Visual Attention

attention and memory

- Attention is also important for encoding information into visual working memory - Not only does attention influence what you experience and remember, experience and memory influence what you attend to - Memory traces of past perceptual interactions bias how attention should be allocated to the visual world

change blindness

- Awareness of the identity and attributes of visual objects can be probed by asking subjects to detect changes made across film cuts , between alternating images (Rensink, , or across eye movements. Results - Subjects perform miserably at detecting changes, even when this involves changing the identity of a real person in the real world asking for your directions to the local library - Thus, although a great amount of detailed information is available in natural scenes, the amount of information that is consciously retained from one view to the next, or from one moment to the next, appears to be extremely low. Conclusion: greatly impaired in their ability to notice changes in any but the currently attended object unless the change alters the "gist" or meaning of a scene

single target search

- Each frame contained 9 or 16 letters each and were presented at rapid rates of 40 to 50 ms. - The task was to detect a single target numeral embedded in one of the frames results: - practiced observers can scan through up to 125 letters per second. - people are very good at temporal selection Implications: - it is possible to extract meaning from visual stimuli at rates much faster than the speed with which these meanings can be stored in any but the most fleeting of memories

Box and line experiment

- Each target is comprised of two overlapping objects, a box or a line. - The box that could be large or small with a gap to the left or right. - The line could be tilted right or left and comprised of either a dashed or a dotted line. Results: - Attending and reporting two attributes from a single object was easier than reporting two attributes, each from different objects. - performance suffered when they had to attend to two objects rather than just one Implications: - because the overlapping to abutting objects occupied the same location, the performance differences must be due to attentional allocation over object-based representations.

factors that decrease search efficiency

- In general, as target-distractor differences get smaller, search becomes less efficient - Increasing distractor inhomogeneity. - Targets defined by conjunctions of two or more basic features ( e.g., color X orientation: a red vertical line among green vertical and red horizontal distractors) - Targets defined only by the spatial arrangement of basic features are, as a general rule, not found efficiently (e.g. search for an "S" among mirror-reversed Ss will proceed at a rate of 20-30 msec per item on target present trials.)

factors that increase search efficiency

- Large target-distractor differences in features such as color, orientation, motion, size, curvature, some other form properties, and some 3-D properties. - Increasing distractor homogeneity - Conjunction targets can be found efficiently if the differences in target and distractor features are sufficiently salient - Difficult searches can become more efficient with extensive practice. However, such perceptual learning is specific to the training stimuli.

token individuation hypothesis

- explains repletion blindness - the hypothesis that a second object file is not created for the second identical event because the visual system fails to treat the second repartition as a different object from the first.

mechanisms underlying search efficiency

- factors that decrease search efficiency - factors that increase search efficiency

What is the neural correlate of this spatial filter or attentional spotlight?

- filtering processes in V4 neuronal responses. Experiment: - presented two stimuli within the receptive field of a V4 neuron being recorded. Now was effective for producing the cell's response, and the other "ineffective" stimulus wasn't. Results: - When monkeys attended to the location occupied by the ineffective stimulus, the cell failed to respond to the presence of the effective stimulus. Implications: - attention modulated the cell's response such that the presence of a competing (effective) stimulus was filtered out.

how is object-based selection achieved

- internal representations known as "object files" support out ability to attend to objects as they undergo occlusion and fragmentation or change over time.

three inhibitory phenomena

- invalid cueing - negative priming - inhibition of return

Event-related potentials

- the massed electrical activity of neurons can be measured through scalp electrodes. - when these electrical events are correlated in time with sensory, cognitive, or motor processing, they are called "event-related potentials" (ERPs). Pros: - non-invasive - high temporal precision Cons: - anatomical precision is limited

measuring search efficiency

Reaction time as function of set size: - The most common method is to vary the number of items in the display (set size) and RT as a function of set size. - The slope of the RT X set size functions is a measure of search efficiency. - A slope of zero msec/item indicates that the target item, when present, is detected without interference from the distractor items. Steeper slopes (bigger) indicate less efficient search and a greater cost for each additional distractor. Accuracy as a function of set size: - Efficient searches produce high levels of accuracy independent of set size even when the display is presented very briefly. For less efficient tasks accuracy declines as set size increases unless exposure time is increased to compensate

Can the spotlight of attention be split into multiple spots?

Test: - have subjects attend to two spatially separate loci and measure attentional effects at intermediate loci Results: - subjects can split focal attention and maintain two attentional foci across hemifields - distractors appearing between two noncontiguous locations did not affect performance for targets. Test: - ask subjects to track the movements of multiple objects Results: - appear to show that subjects can allocate attention to something like four or five objects moving independently amongst other independently moving distractors Implications: - Subjects can divide the spotlight into four to five independently targetable beams

Assuming that one has allocated one's full attention to a particular location, object or event, how focused is selection at that spot?

Test: The resolution of attention is studied by measuring the effects of distracting items on target processing. - Flanker task (interference task): the resolution of attention is revealed by examining the distance at which distractors start to impair target discrimination performance Results: - the acuity of attention is of coarser spatial resolution than visual acuity - overcrowding effect: items spaced more closely than the resolution of attention cannot be singled out (individuated) for further processing.

Guided search explained by second class

a simple feature search is efficient because preattentive processes can direct the first deployment of attention to the likely target item.

object files

are episodic representations that "maintain the identity and continuity of an object perceived in a particular episode"

A search for an S among mirror-Ss explained by second class

are inefficient because no preattentive information is available to distinguish one item from the next.

conjunction searches explained by second class

are of intermediate efficiency because preattentive feature guidance is available but it is not as strong as in a simple feature search.

overcrowding effect

items spaced more closely than the resolution of attention cannot be singled out (individuated) for further processing. - check out figure 9.2

dwell time

measures the amount of time individuals spend engaged on a stimulus - there are two dwell times

attention theory

proposes that rejected information is attenuated rather than completely filtered or completely identified - unattended information is not completely filtered, but it is not processed to the same degree as attended information either

visual search tasks

subjects look for a designated target item among a number of distracting items

reaction time (RT) tasks

the display is usually present until a response is made

accuracy tasks

the display is usually presented very briefly, followed by an interfering visual mask

spatial selection studies

typically have subjects focus on a subset of the spatial array, allowing for selective report of information at the focus of attention - spotlight metaphor: the feeling that attention can be deployed, like a bean of mental light, to reveal what was hidden in the world - cueing experiments

early selection theory

unattended, filtered information is not processed beyond its physical attributes

attention and implicit seeing

- As reviewed earlier, unreportable items in the attentional blink are nevertheless identified - In the inattentional blindness paradigm, Mack has shown that people are "less" blind to stimuli such as one's name or faces, suggest- ing that some meaning is extracted from those apparently unattended objects. In sum, attention limits what reaches conscious awareness and what can be reported through ex- plicit seeing, but sophisticated implicit perception may proceed for unattended, unreportable visual stimuli.

repetition blindness

- The AB is typically measured for two visual events that are different from each other, so what would happen if the two targets were identical? - Performance is worse for repeated targets (repetition blindness) - RB is the result of a failure to create separate object files for the second of two repeated items. The visual system fails to treat the second repetition as a different object from the first. (token individuation hypothesis)

N2pc

- The N2pc reflects the focusing of attention onto a potential target item in order to suppress competing information from the surrounding distractor items. - N2pc may serve as a marker of where attention is focused and how it shifts across space. - N2pc component rapidly shifts from one item to the next during visual search. Lends support to theories that propose attention moves in a serial manner between individual items rather than being evenly distributed across items in the visual field.

invalid cueing

- When a cue stimulus appearing before the target is informative, it will facilitate target performance compared to a baseline in which the prime is neutral. - What if the prime is an invalid cue to the target? This should generate a negative expectation that slows down performance to the target. - found that inhibitory effects are only observed for targets appearing beyond 400 ms after the prime presentation (fake cue). i.e if the target appears 350s after fake cue, the participant will respond quickly.

first class of models -limited-capacity models

- abandons the serial/parallel distinction altogether - argue that all items in a search are processed at once, or perhaps, in groups. - Differences in search efficiency arise because different types of items make different demands on a limited processing resource.

PET and fMRI

- allow non-invasive imaging of brain activity during performance of sensory, cognitive, and motor behaviour - PET measures cerebral flow (rCBF) and fMRI measures deoxygenation signals in the brain pros: - non-invasive nature - ability to measure brain activity across the entire brain with high spatial resolution cons: - limited temporal resolution (slowness of blood flow changes)

Flanker Task

- an experimental procedure in which participants respond to the direction of the central arrow in an array, regardless of the direction the other arrows are pointing - the resolution of attention is revealed by examining the distance at which distractors start to impair target discrimination performance

exogenously

- an external stimulus or event that automatically draws attention to a particular location - bottom-up - stimulus-driven - draws attention automatically and has a rapi, transient time course

object-based representations

- are "sophisticated" in the sense that they represent more than the raw visual input - object-based deployment of attention is influenced by factors that determine perceptual grouping

attentional blink paradigm

- asked subjects to report two targets presented amongst an RSVP stream of distractors - The temporal lag between T1 and T2 was varied systematically across a range of intervals from 80 to 320ms Results: - there was an impairment in detecting T2 if it appeared within half a second of T1. - the inability to report T2 for an extended time after T1 has come to be known as the attentional blink (AB) - an attentional effect rather than a sensory making effect - AB is dependent on the presence of a distractor or mask in the position immediately after T1 (called the +1 position) When this item was removed and replaced with a blank interval, AB disappeared. Although AB is not a masking effect itself, perceptual and/or conceptual interference with T1 is important. Interestingly, when T2 appears in the +1 position, it may be processed together with T1 , allowing it to be reported at relatively high accuracy. Thus, the AB reveals limitations in the rate at which visual stimuli can be processed,

attention and the phenomenology of conscious perception

- attention did not affect the perceived intensity or clarity of a stimulus and had only a small, inconsistent effect on the veridicality of the perceived color or location of a stimulus. - Although attention does not change the experienced clarity and intensity of stimuli, it may determine how you perceive stimuli, especially ambiguous ones. (Consider Rubin's ambiguous figure (Rubin, 1915/1958) which induces a percept that oscillates between two faces or a vase. Attention appears to determine which figure is perceived. )

endogenous attention

- attention is presumed to be under the overt control of the subject - voluntary, effortful, and has a slow (sustained) time course - top-down - goal-driven

Hillyard and colleagues- ERPS - early attention evidence

- auditory modality - using a dichotic listening paradigm in which subjects attended to information from one ear vs the other, they demonstrated that early sensory components(deflections) beginning with 100ms post-stimulus were enhanced for attended stimuli. - these results generalize to visual selection in which subjects were required to attend to one of two spatial locations. Early components of the ERP waveform (P1 and N1) were typically larger for stimulus presented at attended locations versus unattended locations. These effects also begin within 100ms of stimulus onset, providing clear evidence for attentional modulation at early stages of visual information processing. - early selection mechanisms also generalize to visual search tasks using multielement displays. The N2pc reflects the focusing of attention onto a potential target item in order to suppress competing information from the surrounding distractor items. N2pc may serve as a marker of where attention is focused and how it shifts across space.

contributions of PET and fMRI to attention modulation

- demonstrated that attention modulates the activity of extra striate cortical areas specialized for feature dimensions such as colour or motion: > motion processing regions (MT/MST): more activity here if attention was focused on the speed of the motion of the objects > dorsal region in lateral occipital cortex and a region between the fusiform and lingual gyri: activated during attention to colour. > fusiform face area: attention to faces - illuminated our understanding of mechanisms that drive attention to different spatial locations > superior parietal cortex: may play an important role in shifting attention around locations in space.

Filtering in spatial search paradigms

- match-to-sample task in which monkeys were first shown a single target stimulus, then asked to make an eye movement to the same target item in a subsequent array which also contained a distractor item. results: - neural activity to the distractor stimulus was initially present, but subsequently suppressed at around the 200ms after the onset of the search array, illustrating a neural correlate of competitive selection.

single-cell physiological method

- measures activity from individual neurons presumed to be participating in perceptual or cognitive operation Pros: - provides the highest spatial and temporal resolution Cons: - invasiveness - only a few neurons can be examined at any given time (makes it difficult to examine how multiple brain areas interact with each other to perform a particular task)

active selection

- might occur early or late in processing - early selection theory - late selection view - attention theory (intermediate theory) - *attention is not a singular thing with a single locus, early or late. Rather, it is a multifaceted term referring to a number of different acts and loci of selection.

Motter's Guided Search by V4 experiment

- monkeys were required to select elongated bar target on the basis of colour and then report its orientation. - V4 neurons whose receptive fields included stimuli of the target colour maintained their activity whereas V4 neurons whose receptive fields contained items of different colours had depressed activity.

selection

- passive selection - active selection

object-based attention experiment

- presented two different movie sequences that overlapped each other in space. People were throwing a ball in one movie and playing a hand game in another. - subjects were asked to attend to only one of the two overlapping movies. Results: - subjects were able to follow actions in the attended movie and make responses to specific event in it, as instructed by the experimenter, - odd events in the unattended movie were rarely noticed. Implications: - since both scenes overlapped each other, this demonstrates a selective attention that cannot be spaced based. Rather selection was based on objects and events

second class of models that don't support integration theory

- preserves the distinction between serial and parallel processes. - hold that the preattentive stages of vision are characterized by parallel processing of basic features and that there is a bottleneck after which processing is essentially serial. - variation in the efficiency of search is determined by the ability of preattentive, parallel processes to guide attention toward candidate targets or away from likely distractors.

inhibitory mechanisms of attention

- reducing ambiguity - they can protect central, capacity-limited mechanisms from interference - they can prioritize selection for new objects

attention and explicit seeing

- researchers have argued that attention is needed for conscious perception - recall that subjects could only remember details from the attended movies in Neisser and Becklen's study - inattention blindness - change blindness - Perhaps many of these findings can be understood by noting that attention is necessary to prevent visual events from being overwritten by subsequent stimuli. - attentional masking

N400 component and attentional blink

- sensitive to semantic mismatch i.e. "He went home for dinner and ate a worm." The last word "worm" would trigger an N400. Thus, the presence of N400 would indicate that a word has been processed up to its semantic meaning. - If blinked items are suppressed early and not recognized, then little or no N400 should be observed for blinked targets. - If AB is produced by capacity limitations after initial identification has occurred, then the N400 should be preserved even for blinked words which could not be reported. Results: N400 was preserved, providing evidence of direct semantic processing without awareness

bottom-up, exogenous visual attributes that draw attention

- spatial ces and abrupt visual onsets (sudden luminance changes) - salient visual features such as feature singletons - under greater volition (conscious) control (When the task requires searching for a target defined by a singleton in one dimension (e.g., orientation), then singletons in other di- mensions (e.g., color) automatically draw attention even when this is detrimental to per- formance )

cueing paradigm

- subjects are required to respond as quickly as possible to the onset of a light or other simple visual stimulus - the target stimulus is preceded by a "cue"(the brightening of an outline object, onset of some simple stimulus, or a symbol, like an arrow) whose function is to draw attention to the occurrence of a target in space. Results: - cues facilitate detection of and response to stimuli presented at the cue location. - attention as a "spotlight that enhances the efficiency of the detection of events within its beam".

negative priming

- subjects are slower at responding to targets (probes) that were distractors - This suggests that the representation of the ignored primes was actively suppressed, and that this inhibition was carried over to the following trial. - negative priming operates at an abstract, semantic level - single trial exposures to novel figures can produce negative priming, suggesting that implicit representations of unknown shapes can be formed and retained from ignored and unremembered events

inattention blindness

- subjects performed several trials of a length judgment task for two lines bisecting each other in the form of a cross at the center of the computer screen. -On one of the trials, an additional test figure was presented along with the cross figure, and subjects were queried of their awareness of this test stimulus. Results: -The remarkable finding is that a large proportion of subjects did not even notice the test figure, suggesting inattentional blindness. Implications: concluded that attention is needed for conscious experience

Bichot and Schall's Guided Search by FEF experiment

- the activity of FEF neurons evolves to discriminate targets from distractors in search tasks, prior to initiating a saccade to the target - the activity of FEF neurons was stronger to distractors that shared visual features to the target, suggesting a neural correlate of Guided Search.

passive selection

- the front end of a sensory system acts as a type of passive selector, admitting some stimuli and not others - essentially, passive selection continues beyond the receptors

inhibition of return

- three outlined boxes - peripheral cues appeared, either validly or invalidly cueing the onset of a target in the peripheral boxes. - The SOA between cue and target was varied and the usual facilitatory effects of cueing were obtained for targets appearing within 300 ms of the Cue in the same spatial location. - When the SOA exceeded 300 ms, target detection performance was slowed, suggesting a transient bias against returning attention to visited locations

attentional masking

- under conditions when attention is not focused on an item, that item is subject to substitution or erasure by other, subsequent stimuli even when those other stimuli do not overlap the contours of the "erased" visual target.

visual search paradigm

- uses visual search tasks - set size - the tarter is presented on some percentage of the trials, typically 50%. - reaction time and accuracy are measured - search efficiency

Questions arising from spotlight metaphor

1) When attention is deployed from one location to another, do such attentional shifts occur in a digital, instantaneous fashion, magically appearing in a new location to be attended? or does attention move from one location to another in an analog fashion, illuminating intermediate locations as it travels across visual space? A: the focus of attention can move instantaneously from one location to the other without a cost for the amount of distance travelled. However, it is unclear whether attention has an effect on intermediate loci as it moves from point A to point B 2) Can the spotlight of attention be split into multiple spots? A: Attention can be split across two locations Subjects can divide the spotlight into four to five independently targetable beams 3) Assuming that one has allocated one's full attention to a particular location, object or event, how focused is selection at that spot? A: items spaced more closely than the resolution of attention cannot be singled out (individuated) for further processing.

limitation of spotlight metaphor

1. attention can be allocated to regions of different size (the spotlight has a variable width of focus) - the speed of a response to a stimulus is dependent on how narrowly attention is focused. The spatial distribution of attention follows a gradient with *decreased effect of attention* with *increased eccentricity from its focus*. 2. focus of attention may be yoked to overall load or difficulty of a task. - for attention to remain focused on a target, the overall perceptual load of the task must be sufficiently high to ensure that no capacity remains to process other non-target events. - in the absence of sufficiently high load, attention spills over to non-target events. 3. Attention can be allocated to 3-D layouts and restricted to certain depth planes defining surfaces in space. - selection occurs after 3D representations have been derived from the 2D input. 4. there are "object-based" attention that be considered independent of spatial selection.

critical roles of attentional mechanisms

1. attention can be used to select behaviourally relevant information and/or to ignore the irrelevant or interfering information (you are only aware of attended visual events) 2. attention can modulate or enhance the selected information according to the state and goals of the perceiver

temporal attention: visual selection over time

A standard technique for studying temporal attention is to present rapidly presented sequences of visual items at rates of up to 20 items per second (rapid serial visual presentation)

reasoning behind attentional blink

If a stage of processing is limited in capacity, then this will take a certain amount of time to complete. This impairs or delays the system's ability to process a second stimulus presented during this busy interval, causing the attentional blink

Treisman's Feature Integration Theory

Individual Feature processing: is done in parallel. Simultaneous processing is done on the whole display and if feature is present-- we detect it. Conjunctive searching : requires attention to the integration or combination of the features. Attention to particular combination of features must be done sequentially to detect presence of a certain combination. - held that efficient feature searches were performed by mechanisms capable of processing all items in parallel, and that all other searches relied on mechanisms that oper- ated in a serial, item-by-item manner. - In particular, attention was required to conjoin or bind multiple features into a single object. Hence, conjunction searches were serial , and withdrawing attention produced errors for binding features, known as "illusory conjunctions"

parts of the visual system that show attentional modulation of activity

V1 - attentional modulation for early selection V4 and MT: modulatory activity for motion processing frontal eye field: neural activity that differs for targets and distractors

inhibition effects

can be measured as a decrement in performance relative to a neutral base- line.

attentional blink

during a brief time after perceiving one stimulus, it is difficult to attend to something else

feature singleton

i.e. red target amongst green distractors or a vertical target amongst horizontal items

explicit seeing

occurs when subjects can explicitly report what visual event had occurred

implicit seeing

occurs when visual stimuli have been identified, as measured by their impact on performance, but can't be explicitly reported by the subject. - masked priming paradigms provide a good example of implicit seeing,

rapid serial visual representation

present rapidly presented sequences of visual items at rates of up to 20 items per second

late selection view

selection occurs only after categorization and semantic analysis of all output has occurred.

When attention is deployed from one location to another, do such attentional shifts occur in a digital, instantaneous fashion, magically appearing in a new location to be attended? or does attention move from one location to another in an analog fashion, illuminating intermediate locations as it travels across visual space?

the focus of attention can move instantaneously from one location to the other without a cost for the amount of distance travelled. However, it is unclear whether attention has an effect on intermediate loci as it moves from point A to point B

stimulus onset asynchrony (SOA)

the time between the cue and the target

set size

the total number of items in the display