The Classic Cocktail Party

This is the result of covert attention (Driver 2001). The ability to attend selectively to auditory stimuli has been studied extensively using the dichotic listening technique (Driver 2001).
Perhaps even more amazing than the ability to perform multiple tasks accurately is the ability to attend to one channel of auditory information while completely ignoring any other auditory information. In real-world situations, this is demonstrated at cocktail parties or restaurants, where different conversations are being carried on all around you, but you are able to selectively attend to one of them by filtering out all the others. At any given time, you can tune out the current conversation and attend to another one simply by switching the focus of your auditory attention. Cherry (1953) conducted one of the first studies to examine the recognition and discrimination of human speech when two spoken messages are presented simultaneously. Cherry (1953) found that subjects had no trouble paying attention to only one ear and repeating the message they heard in that ear. An interesting finding for the unattended ear (the rejected ear) was that subjects had little or no memory for the message in that ear. Even when the speech in the unattended ear was reversed, subjects could not identify what they were hearing, although some of them did notice something queer about it. Most of the information that was retained from the unattended ear dealt with the physical properties of the speech sound (whether the speaker was male or female, whether the voice was high- or low-pitched, etc.). No information about semantic content was retained (Cherry 1953).
Wood and Cowan (1995) extended Cherry (1953)’s findings to gain a closer at what was occurring in the backward speech condition. It was believed that this condition would be near a threshold for detection, since it involves subtle physical changes and dramatic semantic changes. Further, they were interested in attention to the unattended channel before and after the switch to backward speech occurred. It was found that attention shifted to the unattended ear due to the fact that something different was happening to the irrelevant stimulus, rather than the change being found coincidentally due to periodic sampling of the irrelevant stimulus (Wood and Cowan 1995). This finding was true only for those subjects who were able to detect that a change occurred, whether or not they recognized it as backward speech. The authors fit their results within the framework of Treisman’s attenuation theory. This theory explains that changes in the irrelevant stimuli will be noticed if they trigger a shift to the unattended ear (Driver 2001). This shift is dependent on the severity of the change in physical characteristics and also the preexisting level of activation of the units in memory that are excited by the stimuli occurring after the change.
An interesting extension of the cocktail party effect to the visual modality was conducted by Shapiro, Caldwell, and Sorensen (1997). They used rapid serial visual presentation, in which the participant had to identify one or more targets. It has been consistently found that participants experience difficulty detecting the second target if it is within a certain temporal window after the first. This deficit in detection is called an attentional blink. It was found that