Attention-Dependent Hemifield Asymmetries When Judging Numerosity

Abstract

Introduction: Previous research has indicated that numerosity judgments are significantly better for stimuli distributed across the left and right hemifields than for stimuli falling entirely within either lateral hemifield (Delvenne, Castronovo, Demeyere, & Humphreys, 2009). Here, we explored numerosity-task asymmetries between the left, right, upper and lower hemifields while requiring two different attentional selection criteria; proximity grouping versus similarity grouping. Method: Twenty Denison University undergraduates viewed briefly flashed (200 msec) displays containing spatially intermingled black dots and white dots across all four quadrants, and a central letter. The task was two-fold. After reporting the central letter, participants made a same / different numerosity judgment about a cued hemifield. Half of the trial blocks required judging numerosity based on proximity (e.g., upper left quadrant versus upper right quadrant). The other blocks required judging numerosity based on color-similarity (black dots versus white dots). The independent variables were hemifield (left, right, upper, lower) and grouping (proximity versus similarity), and were completely counter-balanced within-subjects. The dependent variable was d′; hits and false alarms were operationally defined as correct and incorrect “different” responses, respectively. Critically, retinal stimulation was identical across all (four hemifield × two grouping) eight experimental conditions. Results: The statistical interaction between hemifield and grouping was significant (F(3,57) = 4.234, p = 0.009, partial η² = 0.182). For proximity grouping there was a significant bilateral effect: d′ was significantly greater in the lower than upper hemifield (t(19) = 4.325, p < .001, partial η² = .496), but statistically indistinguishable in the left and right hemifields (t(19) = 0.365, p = 0.719, η² = .007). For similarity grouping there was a significant unilateral effect: d′ was significantly greater in the left than right hemifield (t(19) = 3.021, p = .007, partial η² = .324), but statistically indistinguishable in the upper and lower hemifields (t(19) = 0.587, p = .564, partial η² = .018). Conclusion: When judging numerosity, hemifield asymmetries vary systematically with attentional selection criteria, i.e., proximity grouping versus similarity grouping.