-
Ch 2 Attention
- Background
- Auditory attention
- Visual attention
- Integrating information in clearly seen displays
- Attention and distraction
- The neurology of attention
- Conclusion
-
Auditory attention
- Disentangling sounds
- Attending to sound
- Eavesdropping on the unattended message
- Summary auditory attention
-
Auditory attention: Disentangling sounds
- Timing and intensity important
- The ear further from the sound is slightly shadowed by the head and receives quieter sound indicating sound direction
- Wavelength = the distance from one wave crest to the next;
- Shadowing: low pitched = long wavelengths longer than head i.e. negligible intensity shadowing; high-pitched = short wavelengths i.e. intensity cues only available for sounds of short wavelength
-
Auditory Attention: Attending to sound
- Concepts
- Source direction: the auditory system is capable of attending to separate superimposed sounds based on where they come from
- Spotlight on attention: we pay attention to a desired part of the environment
- ------------------------------------------------
- Research: dichotic listening
- Broadbent: 1952
- Description of Procedure:
- Different sounds/speech in each ear through headphones
- Findings:
- Participants could not remember message from unattended ear unless very short (three words); may be due to short lived store
- If the attended message lasted more than a few seconds then unprocessed material in the other ear would be lost
- Hanging on to sound for a short time= echotic memory (like dying echo)
- Further study: male and female voices, each word alternating to each ear; able to follow a particular speaker i.e. using pitch of voice vs location as cue for attention
- Conclusion:
- Speech processing - brain compared to computer
- Serial processing: brain needs to attend and select sound, thought that it can only process one thing at a time
- However: earliest stage of processing needs parallel processing; can only use simple procedures i.e. attention directed on basis of direction of sound and pitch
-
Auditory Attention: Eavesdropping on the unattended message
- Research: the cocktail party effect
- Treisman: 1960
- Description of Procedure:
- Hearing own name mentioned in crowded room while focusing on a conversation
- Making use of shadowing (dichotic listening technique requiring participants and repeat aloud what is heard in one ear following like a shadow behind the message)
- Requires concentration
- Findings:
- Participants appear completely ignorant of what was said in the other ear
- Conclusion:
- Attenuation process: a filter turning the volume down for all sound except attended signal; signals to which we are sensitive (e.g. own name) would attract attention
- Further study: two stories 1 in each ear, message swaps in the middle of what was being said
- Finding: participants attended sense of the story
- Conclusion: Story temporarily sensitised the listener to the next expected words this is known as priming e.g. lexical decision task: participants identify words quicker when preceded by a related word
- *
- Further study: mild shock given with some words, participants became sensitive to these words
- Findings: galvanic skin response (perspiration) seen in sensitive words used in unattended message
- --------------------------------------------
- More research followed
- Deutsch & Deutsch 1963: all messages received same processing whether attended or not
- Norman 1968: unattended information must receive sufficient processing to activate relevant semantic memories
- Would require brain functioning to be more parallel than previously supposed
-
Summary auditory attention
- Auditory system processes mixed sound allowing us to attend single message
- Unattended material is unprocessed
- Listener unable to report detail
- Only most recent material available (echoic memory )
- Suggests parallel acquisition of information followed by serial processing to determine meaning of attended message
- Despite little awareness, unattended material may receive more processing
- Words in unattended ear can produce priming effect and physiological change e.g. perspiration
- Shadowing: following a message from one ear to the other
-
Visual attention
- Knowing about unseen information
- Towards a theory of parallel processing
- Rapid serial visual presentation
- Masking and attention
- Summary: visual attention
-
Visual attention: Knowing about unseen information
- Seeing requires time to capture and analyse information
- Research: iconic memory
- Sperling: 1960
- Description of Procedure:
- Grids- letters three rows of four letters each show for 50 ms; typically three or four named
- Also, using three different sounds after presentation to indicate row required
- Findings:
- Participants still able to report letters from correct row - called: "partial report superiority effect"
- Participants seem to know equally much when two different coloured letters are used & telling participants which colour to report on
- Conclusion:
- Implies that participants must have registered most of the letters presented
- Implies there was a moment when they had access to all the letters and choose where to direct attention (equivalent to echoic memory associated with dichotic listening) and was named iconic (icon = image) memory
- ------------------------------------------------------
- Masking:
- One stimulus (the target) rendered undetectable by presentation of another (the mask) 10 - 50 ms later
- Time between onset of target display and onset of mask = stimulus onset asynchrony (SOA)
- -----------------------------------------------------
- Research: backward masking
- Humphreys: 1981
- Description of Procedure:
- Sequence of two words; both masked; first supposedly undetectable; second difficult to see
- Findings:
- Reporting more likely accurate when second word related to the first e.g. Tiger & lion
- Criticism:
- Cheesman and Merikle1984: although participants claimed words not seen; when asked to guess they did better than chance
- Therefore - meaning can be extracted from material of which participants were unaware
- -----------------------------------------------------
- Research: ?
- Pecher et al: 2002
- Description of Procedure:
- Modification of Evett and Humphreys 1981
- 1. priming words could be displayed very briefly Allegedly undetectable); or shown for 1 second to guarantee priming effect
- 2. two sets of trials
- 90% words related to prime
- 10% related words
- Findings:
- Related words better detected,
- Can only be done when prime word was clearly visible
- When there is a high chance that words will be related participants spot to connections and guess target often correctly
- Conclusion:
- Participants were unable to guess in the briefly shown condition, presumably not been able to identify the prime words
- But those words did produce small priming effect therefore sufficient analysis to achieve meaning must have taken place
-
Visual Attention: Towards a theory of parallel processing
- Coltheart 1980:both semantic and episodic details required to reach conscious awareness
- Electrophysiological studies show that visual item identification occurs in a different region of the cortex than where colour and location processing takes place: these need to be united and processed taking time and attention
- Therefore attention is no longer described as word identification (selecting material for complex serial processing) instead attention is required to join the products of the processes of identification and episodic characterisation
-
Visual Attention: Rapid serial visual presentation
- Interruption: longer SOAs; occurs even when target presented to one eye and mask two other (stimulus onset asynchrony (SOA))
- • This dychoptic (two eyed) interaction cannot take place before information from both eyes is combined in the brain
- Integration: where SOA between target and mask is very short i.e. two items perceived as one; result = target difficult to report
- • Does not occur dychoptically when target and masks are presented to separate the eyes; probably occurs early in analysis
- Therefore - Turvey 1973 describes integration as peripheral masking and interaction as central masking (at level of more complex information extraction)
- ------------------------------------------------------
- Research: rapid serial visual presentation (RSVP)
- Broadbent: 1987
- Description of Procedure:
- Participants are asked to respond (RSVP) by reporting what they had seen
- Series of up to 20 stimuli in rapid succession; each acting as backward masked for preceding item
- Participants are asked to report on only two items
- 1. One white letter (target 1) in a series of black letters,
- 2. After presentation of all stimuli, participants are asked to name the white letter and report if another letter (e.g.X) (target 2) was present in the series
- Findings:
- Target 2 might be spotted if immediately following target one;
- Less likely after that unless more than five items separated the two targets
- Conclusion:
- Something may be happening as a result of processing target 1 making awareness of target to temporarily difficult
- The AB and effect: System seems to require time for being able to process something new (refractory period) in this context called attentional blink (AB)
- Further study: two-stage model of processing
- Giesbrecht & Di Lollo: 1998
- Making target 2 last item in the list
- Findings: AB effect disappears
- Conclusion:
- 1. Stage 1: information about target characteristics captured in parallel
- 2. Stage 2: serial processing on information (preparation and report)
- • Later information cannot be processed and remains at stage 1
- • Disruption to target 1 i.e. masking = longer to process and target 2 waits longer
- • Little effect on target 2 unless masked by another stimulus when targets 2 will be overwritten
- • Overwriting process will damage mainly episodic information; semantic information may survive if priming effect present
- • I.e. similar to Coltheart 1980 need to join semantic and episodic detail
-
Visual Attention: Masking and attention
- The relevance of masking studies to attention-what happens to material not at the focus of attention?
- Also what happens to material that has been attended to but did not have time to process?
- -------------------------------------------------
- Research
- Merikle & Joordens: 1997
- Two issues:
- 1. Perception without awareness i.e. masking studies
- 2. Perception without attention i.e. dichotic listening
- Findings:
- Results entirely comparable in both studies
- Conclusion:
- Same underlying processes at work
-
Summary: visual attention
- Attention can be directed towards different areas without refocusing
- The masking effect seems to be linked to the need to assemble information
- Visual information captured in parallel but assembly is a serial process
- Episodic detail (colour, position)vulnerable to time and to being overwritten by mask
- Semantic information (identity, meaning) relatively enduring but only reaches conscious awareness when bound to episodic information
- Attention is the process of binding the semantic (meaning) with the episodic (context/detail) characteristics of memory - needs time for it to happen
- Unbound semantic activation can be detected by priming and electrophysiological techniques
-
Integrating information in clearly seen displays
- Serial and parallel search
- Non-target effect
- The "flanker" effect
-
Integrating information in clearly seen displays: Serial and parallel search
- Binding of features is a significant process when displays are brief
- Detection and integration of features are necessary to perception and integration requires attention
- ----------------------------------------------
- Research: feature integration theory
- Treisman:
- Description of Procedure:
- Odd item = target; others = distractors
- Display of series of characters containing varying numbers of targets
- Measure time taken to determine presence of target
- Findings:
- No difference in time when targets and distractors same colour and print or when target different colour
- When target different colour and different letter, longer time taken for more distractors (each added approx 60 ms)
- Conclusion:
- Quicker to make decision when considering only one dimensional
- Different types of analyses (e.g. shape or colour) take place in different parts of brain; when differences on one-dimensional (shape or colour) processing takes place in parallel; when differences on two-dimensions (shape and colour) processing is serial and takes longer for more items
-
Integrating Clearly Seen Information: Non-target effect
- Feature integration theory has been very influential but there may be other explanations for the feature binding process
- -----------------------------------------------------
- Research
- Duncan and Humphreys: 1989
- Description of Procedure:
- Searching for the letter L (the target) amongst a number of T's (the non-targets)
- Findings:
- It took longer to find the target
- Conclusion:
- Part of finding the target includes rejecting the non-targets and is a harder task when they come in greater variety
- Suggests that non-targets as well as targets need to be recognised
-
Integrating Clearly Seen Information: The "flanker" effect
- Potential problem for feature integration theory is that understanding the meaning of the printed word can be influenced by other nearby words i.e. one word can prime the other
- Theory:
- Priming effects found even when the words presented so that they eliminated priming effect
- --------------------------------------------------
- Research: the flanker effect
- Schaffner and LaBerge: 1979
- Description of Procedure:
- Three words presented on screen; words above and below target called "flankers"
- Participant to decide category of target as quickly as possible e.g. animal or vegetable and press a different button for each
- Findings:
- Takes longer to determine category of target when flankers belong to a different category
- Conclusion:
- Likely that flankers are being analysed in parallel to target information; when pressing the "other" button response is slower
- Similar to Stroop effect
- Criticism:
- According to feature integration theory (Treisman)serial attention is required to join features together, therefore it should not be possible to process three words simultaneously
- Addressed by Broadbent
- -------------------------------------------------
- Research: Proof of flanker effect
- Broadbent & Gathercole: 1990
- Suggest: target word primed the flankers as to be detected with min attention i.e. not unrelated flankers that make response to target harder but related flankers that make it easier
- Description of Procedure:
- Target appeared first followed by flankers after processing of target started
- If Broadbent correct & target primed flankers then priming effect would be stronger
- If Treisman correct, making flankers late impact should be reduced
- Findings:
- Strong flanker effect
- Conclusion:
- Suggesting priming idea correct
- Criticism:
- Well known that an item appearing in visual field attracts attention, could explain strong effect of flankers
- A convincing demonstration of parallel processing requires simultaneous presentation of stimuli
- ------------------------------------------------
- Research: Stroop effect
- Stroop: 1953
- Description of Procedure:
- A List of words in colour & colour related words in different colours than name e.g. green printed in red ink
- Patient required to name ink colour as quickly as possible
- Findings:
- Took longer to name ink colour on colour related words
- Conclusion:
- Processing one source of information interfered with by presence of another
- Variant: emotional Stroop task
- Severe depression produces cognitive impairments in elderly
- Depressed people showed attentional bias towards depression related material - helps in diagnosing
-
Summary: integrating info in seen displays
- When consciously perceiving complex material i.e. looking for particular letter of particular colour
- Perception requires attention
- Attention must be focused on one item at a time therefore: processing is serial
- Some parallel processing
- Indirectly detected i.e. influence of one word over other
-
Attention and distraction
- Attention captured by novel even advantageous from survival point of view but acts as distraction to attention
- The effect of irrelevant speech
- Attending across modalities
-
Attention and distraction
- Attention captured by novel event advantageous from survival point of view but acts as distraction to attention
- The effect of irrelevant speech:
- Recall/attention drops when speech occurs nearby but not with white noise
- Dichotic listening showed that ignored auditory material may be processed
- However, no impact seen when speech interferes with visual material
- Possibly due to speech not being constant i.e. broken into sounds
- Series of rhyming words less disruptive
- ------------------------------------------------
- Theories
- Jones 1999: listening requires strings of sounds to be organised into perceptual objects; segments of string must be identified, kept track of and ordered; process is automatic therefore interferes with visual memory task; when strings are simple repetitions, ordering task simpler i.e. less disruptive
- *
- Research
- Jones et al: 1999
- Description of Procedure:
- Remember viusally presented list while listening through headphones to repetitive sequence of 3 syllables
- Condition 1: " k...l...m...k...l...m..." - continuously changing stream of sounds
- Condition 2:
- • "k" in left ear
- • "m" in right ear
- • "l" both ears
- • I.e. 3 streams of constant sounds
- Findings:
- Condition 2 less disruptive than condition 1
- Conclusion:
- Irrelevant speech shows that parallel processing not always possible when demands are made on the same process
- Supports "bottleneck" theory of attention as suggested in early theories of attention
-
Attention and Distraction: Attending across modalities
- Attention across two modalities seem to be distracting, however, they can be mutually supporting e.g. lip reading
- Research: ventriloquist effect
- Driver: 1996
- Description of Procedure:
- Two messages from same speaker, in same voice, one shadowing the other
- Condition 1: TV monitor placed above speaker showing person speaking shadowed message
- Participants coped somewhat by lip reading
- Condition 2: monitor to side away from loudspeaker making sound appear to come from person's lips i.e. spatially separated, although sound did not actually change
- Findings:
- Shadowing became easier
- ------------------------------------------
- Application: mobile phones: Spence 2002: if sound delivered to one ear only e.g. earpiece, sound seems to pull visual attention to lateral message thus drawing attention away from events ahead and reducing driver's responsiveness
- *
- Summary: Cross modality attention
- Attention works hard to unite information
- • Spatially separate visual and auditory system can be synchronous (ventriloquist effect)
- • When not synchronous, system tires to order segments independently, result: distraction & loss of info
- • Can cause 'bottleneck' in the ordering system where processing one stream interferes with processing the other
-
The neurology of attention
- The effects of brain damage
- Event related potentials
- Summary: neurology of attention
-
The neurology of attention: The effects of brain damage
- fMRI functional magnetic resonance imaging: shows how area responds to sensory input
- Shows speech activates one area of brain
- Same area activated when lip reading without any auditory sound
- Implies that there must be connections between parts of visual and auditory areas
- *
- Damage:
- Parietal region is part of dorsal stream
- --------------------------------------
- Damage to single parietal lobe leads to sensory neglect: completely ignoring what happens on side opposite damage
- E.g., only eating from 1 side of plate; drawing 1/2 clock or 1/2 vase of flowers, and each flower only drawn 1/2 petals
- Sensory neglect:
- Half object neglected vs half field of view
- Research: tilted chess pieces
- Driver & Halligan: 1991
- Description of Procedure:
- Pictures of chess pieces, added piece to one side of object
- Testing to see if neglect was related to page or object
- Tilting picture so that top part (with the difference) was on right of page
- Findings:
- Unable to see difference if it was on neglected side
- Conclusion:
- Neglect was object related
- *
- Research: drawing cathedral square
- Bisiach & Luzzatti: 1978
- Description of Procedure:
- Asked patient to imaging looking at cathedral in his town and describe what he saw
- Then asked to stand on steps of cathedral and describe again what he saw
- Findings:
- Did describe very well but failed to mention half side of square
- When looking back he described other half of square i.e. previously ignored buildings
- Conclusion:
- Intact memory but not entirely accessible
- Attentive processes are involved in assembly of remembered material as well as in physical present material
- ---------------------------------------------
- Balint's syndrome: extreme form of neglect; Damage to both parietal lobes
- Extremely difficult to shift attention from one object to other e.g. lighter to cigarette
- Extreme difficulty in spelling out words but able to read i.e. each letter was small object but able to see a word as a whole
- --------------------------------------------
- Conclusion: we may all experience the world as a series of objects; however we can shift attention quickly from one object to other, so that we perceive them simultaneously, unless attention processes are damaged
- --------------------------------------------
- --------------------------------------------
- Damage to ventral stream = different integration problems
- Patients aware of all objects but segregated into small parts e.g. paintbrush= wooden stick and black objects(bristles)
- Humphreys: 2001
- Many different problems show that combining features takes place in different stages and brain locations
-
Neurology of Attention: Event related potentials
- ERPs: Event Related Potentials (event triggered by sensory signal)
- 1. Sensory info received
- 2. Series of events detected
- • Neurons fire in receiving part of brain
- • Then in processing parts of brain
- 3. Electrodes attached to scalp record electric signals
- --------------------------------------------------------
- Research: EPRs from sound
- Woldorff: 1993
- Description of Procedure:
- Recording signals with electrodes on scalp
- Findings:
- Signals from brain stem (1st relay betw ear and auditory cortex)10ms after event
- 20-50ms earliest registration in auditory cortex
- • Magnitude of activity affected by sounds played in other ear
- Conclusion:
- At very early stage of cortical analysis, attention from other stimulus reduces signal to brain
- Attention controlled by filter early on in processing, however, signal is attenuated not eliminated
-
Summary: neurology of attention
- Attention associated with the generation of perceptual objects
- Attention influences remembered experiences
- EPR shows cortical signals derived from unattended external stimuli are attenuated
-
Conclusion of Attention
- Attention is vital in joining features of objects
- We generally perceive objects consciously vs their separate parts
- Attention controls cross-modal input
- Attention deals with competition of received stimuli
- Analysis takes place in parallel
- Complex analysis is serial
- System must select from competing stimuli
- Attention selects what reaches conscious awareness - not all results become consciously available
- Allport: attention directs actions & saves us from carrying out simultaneous incompatible actions
- However: issue of consciousness relevant e.g. experienced drivers can perform many actions, conscious awareness of each action has disappeared and they have become automated
- Sense of time changes under hypnosis
- Attention is focused on internal feelings, mental images are self-generated and only change when wanted, therefore one is less aware of passage of time then when sampling from constantly changing environment
- Ultimately all facets of attention lead to conscious awareness
|
|