INPUT SENSORY register SHORT TERM memory LONG TERM memory
LECTURE 2: ATTENTION Alertness/concentration
Selectivity
Control (automaticity)
Focused attention
Divided attention
Selectivity: Focused attention Studies of auditory selective attention:
Dichotic listening task and shadowing (early studies)
Able to focus on one conversation – selectively attending to a convo
Experimental setting = 2 different messages to each ear, subject is asked to attend to one message
Cherry’s finding
Subjects shadowed message presented to one ear, asked later about unattended message
Subjects did not notice it was foreign speech, reversed speech
But noticed it was pure tone, male/female voice
Only physical characteristics are processed in unattended message
Moray’s 1959 cocktail party phenomenon
Subjects did not notice repetition of the same word 35 times
Subject noticed own name mentioned in unattended ear Incompatible with the idea that only physical characteristic are processed in the unattended message
Bottleneck models of attention Multistory model of model
INPUT
SHORT
SENSORY
TERM
register
memory
Sensory register
is
capacity (can
hold a lot)
LONG TERM memory
assumed to have large
STM store is limited
in capacity (can hold little)
All models assume transfer of information from sensory register to STM store
Bottleneck somewhere in-between sensory and STM
They differ in where they regard the bottleneck is and the nature of the
bottleneck 1. Broadbent’s filter model: Assumptions
Evaluation
Stimuli gain access in parallel to a
Consistent with Cherry’s finding:
sensory register
Selection on the basis of
Physical characteristics of unattended info remembered Meaning info was not remembered
physical/perceptual characteristics
e.g. spatial location: left/right ear
THUS:
Selective filter prevents
Unattended info undergoes minimal
overloading of limited-capacity
semantic processing before being
STM store
filtered
Inputs remaining in the buffer after
Model is inconsistent w. Moray’s
filter undergoes semantic
cocktail party phenomenon
processing
2. Treisman’s attenuation model: Assumptions
Instead of an all-or-none filter, an
Evaluation
Own name is processed in
attenuator turns down the
unattended channel
processing of unattended info
(Moray’s) because it has low
Thresholds of all context-appropriate
threshold
stimuli are lower a.) Partially processed stimuli sometimes exceed threshold of conscious awareness b.) Leading to ‘breakthroughs’
Due to HIGH SALIENCE
Meaningful context also reduces threshold
3. Deutsch & Deutsch’s (1963) late selection model: Assumptions
Evaluation Info is analysed FULLY
attenuation model can
without attention
explain
Argued that attenuator The idea of different thresholds necessary
Bottleneck is late ; at
selection for action (response)
CAN explain what the
(physical, semantic)
is redundant
E.g. cannot say 2 things at the same time
Selectivity
Control (automaticity)
Focused attention
Divided attention
Selectivity: Focused attention Studies of auditory selective attention:
Dichotic listening task and shadowing (early studies)
Able to focus on one conversation – selectively attending to a convo
Experimental setting = 2 different messages to each ear, subject is asked to attend to one message
Cherry’s finding
Subjects shadowed message presented to one ear, asked later about unattended message
Subjects did not notice it was foreign speech, reversed speech
But noticed it was pure tone, male/female voice
Only physical characteristics are processed in unattended message
Moray’s 1959 cocktail party phenomenon
Subjects did not notice repetition of the same word 35 times
Subject noticed own name mentioned in unattended ear Incompatible with the idea that only physical characteristic are processed in the unattended message
Bottleneck models of attention Multistory model of model
INPUT
SHORT
SENSORY
TERM
register
memory
Sensory register
is
capacity (can
hold a lot)
LONG TERM memory
assumed to have large
STM store is limited
in capacity (can hold little)
All models assume transfer of information from sensory register to STM store
Bottleneck somewhere in-between sensory and STM
They differ in where they regard the bottleneck is and the nature of the
bottleneck 1. Broadbent’s filter model: Assumptions
Evaluation
Stimuli gain access in parallel to a
Consistent with Cherry’s finding:
sensory register
Selection on the basis of
Physical characteristics of unattended info remembered Meaning info was not remembered
physical/perceptual characteristics
e.g. spatial location: left/right ear
THUS:
Selective filter prevents
Unattended info undergoes minimal
overloading of limited-capacity
semantic processing before being
STM store
filtered
Inputs remaining in the buffer after
Model is inconsistent w. Moray’s
filter undergoes semantic
cocktail party phenomenon
processing
2. Treisman’s attenuation model: Assumptions
Instead of an all-or-none filter, an
Evaluation
Own name is processed in
attenuator turns down the
unattended channel
processing of unattended info
(Moray’s) because it has low
Thresholds of all context-appropriate
threshold
stimuli are lower a.) Partially processed stimuli sometimes exceed threshold of conscious awareness b.) Leading to ‘breakthroughs’
Due to HIGH SALIENCE
Meaningful context also reduces threshold
3. Deutsch & Deutsch’s (1963) late selection model: Assumptions
Evaluation Info is analysed FULLY
attenuation model can
without attention
explain
Argued that attenuator The idea of different thresholds necessary
Bottleneck is late ; at
selection for action (response)
CAN explain what the
(physical, semantic)
is redundant
E.g. cannot say 2 things at the same time
