Rapid Serial Visual Presentation

Producers of music videos have often been known to include sequences where the images change 10 times per second or more. Despite the fact that each image is only present for about a 10th of a second, we typically have the impression of having seen the image clearly, even if we were asked to provide a complete list of all the images that were shown. This example leads to the topic of this entry, rapid serial visual presentation (RSVP), which demonstrates the remarkable ability of the human visual system to process rapidly presented visual information.

The use of RSVP as a research tool started in the early 1970s. In typical experiments, stimuli such as letters, digits, words, or pictures are shown in rapid succession at rates of up to 20 per second. In some cases, a characteristic such as color is used to specify the target, and the subject's task is to identify the target. One of the original motivations was to determine to what extent the natural rate of exploration of a visual scene using saccadic eye movements (about three fixations per second) is limited by the processing speed of the visual system, or whether the time taken to program and execute the eye movements is a significant factor. The fact that text can be presented at rates well above the three fixations per second typical of reading seems to imply that the bottleneck is probably not in the visual systems ability to process information. In recent years, there has been a great deal of interest in the possibility of using RSVP as a way of presenting text information on small displays, such as the ones used for mobile phones and personal digital assistants (PDAs). By automatically presenting a text as short strings of characters at a rate that can be adjusted to the users reading speed, even long documents can be read using a small display. Indeed, there is now a wide range of software applications available that allow users to try reading text without moving their eyes.

In the lab, RSVP techniques have been used to study a number of intriguing phenomena. These include repetition blindness, which is the observation that when the same target item appears twice in the same sequence, we often fail to notice the second presentation. Another intensively studied phenomenon is the so-called attentional blink, which occurs when two different targets occur close together in time (200–500 milliseconds, ms): Detection of the first target is followed by a transient drop in the ability to detect the second one. One might think that the first stimulus is somehow masking processing of the second one. However, it has recently been found that if the second target appears immediately after the first, performance is virtually intact (a phenomenon known as lag-1 sparing). This implies that the first target does not saturate visual memory, and that more than one target can be processed in one packet, as long as there is no distractor between them. One possibility is that the distractor appearing immediately after the first target segments the stream into chunks that are stored separately.

...