Abstract
Categorical information is known to benefit visual short-term memory (VSTM); however, the degree to which memory load influences the exploitation of categorical information is unclear. In this study, 20 human adults completed an object/position recall experiment with two conditions, in which the memory display comprised a sequence of two (low memory load) or four (high memory load) Snodgrass stimuli presented sequentially, each for 400 ms, at one of 64 random locations. Stimuli belonged to one of 14 semantic categories (four-footed animals, insects, fruits, etc) tested equally often. A concurrent verbal loading task was performed. At test, in the object memory condition, a spatial marker appeared at the location of one of the memory stimuli, selected at random. Observers reported (verbally) the name of the stimulus shown at this location. In the position memory condition, a test stimulus was shown above the memory display area, together with numbered markers at each memory stimulus location used. Observers reported the number corresponding to the location of the test stimulus in the preceding memory display. Recall rates were analyzed using one-way repeated measures ANOVA (with stimulus category as a within-subjects factor) for each memory load. At high memory load, a significant effect of stimulus category was observed in both experimental conditions [object memory, F(13, 247) = 4.87, p <0.01; position memory, F(13, 247) = 2.67, p <0.01]; at low memory load, a significant effect of stimulus category was observed only in the object memory condition, F(5.18, 97.23) = 4.87, p = 0.02. These findings suggest that categorical information is exploited for both object and position memory tasks at high memory loads, but, at low memory loads it is exploited only for object memory. This may reflect the propensity to recruit additional (non-visual) stimulus information only where VSTM is stretched by task demands.
Meeting abstract presented at VSS 2012