Abstract
Does iconic memory require attention? Pinto et al. (2013) suggest it does not. Persuh et al. (2012) suggest it does. The study we present extends and confirms the conclusion that iconic memory requires attention. Trial by trial, 30 observers were asked to report either one of the 2 rows in a 3x2 letter matrix located at fixation or whether there was an odd circle among 4 circles located at the corners of a notional square around the matrix. Arrays were visible for 250 ms followed by the cue indicating what to report. Attention to the matrix was manipulated in two ways: 1) By adjusting the perceptual load of the circle task; easy or hard, (within participant manipulation), 2) By changing the probabilities of whether the circle or matrix task was performed on a given trial, (between participant manipulation). The ratios of circle to matrix trials were 40/60, 60/40, 80/20. Ten subjects were tested in each ratio condition. An additional 30 observers were tested with the same procedure but gave whole matrix reports. We predicted that if attention is necessary for iconic memory the following should be true. The number of matrix items reported should be fewer when the circle task is hard and should decrease as the probability of having to report the matrix decreased, because in both cases attention should be drawn away from the matrix. Both predictions were confirmed. Performance on the matrix task decreased as a function of the difficulty of the circle task (F(2,108)= 38.52, p=.00, np2=.832) and as the probability of matrix reports decreased (F(4,108)= 28.45, p=.00, np2=.513). We also found the predicted significant advantage of partial over whole reports (F(2,108)= 9.99, p=.00, np2 =.156) indicating that in fact we were manipulating iconic memory. The results strongly support the conclusion that iconic memory requires attention.
Meeting abstract presented at VSS 2014