Selective Attention PDF

Unit 2 Selective Attention Attention & Perception Dr Harry Moore Bachelor’s Degree in Psychology INDEX CONTENTS 1. Introduction 2. Dichotic Listening Tasks 3. Models of Selective Attention 4. Comparing Models of Selective Attention 2 ...

Unit 2 Selective Attention Attention & Perception Dr Harry Moore Bachelor’s Degree in Psychology INDEX CONTENTS 1. Introduction 2. Dichotic Listening Tasks 3. Models of Selective Attention 4. Comparing Models of Selective Attention 2  1. Introduction The Cocktail Party Problem Imagine: You’re at a party, in conversation with a Real Murcia mega-fan, and they explain to you in excruciating detail every game they’ve had this season, the stats of each player, and the state of its finances. As you’re listening to them, you become aware of a conversation over your shoulder, which is much more interesting: they’re gossiping about someone you know. You end up trying to seem like you’re listening to the mega-fan, but you’re actually tuning in to the gossipy conversation The Cocktail Party Problem demonstrates selective attention: being able to follow one conversation and not be distracted by others Image 1 Party Note. Adapted from Group of People Standing Waiting Outside The Bar, by mentatdgt, 2018, 3 https://www.pexels.com/photo/group-of-people-standing-waiting-outside-the-bar-1185440/ (link) 1. Introduction Selective Attention Experiments Image 2 Auditory stimulation Selective attention was originally studied in controlled experiments Auditory stimuli (e.g. hearing a series of words spoken out loud) were used to study selective attention as they offered a relatively more ‘pure’ reflection of how the nervous system produces attentional processes compared to other types of stimuli, like visual stimuli With auditory stimulation there’s no need for mechanical movements such as eye/head movements Note. Adapted from Photo Of Man Using Headphones, by Andrea Piacquadio, 2018, https://www.pexels.com/photo/photo-of-man-using- headphones-3758102/ (link) 4  1. Introduction Bottleneck Theories Our understanding today of how selective attention works is built upon the early models of selective attention All models described selective attention as a single, limited capacity channel that acted like a ‘bottle-neck’: before the bottle-neck, we are able to process many stimuli in parallel although without paying them much attention, and after the bottle-neck, we are only capable of processing one thing in the focus of attention These early models shaped the early-late debate: researchers argued “when” in attentional processing the bottle-neck occurs Image 3 Bottle Note. Adapted from Corona Beer Bottle Across Sands, by Marianna OLE, 2018, 5 https://www.pexels.com/photo/corona-beer-bottle-across-sands-1727829/ (link) 2. Dichotic Listening Tasks Selective and Limited Nature of Attention Image 4 Dichotic Listening Tasks are widely used to test Headphones selective attention models In these tasks, two different messages are presented to the subject via headphones: one message in each headphone. When studying selective attention, subjects are normally instructed to pay attention to just one of the messages and ignore the other These studies have taught us the selective and limited nature of attention: we select just one of the messages to process at a time, and our Note. Adapted from Portrait..., by Andrea Piacquadio, 2019, capacity to attend to stimuli is limited https://www.pexels.com/photo/portrait-photo-of-smiling- woman-in-black-top-and-glasses-wearing-white-headphones- 3765147/ (link) 6 2. Dichotic Listening Tasks Shadowing Task The Shadowing Task (Cherry, 1953) is a dichotic listening task in which the subject is presented two messages simultaneously (one in each headphone) and must attend to only one of the messages (in the shadowed ear) and repeat it out loud as it is being heard Subjects are later asked to recall the content of both messages Image 5 Shadowing Task Note. Adapted from Figure 4.8, by Sternberg & Sternberg, 2012 7 2. Dichotic Listening Tasks Shadowing Task In some of the experiments, the experimenters altered the message in the unattended ear during a trial. Subjects only noticed certain changes: When asked about what they could recall of the unattended message, they said they noticed when the voice changed from male to female, but not when the message changed from English to German This suggests that when ignoring stimuli, only sensory information (i.e. physical characteristics) reaches our attention, not semantic information (i.e. the meaning of the message) Image 5 Shadowing Task Note. Adapted from Figure 4.8, by Sternberg & Sternberg, 2012 8 2. Dichotic Listening Tasks Split Span Task Image 4 Pair 1: 7 & 9 Headphones Pair 2: 1 & 3 The Split Span Task (Broadbent, 1954) is a Pair 3: 2 & 4 dichotic listening task in which the subject is presented a sequence of 3 pairs of digits, one of each pair in each headphone. 7 9 Subjects must try and attend to both digits 1 3 of each pair. 2 4 After the presentation of all 3 pairs, the subject must recall as many digits as possible Note. Adapted from Portrait..., by Andrea Piacquadio, 2019, https://www.pexels.com/photo/portrait-photo-of-smiling-woman-in- black-top-and-glasses-wearing-white-headphones-3765147/ (link) 9 2. Dichotic Listening Tasks Split Span Task Image 4 Pair 1: 7 & 9 Headphones Pair 2: 1 & 3 Broadbent found that when subjects were Pair 3: 2 & 4 able to correctly recall all the digits they heard, they tended to recall the 3 digits they heard in one ear, then the 3 digits they 7 9 heard in the other ear (i.e. they did not recall 1 3 the digits in pairs) 2 4 Broadbent’s conclusion - When presented with multiple stimuli, we selectively attend to one channel and then to the other channel: you first attend to one ear, then Note. Adapted from Portrait..., by Andrea Piacquadio, 2019, you attend to the other, not both at the https://www.pexels.com/photo/portrait-photo-of-smiling-woman-in- black-top-and-glasses-wearing-white-headphones-3765147/ (link) same time 10 3. Models of Selective Attention Sensory Processing and Selective Filter Image 6 Colander/filter These early experiments gave rise to the idea of attention as a cognitive mechanism of limited capacity: we cannot pay attention to everything Early researchers proposed that in attentional processes there is early sensory processing (detecting sensory characteristics) of multiple stimuli in which we are only capable of detecting their sensory characteristics (e.g. is the voice male or female) After early sensory processing, the information then passes through a selective filter, after which only serial processing is possible: our attention is limited to processing one thing at a time Note. Adapted from Person squeezing cottage cheese in colander, by Los Muertos Crew, 2021, https://www.pexels.com/photo/person- squeezing-cottage-cheese-in-colander- 8064393/ (link) 11 3. Models of Selective Attention Bottleneck and Filter Theories Image 7 Multitasking This concept of attention gave rise to bottle-neck and filter theories of attention (early and late filter theories) Concept of a bottleneck: we initially process multiple stimuli in parallel, and then process individual stimuli serially and more slowly Concept of a filter: not all information reaches the spotlight of our attention These theories explain the ability to seemingly pay attention simultaneously to multiple tasks (i.e. multi- Note. Adapted from Crop..., by Ketut Subiyanto, 2020, tasking) as rapidly switching between two tasks, https://www.pexels.com/photo/crop-faceless-female-entrepreneur- doing-multitasking-work-on-different-devices-4474033/ (link) rather than doing two things at a time 12 3. Models of Selective Attention Broadbent’s (1958) Model Broadbent’s model predicts that in a dichotic listening task, sensory information (e.g. auditory stimuli) is detected at both ears, even though we are trying to ignore stimuli heard in one ear The detection of sensory information in multiple channels (i.e. both ears) occurs in parallel (at the same time) This sensory information then reaches an attentional filter: the sensory information detected at one of the channels reaches short-term memory and becomes conscious Image 8 Broadbent’s Model Note. Adapted from Figure 5.1, by Eysenck & Keane, 2020. 13 3. Models of Selective Attention Broadbent’s (1958) Model The sensory information detected in the ear we are trying to ignore is filtered out before it reaches conscious short-term memory A small amount of unattended stimuli may receive low level sensory processing: subjects in Cherry’s (1953) study were able to notice when the unattended voice changed from female to male This suggests the attentional filter can only select one channel or input at a time: we cannot process two things at once. This has an adaptive function as it prevents us from being overwhelmed by large quantities of information Image 8 Broadbent’s Model Note. Adapted from Figure 5.1, by Eysenck & Keane, 2020. 14 3. Models of Selective Attention Broadbent’s (1958) Model Image 9 Speaker According to Broadbent’s model, unattended information is filtered out before it reaches conscious short-term memory, suggesting that information is selected based on its sensory/physical properties, not on its meaning This is bottom-up processing – attention is guided by the characteristics of the stimulus Note. Adapted from Closed..., by Anthony, Year, https://www.pexels.com/photo/closed-up- photography-of-brown-wooden-framed-sony-speaker-157534/ (link) 15 3. Models of Selective Attention Broadbent’s (1958) Model However, evidence soon began to emerge that Image 10 Meaningful Information Broadbent’s Model had difficulty explaining certain effects: Moray (1959) found that when the unattended message in a dichotic listening task contained meaningful content (e.g. the subject’s name), subjects often remembered the content of the unattended message This suggests that messages of high importance must somehow break through the selective filter: information doesn’t pass through the filter based only on sensory properties, but also based on Note. Adapted from Students..., by Keira Burton, 2020, https://www.pexels.com/photo/multiracial- their semantic content students-gossiping-about-black-man-with-notepad-6147394/ (link) 16 3. Models of Selective Attention Treisman’s Attenuation Model (1960) Image 4 Headphones Treisman’s Attenuation Model (1960) of selective attention aimed to explain how some unattended information gets through the filter Using a shadowing experiment, Treisman (1960) found evidence that unattended information may pass through the filter, especially is if there is a meaningful relation between what is currently being attended to and what is not being attended to Note. Adapted from Portrait..., by Andrea Piacquadio, 2019, https://www.pexels.com/photo/portrait-photo-of-smiling-woman-in- black-top-and-glasses-wearing-white-headphones-3765147/ (link) 17 3. Models of Selective Attention Treisman’s (1960) Shadowing Experiment Image 4 Headphones In Treisman’s (1960) Shadowing experiment, subjects were told to shadow story 1 that they Unattended Attended heard in the left ear, and ignore story 2 in the ear ear right ear Half-way through story 1, story 1 replaced story Story 2 Story 1 2 in the right ear, and the new story 3 replaced story 1 in the left ear (all stories were told by the Story 1 Story 3 same actor) If Broadbent’s model is true, participants should filter out almost all information from the unattended ear: they should attend to the first Note. Adapted from Portrait..., by Andrea Piacquadio, 2019, https://www.pexels.com/photo/portrait-photo-of-smiling-woman-in- part of story 1, then attend to the new story 3 black-top-and-glasses-wearing-white-headphones-3765147/ (link) 18 3. Models of Selective Attention Treisman’s (1960) Shadowing Experiment Image 4 Headphones However, results showed that when story 1 switched from the attended ear to the Unattended Attended unattended ear, subjects carried on shadowing a ear ear few lines of story 1 After a while, subjects select the attended ear Story 2 Story 1 again and shadowed story 3 Information being presented in the unattended Story 1 Story 3 ear seemingly passed through the selective filter, which suggests that the semantic content of the unattended information is being analysed rather than being filtered out (given that story 1 Note. Adapted from Portrait..., by Andrea Piacquadio, 2019, https://www.pexels.com/photo/portrait-photo-of-smiling-woman-in- and story 1 are semantically related) black-top-and-glasses-wearing-white-headphones-3765147/ (link) 19 3. Models of Selective Attention Treisman’s Attenuation Model (1960) Treisman argued that the selective filter doesn’t block out stimuli, but weakens (attenuates) the strength of the unattended information. This means that both attended and unattended information are processed but to different degrees: Unattended information receives low level processing and attended information receives high level processing. Therefore, unattended information is also able to reach consciousness, especially if it is particularly important Image 11 Treisman’s Model Note. Adapted from Figure 5.1, by Eysenck & Keane, 2020. 20 3. Models of Selective Attention Treisman’s (1960) Shadowing Experiment Image 4 Headphones Unattended Attended Treisman’s model allows for more top-down ear ear control than Broadbent’s model: How much processing an input receives depends on the Story 2 Story 1 expectations of the subject E.g. In Treisman’s (1960) shadowing experiment, Story 1 Story 3 subjects were expecting to continue listening to Story 1, which influenced how intensely the unattended information was processed (i.e. the unattended information received a greater level Note. Adapted from Portrait..., by Andrea Piacquadio, 2019, https://www.pexels.com/photo/portrait-photo-of-smiling-woman-in- of processing) black-top-and-glasses-wearing-white-headphones-3765147/ (link) 21 3. Models of Selective Attention Treisman’s (1960) Shadowing Experiment Image 4 Headphones Unattended Attended More top-down control also implies that we can ear ear selectively attend to stimuli in a flexible way, according to our goals Story 2 Story 1 For instance, despite the fact the thing we are trying to ignore is relatively important to us (e.g. Story 1 Story 3 hearing our name, or being in a familiar language), we can use top-down control to ignore the stimulus and selectively attend to another stimulus according to our objective Note. Adapted from Portrait..., by Andrea Piacquadio, 2019, https://www.pexels.com/photo/portrait-photo-of-smiling-woman-in- black-top-and-glasses-wearing-white-headphones-3765147/ (link) 22 3. Models of Selective Attention Late Filter Model (Deutsch & Deutsch, 1963) Deutsch & Deutsch’s (1963) model argues that when trying to selectively pay attention to something, information is filtered out, although this filter occurs later than in other models Whereas other models argue that stimuli are selected based on their sensory characteristics, and to a lesser extent, their semantic characteristics, Deutsch & Deutsch argued that all the information that we detect receives sensory and semantic processing before it is selected/filtered out The most important/relevant stimulus passes through the filter and reaches short-term memory Image 12 Deutsch and Deutsch’s Model Note. Adapted from Figure 5.1, by Eysenck & Keane, 2020. 23 3. Models of Selective Attention Norman’s (1968) Addition to the Late Filter Model Norman (1968) built on the Late Filter Model by Deutsch & Deutsch (1963) and Treisman’s Attenuation Model (1960) by adding the role of long-term memory: Like the Late Filter Model, all sensory inputs are initially processed During this initial analysis, the semantic properties of the inputs are compared with long-term memory (LTM) stores E.g. if a nonsense word is heard, no match would be found in LTM, whereas a meaningful word would match a LTM store Input is compared with LTM Image 12 Deutsch and Deutsch’s Model Note. Adapted from Figure 5.1, by Eysenck & Keane, 2020. 24 3. Models of Selective Attention Norman’s (1968) Addition to the Late Filter Model After this semantic analysis, the inputs are assigned importance according to their semantic content and according to the context (e.g. expectations of the subject) If the semantic content matches what the subject expects to hear, greater importance is given to the input: This also explains the findings of Treisman’s (1960) multiple stories experiment Therefore, the most important input passes through the filter, enters short-term memory and becomes conscious Input is compared with LTM Image 12 Deutsch and Deutsch’s Model Note. Adapted from Figure 5.1, by Eysenck & Keane, 2020. 25 4. Comparing Models of Selective Attention Automatic Processing Image 13 Multiple inputs Surely it’s uneconomical to process all incoming information and then apply a filter? Late filter models assume that all processing that occurs before the filter is automatic, and therefore does not use up the limited capacity of attention All processing that occurs after the filter is more controlled and conscious, and consumes attentional resources Note. Adapted from An Illustration of a Person's Mind, by Tara Winstead, 2021, https://www.pexels.com/photo/an- illustration-of-a-person-s-mind-8378740/ (link) 26  4. Comparing Models of Selective Attention Preattentive vs. Attentive Processing This is reflected by Neisser’s (1967) two-step perspective of attention: Attention involves two types of processing: preattentive and attentive Preattentive processes comprise the initial, automatic processing of many sensory inputs to analyse the physical characteristics of the stimulus Attentive processes are much more controlled, in that they occur later in time, are slower, and consume our attentional resources (i.e. we can only pay attention to a limited amount of things at one time) Note. Adapted from Photo of Woman Lying on Bed While Using Laptop, by Elijah Image 14 O'Donnell, 2020, https://www.pexels.com/photo/photo-of-woman-lying-on-bed- 27 Controlled attention while-using-laptop-4066041/ (link) 4. Comparing Models of Selective Attention Which Model Has the Most Compelling Evidence? Both Treisman’s and Deutsch & Deutsch’s models are supported by the same evidence, but Treisman argued that the Attenuation Model was a better fit to some key findings than the Late Filter Model: Moray (1959) found that when the unattended message contained a meaningful word, participants would often remember the content of the unattended message, providing support for both Treisman’s and Deutsch & Deutsch’s models However, Treisman (1960) found that this kind of effect only happened on 6% of trials: on 94% of trials participants did not remember the unattended message Therefore, although this evidence suggests that the unattended message is not completely filtered out (as in Broadbent’s model), it’s more likely that the unattended message was weakened rather than receiving just as much semantic processing as the attended message (as in Deutsch & Deutsch’s model) 28 Summary Broadbent’s theory predicts little or no processing of unattended auditory information Treisman’s theory predicts flexibility in the processing of unattended information (it is processed, although less intensely that the stimuli we’re trying to pay attention to) Deutsch & Deutsch’s theory predicts that both attended and unattended information receive equal processing before any information is filtered out Common element between the models: selective attention results in our attention being focused on a limited amount of information – like a bottleneck (regardless of whether the filter is early or late...) 29 BIBLIOGRAPHIC REFERENCES Broadbent, D. E. (1958). Perception and communication. Oxford, UK: Pergamon. Cherry, E. C. (1953). Some experiments on the recognition of speech with one and two ears. Journal of the Acoustical Society of America, 25, 975-979. Deutsch, J. A., & Deutsch, D. (1963). Attention: Some theoretical considerations. Psychological Review, 70, 80–90. Eysenck, M. W., & Keane, M. T. (2020). Cognitive psychology: A student’s handbook (8th ed.). Psychology Press. Moray, N. (1959). Attention in dichotic listening: Affective cues and the influence of instructions. Quarterly Journal of Experimental Psychology, 11, 56–60. Neisser, U. (1967). Cognitive psychology. New York: Appleton-Century-Crofts. Sternberg, R. J., & Sternberg, K. (2012). Cognitive Psychology (6th ed.). Cengage. Styles, E. A. (2006). The Psychology of Attention (2nd ed.). Psychology Press. Treisman, A. M. (1960). Contextual cues in selective listening. Quarterly Journal of Experimental Psychology, 12, 242–248. 30  Dr Harry Moore [email protected] UCAM Universidad Católica de Murcia © UCAM © UCAM