The ‘Deployment of Extra Processing’ Account of Attention Peter Fazekas (fazekas.peter@gmail.com) Centre for Philosophical Psychology, University of Antwerp, Prinsstraat 13, 2000 Antwerp, Belgium Philosophy & Cognitive Neuroscience Research Unit, Aarhus University, Jens. Chr. Skous Vej 7, 8000 Aarhus, Denmark Bence Nanay (bn206@cam.ac.uk) Centre for Philosophical Psychology, University of Antwerp, Grote Kauwenberg 18, 2000 Antwerp, Belgium Peterhouse, University of Cambridge, Cambridge CB2 1RD, UK Abstract Attention as Selection The paper formulates an alternative view about the core function of attention claiming that attention is not selection According to SV, the core function of attention is selection: but the deployment of extra processing capacity. This way of it functions over information processing channels and thinking about attention has greater explanatory power, since decides which one of these can reach further processing. it proposes a common implementation both for selection and modulatory effects, and it offers a unificatory perspective on This view is in the centre of the early vs. late selection the workings of perception and cognition. debate that dominated much of the research on attention in the 20th century. Early selectionists claim that basic Keywords: attention; selection; modulation; working memory physical features of all stimuli are detected and processed pre-attentively, and attention selects a few of these channels Introduction for categorical processing (Broadbent, 1958). Late selectionists claim that all stimuli are processed pre- In recent years, the theoretical and empirical literature on attentively even to a categorical level, and attention makes a attention has gone through an exponential growth not just in few of these channels available for post-perceptual (e.g. sheer volume but also in diversity, which has lead many to working memory) processing (Deutch & Deutch, 1963). claim that attention, after all, is not a unitary mechanism but It needs to be acknowledged that SV is not a homogenous rather a feature of “multiple perceptual and cognitive position—there are many varieties, even beyond the control mechanisms” (Chun, Golomb, & Turk-Browne, question of what point of perceptual processing attention 2011, p. 74) operating at almost all stages of the perceptual- (that is, selection) presents a bottleneck. Originally, SV cognitive system. From this perspective, it becomes a pictured attention as a single mechanism dividing crucial question whether there is a common core function of perceptual-cognitive processing into a pre- and a post- the different forms of attention. attentive stage. However, theorists nowadays argue that Traditionally, the core function of attention is claimed to instead of one single or a few major stages of selection be selection: attention selects the relevant ones from the filtering effects occur throughout the processing stream pool of concurrently present stimuli (Broadbent, 1958). This (Driver, 2001). Accordingly, selection mechanisms operate view goes back as far as William James’s account (James, at many different levels of the perceptual-cognitive system, 1890/1983), and is the ‘received view’ in contemporary making it, at least, prima facie unclear what attention is literature (Petersen & Posner, 2012; Chun, Golomb, & selection for. Some argue that attention selects for later Turk-Browne, 2011). stages of perceptual processing (Lavie, 1995), others claim We call this view, according to which attention is a form that attention selects for working memory (Knudsen, 2007), of selection, the Selection View (SV). Our aim in this paper still others talk about attention as selection for action is to challenge this traditional understanding. In what (Allport, 1987). follows we shall argue that both the selective aspect of These many possible stages or forms of selection are attention, and its modulatory effects (Reynolds & Chelazzi, sometimes brought under the same umbrella by the general 2004) that are often claimed to realise the function of characterisation that attention is selection for further selection, are, in fact, implemented by the same mechanism, processing, where the processing in question takes place namely the continuous and flexible re-allocation of further up in the perceptual-cognitive hierarchy (Chun, processing resources, and that the features of this Golomb, & Turk-Browne, 2011). However, what really mechanism are the common core characteristics of all forms serves as the common denominator of these very different of attention. We call our approach the Deployment of Extra versions of SV is that they all take attention to be a form of Processing View (DEP). It is motivated by the ideas that selection: what attention is is selection. Further processing have been implicit in much of the current research in vision of stimuli is an optional consequence that may or may not science (Carrasco, 2011, 2014). follow attention. That is, according to SV, attention, at any given level where it is in operation is, so to speak, a 756 gatekeeper—a separate mechanism controlling the flow of disengagement of attention consist in the active information through the perceptual-cognitive system. management of processing resources. Attention as DEP The Allocation of What Resources? Thinking of attention as selection has been the mainstream Before moving on, a clarification is in order. So far, we view. But there is an alternative way of thinking about have been talking about the allocation of certain resources, attention, one that is often implicit in works of vision but we haven’t elucidated what resources we have in mind. scientists. It has long been known that the focus of spatial Here we propose that processing resources are best to be attention is able to enhance the processing of visual stimuli understood as the natural processing capacity of the (Posner, 1980). For example, when attention is focused on a perceptual system, i.e. representational encoding via spike region of the visual field processing efficiency is increased generation. Stimuli are processed through a hierarchy of compared to cases when attention is distributed over larger neuronal circuits that encodes stimulus-features as specific regions. In recent years, studying early vision, and utilising firing patterns. Spike generation has a particular energy cost very simple displays, Marisa Carrasco and colleagues have (Lennie, 2003), so ultimately, the resources in question are shown that the increase in processing efficiency is due to energy resources. By allocating more resources, attention attention’s ability to affect very low level perceptual improves the quality of the representations of the target processing like spatial resolution (Yeshurun & Carrasco, features. Note that this is in line with how Carrasco thinks 1998), contrast detection (Carrasco, Ling & Read, 2004) about the resources attention distributes. As she puts it: and even saturation detection (Fuller & Carrasco, 2006). “attention augments perception by optimizing our According to these studies, attention facilitates these low representation of sensory input and by emphasizing relevant level perceptual processes—it improves performance in details”  (Carrasco, 2014, p. 208). several tasks by signal enhancement, i.e. by enhancing spatial resolution, and increasing (even apparent) contrast Comparing Explanatory Power and saturation. One might wonder at this point whether the difference These studies suggest that attention is able to directly between SV and DEP is merely verbal. We believe that the affect how much processing capacity is to be allocated to two views are substantially different: they have different different stimuli (even right at the very entry level of the explanatory power (see this section), and integrative perceptual system). From this perspective, attention is not a potential (see next section). On these grounds, we argue that gatekeeper but rather an information processing booster that DEP is preferable to SV. is able to modulate perceptual processing by affecting the allocation of processing resources. This reinterpretation Selection or Modulation? resonates quite well with how Marisa Carrasco herself As it has become increasingly evident in recent years, the summarises the moral of her research: “attention is involved modulatory effect of attention on neural activity is a general in distributing resources across the visual field”   (Carrasco, phenomenon. As it is sometimes put, nowadays it is 2014, p. 184). “overwhelmingly apparent” that attention modulates This is the way of thinking about attention that we call neuronal responses across many stages of the perceptual- DEP. According to this alternative account, the core cognitive system (Squire, Noudoost, Schafer & Moore, function of attention is the allocation of extra processing 2013, p. 452; see also Reynolds & Chelazzi, 2004). capacity: when one voluntarily attends to a specific spatial That is, the ability to modulate neural activity seems to be location or when a particular stimulus automatically a key feature of attention—a mark of the operation of captures one’s attention what happens is that extra attention that, arguably, is just as widespread and processing resources get allocated to the specific spatial fundamental as selection itself. It seems as though location or particular stimulus. That is, attention increases modulation was just as common and just as core a function the allocation of processing resources to the attended region of the many different forms of attention as selection is often or stimulus. It might be the case that a particular stimulus argued to be. has already been processed before a novel engagement of Note, however, that SV sees modulation and selection as attention, so a certain amount of processing capacity has two distinct features of attention that require two distinct already been allocated to it. However, when attention shifts independent explanations. As some of the proponents of SV to this stimulus extra resources get deployed facilitating the explicitly acknowledge (Chun, Golomb, & Turk-Browne, processing of the stimulus. Similarly, when one voluntarily 2011, p. 75), it is not a necessary feature of attention that a withdraws one’s attention from a specific region or when selected stimulus must be processed in an enhanced, or in an one’s attention gets automatically disengaged from a any way modulated manner. Following the gatekeeper particular stimulus the processing capacity allocated to the analogy, it very well might be the case that attention selects region or stimulus in question decreases. That is, according certain information processing channels by simply blocking to DEP, the voluntary and automatic engagement and competing channels, without affecting in any way the 757 working of the processing channel left intact. Consequently, capacity that can be deployed elsewhere—this is how the for SV—that puts selection effects into the centre of particular stimulus in question gets selected and why thinking about attention—modulation requires additional unattended stimuli ‘fade away’. explanation (Chun, Golomb, & Turk-Browne, 2011, p. 76). Note that although in the case of near-limit operation all Here we propose that DEP is preferable over SV because novel instances of allocating extra processing resources to contrary to the latter the former is able to account for both certain stimuli co-occur with the withdrawal of some selection and modulation on the basis of a single processing capacity from other stimuli, this does not mechanism, namely the eponymous active management of necessarily mean the full withdrawal of attention from processing resources. unattended stimuli (i.e. full-blown selection). There very As we have seen, DEP is motivated by the modulatory well might be cases where only part of the processing effects of attention, and therefore accounting for modulation capacity already allocated to some stimuli gets withdrawn within this framework is quite straightforward. If what with a novel engagement of attention at another location. In attention is is the deployment of extra processing capacity, these cases some (residual or excess) processing resources then the processing of attended information is enhanced, remain allocated to the original (now unattended) stimuli compared to the baseline level of processing preceding the due to which these stimuli are still processed to a certain allocation of attention. Similarly, if the disengagement of degree. These are the cases of partial or distributed attention attention consists in a drop of the available processing (Cohen, Cavanagh. Chun & Nakayama, 2012) with capacity, then the processing of newly unattended incomplete or inefficient attentional selection (Benoni & information diminishes compared to the pre-disengagement Tsal, 2013). level. That is, attention as the deployment of extra Full-blown selection (i.e. the total blocking of the processing directly affects the neural activity relevant for processing of unattended stimuli) only happens if all the encoding of quality representations. processing resources are consumed by a single stimulus (or According to DEP, selection is also a consequence of the a single set of stimuli—the attended one) and thus no active management of processing resources. In fact, within processing capacity could be allocated to other stimuli. this new framework there is a gradual shift between modulation and selection, with modulatory effects coming Is Selection Necessary? in various degrees and (full-blown) selection being at the far According to DEP, thus, the core function of attention is the end of the same continuum. amplification of representational encoding of salient or task relevant stimuli via the allocation of extra processing Selection and DEP resources. Selection effects are only by-products of this It is a fundamental characteristic of attention that it is not amplification and especially the corresponding attenuation possible to attend to too many things at once or to distribute of the representational encoding of inconspicuous or attention over a large region of the visual field without a irrelevant stimuli that results from the fact that processing decrement in processing efficiency. That is, selection seems resources are of limited capacity. to be inevitable. And indeed, given our understanding of Note that according this view, selection is far from being processing resources, selection is a consequence of the the core function of attention—it is not even a necessary allocation of processing capacity, since energy resources consequence of the mechanisms underlying the allocation of required for spike generation and representational encoding attention. In systems without resource-limits, or (more are very limited. Lennie (2003) argues that the cost of realistically) in systems operating far from their limits, there spikes in the brain is high compared to the known energy can be spare capacity at the system’s disposal to be consumption of the cortex, which severely constrains the allocated to new stimuli. In these cases, though attention— activity that can occur concurrently. Lennie concludes that as the deployment of extra processing resources—is very due to this limit in the energy resources available, the much in operation, no selection effects occur, since no energy resources need to be flexibly re-allocated again and processing capacity needs to be withdrawn from unattended again in accordance with actual task demand. stimuli. So when the capacity limit is reached and all the resources This way of thinking about selection, however, might be are allocated then any novel act of voluntarily attending or called into doubt by the very low level studies that have episode of automatic attentional capture must necessarily originally motivated DEP. In a series of studies, Carrasco co-occur with a withdrawal of some processing capacity and colleagues show that even in the case of very simple, from unattended regions. That is, the active management of non-cluttered displays with only two stimuli, when attention processing resources, when operating near limit, consists in facilitates contrast sensitivity and acuity at the attended the joint allocation and withdrawal of resources: shifting location, trade-offs (decreased contrast sensitivity and attention is shifting the allocation of (i.e. re-allocating) acuity—compared to the baseline) appear at the non- processing capacity. Allocating extra resources to a attended location (Pestilli & Carrasco, 2005). particular stimulus decreases the amount of processing On the face of it, this finding is in tension with the idea 758 that selection effects occur only when the system operates accompanied by the re-distribution of the limited processing near limit (i.e. when perceptual resources are fully capacity of low level visual perception—it will be fully allocated). The problem is this: in the experimental allocated all the time, but slightly differently: the actually paradigms used in the studies above, the displays, typically attended location will receive more resources, the actually containing only two Gabor-patches, are so simple that their unattended locations will receive less resources. capacity requirement in the baseline condition could hardly Finally, note that this picture is compatible with the active exceed, or even get close to, the capacity limit. So when one inhibition of unattended stimuli (Cerf, Thiruvengadam, attends to the location of one of the stimuli extra processing Mormann, Kraskov, Quiroga, Koch, & Fried, 2010), since capacity could very well be deployed without the necessary active inhibition is, in fact, a tool for resource withdrawal. withdrawal of processing resources from other locations. That is, this finding seems to show that selection effects can DEP, Modulation, and Selection: an Overview occur even if the full processing capacity is only partly According to DEP, attention is the deployment of extra allocated. processing capacity. Since the processing capacity in We, however, think that these results are, in fact, perfectly question is, in fact, the set of available energy resources compatible with DEP. What they suggest, is that the limited required for spike generation, the deployment of extra processing capacity (at least of early visual processing, see processing capacity results in enhanced representational below) is always fully allocated. That is, we interpret the encoding. Enhanced representational encoding is a form of Carrasco studies in question as indicators that early visual modulation, so modulatory effects straightforwardly follow processing always operates near limit. The low levels of the from the deployment of extra processing capacity. visual processing stream seem to be unable to reserve spare If the system operates near limit (i.e. if the resources are (unallocated) processing capacity that could be deployed always fully allocated), then the deployment of extra without any occurrent cost. Therefore, processing capacity processing capacity will lead to the withdrawal of is fully allocated even when subjects are faced with the processing capacity from unattended stimuli (via, say, active simple displays in the Carrasco studies. When attention is inhibition). The withdrawal of some processing capacity captured at a certain location overlapping with one of the from unattended stimuli decreases the quality of the stimuli, and extra resources get allocated to the processing representational encoding of these stimuli. Diminished of this stimulus, then resources from other locations must be representational encoding of unattended stimuli is equally a withdrawn. form of modulation, but it is also a form of selection: it is The claim that the limited processing capacity of low what is called partial (or incomplete, or inefficient) level visual perception is always fully allocated finds plenty selection. of support in the literature. For example, Treisman (1969) Finally, if the attended stimulus consumes all processing suggests that  “we tend to use our perceptual capacity to the capacity then there is no residual capacity that could spill full on whatever sense data reach the receptors” (Treisman, out to unattended stimuli, so unattended stimuli are not 1969, p. 296). The idea here might be that active inhibition processed at all. They are stopped being representationally is unavailable for low level perception, so “the nervous encoded. This is the case of full-blown selection. system is forced to use whatever discriminative system it In other words, if we accept DEP and the further has available, unless these are already fully occupied with assumption about the near-limit operation of at least certain other tests or inputs” (Lavie, 1995, pp. 452-453). Indirect perceptual sub-systems, then we become able to explain evidence might further be provided by Lennie (2003) who both the modulatory effects and the selective nature of claims that the sensory cortex is “among the most active attention. Therefore, since SV fails to explain why attention metabolically”   (Lennie, 2003, p. 496), i.e. it seems to be entails the modulation of signal processing, DEP is much more active than other parts of the cortex with preferable because of its greater explanatory scope. comparable number of neurons. A possible source of further support might be the observation that even in no stimulus Perception and Cognition in the Light of DEP conditions the spontaneous activity of the primary visual In this section, we would like to point out that attention, as cortex is very strong and coordinated, resembling stimulus defined by DEP, is a natural ally of modern approaches to (natural scene) evoked activity (Berkes, Orbán, Lengyel, & higher cognitive processes, especially the functioning of Fiser, 2011). working memory. Working memory is generally thought of To sum up, selection effects do not necessarily follow as the site where information conveyed by the senses and from the deployment of extra processing, but whenever a processed by the perceptual system is brought together and (sub-)system operates near its capacity limit selection gets evaluated in accordance with the needs of ongoing effects will always occur. As the Carrasco studies, energy tasks (Baddeley, 2003). Within this context, attention is considerations, or observations related to spontaneous often seen as the gateway to this site—as the set of activity suggest, early vision might very well be such a sub- mechanisms responsible for selecting which bits of system. If so, then shifts in attention are always information can gain access to working memory (Knudsen, 759 2007). Moreover, it has recently been argued that there is a repositioned to the corresponding information processing significant overlap between attentional and working channel passing on information about a new feature and memory functions (Awh, Vogel & Oh, 2006; Gazzaley & blocking others—providing just the right sort of input to fill Nobre, 2012). In the light of these results, it is timely to up another slot of working memory. That is, slot-based reconsider whether the way we think about attention accounts of working memory and attention as defined matches our most up-to-date accounts of working memory. within SV are natural allies. Attention as it is jumping in shifts from one feature or object to another delivers exactly Slot vs. Resource Models of Working Memory that kind of information that is going to end up in working Perhaps the most characteristic feature of working memory memory. Slots store what attention selects. functions is that they are of limited capacity. In recent years, Resource models of working memory break with the the nature of this limited capacity has hotly been debated notion of all-or-nothing representational encoding. They (Luck & Vogel, 2013; Ma, Husain & Bays, 2014). claim that working memory can and often does maintain Classically, the limited capacity of working memory is partial (low quality) information about items that are not interpreted as a limit in the number of discrete necessarily given priority in accordance with actual task representations that can actively be maintained to serve the demand. Note that by making this claim, and especially by needs of ongoing tasks. The observations according to anchoring the notion of the limit of working memory to which working memory is able to hold only 3-7 items at resources required for creating quality representations, once (Cowan, 2001) is typically modelled by a slot-based resource models of working memory commit themselves to account: working memory has a fixed number of slots that the very same principles that underly DEP. Also note that can store object or feature representations such that when an the emphasis of DEP on the continuous and flexible re- object representation gets into a slot the object will be allocation of these resources is also shared by the resource remembered, and when it does not the object will not be models of working memory. In fact, it seems that the very remembered at all (Luck & Vogel, 2013). That is, if the mechanism that DEP points out as the implementation of all number of items in the input is greater than the fixed attentional effects, is also able to implement the functioning number of slots then no information will be stored in of the working memory as recent resource models describe working memory about the items that do not get into one of it. Reported overlap between attentional and working the slots. memory functions (Awh, Vogel & Oh, 2006; Gazzaley & Contrary to this, resource models conceptualise working Nobre, 2012) thus might very well be due to a common memory as a limited resource that can flexibly be mechanism responsible for the implementation of both sets distributed between all the items in the input (Bays & of functions: as attention enhances the quality of Husain, 2008). The more resource is allocated to a particular representational encoding of certain features via the item, the better the quality of the working memory allocation of extra processing capacity, it becomes possible representation of the item will be. That is, resource models for working memory to maintain more detailed and less shift the emphasis from the number of items that can be noisy representations of these features, which, then, results stored in working memory to the quality or precision of the that they can be recalled with more precision. memory trace of the items in question. The flexibility of That is, whereas SV complements slot-based accounts of resource allocation makes it possible to store enhanced working memory quite well, it doesn’t really match modern quality representations of prioritised items while resource models. However, DEP re-conceptualises attention maintaining low quality representations of other items (Bays in a way that makes it a perfect fit for this latter approach to & Husain, 2008; Ma, Husain & Bays, 2014). working memory that has become increasingly dominant in recent years. Attention and Working Memory Note that SV with its item-based attentional shifts is a quite Attention, Perception, Cognition good match for classical slot-based accounts of working Finally, let’s step back and take a bird-eye-view of what has memory. Attention, according to SV, operates over been argued for so far. Seeing the re-allocation of information processing channels that process characteristics processing capacity as the fundamental mechanism of individual physical features or objects. In this context, implementing attentional effects re-defines the relationship selecting a channel and blocking others amounts to passing between attention, perception, and cognition. on information about certain features or objects while First, it is an essential feature of perceptual processing filtering out others—providing ideal input for the fixed slots that its resources get continuously re-allocated. In the course of working memory. The feature or object the information of this re-allocation the processing of certain stimuli receive of which can get through the bottleneck will be represented extra capacity making these stimuli the attended ones. in working memory, whereas those that are screened off will Attention, thus, is not an extra mechanism working on top not be remembered. Similarly, shifting attention from one of perception (as the traditional SV claimed it to be), but feature to another means that the filtering mechanism gets rather an aspect or a result of how perception itself works. 760 That is, attentional effects are just aspects of the natural Cohen, M. A., Cavanagh, P., Chun, M. M. & Nakayama, K. unfolding of how the perceptual system does what it does. (2012). The attentional requirements of consciousness. Second, recent approaches to working memory teach us Trends in Cognitive Sciences, 18, 8, 411-417. that even higher cognitive functions share the common Cowan, N. (2001). The magical number 4 in short-term characteristic of flexibly allocating processing resources. It memory: a reconsideration of mental storage capacity. seems, thus, that the fundamental principles our proposal Behavioral and Brain Sciences, 24, 87-114. points out underlie a broad range of mental processes from Deutsch, J. A. & Deutsch, D. (1963). Attention: Some attention through perception to cognition. Theoretical Considerations. Psychological Review, 70, 80–90. Acknowledgments Driver, J. (2001): A selective review of selective attention This work was supported by the FWO Postdoctoral research from the past century. British Journal of Fellowship 12B3914N and the DFF - EU MCA - COFUND Psychology, 92, 53-78. Mobilex grant 1321-00165 (PF), and the EU FP7 CIG grant Fuller, S. & Carrasco, M. (2006). Exogenous attention and PCIG09-GA-2011–293818 and the FWO Odysseus grant color perception: Performance and appearance of G.0020.12 (BN). saturation and hue. Vision Research, 46, 23, 4032–4047. Gazzaley, A. & Nobre, A. C. (2012). Top-down modulation: References bridging selective attention and working memory. Trends in Cognitive Sciences, 16, 2, 129-135. Allport, A. (1987). Selection for action: Some behavioural James, W. (1890/1983). The Principles of Psychology. and neurophysiological considerations of attention and Boston, MA: Harvard University Press. action. In: A. Sanders & H. Heuer (Eds.) Perspectives on Knudsen, E. I. (2007). Fundamental components of perception and action (pp. 395–419). Hillsdale, NJ: attention. Annual Review of Neuroscience, 30, 57-78. Lawrence Erlbaum Associates. Lavie, N. (1995). Perceptual load as a necessary condition Awh, E., Vogel, E.K. & Oh, S.-H. (2006). Interactions for selective attention. Journal of Experimental between attention and working memory. Neuroscience, Psychology: Human Perception and Performance, 21, 3, 139, 1, 201-208. 451-468. Baddeley, A. (2003). Working memory: looking back and Lennie, P. (2003): The cost of cortical computation. Current looking forward. Nature Reviews Neuroscience, 4, 829- Biology, 13, 6, 493–497. 839. Luck, S.J. & Vogel, E.K. (2013). Visual working memory Bays, P.M. & Husain, M. (2008). Dynamics shifts of limited capacity: from psychophysics and neurobiology to working memory resources in human vision. Science, 321, individual differences. Trends in Cognitive Sciences, 17, 851-854. 8, 391-400. Benoni, H. & Tsal, Y. (2013). Conceptual and Ma, W.J., Husain, M. & Bays, P.M. (2014). Changing methodological concerns in the theory of perceptual concepts of working memory. Nature Neuroscience, 17, load. Frontiers in Psychology, 4, 522, 1–7. 3, 347-356. Berkes, P., Orbán, G., Lengyel, M. & Fiser, J. (2011). Pestilli, F. & Carrasco, M. (2005). Attention enhances Spontaneous Cortical Activity Reveals Hallmarks of an contrast sensitivity at cued and impairs it at uncued Optimal Internal Model of the Environment. Science, 331, locations. Vision Research, 45, 14, 1867–1875. 83-87. Petersen, S. E. & Posner, M. I. (2012). The attention system Broadbent, D. E. (1958). Perception and communication. of the human brain: 20 years after. Annual Review of Elmsford, NY: Pergamon Press. Neuroscience, 35, 73-89. Carrasco, M. (2011). Visual attention: The past 25 years. Posner, M. I. (1980). Orienting of attention. Quarterly Vision Research, 51, 1484-1525. Journal of Experimental Psychology, 32, 1, 3–25. Carrasco, M. (2014). Spatial covert attention: Perceptual Reynolds J. H. & Chelazzi L. (2004). Attentional modulation. In: A. C. Nobre & S. Kastner (Eds.) The modulation of visual processing. Annual Review of Oxford Handbook of Attention (pp. 183-230). Oxford: Neuroscience, 27, 611–47. OUP. Squire, R. F., Noudoost, B., Schafer, R. J. & Moore, T. Carrasco, M., Ling, S. & Read, S. (2004). Attention alters (2013). Prefontal contributions to visual selective appearance. Nature Neuroscience, 7, 3, 308–313. attention. Annual Review of Neuroscience, 36, 451-466. Cerf, M., Thiruvengadam, N., Mormann, F., Kraskov, A., Treisman, A. M. (1969). Strategies and models of selective Quiroga, R.Q., Koch, C., & Fried, I. (2010). On-line, attention. Psychological Review, 76, 282-299. voluntary control of human temporal lobe neurons. Yeshurun, Y. & Carrasco, M. (1998). Attention improves or Nature, 467, 1104-1108. impairs visual performance by enhancing spatial Chun, M., Golomb, J. D. & Turk-Browne, N. (2011). A resolution. Nature, 396, 72-75.. Taxonomy of External and Internal Attention, Annual Review of Psychology, 62, 73-101. 761