Courtney Turner¹ (), Ella Williams², Roger Remington³, Anna Grubert¹; ¹Durham University, ²Oxford University, ³University of Minnesota

The ability to temporarily memorise visual information is an essential cognitive ability used in everyday life. Such information is held in visual working memory (vWM) and the capacity of this short-term storage differs individually. Importantly, higher levels of vWM capacity have been linked to higher levels of fluid intelligence and enhanced cognitive abilities such as logical reasoning, problem solving and comprehension. However, such correlations are only meaningful if individual vWM capacity is stable and the measures to assess vWM capacity are robust. We tested this assumption by means of five change detection tasks in which we manipulated visual/perceptual stimulus parameters between conditions and measured the effects of these manipulations on individual vWM capacity as measured with Cowan’s K. In each trial, participants were presented with a memory display (100ms) with differently coloured squares. After a retention period (900ms), they were shown a test display (100ms) and had to decide whether it was identical to the memory display or contained a colour change. In different tasks, we manipulated the number of memory items (Experiment 1), the eccentricity of the memory items from fixation (Experiment 2), the figural organisation of the memory items (Experiment 3), the featural identity of the memory items (colour, shape, alphanumerical category; Experiment 4), or the predictability of the stimulus locations in the test display (Experiment 5). Individual K values changed significantly between task conditions in all experiments (apart from Experiment 3), suggesting that individual vWM capacity measures are not robust against changes in visual/perceptual task parameters. However, in each experiment, the different K values produced in the different task conditions were positively correlated. Taken together, these findings suggest that K values may lack validity as absolute measures of vWM capacity, but that they seem to be reliable measures of relative individual vWM abilities.

Acknowledgements: This work was funded by research grants of the Leverhulme Trust (RPG-2020-319) awarded to AG.