Lecture 5 - 02-04-13
Definition of Attention
Attention is the process that, at a given moment, enhances some information and inhibits other
information. The enhancement allows us to select some information for further processing, and
the inhibition allows us to set some information aside
We all intuitively understand that it means to 'pay attention' - it is to focus on something, and
But attention has been notoriously tricky to define.
It has been used synonymously with terms such as arousal, control, and consciousness.
These are all clearly very different things.
Selective attention refers to the skill through which one focuses on one input or one task while
ignoring other stimuli
You ignore other things that are around. Once you see the face in the coffee beans, you can
never not see that face
Shadowing (Repeating an audio track)
Dichotic listening - different messages to each ear; attention is paid to the voice in for instance
your left ear.
attended channel - listen to this one
unattended channel - ignore this one
e.g. counting the passes (white/black teams, basketball, gorilla)
Their attention is driven by the task at hand, counting the passes is what the participants are
people claim that they don't hear anything in the unattended channel; words can change to
nonsense words and they still would not hear them; if the gender of the voice changes,
sometimes the participant will be aware of that; if the name of the participant is said in the
unattended channel, they will actually hear that
Cocktail party effect
other conversations tuned out and you just pay attention to the person you're talking to; if
someone in close proximity says your name, it pulls your attention away; i.e. friend's name, your
interests; how are we able to ignore and tune out all this information when it's not relevant but
something can pull our attention back? Unattended channel can be noticed
The participant’s own name
Words of high personal significance
Boost attended items and dampen unattended items
The passes example is the visual version of the audio task.
Sometimes effects of attention are so strong that we fail to see stimuli that are directly in front
of our eyes.
Participants look at a screen, looking at a fixation target, some task to do, they can do the task
without issue. If that target changes, so rather than a plus it becomes a shape, people don't
report seeing that change. As soon as you tell them about this change, almost all the time they
can report the change. When you bring your attention to something, why can't they notice it
even though it's directly in front of them?
There's also the refrigerator example. You look for something in the fridge and it's right in front
of you. You're thinking of other stuff that causes you to miss something that's in front of your
eyes. What happens to attention that makes you unable to see this consciously?
Early versus Late processing
These are competing but also complementary. Early processing model versus late is basically
saying whether or not the information is coming in or that we block it before it even comes in.
What type of resources do we need? The example of the fridge, maybe it's not coming to
consciousness because the resources is used up
Person Swap example
This is basically the person swap example from Just For Laughs. Usually is more robust when it's
the same gender and race. We can't have all our attention attuned to everything going on at
one time. There has to be mechanisms for us to attend to the thing we're interested in and
general bits of information of things in the rest of the world. We're doing that so that we can
use our mental resources of that one task at hand.
Inattentional blindness: no perception without attention
unconscious perception can still occur in the absence of attention
perhaps it comes in and it's forgotten or something else comes in and cuts off and you lose it
out of consciousness
it's perhaps because some of this information can be unconsciously perceived. Unconscious
perception means that maybe reflects what was seen and you can do something with that
information without knowing you're doing so, i.e. priming. Even with conscious perception, the
example of the cocktail effect of hearing your own name, that does reach consciousness, why
does this happen? Movie perception test example
Besides the guy answering the phone, the clothes and the person changed. Movies tend to play
a lot on our ability to keep track of the details in movie scenes.
Rather than ignoring something that's in front of you, something is changing and it isn't always
easy at ID. The example he used here was the plane with the propeller present in one and not in
the other. The other example was the graveyard one where at the left part, in one image it was
a part of the tomb and gradually it changes into grass. Another example was the colour changing
card trick where the four changes were the guy and the woman's shirt changed, the table cloth
changed, the background wallpaper changed.
The two bars, which is longer example.
In the first image, if someone asks to judge the length of the two bars, most people say with
certainty that the top one is longer than the bottom. In the second picture, the two bars are
exactly the same size where it appears that the top bar is longer than the other but are actually
the same length, all perceived differences are caused by the fins which aren't
attended/consciously seen. Some information is being processed. Those fins that are
consciously seen are influencing your perception of the length of the line.
Information is attended to. There's some attended channel and everything else not attended to
is lost. Only the attended input reaches consciousness.
Everything comes in but at some point only what's being attended to makes it into
consciousness and the unattended are dropped therefore doesn't make it into consciousness.
How do you combine both models that fits all sets of data
Stimuli that aren't attended to have effects on perception, which indicates late selection
electrical brain activity for attended inputs differs within 70 ms from unattended inputs, which
indicates early selection.
Selection depends on resources
complex stimuli involve more effort, leading to early selection
easy stimuli involve less effort, leading to late selection
Selection as a form of priming
lower threshold leads to easier recognition
attended channel has lower threshold
your name is frequent and is primed even when unattended (cocktail party effect. The threshold
is low and is primed over and over; it's more likely even you're not attending to it/self-directed,
that can pull our attention, something that is primed over time)
Posner and Snyder (1975) two kinds of priming related to attention
1) stimulus based priming does not involve effort
our attention is pulled in that direction (of the stimulus); stimulus primed through time
2) expectation based priming does involve effort
if we're thinking about somebody showing something to us on the screen, the fixation points
(that changes), why does that all of the sudden allow people to see it even if not told specifically
what changes? We're diverting resources to our expectations of change and doing that allows
something to be primed, we're lowering the threshold to see some change. We have mental
resources towards the thing we're paying attention to.
High versus low validity condition, the task given, the low validity, there can be three types of
trials. The participants were supposed to ID a pair of letters. Neutral trials had a + sign, yes two
letters are the same or not. Primed condition had the letter that would appear in the test
stimuli. In the misleading trial, the letter was different than the test stimuli. In the low validity
condition, that prediction in the primed experiment only predicted accurately the letters 20% of
the time, the high validity condition predicted it 80% of the time. The interesting thing is that in
the high validity condition, that priming is actually giving somebody information. The AAs are
the information you have and you use. In the low validity, people don't trust the trial since it's
only 20% of the time. The misleading trial actually leads to incorrect expectations. In the low
validity condition, there's little different between neutral and primed. On the high validity
condition, because the priming one provides what the correct answer will be, because the
misleading provides the incorrect expectation. Whether they can ID that there's two
consecutive letters or not, it's relatively easy. The main thing is about the time they take to ID
Low Validity condition
primed condition fast than neutral
misled same as neutral
We have limited amount of resources available to us.
primed faster than neutral
expectations augments repetition priming
Misled slower than neutral
expectation limited in capacity
wrong expectation interferes the performance to correctly detect the letters
Looking at a fixation point, changes to an arrow facing left or right, make a judgment call or
which side the stimulus is on and whether the arrow correctly points at the correct direction.
Stimulus based priming is fast like our name being said. It's easy to grab our attention and it
comes at little cost meaning we can have in our example, something that primes us, if it helps because it's been repeated overtime, we actually perform faster, we can detect the stimuli
faster; in the misleading trials, no cost, no expectations there.
Expectation based priming is slower and also has resource cost. If we're going to form some idea
that this correctly predicts the location of the stimulus, and it's wrong, and we already have that
expectation driven, it takes time and effort to change, we divert to change our resources.
Skeleton and mirror image
We don't shift our gaze or our eye movements but we can definitely shift our attention between
the skeleton or the lady in t