week 7 - attention Flashcards
Attention
Everyone knows what attention is. It is the taking possession of the mind in clear and vivid form, of one out of what seem several simultaneously possible objects or trains of thought. Focalization, concentration of consciousness are of its essence. It implies withdrawal from some things in order to deal effectively with others…
Attention - top down or bottom up
passive attention - bottom up - involuntary stimulus directed - e.g loud noise
active attention - top down - voluntary - goal direction
Active attention - 2 components
focused (selective) attention - attending to only one source of info, ignoring everything else
Divided attention - 2 or more tasks performed simultaneously
When do we attend? Early or late selection processes
How do we attend? space-, object-, modality-selective attention – temporal and parietal cortex.
Different sensory modalities (auditory, visual, etc.).
these are both active attention
focussed auditory attention
cocktail party problem - colin Cherry
in a party - many sounds at once, how do we pick sounds that are relevant
what level of processing of those sounds
Cherry - dichotic listening task
different message presented in each ear, P’s asked only to listen to one
verbal shadowing - asked to repeat attended message out loud
Findings - P’s efficient at screening out unwanted info.
notice physical characteristics - pitch, sex of voice etc
When physical differences eliminated, tasks harder to solve
Moray - ps didnt remember the repeated 35 words in the unattended ear
dichotic listening -interpretation
The fact that only distinct physical differences between two concurrent messages were noticed (male vs. female voice, differences in pitch), and that unattended information did not enter LTM suggested that we have a processing bottleneck.
early selective process - Broadbents filter theory
Proposed that filtering occurs early in processing based on physical characteristics.
He proposed 2 phases of perceptual processing.
stage 1 - physical properties (e.g. pitch / location)
2nd stage: More complex psychological properties such as meaning.
Limited capacity – could not deal with everything
Serial
Broadbent’s theory is bottom-up driven; information is filtered solely on the sensory analysis of physical characteristics.
Inconsistencies in early selection
Dichotic listening - more processing than expected if words hold significance to listener. e.g. if women dissatisfied with weight processed unattended weight related words (fat, chunky)
Evidence of top-down influences
Treisman’s attenuation theory
Subjects deemphasise but don’t filter out unattended message. based on findings that:
Listeners engaged in verbal shadowing occasionally say words presented to the unattended ear.
Some subjects switched ears even when told not to, following the semantic content.
These breakthroughs occur when unattended messages are relevant (e.g. fit semantic context or expectations).
Top-down driven.
Treisman’s attenuation theory (2)
All info passes through sensory register but some is weaker (attenuated).
Different levels of processing – starting with basic physical features, then syllables, words, syntax, meaning. If processing capacity is limited, then the later processes are attenuated.
Support for Treisman’s theory
Dichotic listening task:
Participants asked to shadow one message but to tap when a target word (e.g. green) was heard from any source.
Attenuation theory predicts less detection in non-shadowed ear.
Results:
87% detection in shadowed ear
8% detection in non-shadowed ear
Evidence for early selection is unlikely.
Late selection theory
Deutsch & Deutsch’s (1963) late selection theory: – the limitation is in the response system, not at the level of perception.
Both messages are perceived in terms of meaning, but only one can be shadowed at a time.
The selection criterion for what to say can change – based on ear or meaning
Late selection theory - explination
All stimuli fully analysed but theres a limited capacity at the response at the END of the process. The person chooses to respond on the basis of what is the most important input
Late Selection theory - inconsistencies
Treisman and Riley (1969) responded by repeating the initial Treisman and Geffen (1967) experiment.
But this time, participants were asked to stop shadowing and tap when they heard a target in either ear (to reduce the ‘importance’ of attending to a specific ear.
Results:
67% detection in attended ear
33% detection in non-attended ear
Evidence for late selection is unlikely
Recent support for Treisman
Neural evidence using EEG:
Dichotic listening task, two conditions: Participants asked to attend to a buzz (non-linguistic) or to syllables (linguistic).
Event-related potentials (ERPs) over primary auditory cortex were greater for attended than unattended messages (irrespective of condition).
- Evidence for signal
attenuation, not
complete filtering
out.
ERPs also larger for non-linguistic than linguistic messages.
Suggests that enhancement is based on importance, larger for physical characteristics than for meaning.
Focussed visual attention
vision is our most dominating sense. When paying visual attention we choose where to fixate our eyes for greatest visual acuity.
Other portions of the visual field are attenuated.
However, visual attention need not be located where the eyes are fixed.
Space based theory - spotlight
we attach our attention to locations in space and process whatever is there.
Move our eyes around to change the spotlight of attention.
Space based theory - zoom lens
Attention to spatial locations can be broad or narrow (we flexibly adjust the spotlight of attention).
Broad areas less well processed.
A narrow focal point gives optimal processing but it takes time to move the focus to other areas of the visual field. –> Trade-off
Object based attention
Attention is focussed on particular objects even if our field of view occupies a larger space
Neural evidence for object based attention
Ps either focussed on face or house (they are overlapping eachother on one image), both occupied same spatial location.
If attention is object based, selective brain areas for either faces or houses should be activated.
If attention is space based, both, face and house brain regions should be activated
was found when ps focussed on face FFA was more active. Supports object based
Behavioural evidence for object and space based attention
Participants presented with a cue on one of two bars (bold edge).
Participants asked to detect the highlight on the subsequent target as fast as possible.
Target could appear on same or different location.
Results:
Fastest response given for valid trials. (Demonstrates space-based facilitation).
IS trials responded to faster than ID trials.
Evidence that attention is predominantly object-based.
Lavie and Tsal 1994 review of studies found
Evidence for late selection in situations of low perceptual load – single target, single distractor e.g. Stroop task.
Evidence for early selection in situations of high perceptual load – more stimuli, more demanding tasks.
Lavie (1995; 2005)
A series of experimental studies testing this idea.
She showed that perceptual processing is automatic but that how much processing occurs depends on perceptual load.
Irrelevant information will be excluded from processing only if relevant processing exhausts all available capacity.
Flanker task - manipulation of perceptual load
Respond to central letter
Ignore all other letters
Distractor letters (flankers) presented above or below
flanker task - results
Is the central letter an X ?
High perceptual load introduced – an entire letter string needs to be processed.
Participants were faster to respond in this condition compared to the low load condition with only one flanker.
However, what could be the problem here?
Flankers are closer to the target/next to the target and might therefore be processed faster
Lavie accounted for this problem by introducing a distant flanker.
Participants were still faster in this condition than in the initial condition with only one distant flanker and one central letter X.
Lavie - perceptual load theory
Perceptual Load theory
Attention is a limited capacity system (as with Broadbent’s original conception), but the amount that gets through the filter depends on how much work the perceptual system has do to.
Perceptual load theory could resolve the early vs. late selection debate.
perceptual load is not the same as cognitive load
Importantly, perceptual load is not equivalent to cognitive load (i.e. task difficulty).
Indeed, if cognitive load increases (e.g. greater WM demands), distraction effects increase too.
Cognitive load requires additional processing mechanisms (WM), which reduce the ability to attend to a target (remember Baddeley’s dual task).
Perceptual load and cognitive load involve different processing mechanisms:
Perceptual load can enhance attention.
Cognitive load interferes with attention.
Visual search
A frequent attentional process with important real-life applications.
Inspecting X-ray images of luggage at airports, looking for dangerous items.
- Visual search involves goal-directed (top-down) as well as bottom-up attention processes.
Visual search - feature integration theory
Treisman & Gelade (1980)
Feature Integration Theory
Visual search involves processing object features (colour, size, line orientation) and objects as a whole.
Two processing stages:
Pre-attentive: rapid, parallel processing across a visual scene (single features ‘pop out’, bottom-up detection.
Attentive: slow, serial processing while searching for conjunctive features that make up an object.
Visual Search experiments
Letters: T and O
Colours: Green, Orange
1 target + distractors (display size: 1, 5, 15, 30)
First task:
Focus on colour: Find a green letter
Display size etc (distractors) didn’t really affect reaction time when just looking for a different coloured letter
Second task:
Focus on shape: Find a T
Display size etc (distractors) didn’t really affect reaction time when just looking for a different letter
Third task:
Focus on colour and shape: Find a green T
(Conjunction search)
The more distractors the slower the reaction size
Feature integration theory - summary
proposes that visual search is a two-stage process:
Pre-attentive: rapid, parallel processing across a visual scene (single features ‘pop out’, bottom-up detection.
Attentive: slow, serial processing while searching for conjunctive features that make up an object.
Binding features together to form an object requires focused attention.
In the absence of focused attention, features are miscombined, resulting in ILLUSORY CONJUNCTIONS.
