The Filter Theory

8 Key excerpts on "The Filter Theory"

• 

  eBook - PDF

  Poetry of Attention in the Eighteenth Century

  • M. Koehler(Author)
  • 2012(Publication Date)
  • Palgrave Macmillan

    (Publisher)

  13 A brief history of filter theory’s emergence in mid-twentieth-century psychology will illuminate its convergences with eighteenth-century models and will, in some cases, provide a more pre- cise vocabulary for phenomena described in earlier models. As the pre- vious chapter mentioned briefly, psychology’s filter model of attention developed from auditory experiments performed by E. Colin Cherry and Donald Broadbent in the 1950s. Cherry sought to explain what he called the “cocktail party problem”: how listeners tune in to one voice in a noisy and distracting setting and whether they can entirely block out competing sounds. He conducted a series of dichotic listening experiments, which asked subjects to track and repeat messages heard in only one ear while wearing headphones that fed a different sound input into each ear. Results suggested that listeners traced the message easily in the attended ear but afterward recalled little from the unat- tended ear. In focusing their attention closely on one sensory input, they had blocked an equally available input. Broadbent’s filter theory explained these findings by likening the human nervous system to a sin- gle communication channel with an inherently limited capacity to pro- cess information. 14 Attention was the filter that admitted one input into the channel for processing and screened out the competitors. The filter model imposes a rigid capacity constraint on human attention—even The Filter of Attention in Mock-Heroic Poetry 65 with effort, a listener simply lacks the attentive resources to track the messages in both ears. Without the filter, this model says, we would be distracted and overwhelmed by the profusion of perceptible infor- mation inundating us at any given moment. By this model, attention promotes efficiency: it prevents mix-ups and enables stimuli to proceed smoothly through the channel.

  Sign up to read

  Learn more about book
• 

  eBook - PDF

  Attention

  Theory and Practice

  • Addie Johnson, Robert W. Proctor(Authors)
  • 2003(Publication Date)
  • SAGE Publications, Inc

    (Publisher)

  Information that fits the filter is then passed along to the limited capacity channel, where it can be iden-tified. The results of this analysis are then sent on to a response system and may also be used to update expectations about what is likely to occur in the given situ-ation. Filter theory is an early-selection theory in that information is assumed to be selected by attention at a relatively early stage of processing. Although the powers of selection are impressive, selection is not perfect. For example, when driving your car along the highway, a particularly flashy or interest-ing billboard might distract you. Such distraction is most likely to occur when something relevant to you or highly practiced, such as your own name, is present. 60 ▲ ATTENTION: THEORY AND PRACTICE Figure 3.1 Broadbent’s early-selection model (filter theory) of attention. In other words, potentially relevant information sometimes gets past the selective filter. This finding (discussed at more length in the following chapter) led Treisman (1960) to propose the filter attenuation theory. According to this theory, the early, selective filter does not completely block out unwanted information, but only attenuates or reduces the strength of unattended stimuli. Under normal condi-tions, this attenuated information does not reach consciousness, but when the information is familiar or fits the context of attended information, it manages to pass the threshold for identification. The Late-Selection View Not long after the development of the early-selection viewpoint, Deutsch and Deutsch (1963) and Norman (1968) argued that there is no serious processing lim-itation up to a categorical level of processing. In other words, they suggested that selection does not occur on the basis of an early-selection filter (which sometimes fails), but after stimuli have already been identified.

  Sign up to read

  Learn more about book
• 

  eBook - ePub

  Einstein, Tagore and the Nature of Reality

  • Partha Ghose(Author)
  • 2016(Publication Date)
  • Routledge

    (Publisher)

  Now that we have discussed some fundamental aspects of the human attentional system, I next take up what impact this has on our very perception of reality. The reach of attention covers all sensory systems, though the visual and auditory systems are the most affected and therefore have been studied in far greater detail. The role of attention in the somatosensory (touch) system is also well known through our common experience of ignoring tactile stimuli, and even mild pain, by attending to alternate stimuli. There is an enlightened community of healers who specialise in storytelling to help people battle chronic pain. The idea is to move the mind away from the insidious effects of constant pain by focusing attention on pleasant and compelling stories. This healing approach is becoming practiced in areas as diverse as children’s care to palliative settings.

  Attentional filtering: early selection

  Attention has been thought of as a filtering process—our mind selects things that are important and concentrates on them, while discarding the vast amount of irrelevant stimuli that we are being constantly bombarded with. Psychologists have spent considerable effort on trying to identify how the filtering effect takes place and where along the hierarchy of sensory processing it is effectuated. Several models have been proposed to account for experimental data. We will consider two rather different models, shown in Figure 9.4 .

  One of the earliest models of the human attentional system proposed that stimuli are filtered at a very early stage, and only those that meet a certain threshold of relevance are passed through.18 This idea, which is illustrated in the left panel of Figure 9.4 , is known as early selection . Physical stimuli captured by our sensory systems are filtered at an early pre-attentive stage in order to cope with the vast amount of signals captured by our senses. Attention, whether arising from bottom-up or top-down processing, acts upon the incoming signal stream to select those to be passed on for higher-level perceptual processing and storage into memory. An attractive feature of this model is that an early filtering process fits in nicely with the idea that the brain would not have evolved to undertake massing information processing only to discard data at a later stage.

  Figure 9.4

  Sign up to read

  Learn more about book
• 

  eBook - PDF

  Handbook of Perception and Action

  Attention

  • Odmar Neumann, A. F. Sanders(Authors)
  • 1996(Publication Date)
  • Academic Press

    (Publisher)

  If the system is assumed to function via top-down mechanisms, then the task is to foster processing of the wanted information. Filter theory, so it seems, was the prototype of the first variant, while Neisser's (1976) schema theory was a straightforward example of the latter. Broadbent's filter was an inhibitory mechanisms for controll- ing bottom-up processing, whereas Neisser's (1976) 'anticipation' was a facilitatory mechanism for the control of top-down processing. Upon closer examination, the situation is somewhat more complicated. While Broadbent (1958) seems to have thought of the filter as a device for blocking Theories of attention 425 ('rejecting') unwanted information, the filter metaphor as such does not carry this meaning. One might as well argue that a filter provides pathways for the processing of the desired information, while leaving the rest of the information unaffected. Indeed, given the wealth of information that impinges upon the sensory surface at every moment in time, it is very hard to think of mechanisms that actively affect the processing of all of them in a negative manner, while leaving the processing of the comparatively small amount of attended information unaffected. Such a system would require an immense amount of inhibition and would therefore work very inefficiently. Thus, although Broadbent (1958) may have visualized the functioning of the filter as an active 'rejection', a (both technologically and biologically) more plausible interpretation of the filter metaphor would be in terms of selective facilitation. Things are similar for the 'attenuation' metaphor. As reported by Neisser (1967, p. 212), Treisman explained that it referred to the signal-to-noise ratio between the selected and the nonselected message, with respect to information content.

  Sign up to read

  Learn more about book
• 

  eBook - ePub

  Attention and its Crisis in Digital Society

  • Enrico Campo(Author)
  • 2022(Publication Date)
  • Routledge

    (Publisher)

  Exactly like telephone filters, attention has the task of distinguishing relevant information from simple noise and so allowing the passage of the former while excluding the latter. The filter is therefore in itself an information processing mechanism that has the task of reducing the flow by excluding irrelevant messages. The attention-filter thus creates a restriction of the flow of information (a “bottleneck” in fact) that is processed at the upper level of the central nervous system. Messages that pass through the filter are treated attentively and retained in the memory, while messages considered irrelevant are lost. Before the filter, information undergoes a rapid parallel analysis; after the filter, the central nervous system works serially, that is, one stimulus at a time. The communication system, as a metaphorical domain of origin, also guides the premises, expectations and inferences with respect to attention, as the target domain. It is therefore not surprising that, exactly like telephone engineering, in the case of the mind the attention-filter is a structure and, for this reason, the Broadbent model, as well as others, are therefore referred to as structural models of attention. 2.4 Early Selection and Late Selection: Where to Place the Filter? The empirical evidence that supported this theory mostly derived from experiments based on dichotic listening. Edward Colin Cherry [1914–1979] (1953) was the first to launch these types of experiments, designed to test the ability to privilege one auditory source over others. 7 In the simplest version of the experiment, different messages were transmitted to both ears of test subjects through a headset who were asked to pay attention only to the messages received in one ear

  Sign up to read

  Learn more about book
• 

  eBook - PDF

  Auditory Cognition and Human Performance

  Research and Applications

  • Carryl L. Baldwin(Author)
  • 2016(Publication Date)
  • CRC Press

    (Publisher)

  These findings provided additional evidence for the existence of a brief sensory store capable of maintaining an echoic trace for a period of several seconds (Moray, 1969). Findings such as these led Donald Broadbent (1958) to propose that attention was subjected to a “filter”-type mechanism resembling a bottleneck. F ILTER T HEORIES The first formal theories of information processing in modern psychology suggested that some type of structural bottleneck keeps people from being able to process all available sensory information at any given time. That is, while we are constantly bombarded by a simultaneous cacophony of sounds, we can only process some lim-ited set of these at any given time. A structural bottleneck or filter prevents all infor-mation from being processed. The early filter theories differed in terms of where in the information-processing stream this filter or bottleneck occurred. Broadbent’s Early Filter Model Broadbent (1958) presented his classic bottleneck or filter theory of attentional pro-cessing in his seminal book Perception and Communication . The model essentially proposed that all sensory stimuli enter the sensory register and then are subjected to an attentional filter or bottleneck based on certain physical characteristics. Broadbent noted that in a dichotic listening paradigm, physical characteristics such as the ear of presentation or pitch could be used to allow one of many auditory messages selec-tively through the filter. However, subsequent research provided evidence that under some circumstances more than one auditory message could be processed, at least to some extent. For instance, Moray (1959) found that sometimes when attending to one of two dichoti-cally presented messages people were able to recognize their name in the unattended message. Moray reasoned that only information that was very important to the lis-tener was able to “break through the attentional barrier” (p. 56).

  Sign up to read

  Learn more about book
• 

  eBook - PDF

  Handbook of Cognition

  • Koen Lamberts, Rob Goldstone, Koen Lamberts, Rob Goldstone(Authors)
  • 2004(Publication Date)
  • SAGE Publications Ltd

    (Publisher)

  This theory also pioneered the use of flow-chart models in cognitive psychology (see Figure 4.1). In Broadbent’s conception, information flows from the senses through many parallel input channels into a short-term memory store. The short-term store can hold the information for only a few seconds. Access to response systems depends on a limited-capacity channel , whose ability to transmit information is much smaller than the total capacity of the parallel input channels. Therefore, a selective filter operates between the short-term memory and the limited-capacity channel. The filter acts as an all-or-none switch, selecting information from just one of the parallel input channels at a time. Broadbent (1958) defined an input channel as a class of sensory events that share a simple physical feature. Except for analysis of such features, stimuli on unattended channels should not be perceived. This conjecture accounted for the main results of the early studies on selective listening, but was soon challenged. Moray (1959) and others showed that subjec-tively important words (e.g. the subjects’ own name) tended to be recognized even if presented on the non-shadowed channel. To accommodate such findings, Treisman (1964a, 1964b) developed a variation of filter theory in which the filter operates in a graded rather than an all-or-none fashion. In Treisman’s attenuation theory, unattended messages are weakened rather than blocked from further analyses. Both selected and attenuated messages are transmitted to a recognition system with word recognition units. Because thresholds of recognition units for important words are lowered, these words tend to be recognized even if appearing in attenuated messages. In the filter theories of Broadbent and Treisman, attentional selection occurs at an earlier stage of 4 Attention C L A U S B U N D E S E N A N D T H O M A S H A B E K O S T processing than stimulus recognition. Such theories are called early-selection theories.

  Sign up to read

  Learn more about book
• 

  eBook - ePub

  Localist Connectionist Approaches To Human Cognition

  • Jonathan Grainger, Arthur M. Jacobs, Arthur Jacobs(Authors)
  • 2013(Publication Date)
  • Psychology Press

    (Publisher)

  CHAPTER TWO

  A Model of Selective Attention as a Mechanism of Cognitive Control

  George HoughtonUniversity College London

  Steven P. TipperUniversity of Wales, Bangor

  The basic phenomenon of all intellectual achievement is the so-called concentration of attention.

  —Wundt, 1904, p. 481

  SELECTIVE ATTENTION AND COGNITIVE CONTROL

  In this chapter, we discuss the problem of selective attention, with the emphasis on neural network theories and the possibility of understanding attention at the neural level. Many ideas about the function of attention and the mechanisms supporting it have been developed in experimental psychology. Although researchers have not traditionally couched these ideas either in neural terms or in terms of formally defined mechanisms, the ideas have been important in the development of neural network theories (at least for the amount of experimental data they have generated). For instance, psychologists have seen attention as a (low-level) perceptual filter (Broadbent, 1958), a higher level competitive filter based on an internal assessment of the relevance of competing inputs (Deutsch & Deutsch, 1963), and a mechanism for perceptual grouping (Treisman & Gelade, 1980). Other related debates have concerned what attention acts on (i.e., what kinds of representations get selected), for instance, locations in space (Eriksen & Eriksen, 1974), whole objects (Duncan, 1980), or goal-relevant (action-based) representations (Tipper, Lortie, & Baylis, 1992; Tipper, Weaver, & Houghton, 1994).

  The question of what attention does has traditionally involved two contrasting positions, the late and early selection theories. Early selection theories have typically postulated that the direction of attention affects the perceptual processes taking place in the attended domain. Perceptual processes are thought to be limited in capacity in some way, and attention acts as a spotlight, so that the attended input receives “further processing” or has more “attentional resources” allocated to it. Such ideas have been frequently criticized for their vagueness and even circularity (e.g., O. Neumann, 1987), but one fairly concrete proposal that has received considerable investigation is that attention is necessary for binding distributed representations of perceptual information into perceptual groups or objects (Treisman & Gelade, 1980). Unattended inputs are represented at a primitive featural level only, and the relations between them (which features belong with which, etc.) are not represented. Other work has suggested that the extent to which attention affects basic (visual) perceptual processes might depend on other factors such as the complexity of the scene (number of inputs to be simultaneously processed: Desimone & Duncan, 1995; Lavie & Tsal, 1994) or the informational requirements of the task for which attention is deployed (Tipper et al., 1994).

  Sign up to read

  Learn more about book