Abstract
Although it is yet unknown what determines which image we come across in our daily lives will be remembered, physical properties of visual stimuli are typically not considered to substantially contribute to memory. It was recently found that during naturalistic encoding image size influences memory such that bigger images are better remembered. Here we hypothesized that during naturalistic encoding images viewed for longer duration would be better remembered. After replicating the image-size on memory effect in an online experiment (n=189), we ran additional online experiments to test the effect of presentation-duration on memory with new naïve participants (n=90). In this experiment participants freely viewed 160 images presented for 250, 500, 1000, or 2000 ms (image onsets always 2500ms apart) without being aware of any memory related aspect of the experiment. They were later given a surprise old-new recognition test (320 images, 50% old). A main effect of presentation-duration on memory was evident by longer presentations leading to better memory (250ms<500ms<1000ms, post-hoc p’s<0.05). As in the previous study, we found a main effect of visual category on memory with faces best remembered and outdoors the least. Since enlarging image size and extending image presentation duration involve different neural mechanisms, we assumed that the contribution of image size and presentation-duration would be independent. To test that we ran a new study (n=93) with small/large images presented for short/long durations during the exposure phase, while at the surprise recognition memory test all images were presented at the same size and for the same duration. As expected from the earlier studies we found effects of size (p=5.8*10^-7), and presentation-duration (p=9.29*10^-6). Importantly, there was no interaction between size and presentation-duration (p=0.906). These results suggest that during naturalistic encoding, both image size and its presentation duration influence memory and their influences may be independent.