FINAL ATENCION

CHAPTER 1: DEFINITION What is attention? - James (1890) told us that: “Everyone knows what attention is. It is the taking possession of mind in clear and vivid form . . . it implies withdrawal from some things in order to deal effectively with others.” - Shiffrin (1988) provides a more precise definition: “Attention has been used to refer to all those aspects of human cognition that the subject can control . . . and to all aspects of cognition having to do with limited resources or capacity, and methods of dealing with such constraints.” - Desimone and Duncan (1995) capture the properties of visual attention, and say that: “The first basic phenomenon is limited capacity for processing information. At any given time only a small amount of the information available on the retina can be processed and used.” - Agreement on limited capacity for processing info and allocation can be intentionally controlled Varieties of attention 1- Attention as a selective process: some aspects (spatial, local, global, object...) are chosen to be attended ("mental set") and the rest are ignored à Capacidad de un organismo para concentrarse en una sola Fuente de información y en la realizacion de una tarea única. 2- Attention as maintaining processes: there appears to be an interrupt process that automatically detects novel, possibly important environmental changes outside the current focus. Anyway, attention can be directed back to the original task. à Tiempo que una persona es capaz de concentrarse en un estímulo simple o una representación antes de que empiece a cometer errores. 3- Attention as distribution of resources: Visual and auditory attentions seem limited, we are not able to monitor many sources of information (but we can combine info from different senses). It's as if we allocate or withdraw attention according to the demands of the task. Thus, it seems that there is a limited amount of resources or a mechanism that selects only some info for further processing. à Mecanismo que una persona pone en funcionamiento para atender simultáneamente varias demandas del ambiente. Distribuimos recursos de atención a cada actividad. - External attention: dirigida a estímulos del exterior - Internal attention: dirigida a procesos mentales/estímulos de recepción interna - Automatic (involuntaria) vs. controlled (voluntaria): "slips of action" often happen when not paying attention (attention as control of action), since the more skilled we are in an action, the less control is needed and the less attention is paid. - Attention and memory: explicit (conscious) vs. implicit (unconscious) memory. Some processes can take place without conscious attention, but in fact "unattended" stimuli might have an effect on subsequent responses. - Attention: [selective, sustained, divided] - [external, internal] – [visual, auditory, tactile] - [overt, cover] - [exogenous, endogenous] - [local, global] How do we measure attention? -Physiological and motor NS responses:

EEG (Electroencephalography): exploracion neurofisiologica que se basa en el registro de la act cerebral en condiciones basales (sueño,reposo) y durante activaciones mediante un equipo de electroencefalografía

ERP (Event Related Potentials): mide la respuesta cerebral del cerebro es el resultado directo de un evento específico sensorial.

GSR and EDR (galvanic/electrodermal response): miedo, ira, sentimiento sexual generan cambios en la resistencia eléctrica de la piel. No ofrece información sobre la dirección o valencia de la emoción (+)/ (-). (Orienting and habituation)

Heart rate: (arousal, effort) EMG (electromyography): actividad eléctrica causada por los músculos. Poderoso indicador de valencia, reacción a

estímulos visuales, olfativos… (Arousal, effort) Dilation of the pupil (orienting, novelty, effort). - Motor responses: body posture (eye movements) pursuit movements (slow, maintaining) saccadic eye movements (fast, to center the fovea over relevant stimuli)

-Performance in cognitive tasks - Reaction times: the more complex the task the larger the RTs - Detection task: detect something with different backgrounds - Discrimination tasks: similar to the previous but compared with something - Memory tasks: it is assumed that the more attention, the less data collected - Visual search tasks: for stimulus properties - Interference tasks: a particular aspects interferes with the target

-Subjective experience What causes attention? - Stimulus properties: size, position, color, intensity, movement, complexity, relevance, novelty - Arousal: Yerkes- Dodson 1909- optimal level of arousal is lower for complex tasks compared to simple tasks. When stressed we cannot complete complex tasks. *ADHD (Attention deficit hyperactivity disorder): distracción moderada. Periodos de atencion breve, inestabilidad emocional y conducta impulsiva. - Interests and expectations: motivations have an influence on attention. *Mental set: internal representation that matches some association (expectations)

CHAPTER 2: HISTORIC EVOLUTION OF THE STUDY OF ATTENTION Historic evolution of study of attention - Posner proposes three phases:

- 1950-1960s: human performance and "human as a single channel processor" - 1970-1980s: cognition, mostly concerned with internal representations, automatic and controlled processes and focusing and diving attention. - Mid-1980s "cognitive neuroscience", taking acount of biology, neuropsychological patients and computing

- Wundt (1832) James (1842) and Titchener (1867). Wundt and Titchener were from the German school, being Titchener pupil of Wundt. They all interested in "consciousness" part, sensations and attention. Used introspection. - Watson (1878) and Skinner (1904) behaviorism gained influence, only focus on behavior. Attention lost interest. - Anyway, some things were done: John Ridley Stroop (1897) studied interference, although he didn't adressed attention - Yerkes (1876) studied arousal - Ramón y Cajal (1852) studied neurons, mainly observing and drawing. Reticular activating system. - Pavlov (1849): described how pupil size and HR are related to attention. Orienting response Cognitive revolution 1950s: - Summary: initial studies of auditory and visual models proposed by Broadbent, Treisman, and Deutch and Deutch. Those models shaped the argument "early-late" debate, assuming a single, limited capacitygeneral purpose processing channel. Prior to the bottleneck, processing was parallel and did not require attention, but after it, is serial and requires attention. Theorists argued about where in the process the bottleneck is. - Information processing approach: metaphor to approach human mind

- Info processing: sequence of activities to give a response (symbolic representations manipulations) - Processing stages: serial, but there are also parallel processes - The processor: has limited capacity

- Welford (1952): when presenting subjects two stimuli serially, and subjects must make a speeded response to both, RT to the second stimulus depends on the SOA. When 2nd stimulus presented after short SOA- larger RT than after a longer SOA. Welford named this delay Psychological refractory period (PRP). He deduced the existence of a "bottleneck" where the processing of the 1st stimulus must be completed before processing of the 2nd can begin.

- Cognition and cognitive neuroscience: internal representations, automatic and controlled processing, focusing and divided attention. We can observe the underlying neuroanatomy, and attention as a system of several brain networks.

Human single channel processor: - Dichotic listening: two channels, attend to one or both, and remember the message or detect specific items.

- Shadowing (Cherry 1953), participants had to repeat the relevant message out loud as it arrived. They reported changes of the unattended channel in physical properties (speed, voice, masks, volume) but not complex ones (language, message...)- Broadbent interpreted it as stimuli that do not need response are discarded ASAP, before being fully processed, and physical cues are effective for distinguishing.

àHasta qué punto recuerda la información el odio desatendido. - Split span: (Broadbent 1954), numbers presented sequentially in both channels, and must recall them when sequential rapid stream has ended. Participants recalled numbers grouping them on the basis of the channel they came from - selection is ear by ear. Info of one channel is stored in the buffer when reporting the other. à Conocer la estrategia utilizada por los sujetos para recordar la información.

Filter models: - Broadbent 1958: Filter theory: new conception of mind. The human information system has limited capacity and needs to be protected from overload. That is why a filter that regulates the entrance of inputs is it thought to exist. Psychological processes as a flow of information through the nervous system. Three main conclusions:

§ We can analyze human functions in terms of the flow of information, without having to know the neural of physical basis of that transmission. Information is whatever reduces the amount of uncertainty present in a situation. Its transmission is maximal when a given stimulus always gives rise to the same response (no uncertainty between stimulus input and answer output). Redundancy is found when there is less than the maximum amount of info (low loss of uncertainty). High redundancy implies low information.

§ The whole NS as a single channel limited in the rate at which info can be transmitted. § Limited capacity would imply a selective filter, which protects system from overload, selects a portion of incoming

info, blocking the rest. There must be also a short -term buffer that precedes the filter and holds (in parallel) unselected info, for short periods of time, but if not recovered, is lost. it is an early selection model.

Senses àshort term store (both are S system: physical properties, parallel processing) à selective filter (rigid: all or nothing) à limited capacity channel (P system: serial processing, conscious, influences short term memory) à long term store of past events and probabilities (influences filter) and regulatory output system à effectors àresponse

- Implications: combining tasks that require continuous parallel processing is not possible, only when one can proceed momentarily without attention and we can switch rapidly. Non attended info is wasted before further than physical processing.

Challenges to Broadbent’s Model - Moray 1959: shadowing task, poor recognition of words in the unattended ear, but could recognize their names. Evidence that filter is not only physical. - Gray and Wedderburn 1960: repeated Broadbent's split span but mixing semantic categories (numbers and words) in the same channel. Subjects grouped them by meaning (against Broadbent's model). - Deutsch and Deutsch 1963: first late selection model. Filter nearer the response, all the info is firstly processed and then selected. Incoming signals were compared and weighted for importance. Anyway, compare all incoming signals would make decision times extremely slow. Their idea was that after fully processing, only the most important signal are responded, whenever it is above the current arousal level. Relevant: Stimuli àsensorial registeràperceptual systemàfilterà response system Irrelevant: Stimuli àsensorial registeràperceptual system - Anne Treisman 1960: shadowing task, switching channel of attended ear. Subject’s ability to shadow attended channel was affected by the content of the material of unattended channel. When switching stories participants shadowed a few words of Story 1 in unattended channel when this was switched from attended channel. (They reported some words or attended channel before focusing again on target channel). Evidence against a filter based on physical properties. He proposed that the filter was not rigid, but attenuates or reduces the strength of unattended channels. Information consistent with current expectations or personally relevant activates those words above level of consciousness.

- Sensory register filter limited capacity system long term store

Other filter models: Multiple selection, flexible filters: selection may take place at any time, early or late. Criticisms to filter's theories: difficulties generalizing to other sensory modalities such as vision. Dichotopic presentation is possible but visual reality is different.

Dual task-models: Divided attention Automatic vs. controlled processing.

Early studies on visual attention - Visual reality is not like auditory one, is spatially organized. Participants were unable to remember infor from unattended eye. - Sperling 1960: presented stimuli using a tachistoscope. Brief visual displays of sets of letters (3x4) by 50ms. Subjects only could report a few letters, despite they reported to "see" the display shortly after stimulus offset. This brief visual memory was shorter than the rate of retrieving of the items. This was evidence for a high capacity but fast decay visual memory store. He later introduced a modification, signaling only a row after the offset, row that subjects must recall. They were able to recall all. This is the Partial report superiority effect – ventaja que se consigue con la presentación de la señal-. The longer the delay offset-cue, the poorer recall (after 500ms no difference). Cue is supposed to allow a subset of items to access later stages of processing. The brief visual info store was later called iconic memory and is equivalent to Broadbent's sensory store. Demonstrates that participants processed all stimuli. - Later experiments similar to Sperling's were done using only two rows and a bar as cue (under the letter) or mixing numbers and letters cueing with the category to be reported. No superiority effect was found in the second, which suggested a physical-properties filter. - Mewhort 1967: similar to Sperling's, but two rows, one of them could be word-like. When non-cued row was word-like, target row was better recalled- iconic memory not purely pre-categorical. (El significado de los elementos inatendidos puede afectar e influir en el procesamiento de la infor atendida. Por lo tanto la memoria icónica o es un almacén exclusivamente visual, sino semántica también). Identity and location - Eriksen and Rohrbaugh 1970: bar probe experiment (two rows, bar under letter to cue). Delay of cue affected probability of correct response and the more probability to fail, the more probability to report an adjacent (to the target) item- mislocation error. Evidence that both identity and location were available in parallel, at the time of selection. Consistent with a model that implies parallel processing to the level of identification and serial selection for response. - More experiments proved independent coding for identity and location of stimuli. - Styles and Allport 1986: target letter (of a primed set of letters) between four distractors. Distractors could be groups of lines, letter-like symbols, digits (meaning but different category), letters from a different target set, or letters from the target set. They had to write down the target and the distractors (if letters) in their correct positions. With shorter SOAs, in the last condition, they identified correctly the letters, not the positions- evidence of independent representation of identity and location. Interference was mostly due to category of distractors rather than physical properties. -Yerkes-Dodson (1909): arousal and complexity of the task -Welford (1952): psychological refractory period (longer SOA- shorter RTs) -Cherry (1953): shadowing. Listening two channels, repeat one out loud. Physical changes in unattended are detected, more complex don't- physical properties as criteria for choosing which info to process. -Broadbent (1954): split span. Number sequentially both ears, reported depending on the ear. Selection based on the channel. Evidence for short term buffer. -Broadbent 1958: filter theory- early selection model, selection based on physical properties. -Moray 1959: challenge to filter theory. Shadowing, subjects recognized their names in unattended channel. -Gray and Wedderburn 1960: split span mixing categories. Subjects grouped words by meaning. -Deutsch and Deutsch 1963: late selection model. Info processed before selection. -Anne Treisman 1960 and 1964: shadowing a channel, message change to the other channel, they reported some words of original message before focusing on target channel again. Against physical properties filter. -Sperling 1960: 3x4 displays of letters, 50ms, recall. Only a few - high capacity, fast decay visual memory. When signaling a row, complete recall - Partial report superiority effect. Physical properties filter. -Mewhort 1967: two rows. When not cued row is word-like, cued row better recalled. Filter not purely precategorical. - Eriksen and Rohrbaugh 1970: bar probe. The larger delay offset-cue more fails, the more probability to fail, the more probability to be a mislocation- evidence of parallel processing of identity and location - Styles and Allport 1986: target set primed, target between four distractors: lines, letter like, numbers, letters of non target set, letters of target set. Report target and distractors (when letter) in position. When short, SOA most errors were mislocations (in last condition) - evidence of independent representation of identity and position.

CHAPTER 3: VISUAL SELECTIVE ATTENTION: - Summary: it focuses on the nature, control and direction of visual attention. It reviews the evidence for a spotlight of visual attention, Posner's work and other effects. Does the focus moves like a zoom lens rather than a spatial spotlight? Is it directed to spatial locations or to the objects in those? How are objects coded? - Helmholtz noted that attention and fixation are not necessarily the same. One of the first metaphors proposed by James was the spotlight metaphor, with a focus, a margin and a fringe (periferia). Only what is inside the focus is processed. Other views see it as a lens with a variable zoom. - Two broad categories of visual attention tasks (differences in both can explain conflicting evidence):

- Selective filtering: most early experiments. Participants had to select a stimulus from a complex environment and a response from a wide choice. Dependent measure: accuracy - Selective set: 1970s- Stimulus set was usually small, simple stimuli and small set of responses. Usual measure: RT

Selective set experiments – Visual selective attention - Eriksen paradigm – Flanker task-: Eriksen and Eriksen 1974, They presented a letter α, flanked by two or four letters β; If the letter in the middle corresponded to a determined set of letters, subjects had to move a lever to the right, if corresponded to other set, to the left. They had to respond as fast as possible. Slower RTs were found in the incompatible (different set) condition compared to the compatible one (Flanker compatibility effect- (FCE)). Since both flankers and target competed for the response - evidence that irrelevant stimuli are also available for the processing system. They also discovered that the distance between the distractors and the target affected the results. Within one degree of visual angle they could not be ignored (notion of a minimum width spotlight), further, interference disappeared. S,C: left / K,G: right à - Feature specific theory of interference: Bjork and Murray 1977, completely opposite results, same task. They found that best inhibitor for B was another B. They explained it saying that we try to suppress the flankers, and when they are the same as the target, an extra effort is needed to avoid suppressing the target too - evidence of an early feature selection* - (like Broadbent's) against the idea of an earlier processing of the meaning. * Interference due to feature-specific competition btw signal and noise elements in a display Para ver a que se debía que los sujetos tardasen más en responder a una letra cuando esta estaba rodeada por otras,: - Jonides and Gleitman 1972: in order to isolate these factors, presented a letter surrounded by 2, 4 or 6 items briefly. Subjects were asked to search for a letter between the letter or number distractors. They used O as a letter but also as Zero, so featural properties were the same, but the category varied. Subjects were asked to search for an O or a zero. Slower RTs were found in the same category search- evidence that items are categorized before selection. Si el efecto estuviera exclusivamente en función de las características, el carácter ambiguo “O” debería haber funcionado igual en ambas condiciones. - Miller 1991 manipulated five factors that thought might influence processing of unattended stimuli:

- Poor spatial resolution of attentional focus - Inability to hold attentional focus on a fixed location - Inability to focus completely on an empty display location - Inability to filter stimuli that onset during the task - Inability to prevent analysis of all stimuli when there is insufficient demand of the attended stimuli

- He was unable to eliminate the FC, so concluded that early selection rarely completely excludes unattended stimuli from semantic analysis. Later reviews and works suggested an inhibitory field surrounding the attended area, which varied with the demand of the task. Perceptual load and selective attention - Due to the differences between selective filtering and selective sets, different perceptual processing is probably required. - Lavie et al. 1995: proposed that the degree of processing achieved by visual stimuli depend on many factors, and therefore a late/early distinction might be inappropriate. They proposed that whether selection is early or late depends on the task demands. When selecting the target does not require full use of attentional resources, there will be interference; if selection requires full use of attentional resources, unattended stimuli will not be processed. Attention is seen as a limited resource rather than a structural limitation (like Broadbent's). To test this, Lavie conducted two experiments. The first one compared the size of possible targets (one to six). Only the low load condition presented significant interference. The second experiment demands varied requiring two kinds of processing. Depending on the color and shape of a cue nest to the target, subjects had to respond or not (go-no go). The low load condition had to respond if the cue was blue, not if red. The high load condition had to respond when the cue was a red circle or a blue square, and not with a blue circle or a red square. Interference was only found in low load conditions. - Kahneman and Chajczyk 1983 also found that Stroop effect was diluted when more info was presented in the visual array. - Lavie 2004: Load Theory of Attention (LTA). Level and type of load involved in the processing of a task also influences to which degree distractors are processed. Frontal lobes are in charge of working memory and dual task coordination, Lavie suggested that if frontal processes are loaded with a list of numbers to remember (for instance), performance in a selective attention task would suffer. An experiment asked subjects the profession of a name along with a distracting picture, while remembering a list of numbers. When the working memory load was high, high was the interference of the incongruent face. They also made experiments switching from a selective attention flanker and a memory task. Perceptual load was manipulated increasing the size of the distractors, while working memory load was manipulated by increasing the number of items to memorize. High perceptual load decreased distractor effects, while memory load increased them. They proposed two mechanisms:

1- (Passive) Perceptual mechanisms that allows distractors to be excluded in high perceptual load conditions 2- Active mechanisms that control behavior according to task priorities, excluding irrelevant stimuli even when they are seen during low perceptual load.

Spatial cueing - Posner 1980: subjects have to detect in which side the target appears. A cue appears before each trial signaling the following position of the target (valid) or the opposite one (invalid). The cue can be central or exogenous, in the fixation point, and requires certain interpretation (arrow or gaze); it can also be peripheral or endogenous, when a signaling cue (light for instance) appears on the possible location of the target (peripheral to the fixation). Lower RTs for valid trials.

Proportion of valid/invalid trials also matter, when only a small proportion is valid, central cues are better ignored, but central cues aren't. Anyway, peripheral cues still will have an impact. - Posner also discovered the Inhibition of return (IOR). Since it takes time to move the spotlight through the visual field, if after a peripheral cue there is a delay of more than 300ms, target detection at that position is slowed down. Once that target location is in our focus, and targets seems to not appear, it's tagged and taken as an invalid trial, so attention moves to the other side. Longer RTs for valid trials. Overt and covert orienting - Overt: moving our focus to a position - Covert: move our attention without moving the eyes' focus. - Jonides 1981 showed that a secondary memory task affected voluntary orienting to central cues, but not peripheral covert orienting. He thought it reflected two different systems. Other explanations propose that exogenous (peripheral) orienting is reflexive (but not completely), while endogenous (central) orienting is voluntary. Both present different time courses for costs and benefits.   - Eriksen and Eriksen 1974: Flanker task. One target letter with two or four distractors besides. If the target corresponds to a set, move a lever to one side, if corresponds to other set, to the opposite side- Response compatibility effect (same set- faster RTs and vice versa). Also separated the flankers and the target. One angle of visual degree is the width of the spotlight. - Bjork and Murray 1977: flanker task, but best distractor was the same target- evidence of early feature selection Jonides and Gleitman 1972: flanker task with distractors as letters or digits. Used target O as letter and as zero. Slower RTs found for same-category search, despite being the same symbol - evidence for categorizing before selection - Lavie et al. 1995: if selection demands a lot of attentional resources, unattended stimuli won't be processed. Two experiments: first, varied the number of possible targets (1-6) second, combined with a go- no go task, signaled by a cue next to the target (only color/ combination of color and shape). Interference only found in low load conditions. - Kahneman and Chajczyk 1983: stroop effect reduced when more info presented (higher load) - Lavie 2004: LTA, two experiments. One memorized numbers while identifying the profession of a name, distracted by an incongruent face. The second involves switching between a memory task and a selective attention task. in both cases memory load increased distraction and perceptual load decreased it. - Posner 1980: cueing the presentation of a target in one of two sides with an exogenous or endogenous cue. Shorter RTs for valid trials, but IOR can appear with delay of target- shorter RTs for invalid trials. - Jonide 1981: exogenous peripheral cues unaffected by secondary memory task, while endogenous central cues are affected. Combining the attributes of objects (CHAPTER 4:– LIBRO STYLES) - Summary: evidence that the brain encodes different attributes of stimuli (color, identity, location...) in parallel. how are these codes accurately combined? Terisman's FIT. Is visual attention object-based or space based? - There is experimental evidence that the visual systems analyses different dimensions of the visual environment. How is all combined is the "binding problem". Plus, objects are perceived with several other objects around, from which must be segregated. - The visual pathway has multiple modules and pathways to analyze those different dimensions. We normally perceive objects with several other objects around and they need to be segregated. Neurophysiological evidence for independent codes - Shallice and Warrington 1977: two patients with tumors in left parietal lobe where testing at reading. They could read perfectly well, but were very inaccurate when asked to report a single letter from a specified position in a word. The errors were mislocations (letter identities were available, but not properly combined with their location). When more than one words where presented, "migration errors" happened (both words mixed, but letters kept their word location). Difficulties in making a within category discrimination based on location. - Examination of abilities of neurophysiological patients with localized brain damage is consistent with specialization of function.

- Achromatopsia: selective loss of color vision, but still can perceive shape and motion - Akinetopsia: loss of ability to perceive motion

- Modern neuroimaging techniques also support the idea of different areas activated when response requires different features like color, shape or motion in the same display. - Dorsal pathway- where / responds to orientation - Ventral pathway- what /Specialized in objects identity, more detailed. Putting it all together - Colheart 1980: identity of an item is early stored during the display. It's relatively stable and decays slower than the physical attributes (color or location). Physical (=episodic) info is unstable and decay rapidly unless further processing. This model poses the existence of a "lexical monitor", which coordinates the semantic and episodic info of an item. Unless both are established together, identity info dies and cannot be reported, but has accessed, which means that still can be a residual unconscious activation that facilitates or interfere semantic activation (priming). - Treisman and Gelade 1980: Feature integration theory (FIT), in continuous evolution. Main assumptions:

- Sensory features such as color, size, and orientation are coded automatically, pre-attentively, in parallel level - Features are coded by different "modules" that work with "maps" to represent such features - Detection of single features takes place pre-attentively, in parallel

When need to detect a particular mix of properties, the separate features must be combined in a conjunction, through three ways:

- Features may fit into predicted objects "frames" according to the stored knowledge - Attention may select within a "master map" of locations that contain all features. Attention allows the retrieval of features currently active and creates a temporary representation in an "object file", that will later be compared to stored knowledge for recognition - Without attention, features may conjoin on their own, sometimes correctly or wrong (illusory conjunctions)

Evidence to support FIT: - Tresiman and Gelade showed that when subjects search a target defined by a unique feature, search time is independent of the number of distractors, it draws attention automatically (pop out effect, consistent with pre-attentive processing). On the other hand, with targets defined by a conjunction of properties, search time increases linearly with the number of distractors. In those cases, targets are searched serially across the display, with focal attention moving.

- Duncan and Humphreys 1989: Attentional engagement theory. Beck 1966 showed that finding a T between Ls (share basic features) was harder than detecting a T between rotated Ts (not share physical features) (the more similar distractors are the more difficult for selective mechanism to segregate the display). Duncan presented a series of Ts and an L, features are the same, while their conjunction is different. The distractors can be heterogeneous (different positions e.g.) or homogeneous (all equal). FIT said that featural properties were pre-attentively, while whether features are arranged to make an L or a T would require serial search. Nevertheless, when distractors were homogeneous, absent responses were faster than present responses (display size effects reduced or absent). They called this selection at the level of the whole display, and suggested that visual search for the target is in this case based on a rapid rejection of the distractor group. Efficiency on target search depends not only o the distractor/target differences, but on similarity between distractors. The computer model SERR models this theory. (*SERR: computer model for attentional engagement theory) A neurophysiological explanation of the binding problem - Crick and Koch 1990 suggested that could be explained by the synchronization of activity over currently active neurons. - Singer 1994 suggested that cell assemblies of individual cells contribute at different times to different representations. Thus, significance of an individual neuronal response will depend on context within which is active. Three requirements to represent objects in assemblies:

- Response of individual cells must be probed for meaningful relations. - Cells that can be related must be organized in an assembly - Once the assembly is formed, its members must remain distinguishable from members of other assemblies

- Shallice and Warrington 1977: patients with tumor in left parietal lobe reading task. Mislocation errors within words. When more than one word- migration errors - Colheart 1980: identity and physical or episodic memory, bond by a "lexical monitor". The first is rapidly set, and slowly decays, the second decays rapidly, unless selected for further processing. - Treisman and Gelade 1980: FIT - Duncan and Humphreys 1989: Attentional engagement theory - Crick and Koch 1990: binding explained by synchronization of currently active neurons - Singer 1994: different cells assembly and contributes to different representations depending on the context

CHAPTER 4: DIVIDED ATTENTION  

(CHAPTER 6: LIBRO STYLES) - Summary: How is attention divided when combining tasks? Resource theory and importance of stimulus response compatibility between tasks. Although in same cases task can combine when they do not compete for the same input/output subsystem, it seems to be a limit at the final stage of processing, when responses are selected. Attentional blink and change blindness are related to the diversion of attention. Background - Welford 1952 and Broadbent 1958 suggested that there is a central bottleneck in processing that limits dual task performance. (Reminder- Welford 1952: experiments over the PRP manipulating SOA/ Broadbent 1958: filter theory, single processing channel preceded by a filter, dual tasks only possible when rapid switch of the filter). Other authors have argued that the bottleneck is not because of the structure of the processing system, but is due to a limited processing "capacity". - Allport et al. 1972 had expert piano players reading pieces that they had not seen before, while shadowing a prose at a rate of 150 words per minute. With only little practice participants were able to perform both tasks combined as fast an accurate as separately. Since two complex task demanding continuous attentional processing could be combined without loss of speed or accuracy it was thought as evidence against the single channel theories. - Broadbent argued that some decrements could be detected, and redundancy in the stimuli could allow attention to shift with some certainty of the next word or note. Capacity theories: - Some theories assume a general capacity limit or resources, others suggest different kinds of resources depending on the task - If the attentional resource or capacity that limits the system is of general purpose, all task that require attention will draw on the same resource. GPLCP: General Purpose Limited Capacity central Processor, if it is engaged in one operation like shadowing, it's not available for other operation.

- But if there are different varieties of resources dedicated to process different types of information, there may be specific limits for each variety. If two task do not compete for the same result, they should not interfere. - Kahneman's theory of attention and effort 1973: he proposes a model of mind rather than a flow of info through a structural system. Attention is seen as a limited power supply, but increased effort can improve performance. The amount of attentional capacity (supply) can vary according to motivation. The amount of effort available is directly related to the overall arousal, as it increases or decreases, so does attention capacity.

- Stimuli- arousal (available, is a limited resource/ capacity, depends on the levels of arousal) evaluative decision process (decides how much energy devoted to each activity) is influenced by 1- enduring dispositions (memories and past experiences) and 2- momentary intentions - "energy" distributed between the possible activities - responses - Criticisms: first, not a perfect linear relationship, performance only improves with arousal up to a certain point, then decreases (Yerkes and Dodson 1908). Second, defining arousal is problematic. Third, complexity of the task also has an influence, but how to measure complexity? Fourth, if we use interference as index of difficulty, what is the independent measure?

- Returning to the GPLCP, all tasks that require attention will draw in the same resource of compete at the same bottleneck. If the GPLCP is engaged in some mental operation, it is not available for another operation, and the same for the response, if one has been selected, no other one can be emitted until GPLCP is free. But if there are different types of resources that process different types of info, there is no reason for them to interfere unless there is competition at common levels. - Bourke, Duncan and Nimmo-Smith 1996 tested the central capacity theory, having subjects performing dual tasks, giving priority to one of them: random number generation, prototype learning, manual task and tone detection (combinations of the four). Despite being so different, all interfered with each other. But number production the most and tone detection the least. Both showed that property when being the one of priority and when not - evidence for a general limiting factor, don't specify whether is a general pool of processing resources, a limit on central executive processes, or a general limit in info processing. Data limited vs. resource limited - Norman and Bobrow 1975: they came up with the performance resource function (PRF). It represents performance and resources for a single task.

- When performance does not improve no matter how much resources are invested, performance is data limited. Data limitations also arise when data input is of poor quality and in memory. (Listen to a Chinese class) - When more resources invested produce similar changes in performance, it is resource limited. (Listening in class rather than chatting)

When two tasks are combined, resources must be allocated between both, depending on the priorities we set. Performance Operating Characteristics (POC) is a plot of performance on one task against performance in the other. The curve generated represents the change in performance level in task A when the same is changed in Task B. Using it we can see how independent two tasks are. If both are resource limited, there will be a complementary relationship between both. Different patterns may indicate that they don't share results or they're not resource limited.

- Criticisms: again, no way to assess independently the resource demands of the tasks and whether these resources are from the same pool or not.

Dual task performance: How many resources? - Posner and Boies 1971: subjects had to do two tasks 1- letter matching, warning signal followed by letter, 500 ms and another letter. Indicate pressing a key whether it was the same. 2- Also monitoring for an auditory tone (press a key with the other hand when detected). RT to the tones was equal when presented before the first letter, slower when simultaneous to any letter, slowest when presented during the interval of two letters - evidence for a general limit on attentional processing (during the time the task requires more attention, less available for tone detection). - McLeod 1977: replicated Posner's experiment but asking subjects to say "bip" when the tone was presented (instead of pressing a key). Interference disappeared. This is contrary to the idea of limited resources on attentional processing, and supports that the limit is specific to the type of response required. - Navon and Gopher 1979: Multiple resources proposes that there are a number of resources that task might require, but they only interfere as long as they compete for the same resource - McLeod and Posner 1979: they suggested that there is a special class of translation between the input and the output in dual tasks. They tested different auditory/vocal transformations combining a number of vocal/auditory tasks with visual/manual pattern matching. Four groups of subjects: three made a vocal response to the auditory task and moved a lever to left or right depending on whether the letters presented in the visual task (similar to Posner and Boies 1971) was the same or not. The fourth group (modality crossover group) responded manually to the auditory task and verbally (same/different) to the visual. G1 had to shadow the auditory stimuli ("up" and "down"), G2 heard the same but had to respond with a semantic association (high/low) and G3 heard "high" or a 400Hz tone, and had to answer "up" or "low" respectively. Auditory stimuli were given in different moments of the visual task. Two main results: when stimuli overlapped there was interference, being the G4 the group with more interference. They suggested a loop separated from the rest of the system that is involved in retrieving the programming of word production by hearing a word (prepares the system to pronounce it). In G4 the subject is prepared to make a vocal response to the visual cues, but when the auditory stimulus comes at the same time, response to the word is activated via this loop, interfering with response to visual task. G1

G2 G3 G4

Visual Task (same-different letters)

lever

lever

lever

vocal

Auditory Task (“up, down”)

shadow

Semantic assoc. “high low”

High= up 400Hz= low

lever

- Psychological Refractory Period (PRP): does PRP arises because of a central bottleneck in processing or because stimuli and responses are similar?

- Pashler 1990: two successive stimuli to which they had to respond, he manipulated similarity between stimuli and their responses. He found PRP even when they were different - evidence of a processing bottleneck Attentional blink (AB):

Rapid Visual serial Presentation (RVSP) requires monitoring of a sting of letters, and respond as soon as they see the target. Letters change at a rate of 100ms, all in the same location. The usual pattern of results is that attentional processing for items that follow the target are impaired for 200-500ms. This is called "attentional blink". It implies that if two targets are involved, when T2 is presented closely after T1 is likely to be missed. Each position is termed "lag". E.g. at the end of the string have to say whether you saw a Y or an X, followed by a C or an N. Lags 2, 3 and 4 were more often missed than the rest (200-500ms after T1). Two separate attentional processes involved:

- Monitoring the sequence - Make a decision about whether there is a Y or an X in the sequence. This process takes resources away from the first.

- There is evidence for both, shared and specific resources, but neither account explains all data - Allport et al. 1972: dual task, play new piano piece while shadowing a text. With practice both were performed as well as separated - evidence against single channel processor - Kahneman 1973: limited attentional resources, but effort and motivation can increase or decrease the available amount of it. - Yerkes and Dodson 1908: arousal and performance on simple and complex tasks. Bourke, Duncan and Nimmo-Smith 1996: subjects perform dual task with four different possible tasks, one is given priority. All interfered with each other - evidence of general limiting factor, don't know where - Norman and Bobrow 1975: PRF and POC - Posner and Boies 1971: letter matching task when monitoring for an auditory stimulus. Slower RTs when responses required at the same time- evidence for general limit on attention capacity - McLeod 1977: replication of the previous, but participants said "bip". No interference- evidence that limit is specific to the type of response required - Navon and Gopher 1979: there's only interference when two task compete for the same resource - McLeod and Posner 1979: 4 groups combining visual/manual and auditory/vocal tasks- the more interference when auditory stimulus interfered with the response to the visual task (verbal) - suggest existence of a "privileged loop" for word production - Pashler 1990: PRP manipulating stimuli similarity- interference in all groups- evidence for a bottleneck in processing

CHAPTER 5: AUTOMATIC AND CONTROLLED PROCESSING

(CHAPTER 7: LIBRO STYLES) - Summary: automaticity, skill and expertise. Two process theory and Neuman's critique. How does expertise emerge with practice? How to distinguish task that do or do not require attention? How is attentional or conscious control implemented? - Norman and Shallice 1986: it seems to be two different modes of controlling information. Controlled processing is deliberate and conscious, and can only deal with a limited amount of info. Automatic control may refer to:

- Actions carried out without awareness (walking) - Actions initiated without conscious deliberation (sipping a drink while talking) - Automatic orienting of attention - Task combining without apparent interference

- Thanks to combination of both, we can perform two tasks, as long as one of them is automatic. - Learned with practice: When tasks can be combined successfully, they seem to be controlled automatically and independently. But when mapping between the input and the output of one task are not independent form the other task, a different type of control is necessary, one that requires conscious attention and seems to have a limited capacity. With enough practice, both task can become independent. Learning to do two things at once: - Spelke, Hirst and Niesser 1976: examined effects of practice on people's ability to combine tasks. Gave two subjects 85 hours of practice during 17 weeks. They had to read stories while writing to dictation. Reading rate was highly slow and handwriting poor. After 6 weeks their reading rate had increased and hand writing improved. However, they had no memory of written words (consistent with dichotic listening), which suggest the need of attention to retrieve memories. With more practice they were able to detect rhymes and semantically related words within the dictated lists, and they even come to write down the category of the dictated word rather than the word itself at the same time as reading and fully understanding the text. Did the task became automatic? How much attention is needed for a task? Did attentional capacity increase? How much attention has increased with practice? Automatic vs. controlled processing - Atkinson and Shiffrin 1968 Model of memory: after Broadbent's model, some questions arose: how is the filter controlled? Who sets the filter? The focus is not placed not only on the structure of the system but also on how is it controlled. Selection, rehearsal and recording in short term memory required "control" processes. Storage and processing

take place in short-term "working" memory. Similar to Lavie's perceptual load: the more demanding the processing, the less capacity available for storage and vice versa, and also impairs the ability to ignore distractors. Later modifications on the concept of working memory have included storage and control aspects. For info in working memory to be "working" it need manipulation by the subject, who "controls".

Senses à sensory bufferàselective filteràshort term (working) memory - Posner and Snyder 1975. To which extent are our conscious intentions in control of the way info is processed in our minds? They made the following distinction:

- Automatic activation processes: - Occur without attention - Without any conscious awareness - Without interference with other mental activity - Cannot be prevented

- Controlled activation processes: - Conscious - Limited capacity

Conscious effort to succeed interference Not the same interference btw categories - Stroop 1935 developed the Stroop task, widely used in this field. The task is to read the words (meaning colors), that might be presented in a congruent (same color than the word) or incongruent (different) condition. RTs are slower for incongruent condition. However, if the task is reversed (name the ink) RTs are even slower, because the word meaning is processed automatically, more conscious effort seems to be needed. Naming the ink is always slower than word reading, but ability to respond to one of the conditions ignoring the other is interfered by the other. Words automatically activate their response, and although conscious control can avoid it to be made, there is a time cost. Other variant is to say how many characters are in a display, being the characters numbers, not necessarily the same as the number of items in the display. When the task involves deciding if stimuli match physically, judgments are faster for colors compared to words, and there's more interference from colors on words. - Priming: present a word or a non-word after a primed word. Subjects have to say whether it is a word or a non-word. When words were semantically related, there was a facilitation in terms of RTs. - Posner and Snyder 1975: letter matching task. Presented a priming stimulus (letter or plus sign) followed by a pair of letters. Task was to decide whether letters were the same or different, as fast as possible. The prime would activate its representation on memory, facilitating congruent trials. They proposed the Spreading Activation Account, suggesting that concepts are stored like in a network, with related concepts close to each other. When a concept is activated, it spreads its activation to the neighbors. - Schiffrin and Schneider 1977: visual search tasks manipulating the frame time, the set size and the frame size. They had 1, 2 or 4 letters as memory set (e.g. KGFM) and had to indicate whether any of them was included in the display (there was one or none). The most important manipulation was the mapping between the stimulus and the responses.

- Consistent: targets always consonants and distractors always digits [1 5 8] - Varied mapping: both memory set and distractors are a mixture of letters and digits [L 5 H] With consistent mapping, search showed to be independent of the number of items in the memory set and the display (automatic processing). With varied mapping: time increased with number of distractors and search remained serial (controlled processing). They also wanted to see whether individuals could, with practice, automate processing for a novel, arbitrary distribution. They divided consonants from B-L and Q-Z. In consistent mapping, memory set was from one of the groups of consonants, and distractors from the other. Only after 2100 trials performance began to resemble letter/digit experiment. Later, mapping between sets were reversed (distractors became targets) and performance became very poor, to gradually improve. It seems as if participants had to unlearn an automatic process.

Neuman's critique of two-process theory - Neuman 1984 summarizes this theory:

- Mode of operation: automatic processes operate without capacity and they neither suffer nor cause interference - Mode of control: automatic processes are under the control of stimulation rather than control of intentions of the person - Mode of representation: automatic processes don't necessarily give rise to conscious awareness

Criticisms: it's extremely difficult to demonstrate that a task that appears to be automatic does not require attentional capacity. E.g. Spelke et al. 1976 (write what heard, read what saw) there was a constant rule between input and output. He also suggests that interference is not always present, and depends on the task demands. Kahneman and Henik 1981separated the color and the letters in Stroop task, and interference was reduced. This is evidence that processing of the unwanted dimension will only be automatic with a constrained set of circumstances, like being related to the intended action. Automatic processing is not an invariant consequence of stimulation independent of subjects' intentions, despite sometimes it might be unavoidable. Neuman believes that automatic processing in not uncontrolled, but controlled below the level of conscious awareness. Three questions:

- Are brain processes not related to ongoing activity unaware? - Yes, unaware of contents of long-term memory and changes during forgetting - Are there processes within execution of a task that escape awareness? - Yes, like "tip-of-the-tongue" - can an action proceed without awareness? - Yes, slips of action

Neuman proposed that the difference between automatic and controlled processing is the level of control required. Practice leads to the development of skills, which are stored in the long term memory as parameter specifications (action to perform, selection of one aspect of the stimulus...). A process is automatic if parameters are specified by a skill in conjunction with input info. If this is not possible, one or more attentional mechanisms must come into action. They are responsible for interference and give rise to conscious awareness. Para Neuman la automaticidad no es una especie de proceso, sino algo que se da cuando las condiciones son las adecuadas. - Gopher 1993: proposed attention management as a skill that can be trained. (CHAPTER 8: LIBRO STYLES) Possible functions of attention - Schneider and Deubel 2002 identified two functions of attention or aspects of processing selectivity:

1- Selection for visual perception (ventral visual pathway; color, shape, category...) 2- Selection of spatial motor action: simple actions usually imply the need to select one movement target among other potential targets (dorsal pathway; spatial info for motor action) Ex. Holding an object with our hands, requires having chosen previously a target among other possible targets.

Selection for action - Allport 1987: searching for the functional and neurophysiological basis of attentional behavior, pointed that although we can encode many objects simultaneously, we can only make one action at a time. There is then a biological need for "selection of action", that designates a subset of the potentially relevant info to have control over a given effector system. (Si todas las acciones potenciales estuvieran intentando controlar simultáneamente la acción, se produciría un caos) - How are actions controlled? Neuman defined action as "sequence of movements controlled by the same internal control structure that is not a reflex". First, we have to recruit the appropriate effector. Then, select it at the right time so only one action is attempted. Other information is provided by the environment and allows us to change the course of action if needed. - How is selection for action achieved? Provided that attributes of different objects are properly integrated, the next problem is for the processing system to ensure that all possible action do not interfere with each other. This is achieved by blocking. Negative priming - Negative priming: slowing of a response when target has been previously presented as an ignored object - Tipper et al.: typical experiment, subjects are presented pairs of overlapping line drawings and had to name one of them. The target is specified by its color. The target may be repeated (attended repetition). First, it is presented in the prime trial as a distractor, and in the probe trial as a target. In those cases higher RTs are shown. The brain inhibits the ignored objects in order to extract easily the targets, depending on our intentions. Reactivating that stimulus previously blocked takes time. Distractors are actively inhibited by goal-depending mechanisms, and those goals influence which attributes are allowed to control action (e.g. color). -In order for the target to be selected, the distractor must be “actively inhibited”. -The goal of the current task influences which stimulus attributes are allowed to control action. -Inhibitory mechanisms are goal dependent and inhibition is directed to different properties of a stimulus depending on which properties are required control response and how difficult selection is. Norman and Schallice SAS: - Norman and Schallice 1986 model of willed and automatic control of behavior. They proposed that deliberate (voluntarios) attentional resources are required when tasks:

- involve planning or decision making - involve components of trouble shooting - are ill-learned or contain novel sequences - are judged to be dangerous or technically difficult - require overcoming a strong habitual response Algunas secuencias de acción que por lo general se producen automáticamente pueden llevarse a cabo bajo control consciente si es necesario, suprimiéndose las acciones involuntarias y facilitando las deseadas.

This theoretical framework centers on the idea that we have action schemata in long term memory that are awaiting the appropriate set of conditions to be triggered. Normally, the most active schema will take the control of action. There are two sources of activation according to the model:

- from the stimulus environment (bottom up) - according to the current goal(top-down) à comes from our intention

Contention scheduling – mechanism to sort out conflicting schemata by interacting inhibition and excitation. For habitual tasks there is a set of conditions processing structures and procedures that allow actions to be carried out without any need for deliberate conscious attentional control. However, when no pre-existing schema, as a novel task or when additional control is required, top-down biasing of schemata is provided by the action of a SAS (Supervisory Attentional System), mechanism to short out conflicting schemata by applying additional inhibition and excitation, changing the probability of selection by the contention scheduling mechanism. - Baddeley 1986: compared SAS to Central Executive, somehow similar to "homunculus" explanations until better explanations are found. Whether it is unitary or a number of autonomous control processes has not been discovered yet. - If the SAS is damaged, patients should be unable to bias the schemata that are intended to control action, or switch from a currently active schema (mental set) to other. This would produce perseveration errors as those shown by patients with frontal lobe damage when using Wisconsin Card sorting test, where the rule to sort the cards is changed from time to time, without telling the subjects, only telling them if they are doing it right or wrong. El paciente puede saber a nivel consciente que tiene que hacer pero a en un nivel inconsciente el sistema de procesamiento de la infor no lo sabe.La información verbal no influye sobre la conducta.

CHAPTER 6: ATTENTION FROM A NEURO-SCIENTIFIC PERSPECTIVE: INTRODUCTION A complete explanation of the mind comes from different types of observations: thoughts, ideas, and feelings…Our mind lies in the brain. When our brain disappears, our mind, feelings…disappear too. Gall: 1. The brain is the organ of the mind. 2. The cerebral cortex contains regions that control specific functions. He believed that the skull had the same parts the brain has. Phrenology: correlate prominent landmarks of skull with psychological processes (e.g. reasoning, language, authority, religiosity... Broca & Wernike: Two cases of patients. Broca had speech problems (production of language). He had a massive hole in the frontal part of the brain. What was concluded was that the area damaged was the speech area of the brain. These patients have problems in understanding and created new words with no meaning. Integrating brain and mind Sechenov: Founder of the Russian School of Reflexology. Reflexes of the Brain: internal thoughts and internal experience caused by sensory stimulation. An animal whose spinal cord was separated from the brain showed rigidity as if muscles remained contracted. So...all sensory input stimulates motor activation unless prevented by the Brain. He thought that our thoughts were reflexes of stimulation (machines that responded). Discard those ideas. We’ve a complex network of different systems interconnected and that gives more info and subjective experience. Charles Sherrington. Integrative Activity of the Nervous System: • Not just a simple physical circuit but due to a complex mesh of interconnected neurons. • Subjective experience required a neuronal organization well beyond the principles of reflexes. All parts of the nervous

system are connected together. Hebb • Provided an exciting integrative theory of how neural circuits might support mental processes such as attention and

memory. • Rejected the ideas of behaviour as simple responses to stimulation. • Repeated stimulation would develop a structure consisting of cell assemblies of neurons spread throughout the brain but

capable of acting together as a close system. • Not located in a single place but spread throughout the brain.

• With learning connections become stronger....on cell triggers all “hebbian synapse” • Hebb postulated a hierachy in the cell assemblies • From the individual aspects (bottom) to complete figure (higher-level) • 1950-1960 Recording electrical activations of Individual cells in awake, behaving animals. Several levels on analysis of

visual input were confirmed. To identify an object we activate ventral pathway, visual system and Broca’s area. Specific areas may be related with motor actions of objects. Grandmother cells: if we continue with specialization of areas, we may encounter a cell that responds in the same way as our grandmother’s. Cognitive psychology • Orienting tasks: Posner task (locate attention in space), Visual search task (locate attention within objects). • Conflict tasks: Stroop effect, Flanker task, Simon task • Vigilance task (monitoring something for hours): alert task. • Mental Chronometry: The use of response time as a measure of cognitive operations. • Substractive method (1868): a) Detection and motor response; b) a+discrimination. If we have a simple process and a

response, and after I add something else (?????). • Additive method: all processes occur in sequence. Dont take in account that adding something can change the strategy of

completing the task. We avoid this with neuroimaging. Neuroimaging • Single cell recording: measure of the activation of individual cells. • Electroencephalography (EEG): measures brain activation on the scalp. We obtain very good temporal resolution and

bad spatial resolution. Gives a lot of information in terms of frequency. Beta: active during cognitive performance.

Gamma: active thought and most complex tasks. Measures how the brain collects all information to perceive something as an individual object (??). • Magnetoencephalography (MEG): measures magnetic fields of the brain form the cortex of the brain. Non-invasive, good

temporal resolution, measures magnetic fields, more resistant to interference and better spatial resolution. Very expensive.

• Event related potentials (ERPs): measured in ms. Measures how the brain activation changes. It has a good temporal resolution. It detects that when we pay attention to something, the activation is higher (de cajon saes o k). It’s difficult to localize the source that produces the activation seen in the electrodes.

• Estructural: CAT, PET, MRI. There are differences among them in terms of contrast. • Functional (hemodinamic measures): PET semi-invasive, a radioactive substance is injected. Photons collapse with

protons and measures blood flow changes resolution during mental activity. Very good spatial produce image. Bad temporal resolution.

• fMRI: see what happens in brain during tasks performance. It adds monitors and devices. Measures blood oxigenation: oxigenation mechanisms produce changes in magnetic fields that can be measured.Very good spatial resolution but poor temporal resolution.

• Substractive method: substract condition of the other. Common aspects A and B but not C (which is a very specific aspect). As we have individual differences, sometimes we don't show the same activation.

Which control condition would you use? • Viewing words: seeing words that aren't real words, but similar. Ex: TABEL (table) • Listening words • Reading words • Producing words Cognitive Neuropsychology • Data come from patients • Single dissociation: when a patient has a lesion in the brain and shows an impairment in one specific task. • Double dissociation. Transcraneal magnetic stimulation (TMS): magnetic currents in the cerebral cortex. It produces transit lesions. It’s about mapping. It has long-term effects. Lesion goes out fast but we don't know if it can be permanent. We can reach deeper areas of the brain. Attention alters orientating processing in the human lateral geniculate nucleus Top-down process with manipulation of attention. Attention helps even from the entrance (?). Geniculate nucleus: part of the visual pathway and part of the hypothalamus.

CHAPTER 7: NETWORKS OF ATTENTION - Three basic assumptions:

- Cognitive tasks are non-unitary processes and require activation of several brain structures that are connected forming networks - There are networks involved in information processing and control of that processing (attentional networks) - Networks of attention change with learning and development and are susceptible of damage.

- Posner: 3 main networks - ORIENTING NETWORK (posterior): Neuromodulator: Acetylcholine à [Brain areas: Posterior parietal lobe, Cx temporoparietal, Frontal eye field, Superior colliculus, Pulvinar]

- Data for orienting network come mainly for single cell firing recording in monkeys. They were trained to attend overtly or covertly to a target location. Superior parietal cells fired when attention was oriented without eye movements. The Pulvinar in the Thalamus and the superior colliculi are also involved. - Parietal lobe- disengage: in humans evidence come mainly from neuropsychology, from neglect patients with lesions on the temporal and parietal conjunction (temporoparietal cortex), who had difficulties orienting visual field to the contralateral side of the lesion. Using Posner task with those patients they discovered that the role of the parietal lobe was the disengagement of attention. The patients presented a deficit: when attention must disengage from the good side (ipsilateral to the lesion) to move to the bad side. They didn't have problems detecting changes in the cued location, no matter whether trials were ipsi- or contralateral to the lesion. The problem appeared whit the invalid trials, for which contralateral trials were highly affected (attention had to be disengaged from the good side to be moved to the bad side). - Superior colliculus- attentional movement and eye movements: Posner and Rothbart studied ocular movements during infancy. They also studied neuropsychological patients with Progressive Supranuclear Paralysis, that had problems with saccadic and vertical eye movements. There's damage to the superior colliculus and the consequence is a loss of voluntary eye movements and they are also slower in covert orienting of attention regardless of cueing (endo/exogenous) in Posner task. - Frontal Eye Field- Inhibition of return (IOR) and eye movements: when SOA is too long in invalid trials of Posner Task, invalid trials facilitate answer (shorter RTs) compared to valid trials. Applying TMS (Transcranial Magnetic Stimulation) over FEF (Frontal Eye Field) produces IOR to disappear. PET (Positron Emission Tomography) shows more activation of these areas before the onset of the target. - Pulvinar- focalization: increases activation to attended targets (focus). It is the gateway to visual cortex (secondary association areas, it is an associative nucleus). Francis Crick defines it as the spotlight that would select for further processing. LaBerge and Buchsbaum 1990 applied PET to normal subjects, presenting a target (red J) among distractors (more Pulvinar activation needed) or more obvious (alone or more salient- less Pulvinar activation)

- EXECUTIVE NETWORK (anterior): modulated by DA (Dopamine) à [Brain areas: anterior cingulate, lateral prefrontal cx, basal ganglia]

- Attention is not only about perception, but also about behavior. Executive functions are involved in organizing and regulating behavior: concentration, effort, coordination, maintaining relevant goals, inhibition, and voluntary control. - Duncan and Owen 2000 used fMRI during five tasks:

- Stroop - Novelty (movement digits) - Working memory (# of elements) - Working memory (recovery delay) - Perceptive load (degraded letters)

- Three main areas were active: anterior dorsal cingulate, dorsolateral prefrontal cortex (PFC) and ventrolateral PFC. - Anterior cingulate: detecting conflicts and monitor responses (avoid errors)- Tested with Stroop, Flanker and Simon tasks. When we realize a mistake in performance we produce an ERN (Error Related Negativity) measured through EEG, it is an error related negative potential from our anterior cingulate. - Right frontal cortex: maintain task demands active in working memory (WM) (monitoring, instructions). Patients with frontal lobe damage who perform the Wisconsin Card Sorting Test present perseveration (lack of adaptation to changes) - Raichle et al. 1994 had subjects perform a verb generation task while using PET. The areas activated where anterior cingulate cortex, left frontal cortex, left temporal cortex, right cerebellum. It can be compared with passive reading of verbs as control group in order to be sure that those areas are actually involved in verb generation. - Lateral PFC: task demands in working memory:

- Dorsal: selection and manipulation of info - Ventral: recovery and maintaining of objects representation - Using task switching (e.g. indicate whether stimulus are different depending on color. It can change to shape. - MacDonald et al. 2000 used the Stroop task with different cues (read word/name color). DLPFC (dorsolateral PFC) was active when had to name color but not when have to read the word. It keeps task goals active. Cingulate cortex active only when target appears. It supervises for conflict: higher activation in incongruent trials. Both have complementary functions and there's interaction between networks.

- Norman and Schallice 1989: SAS a model of executive control: - Perceptual system- schema control units (influenced and influence SAS) - contention scheduling - It is used when:

- Situation requires planning or decision making

- Links between inputs and outputs are novel or not well learned - Situation requires a response that competes with string habitual responses - Situation requires error correction -situation is difficult or dangerous

- The cingulate cortex: - Blood flow increases during divided vs. focus attention - Blood flow increases during word generation - Strop - ERN

- VIGILANCE/ALERT: modulated by NE (norepinephrine) à [Brain areas: locus coeruleus, frontal and parietal cx, thalamus]

- Vigilance, sustained attention and arousal are closely related. Vigilance refers to ability of organisms to maintain their focus of attention and remain alert to stimuli during prolonged periods of time. HR slows, flood flow over right frontal and right parietal increases. Physiologically, correlates with a modulation of P300 in ERP and lateral readiness potential. - Background: experiments by Norman Mackworth 1950 based on WWII specially focused on radar vigilance. With a task consisting on a clock's hand moving. They had to respond when it jumped a position. The number of changes detected decreased with time (until 2h). It seems that there's no effect on early performance. - Sustained attention: is based on behavioral performance. Variables that might affect sustained attention tasks:

- Simultaneous or successive presentation of target/noise signals - Event rate - Spatial uncertainty - Dynamic properties - working memory demands (many targets, or doing something while remembering things) - Symbolic or exogenous signals (like arrows that require interpretation or a cue in the location, respectively)

- Signal detection theory: with signals that are very similar to the noise, a lot of attention is needed to track the differences. This theory proposes that two factors affect performance (vigilance decrement): δ': sensitivity and β: participant’s criteria, instructions, and willingness. - Vigilance network:

- Right frontal and right parietal: blood flow over them increases. The reason for asymmetry is unknown. Patients with lesions in frontal lobe perform worse in task requiring vigilance and their HR does not slow down. Neuroimaging studies show a decline in activity over fronto-parietal regions has been suggested to mediate vigilance decrements. - Locus coeruleus: releases the NE in the fronto-parietal circuit. It's related to emotional arousal and sleep. It boosts signal to noise ratio and keeps the state of alertness. It has strong connections with the parietal lobe, Pulvinar and colliculus. - fMRI studies in SART (Sustained Attention to Response Task), with targets occurring a 11% of the time showed fronto-parietal activation but introducing an alert tone that helps detection of targets reduces frontal activation.

- PSP: Progressive Supranuclear Paralysis - IOR: Inhibition of return - SOA: Stimulus Onset Asynchrony - TMS: Transcranial Magnetic Stimulation - FEF: Frontal Eye Field - PET: Positron Emission Tomography - fMRI: functional magnetic resonance imaging - PFC: prefrontal cortex - WM: working memory - ERN: Error Related Negativity - EEG: Electroencephalography - DLPFC: dorsolateral prefrontal cortex - SAS: Supervisory Attentional System - HR Heart rate - Even related potentials - WWII: World War II - LC: locus coeruleus - SART (Sustained Attention to Response Task)

CHPT 9: CONSCIOUSNESS AND ATTENTION - According to James' definition of attention, it involves "focalization, concentration of consciousness". It is related to conscious control, intentional control and willed behavior. For skills attention is not required, since performance becomes automatic or unconscious. But, what is consciousness? How does it arise? What is it for? - Evidence from normal subjects is that automatic= unattended= unconscious. Anyway, unattended information can bias responses, so regarding methodology, what criteria should be used to determine consciousness or awareness? Experimental paradigms

- Dichotic listening tasks. Corteen and Wood (1972) conditioned their subjects to expect an electric shock after particular words. Then, while participants completed a shadowing task, they presented those words in the unattended channel. Participants claimed to be unaware of anything from the unattended channel but they showed a clear galvanic response to the shock related words and their associates. - Parafoveal vision experiments: participants are instructed to focus attention on a central stimulus (foveal) and ignore the rest (parafoveal). Underwood 1976 claimed to have demonstrated unconscious semantic processing of unattended words that flanked the target. Participants reported not reading the flankers. - Allport 1977: used a masking experiment (with two masks, one before, other after the masked word). When subjects could not report the meaning of the word, they would sometimes produce interesting errors like "blues" instead of "jazz". Marcell 1983 also made a masking experiment, masking a word and then presenting a target word or non-word. Participants had to distinguish whether it was a word or not. The masked words could be semantically related or not to the target word. Related trials produced facilitation. Both previous experiments support the idea of unconscious processing so, is it necessary attention for processing meanings? - Holender 1986 made a critique of some of those studies, saying that it was not safe to rely in the subjects ignoring the info of one channel in dichotic listening, nor in their criteria of unawareness, they could have briefly perceived the stimuli. - Cheesman and Merikle 1984 proposed two kinds of thresholds:

- subjective: experimenter has to rely on participant's information - objective: can be set independently by the experimenter and should be set at chance level (performance around chance level would be considered as unaware of the stimuli)

- There is some evidence from neuropsychological patients with deficits of attention and consciousness that despite their unawareness, make responses above chance level. Blind sight patients, visually neglect, prosoagnosia (fail to recognize familiar faces) and amnesia prove that info is available for their systems, but below the level of consciousness. - Milner andGoodale 1995 studied visual agnosia in a patient, with bilateral (both hemispheres) damage to occipital cortex. The ventral stream was damaged (the one involved in processing "what", conscious identification), while the dorsal, involved in "how", in action was intact. Participants could not recognize objects, but approximated them as if that info was available for them. Consciousness - Allport and Colheart suggested that is the act of integrating physical and semantic info that gives rise to conscious awareness. (Crick and Koch's neuronal synchronization) - Umiltá 1988 suggest that the role of consciousness is to allow voluntary organization of conscious operations that occur in our minds (control depends on consciousness) - Chalmers 1996 argues that being conscious of a stimulus allow the info perceived to be used for the basis of action. - Baars 1997 suggests that consciousness is a workplace in which attended info is made available Inattentional blindness: requires participants to focus attention on one part of the display, but are not told to ignore nothing. In one of the trials, an unexpected stimulus is presented out of the zone where they had to focus. In most of the cases, participants did not see that stimulus (a word) Change blindness: fail to notice changes in a display