Attention II Theories of Attention Visual SearchSpotlight Theory of AttentionPosner Cueing TaskSlide 4Endogenous Cueing TaskSlide 6Components of AttentionProblem (1) for spotlight theoryProblem (2) for spotlight theoryEvidence for object-based attentionWe can select a shape even when it is intertwined among other similar shapesObject-Based AttentionSlide 13Slide 14Visual SearchExamples of Visual SearchSlide 17Slide 18Slide 19Slide 20Slide 21Typical ResultsFeature Integration Theory (FIT)Interpretation in FITFeature search asymmetriesIllusory conjunctionsProblem for Feature Integration TheoryGuided SearchEarly vs. Late Selection TheoriesWhen listening to somebody, what else might get noticed?Dichotic listening/ Shadowing tasksWhat gets through?Early Selection Theory (Broadbent)Problem (1) for early selection theoriesProblem (2) for early selection theoriesTreisman’s Attenuation theoryAttention IITheories of AttentionVisual SearchSpotlight Theory of Attention•Theory which holds that we can move our attention around to focus on various parts of our visual fieldPosner Cueing Task•Posner presented either a central cue or a peripheral cue before a target appeared•Cues valid on 80% of the trials, invalid on 20% of the trials or vice-versacuetargetISICentral cueperipheral cuePosner Cueing TaskcuetargetISICentral cueperipheral cueCentral Cue condition triggers endogenous attention: voluntary attention Top-downPeripheral Cue condition triggers exogenous attention: Bottom-up i.e. stimulus drivenEndogenous Cueing TaskEndogenous Cueing TaskComponents of Attention•The findings from patients with brain damage led Posner to construct a model for attention that involves three separate mental operations: •disengaging of attention from the current location•moving attention to a new location•engaging attention in a new location to facilitate processing in that location.Problem (1) for spotlight theory•Attention does not appear to move around in continuous fashion (like a moving spotlight).–Moving attention isn’t slowed down by intervening stuff–Distance splotlight needs to travel does not affect response time (when other artifacts are controlled)•Quantal theories of attention (Sperling & Weichelgartner, 1995)–attention jumps from place to placeProblem (2) for spotlight theory•Attention appears to be object-based, not location basedlocation-based attentionobject-based attentionEvidence for object-based attention•Experiment: are number of bumps on the ends of the objects the same?•Faster judgments when bumps are on the same object (in spite of slightly larger distance)•compatible with an object-based attention theoryWe can select a shape even when it is intertwined among other similar shapesAre the green items the same? On a surprise test at the end, subjects were not able to recall shapes that had been present but had not been attended in the task Evidence for object-based attentionObject-Based AttentionSubjects attended either the moving or the static objectObject-Based AttentionFFA = fusiform face areaPPA = parahippocampal place areaObject-Based AttentionVisual SearchExamples of Visual SearchFinding objectsFinding a faceDisjunctive SearchWhere’s Waldo?Conjunctive Searchdisjunctive feature searchLook for an “O”TTTTTTTTOTTTTTTTTTTTTTTTTTTTTTTTTTdisjunctive feature searchLook for something redTTTTTTOTOOOTTOTOOTOOOOOOOTTTTTTTConjunctive feature searchLook for something red AND “O”TTTTTTOTOOOTTOTOOTOTOTOOOOOOOTTTTTTTTSearch times can be influenced by set sizeis there a black circle?Number of Stimuli in DisplayResponse TimeConjunctive SearchDisjunctive feature searchTypical ResultsFeature Integration Theory (FIT)Number of Stimuli in DisplayResponse TimeConjunctive SearchDisjunctive feature search(serial processing; attention needs to deployed to each region)(parallel processing; target pops out; search is preattentive)Interpretation in FITFeature search asymmetriesIt is easier to find X among Ys than Y among Xs if X has an extra feature compared to Y.QQQQQQQQQQQQQQQQQQQOO OQOOOOOOOOOOOOOOOOOFind the OFind the QIllusory conjunctions•Prediction of theory: if attention can conjoin features correctly, the lack of attention can lead to incorrect (illusory) conjunctions?Read the vertical line of digits in the following displayFor unattended locations, subjects might report illusory conjunctions of features, e.g. blue “O”Snyder, 1972; Treisman & Schmidt, 1982Problem for Feature Integration Theory Some conjunctions are easy and produce fast search times.XXXXOXOOOOOXX(e.g. Theeuwes and Kooi, 1994)Guided Search•Guided search model is a modification of feature integration theory•Separate processes search for Xs and for white things (because they are the target features), and there is a consequent area of double activation that draws attention to the target. XXXXOXOOOOOXX(Wolfe, 2003)Early vs. Late Selection TheoriesWhen listening to somebody, what else might get noticed?Dichotic listening/ Shadowing tasksdemo: http://www.linguistics.ucla.edu/people/schuh/lx001/Dichotic/dichotic.htmlWhat gets through?What happens to unattended message?Not much, we seem to remember mostly low-level information (human voice or not, changes in gender, not a change in language)The same word can be repeated without being noticedEarly Selection Theory (Broadbent)•Sensory information is processed until a bottleneck is reached•One of the inputs is then allowed through a filter on the basis of its physical characteristics, with the other input remaining in the buffer for later processingProblem (1) for early selection theories•People notice their own name at parties: cocktail party effectTreisman (1960)“Mary had a little lamb”“...[John Smith] you may stop now”“Mary had a little lamb”Resultw/o name: 6% noticew/ name: 33% noticeshadowed ear:unattended ear:Problem (2) for early selection theories•Some semantic processing in unattended ear•Treisman experiment:Treisman (1960)Treisman’s Attenuation theory•Messages are attenuated but not filtered on the basis of physical characteristics•Semantic criteria can apply to all messages, attenuated or not•Semantic criteria are harder to apply to attenuated messages, but still