch size and density, whereas density and patch-size aftereffects showed no significant correlation with each other.Taken together, these results suggest that density and patch size are normally perceptual precursors of numerosity and that numerosity perception occurs at a later stage of processing. Public Significance StatementMathematical ability has been linked to perceptual abilities related to visually perceived number.The claim that visual number is an adaptable dimension has received widespread acceptance, despite a poverty of data dissociating number adaptation from other possible forms of explanation, including aftereffects of density and size.Whereas most studies have measured only a single dimension at a time, holding others constant, our approach here has been to vary multiple dimensions during measurement to clearly differentiate judgments of number from judgments of density.The aftereffect data support the contention that numerosity can typically be derived from area and density and that aftereffects on perceived numerosity are caused by adaptations that primarily affect encodings of density and patch size.