Lecture 9 - Attention Flashcards
Span of apprehension
asking how do we (or do we) filter info as it comes in?
we have this great big world of sensory data: where does the filtering take place?
one of the first ways we have of probing and testing those questions
Selecting channels
what it means for a channel (sensory or information) to exists
William James
father of american psychology
helped develop pragmatism in philosophy
wanted to know how people made decisions: what factored into it? what was important to them?
how can we quantify information? (airplane pilots, driving)
How much can a person track and make sense of?
cause there’s a lot there
how do you make your experience the most important thing going on?
attention limits what’s coming in
Your attention system(s) help bias/constrain what
you process at any one time.
• This helps overcome the problem of computational
complexity (too much happening in our environment to process at any one time: we can’t attend to it all, we can’t remember it all, we can’t learn it all) in the environment.
- Processing and resource limitations
- Action limitations
• Attention is not a single attribute, but rather the result of
many abilities.
Processing and resource limitations:
You can only take
in and process so much information at one time.
only so much of the world comes into your field of view (fovea - high detailed vision)
we have to have some means of limiting what’s coming so that we know exactly where to focus to get the most info and what we don’t want to pay attention to (take up finite resources)
Action (motor) limitations:
You are spatially and temporally
limited in what you can physically do - how you can interact with your environment
limitations on where I am and what I can do
Attention is not a single attribute,
but rather the result (effect) of many abilities (processes).
it’s not a single capacity or attribute, not just one place in the brain where attention lives, it’s the effect of a lot of different processes
we’re trying to modulate the degree to which we process different things
attention as the modulation of some kind of information:
focus attention and increase processing “over there” or you ignore something, decrease the processing “over there”
Attention can be divided into (many) categories.
how we direct attention, where it is going
Exogenous orienting of attention
Endogenous orienting of attention
Exogenous orienting of attention:
exo: outside, external: something grabs your attention
The degree to which outside
stimuli (take control) make you shift you sensory systems toward them.
• Example: A loud sound or flashing light may capture your
attention - novel and extreme [FIRE!] [Automatic, bottom-up: we don’t have control over it, we can’ help but notice it.]
Endogenous orienting of attention
how you choose to focus your processing power
The degree to which internal goals and desires direct your attention toward the
environment.
• Example: Searching for a person you know or trying to
hear someone in a crowded room. [Intentional, top-down]
Earlier in the course we talked about stimulus salience. By
its nature, a highly salient stimulus is more likely to
engage which kind of attention?
Exogenous attention or Endogenous attention
it depends!
a really strong stimulus can pierce through whatever you’re doing (exo) but in most cases that’s not the world we live in
endo: you’re so focused on something that you don’t notice what’s going on around you - concentration can block out environment
Attention can be divided into many functions.
- Focusing
- Perceptual enhancement
- Binding
- Sustaining behavior
- Action selection (central executive)
salience
can be something important only to you or can be something about a stimulus that grabs you
Focusing
Limiting the number of items being processed and increasing the amount of processing dedicated to that thing
Perceptual enhancement
when you’re really starting to get more information out of a stimulus or object
(greater feature resolution)
When we “focus” on something in the environment, how much of the outside (exogenous) world actually gets shut out?
how good are we at this?
Recall Broadbent’s (1958) early filter model
input (sensory data) -> filter -> detector -> to memory
Do our sensory systems filter out information as it comes in? ( I can only see so many things or hear so many things)
Early psychological experiments with attention
asked subjects to view an array of letters. Subjects would then
report the letters seen (asked: what were the letters?).
• The number of items subjects could report was called the span of apprehension (or span of perception - what you can perceive from one brief exposure).
•When using full reports, people will identify 4-5 of the letters.
The hypothesis was that the visual system didn’t allow more than 5 letters to be
perceived at any one time.
An early sensory “bottleneck” (funnel) would select a few items to be processed at a time.
eyes as a bottleneck or ears (only see so much or ear so much at one time and only so much info gets through to the brain to be processes further)
A competing hypothesis suggested that
more than 5 letters are processed.
the letters don’t get transferred to short-term memory if they aren’t used quickly
maybe you do hear or see more but that info doesn’t get transferred (you don’t do anything with it) to short-term/working memory
George Sperling (1960) tested these competing hypothesis:
- Subjects would be shown an array of letter as before.
- One group would be asked to use full report (as before).
- Another group would use partial report: After the letters are flashed, a tone (high, medium, or low) would indicate which row to report.
• When people use partial report, they seem to have access
to more of the visual display.
- probes or cues the different parts of the memory system to show that we have more access than we originally thought
- Full report yielded 4-5 total letters (as before).
- Partial report yielded (on average) > 3 letters per row
* This suggests that subjects actually had access to at least 9 of the 12 letters.
Sperling concluded that people have a high-capacity iconic
(sensory) memory, it just needs to be transferred to short-term
memory before it decays.
iconic (sensory- visual) memory
ex: looking around a room for a second you have this stimuli burned into the retina
if you probe it correctly then you can pull the info out of there and move it to short term memory BUT if you don’t it goes away
the important limiter doesn’t seem to be on
where isn’t the bottleneck?
the important limiter doesn’t seem to be on what’s actually sensorially available to us (because that’s more available than what we’re able to process in real time)
so the bottleneck isn’t at the initial sensor filter
It appears we can use attention to focus on (transfer) some
information from
sensory memory for additional processing.
If there is a bottleneck on sensory processing, it must be after
sensory (iconic) memory
How do we focus on (pay attention to) specific elements?
The tone in Sperling’s experiment told us what was important.
the tone was the cue that told us what we need to process out of sensory memory
We use our attention to
endogenously orient to what is most
important to us in the immediate environment. Those are the
things that need to be fully processed
We focus our attention on the
relevant channel…
we seem to have different pathways for dealing with sensory information
What is a channel?
• The concept comes from information processing and started with air traffic
monitoring.
• Channels can be divided by sensory modality (e.g. auditory channel, visual channel, tactile channel, etc.).
- any way that info comes to you that constitutes a distinct channel
• Within a modality, you can also have multiple channels (e.g. two speakers, two screens, etc.).
- different ways in which the info is coming to you
• A channel is any coherent sensory stream, object, location, etc. that demands or uses attention.
slippery!!
interface between exogenous memory and endogenous memory
tone was an exogenous cue (salient, relevant but only because endogenously we were listening to it and we knew what to do with it)
tone can from outside (exogenous) and was salient only insofar as we said it was (endogenously it was part of the task, part of the goal )
it was a goal, but the cue was distinctive and exogenous
why are channels slippery?
what constitutes a channel?
sensory channels (vision vs. audition) but within those modalities you can have several channels (things you are paying attention to) but more broadly a channel is any coherent sensory stream so that you can have an object as it’s own channel or a location in space as it’s own channel (if a lot of things are happening there)
a channel becomes a general thing that engages your attention: it’s what your attention can focus on for some period of time
- hard to draw that line
how do we switch back and forth between channels?
We tend to select the sensory
channel that is most relevant at
any one time.
selective attention
We tend to select the sensory
channel that is most relevant at
any one time.
• This is a dynamic process. You focus on some items, but a strong exogenous cue (e.g. someone yells “Fire!”) will capture our attention.
• Some stimuli (e.g. your name)
appear to be inherently
relevant.
Dichotic listening tasks
are used to test how well we attend to specific channels.
• Subjects are presented with two auditory channels (one in each ear).
• They are instructed to attend to one channel and shadow the content (repeat it out loud).
• Afterwards, the participants are
tested to see what they recall from both the attended and unattended channels.
Based on the Sperling ‘span of apprehension’ test, what
do you expect the results would be of the dichotic listening task?
they should recall the attended content only
if you had full report (reporting everything) then not everything could get through (only the things you’re actively attending to could make it)
the listening task is full report (only report things from the attended channel)
Your attention system(s) help
constrain complexity to manage
processing, resource and action limitations.
Endogenous attention is directed by
you
exogenous
attention is captured by
salient stimuli
The ‘span of apprehension’ experiments showed that
the
processing bottleneck seem to occur after iconic and echoic
memory.
We use selective attention to
process information in one
channel and ignore other channels.
Dichotic listening experiments show that
we can reliably do this, but that highly salient information still captures attention.
Note that dichotic listening (in this case) is a full report paradigm.
Subjects should lose what was unattended in
echoic memory.
the cocktail party effect
Dichotic listening experiments
Dichotic listening experiments typically show that people can easily shadow the attended channel (ear) without interference
from the other channel.
two physical stimuli coming in but you’re able to exclude part of it because you’re not attending (focusing) on it
Dichotic listening experiments
Colin Cherry (1953)
• found that people could not report content
from the unattended channel.
Dichotic listening experiments
Neville Moray (1959)
found that subjects couldn’t recognize a word repeated 35 times in the unattended channel.
selective attention
can be really affective!
Unattended items usually
decayed from echoic memory
they don’t get used so they’re filtered out - buff gone.
Dichotic listening experiments
If not shadowing, subjects tend to
report items from different channels separately (grouped by channel).
Some things do ‘break through’ from the selection filter of the unattended channel:
- Cherry (1953) found that people could report gender, pitch changes, or abrupt tones in the unattended channel. ((but not content)
- Moray (1959) reported that names and other ‘important information’ could get through (~30% of the time).
- In each case, highly salient information seems to capture attention (however briefly). These items were processed before they decayed from sensory memory.
Given that some salient, exogenous cues capture attention, this was a problem for Broadbent’s early selection model:
messages –> sensory memory –> fliter (attended or unattended channel –(attended message)–> detector –> to memory
How do unattended (but salient) stimuli get through the filter?
Perhaps some meaning is being processed?
Suppose you are attending to one auditory channel and ignoring the other. If you habituate to the stimuli in the unattended channel and then it suddenly changes in pitch,
this would cause:
Dishabituation
