Hack39.Avoid Holes in Attention

Hack 39. Avoid Holes in Attention

Our ability to notice things suffers in the half-second after we've just spotted something else.

A good way to think about attention is as the brain's way of paring down the sheer volume of sensory input into something more manageable. You can then concentrate your resources on what's important (or at least perceived to be so on first blush) and ignore the rest. If processing capacity weren't limited, perhaps we wouldn't need attention at all-we'd be able to give the same amount of concentration to everything in our immediate environment, simultaneously.

Another reason we continually pare down perception, using attention as a final limiting stage before reaching conscious awareness, could be that perception causes action. Maybe processing capacity doesn't intrinsically need to be limited, but our ability to act definitely is: we can do only one major task at a time. Attention might just be a natural part of conflict resolution over what to do next.

M.W.

Attention isn't the end of the chain, however. There's conscious awareness too. The difference between the two is subtle but important. Think of walking down a street and idly looking at the faces going by. Each face as it passes has a moment of your attention, but if you were asked how many brown-haired people you'd seen, you wouldn't have the slightest idea.

Say somebody you recognize passes. Suddenly this semiautomatic, mostly backgrounded looking-at-faces routine jumps to the foreground and pushes the face into conscious awareness. This is the act of noticing.

It turns out the act of noticing takes up resources in the brain too, just as paying attention does. Once you've noticed a face in the crowd, there's a gap where your ability to consciously notice another face is severely reduced. It's a big gap tooabout half a second. This phenomenon has been dubbed the attentional blink, drawing a parallel with the physical eye blink associated with visual surprise.

Attentionjust like vision, which cuts out during eye movements [Hack #17] is full of holes that, as a part of everyday life, we're built to ignore.

3.7.1. In Action

There's a standard experiment used to induce the attentional blink, using a technique called rapid serial visual presentation (RSVP). RSVP consists of projecting black letters onto a gray screen, one at a time, at about 10 letters a second.

You're instructed to watch the stream of letters and be on the lookout for two particular targets: a white letter and the letter X. Spotting either on its own is easy enough. One-tenth of a second (the length of time a letter is on-screen) is enough time for recognition and awareness. Spotting the targets when they're close together in time, however, is much harder.

If the letter X follows the white letter by five places or fewer, you'll probably miss it. Spotting the white letter, the first target, stops the second target, the X, from reaching conscious awareness. That's the attentional blink.

Obviously this isn't an easy test to do at home, but we can approximate it using speed-reading software. Speed-reading software often has a function to run through a text file, flashing the words up sequentiallyand that's what we'll use here.

You can use whichever software you like. I used AceReader Pro (http://www.stepware.com/acereader.html; $49.95; trial version available).

Although the AceReader Pro trial version is suitable for this small test, it is available only on Mac and Windows. FlashWare (http://www.flashreader.com) is a simple, freeware Java applet that takes a file as input for rapid serial visual presentation. GnomeRSVP (http://www.icebreaker.net/gnomersvp) and kRSVP (http://krsvp.sourceforge.net) are speed-reading applications for the Gnome and KDE Linux desktops, respectively.

Whichever piece of software you choose, you'll need it to have a mode that lets you load an arbitrary file and step through it at about 300-400 words a minute. For AceReader Pro, that means choosing the Online Reader & Expert Mode option.

You'll need a text file, preferably one you haven't read. Ask a friend to choose two relatively unusual words for you from a random place in the text, making sure they're only two or three words apart. These are the words you have to look out foryour targets.

Now load the text into the speed-reader software (in AceReader Pro, choose File Load File), set the words per minute (WPM) to 400, and click the Play button (the green triangle) to begin. What you're expecting to experience is that you'll spot the first word easily and miss the next one completely. Figure 3-7 shows AceReader Pro in action; you'd notice the first (left), but the second (right) would go utterly without notice.

Figure 3-7. AceReader Pro presenting target words

With this particular experiment, nobody experiences the attentional blink every time. For instance, if you're already good at speed-reading or it's easy to guess the sentences in the text document as they come up, it probably won't work. We're using this software to simulate the controlled RSVP experiment, which uses random letters. Doing it this way isn't as reliable.

That said, it worked for me about half the time, and I can only describe the attentional blink itself as a peculiar experience. At about five words a second (300 words a minute), I wasn't overwhelmed by having to read every word and decide whether it was one of my targetsbut I was certainly on the cusp of being overwhelmed. I had to sustain a high level of concentration on the screen.

The first word jumped out at me, as I expected it would. OK, I'd recognized that one; now I could look out for the next. But the next word I remember reading properly was four places after. I'd somehow missed my second target. What had occurred in between was my attentional blink. Thinking back, I could remember the sensation of having seen my second word on the screen, but somehow, although I'd seen it, I hadn't twigged that it was actually my target. My memory was distinctly less visual and sure than for the first word, and all I could really remember, for the duration of the blink, was the feeling of doing two things at once: processing the first target and trying to keep up with the fresh words on-screen. If I hadn't been able to stop and figure out why I hadn't noticed my second target, knowing it had to have flashed up, I would've missed it completely.

3.7.2. How It Works

Clearly the attentional blink does exist. The half-second recovery time after noticing a target has been shown many times in experiments. Like attention in general, however, precisely how it arises in the brain is still subject to research.

One strong theory assumes there's a limited amount of attention to go round, which is rapidly transferred from one letter to the next in the rapid serial visual presentation task. Due to the amount of processing each letter needsto see if it's white or if it's the Xand the speed of change of letters, attention is forced to operate at maximum capacity. When the white letter, the first target, is spotted, additional attentional resources are suddenly needed to lift it to a level of conscious awareness. These extra resources have to come from somewhere, and the process of raising one's awareness takes time; for that period of time, new incoming letters aren't given as much attention as they really need.

That's not to say new letters aren't given any attention at all, and that's where the analogy with eye blinking breaks down. Eye blinks shut off vision almost completely, but attentional blinks just reduce the probability of spotting a target during the blink. The success rate for spotting the second target, the X, dips to its minimum of 50% if the second target occurs a quarter of a second (250 ms) after the first target and then gradually recovers as the half-second plays out.

In this view, it's not so much that the second target doesn't get seen at all, it's that it gets processed but there just isn't enough attentional resource to go around and so it isn't brought up to conscious awareness. Additional, random letters keep coming in and claim the processing resource for themselves, and so you never notice that second target.

Two pieces of evidence back this up. First, the processing demand contributed by the random letters is essential for the attentional blink to show up. If the letters aren't there, or instead something that is easily ignored is used (like blocks of random colors, perhaps), they don't act as a processing drain. The second target is seen as easily as the first target in that case.

Second, although the second target may never reach conscious awareness, it can still influence the subconscious mind. There's an effect called priming, in which seeing a word once will make it, or a related word, easier to notice the second time [Hack #81] . So, for example, in the RSVP task, if shown the word "doctor," the subsequent word is faster and easier to spot if it's the word "doctor" or "nurse."¹ It turns out that the second target, even if it isn't consciously noticed, can prime the next item. This means that the items shown during the attentional blink reach the level of processing required for meaning, at least, and aren't just discarded. The limited-resources-for-attention theory appears to be a good one: there's just not enough attention to lift two items to awareness in quick succession.

There's one exception to the attentional blink, and that's when the second target, the X, immediately follows the first one, the white letter, with no random letters in between. Curiously, this enables both to be lifted to awareness together.

Think of the attentional blink next time you're looking along a bookshelf for particular titles or down a list of names for people you know. I've had experiences looking down lists when I miss one of the names I'm after time after time, only to look againslower the second timeand see it was shortly after another name that had jumped out at me each time for some other reason.

3.7.3. End Note

1. An excellent review paper on the subject, especially the priming effect, is: Shapiro, K. L., Arnell, K. M., & Raymond, J. E. (1997). The attentional blink. Trends in Cognitive Science, 1(8), 291-296.

3.7.4. See Also

• Two good introductions to the general topic of attention are: Styles, E. A. (1997). The Psychology of Attention. Hove: U.K.: Psychology Press. And: Pashler, H. (1998). The Psychology of Attention. Cambridge, MA: MIT Press.