Lecture 4

a problem with feature integration theory

1- pop can be dependent upon complex features 2- high level qualities not low level qualities predict pop out 3- this is inconsistent with the idea that pop to is only relevent to simple feature tasks.

what are the problems with feature integration theory FIT?

1- pop can be dependent upon complex features 2- high level qualities not low level qualities predict pop out 3- this is inconsistent with the idea that pop to is only relevent to simple feature tasks.

Which search task needs attention?

Conjunction tasks. Serial search doesnt require feature binding.

What did wolfe find?

He found that when comparing parallel versus serial search, once cannot merely have a direct straight line for distractors versus reaction time. In fact the line actually curves around. this is saying that it isn't straight parallel or straight linear search. it is actually inefficient versus efficient search. Wolfe believes that this actually curves. it isn't a clear straight line pattern. ok so this straight line that other people have proposed says that as you add a new item the time taken to process each new item diminishes. but wolfe said lets talk about efficiency of search... so in this way we are talking about the efficiency of the search. so a parallel search will be more efficient than a serial search. because the time taken for the parallel search isn't 'effective'.

Please describe the slope for reaction time in target absent and target present trials.

In target absent trials the line is twice as steep. 2:1 slope. the slope is steeper because reaction time increases as the amount of items increase in a display search. this is because this display needs to be serially searched. Item by item.

The main advantage of studying visual search using letters and digits as stimuli instead of objects in natural scenes is:

It allows quantification of the overall processing load

Whats the task used to locate relevant stimuli in the environment?

It is called the visual search task. -We show a target, then we ask if its present or not -a target is only pop out when it can be defined by a single feature -We usually do visual search tasks by measuring the reaction time. -Easy search tasks/unique items(require parallel search)= pop out effects (no effort required to see the search task→ effortless) -------With pop out type stimuli, the RT doesn't change, even when the distractors increase Conjunction targets (require serial search) do not pop up (it is called a conjunction target as it has numerous features that are combined that have to be searched for. Both properties must be identified to be located) -------RT increases, RT increases with more distractor items. → there are a number of features (hence conjunction). This task is more effortful, it takes much longer to complete. Each item (distractor)that is put in the search frame, increases search time. This is because it increases time of scanning. Search tasks are self terminating. Once we locate the target, we terminate the search. On target absent trials, we have to scan the entire search screen so we can 'exhaust' the search as we cannot find it.

In the attentional blink paradigm:

Performance on the second target is worse if the first target must be processed

Who looked at automaticity in the sense of visual search?

Shiffrin and schneider. -presented a memory set -presented frames and asked if any of the targets were presented in these memory sets. -two things make this hard (1-size of memory set) (2-number of items simultaneously present in the display) -there were two versions of this. once memory set had been presented. there was the presentation of the display items. *version 1- consistent mapping (targets always used as targets, distractors always used as distractors) By "consistent mapping between target and distractor categories" we mean that the targets were always consonants the distractors were always digits. *version 2- varied mapping "consistent mapping between target and distractor categories" we mean that the targets were always consonants the distractors were always digits. By "varied mapping between target and distractor categories" we mean that the targets distractors were both mixed sets of consonants and digits. the findings-- -performance under consistent mapping became automatic with practice. -performance never became automatic under varied mapping. -the idea that automaticity is achieved is that this must have a parallel search, and be capacity free.

According to Treisman in FIT theory, what does it mean when there is no attention?

That objects are free floating. Without having attention onto something things are free in space. This is called illusory conjunctions.

the attentional blink

The attentional blink is suggested that is a similar phenomena. -present rapid sequence of characters at the centre of the display 100ms exposure per letter. each new item masks the one that precedes it - report two targets 1. report the white letter (T1) 2. Is an X presented (T2) -the lag is measured in frames= time. a lag of two means there is one intervening item. -We want to know how well people do on T2, on a function of the occurrence of T1. So seeing how much the lag between T1 and T2 affect the outcome of T2. Result→ there is a dramatic decline in T2 performance, there is a lag at 3 items, and then it recovers. -if you tell people to look for one item. there is no attentional blink. -The attentional blink shows failure to divide attention across two tasks. If people make an attempt to process the first target there is a massive cost on whether they recall the second target T2. -performance depends on time lapsing -the worst performance occurs at a lag of three or four. recovers after three items. -effect takes time tp build up -this highlights another cost of dual attention

What is another example of an experiment that processes irrelevant stimuli? Please explain the experiment.

The erikson flanker task -there are three strings of letters. -People are detecting a particular target letter -People ask which of two targets letters is present? the respondent must respond as quickly as possible. -This target letter can be shown under three conditions • Compatible condition→ flanking distractor letters are the same as the target • Mutual→ central target is flanked with mutual distractors, such as all X's when the target is an E • Incompatible target→other flanker alternatives results -there is involuntary processing of flankers even when attempting to ignore them. -there is a cost from incompatible targets and flankers -there is benefit from compatible items same as strop effect -Reaction times are faster in the compatible condition -The results= RT fastest in compatible, neutral and then incompatible. -failure of focused attention -the identity of the flankers have an effect on performance even when you try and ignore them (failure of focused attention) -there is some parallel processing of conjunction stimuli -the effect of the compatibility is dependent upon a few different things a) decreases with spatial separation (the effect is most prominent when the letters are close together) =Tells us something about the attentional spotlight. This is consistent with the idea that anything that falls within the attentional spotlight is processed involuntarily/ automatic. things outside the spotlight can be ignored.

what theory did welfard put forward?

The single channel theory. -this theory proposes that you are only able to make one decision about a stimulus at a time. -decision making can cause a bottleneck on performance -we don't process all stimuli together. we can reject a lot of stimuli (non targets rapidly) (parallel) -complex decisions take moire time than simple decisions. we don't process all stimuli equally.

Name an experiment and describe it where we cannot help but process distractor stimuli.

The stroop effect. -This is an experimental task. You show people strings of letters that are presented in colour ink.. the task is to ignore the word and name the colour of the ink that the word is written in. • Fast=compatible • Neutral=intermediate • Slow=incompatible -The effort of attempting to avoid distractors when completing a task requires much effort. In essence this increases the reaction time. -The effect is you process an irrelevant part of the stimulus involuntarily. -Two processes that both have different efficiencies. 1. Fast process→ reading 2. Slow controlled process→word name available -The process associated with irrelevant attribute is the fastest process and it will continue to make interference. -The stroop effect is a failure of focused attention. -Conclusion Word reading= fast and automatic process Colour naming=slow and controlled process -the reason why you are influenced by the word reading (interference) is because this is an automatic process. there is parallel processing of naming colour and word reading. the word name is available before the colour name! this creates interference. the process attributed to the irrelevant process is the faster process. The strop effect is a failure of focused attention. Cannot ignore an irrelevant attribute. Because it is the most automatic task. (therefore word reading is completed more rapidly than naming the colour of the word)

The role of attention is to bind features. what is this theory called? who is it by? define this theory.

The theory that attempted to bring this concept together is called the feature integration theory (FIT) Treisman and Gelade. -Bind different features together into a coherent perceptual object. 'The perceptual hole'. Put all these features into one location. -The idea is that we have separate maps for color, diagonal, vertical all the elementary features that the visual system is sensitive too. The content of these various maps are not related to each other -The role of spatial attention is to bring together these different maps -By giving attention to things we put the visual spotlight on something -this is why there is no change in reaction time in serial search -there is an increase in response time as a result of conjunction targets. this is because as distractors increase, so to does the reaction time.

Shiffrin and Schneider (1977) found that automaticity was acquired when:

There was a consistent mapping between target and distractor categories -explanation The relevant experiment is found in Shiffrin and Schneider (1977), not to be confused with another prominent paper Schneider and Shiffrin (1977). [An aside: before moving to Melbourne Dan was a Postdoctoral researcher in Shiffrin?s lab!]. In their experiment the stimuli were digits and capital letters (actually, they used only capital consonants - you can imagine that ?O?/?0? and ?I?/?1? might have become annoying if they allowed vowels). Participants would view a memory set containing a certain number of items. The memory set is consonants only in the ?consistent mapping? condition, but in the ?varied mapping? condition the memory set can contain consonants and digits. Those memorized items would become the search targets in the next phase of the experiment, in which the participants searched test frames. Then they?d be shown a memory set again, and then search a test frame again, and so on for a large number of trials. By "consistent mapping between target and distractor categories" we mean that the targets were always consonants the distractors were always digits. By "varied mapping between target and distractor categories" we mean that the targets distractors were both mixed sets of consonants and digits.

Whats the point of spatial attention?

To locate stimuli among distractors.

an alternate model to the use of only serial processing. or only parallel processing. Who provided this model?

Wolfe provided this model. It is called the guided search Theory. The basic idea of this theory is that search involves two consecutive cognitive stages. 1-parallel, 2-serial. stage 1- parallel search (provides a candidate list of possible targets) MAKE LIST stage 2- serial search -CHECK LIST (Serial search of things that might be targets) The efficiency (the measured reaction time depends on the time taken to complete both these stages). the second stage timing is dependent on the size of the list. so (large candidate list= more time= inefficient), (small candidate list= less time= efficient). Th e consequences of this two stage theory depends on the reaction time taken for the intial fast parallel search and then the second serial search. This depends on the distractors and targets (if large candidate list= inefficient search) if small candidate list= efficient list)

The 2:1 slope ratio in mean RT found in conjunction search tasks has been taken as evidence for:

a serial self terminating search Self-terminating means the searcher is stopping when the target gets found. That means that in the target-present condition on average a searcher will search half the display before finding the target. However, in the target-absent condition the searcher will always search the whole display. That predicts the 2:1 slope ratio described in the question.

what makes a process automatic?

parallel, fast, effortless, doenst require capacity learned stimulus-response association -automaticity enables us to chain together hard sequences of complex skills. -learning a task can initially be cognitively demanding, but then practice can make this an automatic process.

what is an illusory conjunctions?

without attention items are free floating in space.