Psi man mascot


Attention is our ability to focus cognitive resources on a particular thing. Focal attention is, for all practical purposes, the same thing as working memory. George Miller discovered in the 1950s that working memory could hold a maximum of seven (plus or minus two) discon­nected items simultaneously.

What is a synonym for attention?

How does the cognitive system deter­mine what material is given attention? That is the problem of explaining selective attention.

In the late 1950s, Donald Broadbent came up with a filter theory of selective attention. He proposed that the human cognitive system had some mechanism (a filter) that could remove unwanted stimuli from attention.

The filter theory seemed to be supported by the cocktail party phenomenon described by E. Colin Cherry. Cherry pointed out that we can tune in on one conversation at a party while tuning out others.

Cherry verified the cocktail party phen­omenon experimentally. Participants could pick out one voice from others, even if all voices came from the same speaker. It helped if the voices were dissimilar, such as male vs. female.

What is the cocktail party phenomenon?

Further studies showed that the brain could use directional information to help separate conversations. If voices come from different directions, we easily tune one in and the other out.

Cherry also cited evidence that filtering was not complete. If one's name was mentioned in a conver­sation to which one was not paying attention, it was noticed.

This indicates that non-attended information is undergoing some processing. Evidently part of the brain monitors information to which we are not paying attention.

What is evidence that part of the brain pays attention to non-attended messages?

Similar findings came out of experi­ments on dichotic listening (which means "two ear" listening) by Anne Treisman (1960). Treisman put headphones on her subjects and played different messages to each ear.

Treisman instructed subjects to pay attention only to the message in one ear, not the other. To enforce this, Treisman had the subjects shadow one of the messages, which meant to repeat the message out loud, word for word.

Subjects had no trouble doing this. They repeated a message played to one ear while a different message (in the same voice) was played to the other ear.

What did Treisman discover, in the dichotic listening experiments?

Next Triesman tried suddenly switching the messages to the opposite ears, in the middle of a sentence. When she did this, the subjects "crossed over" to other ear for a few seconds to complete the sentence.

For example, if the voice going into one ear said, "To be or not to be..." and a moment later the same voice said, in the other ear, "...that is the question," subjects would shadow the whole sentence.

That violated their instructions. They were supposed to pay attention only to one ear.

Treisman thereby confirmed Cherry suspicion. Subjects do monitor information outside of attention, even if they are usually not aware of it.

After Treisman published this finding, other researchers confirmed that information in the nonattended ear could influence cognition. For example, MacKay (1973) found that information in the nonattended ear could disambiguate a word presented to the attended ear.

If the word in the right ear (which the person was shadowing) was bat and the voice in the unattended ear was talking about caves, bat was interpreted as referring to a flying mammal, not a wooden bat.

How did Mackay document an effect from messages in the unattended ear?

Findings like this led to the demise of the filter theory. If words in the nonattended ear were analyzed for meaning, there was something more complicated than a filter accounting for selective attention.

Broadbent (1962) modified his theory. Now he hypothesized a parallel process that came before attention. Its purpose was to direct attention toward something that might require conscious processing.

When we discussed driving on autopilot, at the beginning of Chapter 3, we mentioned something similar. A driver will snap out of the autopilot state if something unusual appears by the side of the road. Broadbent was proposing a monitoring stage prior to attention to enable this sort of intervention by consciousness.

How did Broadbent modify his theory in the 1960s?

Shiffrin and Schneider (1977) came to a similar conclusion after analyzing visual search experiments. When a person scanned a long list of items, something different or unusual drew attention automatically.

Shiffrin and Schneider referred to this as a pop-out phenomenon. For example, the letter "O" in a page full of X's will just pop out. One can hardly avoid noticing it. The same was true of hearing one's name, in the Cherry's cocktail party scenario.

Shiffrin and Schneider summarized this and much other evidence pointing toward two types of processing in human cognition. There is controlled processing fed by automatic processing.

Automatic processing calls attention to patterns which might be important, among other functions. Controlled processing is used to evaluate situations in more detail.

What "pop out" phenomenon did Shiffrin and Schneider discover?

This is our now-familiar two modes of processing again. Neisser pointed out in 1963 that unconscious (parallel) vs conscious (serial) distinction was discovered repeatedly by different researchers.

Broadbent, and then Shiffrin and Schneider, discovered it again. They described one system as automatic and operated with parallel processing, to screen large amounts of data. The other system was focused, serial or step-by-step in nature, and it corresponded to focused attention.

What distinction, repeatedly discovered, is relevant here?

A similar arrangement appears to exist in the visual system. We discussed how the mysterious phenomenon of blindsight is probably due to the existence of two distinct visual systems.

One visual system, the conscious one, goes from the optic nerves through the occipital lobes at the back of the brain. That supplies information for us to construct our visual world.

The other visual system goes through the superior colliculus. It is devoted mostly to guiding eye movements, and it is unconscious.

The collicular visual pathway has the ability to trigger eye movements. It can put a visual stimulus at the center of focus in the visual system.

That resembles the function of the early, parallel processing layer proposed by Broadbent, then again by Shiffrin and Schneider. The parallel processing system can pop things into attention.

What is evidence that the collicular visual system can draw attention to odd events?

Rudimentary pattern recognition would be helpful in such a system. If the second visual system can evaluate patterns in the periphery, it can decide which are unfamiliar and need further analysis.

Sure enough, Boehnke, Berg, Marino, Baldi, Itti, and Munoz (2011) located neurons in the superior colliculus of rhesus monkeys that responded to "oddball" events. The function of such neurons, they suggested, was "to support efficient selection in a cluttered dynamic world."

Analysis by synthesis

In his influential book Cognitive Psych­ology (1967), Ulric Neisser argued against the whole idea of a filter for attention. It was too passive, suggesting the cognitive system received inform­ation without seeking it out.

If we view the thought process as goal-oriented, attention is a construction aimed at accomplishing a purpose. Selective attention results from what we seek, not from what we fail to filter out. Neisser called his proposal analysis by synthesis.

What was Neisser's "analysis-by-synthesis" proposal, and how does the apple picking analogy clarify it?

Neisser explained the analysis by synthesis model with an analogy. If we see a man picking apples in an orchard, we assume his activity is determined by what he is seeking (ripe apples) not by what he is filtering out or choosing not to select (unripe apples, twigs, bugs).

We make this assumption because we recognize that apple picking is goal-directed activity. If attention is also seen as a goal-directed activity, then the problem of selective attention is the problem of explaining what is included, not what is left out.

How does analysis-by-synthesis deal with the cocktail party phenom­enon? Suppose you are at a party, straining to pay attention to what the person in front of you is saying.

This is a goal-directed activity. You are trying to construct the meaning of that person's speech, perhaps using direc­tional information or tone of voice to help you perceive the person's voice.

In short, your attention is an act of positive selection or synthesis, not an act of negative selection or filtering. Every­thing not included in your synthesis is left out of attention.

With a few exceptions–such as having our gaze drawn toward movement in peripheral vision, or having our attention activated when our names are called–the material that enters attention is that fitting the goals of cognitive activity.

This does not negate the idea of infor­mation passing through an early, diffuse, parallel processing stage. It suggests that the criterion for selection of that information, promoting it to attention, is its goodness of fit with goal-directed activity.

If non-attended information is relevant (if it finishes a sentence you have been shadowing, or involves your name, or represents a danger while driving) then it is positively selected. Then bottom-up, data-driven activity is allowed to integrate with ongoing activity in focal attention.

What is concentration?

The word concentration is commonly used to label a state of focused atten­tion: a rallying of cognitive resources. Concentration can be focused inward or outward, like attention in general.

When concentration is focused inward, the result can be an appearance of distraction. This is the source of stereotypes about absent-minded professors, lost in thought, barely noticing the surroundings.

When concentration is focused outward, people such as talented athletes develop a sort of "tunnel vision." They ignore all distractions. Distractions do not contribute toward their goals.

Details that contribute to the goals are allowed to penetrate. In fact, sensitivity to subtle cues is enhanced by concen­tration, as long as information relates to the focus of concentration.

Cognitive Reshuffling: Derailments in the Train of Thought

Thought processes sometimes unfold smoothly. That is when we use phrases like a stream of thought. However, the thought process can also be interrupted by unexpected events.

Researchers find evidence of reshuffling under the surface when something unexpected happens. Cognitive reshuffling was documented in EEG studies by Kutas and Hillyard (1980).

Kutas and Hillyard studied changes in the EEG caused by reading sentences such as "He spread the warm bread with..." The final word was either predictable (like "butter") or slightly unusual (like "mashed avocado") or totally weird (like "socks").

What was the Kutas and Hillyard research on "reprocessing"?

When subjects read sentences of the slightly or totally unusual variety, Kutas and Hillyard found a distinctive wave in the EEG, the N400 potential (a negative potential 400 milliseconds after the word was presented). They suggested this was due to "reprocessing" or a "second look" required by the unexpected sentence ending.

What did researchers observe, when they studied reactions in Candid Camera?

Subtle signs like an eyelid fluttering or the corner of the mouth twitching may reveal cognitive reshuffling. Barker (1968) found that people twitch and their EEGs may shown minor seizure-like disturbances when they confront a difficult thought.

Babies several months old often display a moderately disordered motor response–hiccups–as a reaction to unexpected events. These bodily reactions may be part of a "reset" reaction similar to what is betrayed in adults by eyelids fluttering.

How do babies sometimes respond to unexpected events?

When adults are confronted with a new and unexpected event, requiring a reshuffling of thought, they momentarily go blank in the face. Scheff (1985) demonstrated this in a unique study where researchers examined old tapes of Candid Camera episodes.

In Candid Camera people were deceived as a joke. Absurd situations would be set up, then the truth would be revealed as the victim was told, "Smile, you're on Candid Camera!"

Slow motion replays of the victim's facial expressions at the moment of truth showed a surprise reaction "manifested by widening of the eyes, lifting of the brows, and, in most instances, opening of the mouth." After this surprise face came an emotional reaction, usually embarrassment, "signaled by hiding the face or mouth" while laughing.

What differences were observed in facial expressions of people who did, or did not, catch on to a prank played on them?

Scheff also analyzed segments in which the person never realized what was going on. Some victims of the practical jokes failed to comprehend what was happening to them, even after being informed they were on Candid Camera. They kept playing along with the set-up situation as if it was real.

Scheff found that these people did not show the split-second look of surprise, nor did they show an emotional reaction. They did not reset their mental activity, and their train of thought continued along the same track as before.


Barker, W. (1968) Brain Storms: A Study of Human Spontaneity. New York: Grove Press.

Boehnke, S. E., Berg, D. J., Marino, R. A., Baldi, P. F., Itti, L., & Munoz, D. P. (2011) Visual adaptation and novelty responses in the superior colliculus. European Journal of Neuroscience, 34, 766-779.

Broadbent, D. E. (1962). Attention and perception of speech. Scientific American, 206, 3-9.

MacKay, D. G. (1973) Aspects of the theory of comprehension, memory and attention. Quarterly Journal of Experimental Psychology, 25, 22-40.

Scheff, T. J. (1985). The primacy of affect. American Psychologist, 40, 849-850.

Shiffrin, R. M. & Schneider, W. (1977). Controlled and automatic human information processing: II. Perceptual learning, automatic attending, and a general theory. Psychological Review, 84, 127-190.

Write to Dr. Dewey at

Don't see what you need? Psych Web has over 1,000 pages, so it may be elsewhere on the site. Do a site-specific Google search using the box below.