Lecture 3: Attention - Structural and Capacity Theories Flashcards
What was the response of the Early Selection to the semantic activation on unattended channel shown by indirect mean?
Doesn’t deny WEAK activation of semantic material on unattended channel. Indirect measures don’t show it occurs to the SAME degree
What is the difference between ES and LS re activation on unattended channel.
ES = Weak semantic activation
LS = Brief semantic activation
Possible to distinguish between this? Study divided attention.
How well can we distribute attention across multiple channels? Explain study.
Cost of Divided Attention (Moray, 1970)
> Dichotic listening using simple tone, some beeps had higher altitude (signal).
> Three different experimental conditions:
(1) Selective attention (SEL trials)- tones on both channels (ignore one ear and only attend to one) 67% correct
(2) Exclusive OR (XOR): Targets could occur either on the left year or the right ear but not both at once. Attend to both ears (divide attention): 54% correct (hard)
(3) Inclusive OR (IOR): Either this or this (OR trials), or both at once (AND trials). Attend to both ears and divide attention. Excluding the AND trials = 52% correct, AND trials = 31% correct
> therefore MODERATE cost of DIVIDED attention (OR < SEL)
> LARGE cost of SIMULTANEOUS detection (AND < OR
How did early selection theory fail at predicting Moray’s results?
OR < SEL - PREDICTS because there is attenuation with divided attention (as unattended gets attenuated)
AND < OR - DOESN’T PREDICT because attenuation shouldn’t depend on identity of stimulus. (Target/nontarget distinction made in limited-capacity system, so if filtering occurs before this, distinction shouldn’t matter)
How did late selection theory fail at predicting Moray’s results?
OR < SEL - DOESN’T PREDICT because if there aren’t two targets, expect no competition
AND < OR - PREDICTS because two simultaneous targets will both be selected by “pertinence” and compete to get through filter
What are structural (bottleneck) theories?
> Some neural structures can only deal with one stimulus at a time
Competition for the structures produces processing “bottleneck” (filter theory)
(ES: bottleneck getting into LTM; LS bottleneck getting out)
What are capacity (resource) theories?
- Information processing is mental work (e.g. pupil dilation)
- Processing requires activation of neural structure
- Limited capacity to activate structure (brain in an energy limited system)
What is the Capacity Theory by (Kahneman, 1973)?
- to do mental work we have to activate structures, overall limited capacity to do this but capacity can be allocated flexibly to simultaneous tasks
- divided attention - we can equally devote attention to task 1 and task 2
- focused attention - we can allocated mostly to task 1 and less to task 2 (performance will be poorer depending on the task’s capacity demand)
Explain results of Strayer and Johnston (2001): Talking on a mobile phone interferes with driving
- looked at hands-free mobiles vs. mobile free only driving
- When on mobile phone 3x increase in missed red lights (from 2% to 6%) and increase in reaction time (by ~50ms)
Does capacity theory say anything more than that performance will be worse when doing 2+ tasks?
- Yes - it suggests ways of studying capacity about how the performance will deteriorate when multi-tasking
- Study both attentional system and capacity
What is Attention Operating Characteristic (AOC)?
> You should be able to continuously allocate capacity to one task over another and you would expect “graceful degradation” of performance as available capacity is reduced (not catastrophic failure)
Progressive performance changes with changes in capacity allocation
e.g. people should be able to devote 80% to one task and 20% to other - and as you vary this ratio (task 1: task 2) - should generate a smooth trade-off which will tell you about the attentional system and capacity requirements of task
What are Possible Tradeoff Functions?
- capacity theory doesn’t prescribe where the tradeoff curve goes (e.g., where 50:50 divided attention point will be)
- Shape of tradeoff curve tells us about capacity demands of two tasks
Experimental example of Tradeoff Functions
- Bonnel & Hafter (1998) did a perceptual task with an auditory and visual version. Two aspects were:
- Detection: Whether a target stimulus (luminance disk/ auditory tone) present or not
- Discrimination: increase or decrease in (luminance or tone)
- Vary capacity fraction on auditory and visual tasks (80% auditory, 20% to visual etc.)
- Results:
> Discrimination task produces a smooth quarter circle tradeoff chart - tradeoff = the task is capacity limited
> Detection task is easier - no tradeoff= no capacity limitation (at intersection of horizontal & vertical axis)
> These two tasks differ in their capacity requirements.
> Detecting may be pre-attentive.
> Understanding meaning requires focal attention. Understanding meaning may be similar to discrimination.
Pros and Cons of Capacity Theory
- Value of capacity theory is new experiments it led to
- Emphasises divided attention, flexibility of attentional control
- Shortcoming is its vagueness (can always come up with a capacity explanation)
What are some Alternative Approaches Based in Signal Detection Theory?
- Theory of how we make decisions about weak or difficult stimuli
- Mental representations of stimuli are statistically variable or “noisy”
- Attending to multiple stimuli increases amount of noise and reduces accuracy
- Makes mathematically precise predictions about divided attention costs
- Many results attributed to capacity limitations can be explained by changes in the amount of noise….
