Lecture 9 & 10 – Attention
The Perception-Action Cycle
You can’t translate every sensory input in a motor output: some selection has to be made. This is
the primary function of attention.
Selective attention has been demonstrated both in the auditory and visual domains:
For sound, the dichotic listening experiment was used: participants have
to focus attention on (and shadow) one of two messages. The limited
capacity of our attentional system doesn’t allow us to understand both
messages at the same time and is the reason why we need selection. In
fact, attention decides which sensory inputs go through the selective filter
(and reach higher-level analysis) and which don’t.
For vision, attending is achieved either via overt attention or covert attention.
Overt attention refers to orienting your eyes/body towards the stimulus (mediated
by the superior colliculus). Covert attention refers to shifting attention without
moving the eyes (without overt signs): this mechanism is especially relevant for
social animals, where gaze direction has a strong meaning (e.g., threat, sexual
attraction).
1
,Again, covert attention also exists in the auditory domain, where one such
example is the cocktail party effect (pretending to listen to someone in
front of you while actually paying attention to what is said in another
conversation).
A related phenomenon is attentional capture: while listening to the person in front of you, your
attention may be suddenly captured by someone saying your name in another conversation.
Bottom line: there is limited capacity in going from perception to action.
Different Types of (overt) Attention
1. Voluntary attention (top-down)
2. Capture attention (bottom-up)
3. Object-based attention
4. Feature-based attention
Top-down and bottom-up are collectively called spatial attention.
Let’s have a closer look at these:
#1 Voluntary attention (top-down)
aka endogenous, goal-driven
Typically measured with the Posner cueing task:
provision of cues to direct attention at a particular location before presenting the target stimulus.
People respond faster in valid trials, compared to invalid and neutral trials. In other words, if the
target is at the already attended location (cued location) reaction times will be faster.
2
, #2 Capture attention (bottom-up)
aka exogenous, stimulus-driven
A suddenly appearing stimulus will automatically capture attention. Here, subjects are instructed
to ignore the prime, as it has no predictive value. Despite this, when there is a valid trial (prime
happens to be at the same location as the target), reaction time is shorter at the primed location.
This is evidence for capture attention: even if I don’t want to focus my attention on the prime
location, I still “have to”. This mechanism is automatically driven by the visual input rather than
our intentions.
However, when the CTOA (cue-target onset asynchrony = time between cue and target) exceeds
200ms, inhibition of return (IOR) occurs. This means that attention at the primed location is
now inhibited, rather than facilitated. In other words, capture attention facilitates processing at
the primed location until 200ms after cue onset, whereas IOR inhibits attention at the primed
location after 200ms from cue onset.
In the experiment above, IOR causes the opposite effect: slower reaction times for valid trials
and shorter reaction time for invalid trials when the target is presented ~200ms after prime onset.
3
The Perception-Action Cycle
You can’t translate every sensory input in a motor output: some selection has to be made. This is
the primary function of attention.
Selective attention has been demonstrated both in the auditory and visual domains:
For sound, the dichotic listening experiment was used: participants have
to focus attention on (and shadow) one of two messages. The limited
capacity of our attentional system doesn’t allow us to understand both
messages at the same time and is the reason why we need selection. In
fact, attention decides which sensory inputs go through the selective filter
(and reach higher-level analysis) and which don’t.
For vision, attending is achieved either via overt attention or covert attention.
Overt attention refers to orienting your eyes/body towards the stimulus (mediated
by the superior colliculus). Covert attention refers to shifting attention without
moving the eyes (without overt signs): this mechanism is especially relevant for
social animals, where gaze direction has a strong meaning (e.g., threat, sexual
attraction).
1
,Again, covert attention also exists in the auditory domain, where one such
example is the cocktail party effect (pretending to listen to someone in
front of you while actually paying attention to what is said in another
conversation).
A related phenomenon is attentional capture: while listening to the person in front of you, your
attention may be suddenly captured by someone saying your name in another conversation.
Bottom line: there is limited capacity in going from perception to action.
Different Types of (overt) Attention
1. Voluntary attention (top-down)
2. Capture attention (bottom-up)
3. Object-based attention
4. Feature-based attention
Top-down and bottom-up are collectively called spatial attention.
Let’s have a closer look at these:
#1 Voluntary attention (top-down)
aka endogenous, goal-driven
Typically measured with the Posner cueing task:
provision of cues to direct attention at a particular location before presenting the target stimulus.
People respond faster in valid trials, compared to invalid and neutral trials. In other words, if the
target is at the already attended location (cued location) reaction times will be faster.
2
, #2 Capture attention (bottom-up)
aka exogenous, stimulus-driven
A suddenly appearing stimulus will automatically capture attention. Here, subjects are instructed
to ignore the prime, as it has no predictive value. Despite this, when there is a valid trial (prime
happens to be at the same location as the target), reaction time is shorter at the primed location.
This is evidence for capture attention: even if I don’t want to focus my attention on the prime
location, I still “have to”. This mechanism is automatically driven by the visual input rather than
our intentions.
However, when the CTOA (cue-target onset asynchrony = time between cue and target) exceeds
200ms, inhibition of return (IOR) occurs. This means that attention at the primed location is
now inhibited, rather than facilitated. In other words, capture attention facilitates processing at
the primed location until 200ms after cue onset, whereas IOR inhibits attention at the primed
location after 200ms from cue onset.
In the experiment above, IOR causes the opposite effect: slower reaction times for valid trials
and shorter reaction time for invalid trials when the target is presented ~200ms after prime onset.
3
