Abstract
Working memory (WM) capacity measurements are typically taken in laboratory settings using 2D stimuli and delay periods devoid of sensory stimulation. However, this situation is almost inexistent in real life where our senses are constantly bombarded with 3D sensory signals. We have therefore designed a 3D videogame (Game of Runes) to measure and train WM capacity, using a commercial videogame engine (Unreal). The engine allows simulation of realistic environments with precise control over events timing and stimulus complexity. On each trial, 2 to 8 sample items were simultaneously presented at random positions on a virtual surface (15 possible locations) for a total duration of 0.5 seconds × # of items. Stimuli consisted of Futhark Runes, a pre-Germanic alphabet containing 24 characters, sufficiently different from the Latin alphabet to prevent word formation. Following stimuli presentation, the player was required to navigate in a rich environment to the test section (variable delay period; median: 5.397 seconds). For successful trial completion, the player had to select each remembered rune amongst 16 randomly chosen test runes. To evaluate WM capacity (N), we fitted probability distribution functions (PDF; N=2-8 items), to each player's performance (percent of correct trials per # of items). Using a least squares algorithm, we determined the average capacity across 8 subjects to be 5 items. Finally, upon investigating the time difference between each rune's selection, we found no major discrepancies amongst rune number 3, 4, and 5 (Ranksum Test; p=0.03, p>0.05, p>0.05 respectively), however, there was a clear increase in selection time for rune 6 (p=0.0001), suggesting that subjects start guessing after remembering 5 items. We further observed, in at least one subject, an increase in N after playing multiple games. Our results show that 3D videogames can serve to measure and train WM capacity, yielding at least similar results as conventional methods.
Meeting abstract presented at VSS 2014