Ward 2011 The Thalamic Dynamic Core Theory of Conscious Experience

Consciousness and Cognition 20 (2011) 464486
Contents lists available at ScienceDirect
Consciousness and Cognition

journal homepage: www.elsevier.com/locate/concog
Review
The thalamic dynamic core theory of conscious experience

Lawrence M. Ward
Department of Psychology, University of British Columbia, 2136 West Mall, Vancouver, BC, Canada V6T 1Z4
a r t i c l e
i n f o
a b s t r a c t
I propose that primary conscious awareness arises from synchronized activity in dendrites of neurons in dorsal thalamic nuclei, mediated particularly by inhibitory interactions with thalamic reticular neurons. In support, I offer four evidential pillars: (1) consciousness is restricted to the results of cortical computations; (2) thalamus is the common locus of action of brain injury in vegetative state and of general anesthetics; (3) the anatomy and physiology of the thalamus imply a central role in consciousness; (4) neural synchronization is a neural correlate of consciousness. 2011 Elsevier Inc. All rights reserved.
Article history: Received 19 October 2010 Available online 23 February 2011 Keywords: Consciousness Awareness Thalamus Dynamic core Neural synchrony Cortico-thalamic Thalamo-cortical NCC
1. Introduction The search for the neural correlates of consciousness (NCC) has been intense and productive in the past two decades since Crick and Koch (1990) focused attention on this project (e.g., Rees, Kreiman, & Koch, 2002). Recent work has emphasized the importance of the thalamo-cortical system of the brain in generating conscious awareness. Within this system three possibilities for the critical brain activity most closely associated with consciousness have been proposed: it occurs primarily within the cortex (e.g., Crick & Koch, 2003; Romijn, 2002), it occurs in the entire system of thalamo-cortical loops (e.g., Edelman & Tononi, 2000; John, 2001, 2002; Llins, Ribary, Contreras, & Pedroarena, 1998), it occurs primarily within the thalamus (e.g., Peneld, 1975). The rst two of these have been and are routinely explored in the recent literature. In this paper I discuss the third possibility in the light of a convergence of studies in cognitive science, neuroimaging, neurophysiology, neuroanatomy, and computational neuroscience, bringing up-to-date the proposal that the neural activity most closely associated with primary consciousness occurs primarily in the thalamus. My proposal is simple: phenomenal consciousness is generated by synchronized neural activity in the dendritic trees of dorsal thalamic neurons: a thalamic dynamic core (see also Ward, 2004 for a very brief treatment of an early version of this idea). This synchronized thalamic activity is driven by the massive cortico-thalamic projections that carry the results of cortical computations (information processing), said cortical activity also, perforce, being synchronized in the same way through thalamo-cortical echoes. The cortex computes the contents of phenomenal consciousness but phenomenal consciousness itself usually represents only the results of cortical computations, not the computations themselves. This paper outlines the reasoning that gives rise to this thalamic dynamic core proposal and presents four crucial evidential pillars that support its plausibility.
Fax: +1 604 822 6923.

E-mail address: lward@psych.ubc.ca 1053-8100/$ - see front matter 2011 Elsevier Inc. All rights reserved. doi:10.1016/j.concog.2011.01.007
L.M. Ward / Consciousness and Cognition 20 (2011) 464486
465
1.1. Explanandum This paper is aimed at a particular aspect of the NCC: that of uncovering the neural correlates of the conscious state itself, in the sense most pointedly described by Searle (2000): What one wants to know is, how is it possible for the subject to be conscious at all? (p. 572) That is, I am addressing the problem of how it is that the human brain (and possibly other animal brains) gives rise to experience of the kind people have when they are awake, and possibly when they are dreaming, and that presumably they do not have when they are deeply asleep (slow-wave), in a coma or vegetative state, or deeply anesthetized with a general anesthetic. What brain activity is necessary and sufcient for the conscious state to be present in a human? According to Searle (e.g., 2000), the conscious state is characterized by three critical properties: qualitativeness (how it feels), subjectivity (it exists only when experienced by some entity, i.e. it has a rst person ontology), and unity (experiences occur within a unied conscious eld), as well as by other properties that may or may not be present in all experiences, for example intentionality, central or peripheral, a mood, pleasantness, Gestaltness, and familiarity. This question, and the properties of the conscious state just listed, are distinct from, although deeply related to, the question of what neural activities are correlated with specic contents of our conscious experience. Other important aspects of consciousness, such as metaconsciousness (consciousness of being conscious and of what one is conscious of) and self-consciousness (consciousness of having a self that is separate from other selves and from the outside world) are viewed as contents of consciousness that are built from recursive application of primary consciousness to conscious contents (e.g., Zelazo, 2004), are contributed by processing in specialized areas of the cortex (such as the precuneus), or arise from emotional (feeling) processing in various cortical and subcortical areas (e.g., Damasio, 1999). Seth and Baars (2005) presented a list of 16 widely recognized properties of consciousness to be accounted for by any theory. Although I wont address them all individually here I believe that the thalamic dynamic core can account for most if not all of them at least as well as the theory they espouse, Neural Darwinism, partly because it shares important elements with the latter theory. One of those necessary properties, however, is of central importance: the notion that . . .consciousness facilitates widespread access between otherwise independent brain functions. (Baars, 2002, p. 47), the conscious access hypothesis. Regardless of through which theoretical lens the problem is viewed, the evidence is overwhelming that consciousness is functionally integrative and that this is the dominant tness advantage it provides for conscious organisms. In other words, the role of this integrative processing is to provide internal representations (models) of the niche-relevant causal structure of the environment, including objects and their surroundings and the events that take place there. Baars (2002) reviews extensive evidence that integrative conscious processing involves more widespread cortical activity than does unconscious processing. Any respectable theory of consciousness must explain this fact (e.g., Dehaene, Kerszberg, & Changuex, 1998), or integrate it into the assumptive base of the theory. The thalamic dynamic core approach accomplishes this task naturally and directly. The dynamic core concept itself is not new. Most recently, Tononi and Edelman (e.g., 1998a, 1998b; Edelman & Tononi, 2000) proposed that the large, brain-wide population of synchronously ring cortical neurons that are associated with conscious awareness of a stimulus forms a dynamic core of neural activity that is in fact the neural correlate of consciousness. In order to support conscious awareness, this dynamic core of synchronous neural activity must be at the same time both integrated (i.e., synchronous) and differentiated (i.e., must be only one of many possible patterns of such ring), and must be of relatively high informational complexity (and thus different from the rigidly synchronous but simple pattern of neural ring typical of epilepsy and its accompanying unconsciousness). Although somewhat more elaborated in its present form, the concept dates back at least to Kinsbournes (1988) proposal of a dominant focus of neural activity that binds together the various aspects of brain activity that support consciousness. It is also related to Dennetts (1991) idea of the dominant narrative in a system of multiple drafts. Indeed some such concept is demanded by Freuds (e.g., 1984/1915) division of the mind into conscious and unconscious dynamics. Freuds ideas were heavily inuenced by those of Fechner (1966/1860), Herbart (e.g., 1824/1961) and Leibniz (e.g., 1973/1714), and of course those thinkers also had their own precursors. The novelty in the current proposal is the assertion that neither a particular locus (e.g., the thalamus, or any other), nor a particular kind of neural activity (e.g., a dynamic core of 40 Hz synchronized ring) is both necessary and sufcient to generate the state of primary consciousness awareness. In what follows I argue that what is required is both a particular locus and a particular kind of neural activity, viz. the thalamic dynamic core. 2. Is conscious experience localized within the brain? The approach taken in this paper nonetheless does raise the issue of whether conscious experience arises from brain activity in some particular brain region, or perhaps in particular neurons, or rather arises from activity that is distributed throughout the brain. This issue has been addressed extensively by Dennett (1991). Dennett argued that the Cartesian Theater of consciousness, the idea that there is a place in the brain/mind where everything comes together with respect to consciousness, a sort of theater where an audience (a homunculus?) watches the play of experience, is fundamentally misguided. Descartes believed that the pineal gland of the brain was the place where brain processes and (nonmaterial) mental processes interact, and thus consciousness is generated. Others have suggested other places in the brain where consciousness occurs, including nearly every part that has been identied. Modern neuroscientic views, however, favor a more distributed process (e.g., Rees et al., 2002), and philosophy has discovered many reasons to reject homunculi, except as convenient rhetorical devices (e.g., Dennett, 1991). Interestingly, psychologists have been reluctant to relinquish the
466
theatrical metaphor. For example, a prominent psychological-level theory of consciousness uses it explicitly while rejecting the idea that the theater is located at a particular place in the brain (Baars, 1997). Crick and Koch (2003) proposed that the front of the brain functions as a homunculus that watches the show put on by the back of the brain (the sensory/perceptual areas), an idea similar to the B-brain watching the A-brain proposed earlier by Minsky (1986) and developed further by Ward (1992). The present proposal retains the idea of distributed computation in the cortex that is espoused by modern theorists, especially Tononi and Edelman (1998a, 1998b; Edelman, 2004; Edelman & Tononi, 2000), and the central importance of synchronized neural activity in cortico-thalamic circuits (e.g., Llins et al., 1998). It gives special importance, however, to thalamic neural activity comprising an active blackboard that displays the results of cortical computations and thereby becomes the locus of the brain activity most closely associated with experience. The proposal that the thalamus is a particularly important locus in the brain involved in generating consciousness also is not new. As Newman (1995) pointed out, views of the locus of the neural correlate of conscious awareness have oscillated around three foci for many years: the cerebral cortex, the reticular activating system, and the thalamus. Peneld (e.g., 1975, p. 19) was perhaps the most extreme proponent of the subcortical view, asserting that The indispensable substratum of consciousness lies outside the cerebral cortex, probably in the diencephalon (the higher brainstem). The thalamus is a major component of the diencephalon (it also includes the epithalamus and the hypothalamus). Penelds ideas were founded upon several major evidential bases: (1) the results of stimulating the brain, especially the temporal lobe, with low-intensity direct current electricity; (2) the results of surgery for epilepsy in which various chunks of brain, especially temporal lobe, were removed; and (3) the structural and functional anatomy of the brain that was known in the early 1970s. More recently, several authors have argued that the thalamic reticular nucleus (TRN) plays a role in consciousness by modulating the local 40-Hz oscillations observed in various parts of the brain via its inhibitory inputs to the dorsal thalamic nuclei (Min, 2010; Newman, 1995). Indeed, abolishing inhibitory interactions among the neurons of the TRN dramatically increases absenceepilepsy-like, low-frequency synchronous oscillations in the dorsal thalamic nuclei (Huntsman, Porcello, Homanics, DeLorey, & Huguenard, 1999), indicating that such inhibition might play a major role in preventing the neural hyper-synchrony that characterizes epilepsy and its accompanying state of unconsciousness. Moreover, the TRN has been implicated in producing the unconscious state seen in absence epilepsy, presumably playing the role of strongly inhibiting thalamic neuron activity under the inuence of cortical excitation (e.g., Steriade, 2005). Another vociferous recent proponent of localizing the critical NCC for the state of consciousness in the thalamus has been Joseph Bogen, the surgeon who developed the split brain operation (e.g., 1995a, 1995b, and see his website for more recent unpublished work). Bogen argued that the intralaminar nuclei of the thalamus form the essential substrate of the state of phenomenal consciousness. His argument is complex, but critical to it is the fact that there are only two places in the CNS where very small bilateral lesions (those involving less than 1 g of neural tissue) abolish the state of consciousness: in the mesencephalic reticular formation and in the intralaminar nuclei of the thalamus. Moreover, the intralaminar nuclei are connected to much of the rest of the brain through diffuse reciprocal connections, making this a candidate for a central clearinghouse or modulator of cortical and subcortical activity. Another proponent of a subcortical locus for a substrate of phenomenal awareness is Merker (2007). He updated Peneld and Jaspers (1954) centrencephalic system proposal and reviewed extensive evidence that the top of the brainstem, and the superior colliculus in particular, forms a system that integrates motivation, the sensory world, and body capabilities to accomplish goal-directed action. His arguments rely on the extensive convergence of inputs from pretty much the entire brain into this region, including to and from parts of the thalamus. Indeed, the thalamus plays an important integrative role in his theory, although the theory does not specify it to be the substrate of consciousness. Merker writes evocatively of how this primitive substrate of consciousness could have evolved when conscious contents were limited to very simple ones because the dramatic forebrain expansion characteristic of mammals had not yet occurred. Among other important facts discussed by Merker (2007) is the remarkable observation that in over 750 operations to cure epilepsy, during which the patient was not anesthetized, Peneld and Jasper (1954) never once observed even an interruption in the continuity, let alone cessation, of a patients consciousness as they removed large chunks of cortex, sometimes even an entire hemisphere. Interestingly, in his later book Peneld (1975) identied the diencephalon with the highest brain mechanism that is directly responsible for consciousness. The part investigated thoroughly by Merker, on the other hand, was termed by Peneld the brains computer and was said to be responsible for sensory-motor integration, as updated and extended by Merker (2007). Both mechanisms acting together were thought to be necessary to explain human behavior because the diencephalon has privileged access to frontal and temporal areas of the cortex, whereas the older system just below this area at the roof of the brainstem has privileged access to sensory and motor mechanisms. It is thus likely that the entire upper brain stem plays a crucial role in human behavior. Although in this paper I emphasize the role of the thalamus in phenomenal awareness, there is little doubt that the integrative system detailed by Merker is critical for efcient, organized goal-directed action, and might also contribute essential components to the substrate of awareness.
3. The thalamus is a miniature map of the cortex One of the most compelling reasons why the thalamus has gured prominently in theories of consciousness is that it represents a central, convergent, compact miniature map of the cortex. Thus, as will be seen in Section 3.1, it is
467
well-positioned in the brain to perform the role of integrating a wide variety of cortical computations as well as gating sensory inputs and integrating both of these with limbic activity from the hypothalamus and other subcortical regions. 3.1. Thalamic anatomy and cortical connections Fig. 1 adumbrates the generally-agreed-upon classical view of the global architecture of the cortico-thalamic and thalamo-cortical circuits (Nieuwenhuys, Voogd, & van Huijzen, 2007; see also Jones (1985, 2007) and Fig. 1.4 on p. 10 and 1.5 on p. 11 of Sherman and Guillery (2006)). The thalamus is parcellated into about 50 nuclei and subnuclei, which do not communicate directly with each other. Rather, each nucleus has reciprocal connections with a specic cortical area, as well as with the thalamic reticular nucleus (TRN). Note the relatively massive projection from each cortical area to a specic dorsal thalamic nucleus (indicated by the thick lines on the left side of Fig. 1) as compared to the much smaller thalamo-cortical reciprocal projection (thin lines on the right side of Fig. 1; Llins & Par, 1991). Such corticothalamo-cortical circuitry is ubiquitous among mammals, and is augmented in the higher mammals by the addition of entire cortico-thalamic units, rather than by elaboration of only cortical circuits. Thus, each new cortical area added by evolution has been accompanied by the addition of its specic nucleus to the dorsal thalamus. Indeed, so close is the relationship between the dorsal thalamic
Fig. 1. Some circuitry of the thalamo-cortical complex as seen in a horizontal slice. Thalamic nuclei are represented by the central colored blobs. Corticothalamic projections are represented on the left, thalamo-cortical projections on the right (with associated color-coded cortical projection areas). The more massive cortico-thalamic projections are represented by thicker lines. Not all nuclei or projections are shown, differentiated, or labeled. Most limbic and other subcortical projections to thalamus are not shown. Note the reciprocal connections between the thalamic nuclei and their associated cortical areas. Note also the lateral projections to the TRN of both cortico-thalamic and thalamo-cortical bers. ATN: Anterior thalamic nuclei; DM: dorso-medial nucleus; LGN: lateral geniculate nucleus; MGN: medial geniculate nucleus; TRN: thalamic reticular nucleus; P: pulvinar nucleus; VN: ventral nuclei. Based on Nieuwenhuys et al. (2007).
468
nuclei and their reciprocally-connected cortical areas that the dorsal thalamic nuclei virtually comprise a seventh layer of the cortex. Moreover, given the recently recognized reciprocal relationship between various thalamic nuclei and many subcortical areas, such as the basal ganglia, the striatum, the amygdala, the cerebellum, etc., the thalamus interacts directly with, or is a target for, nearly every other part of the brain. It is thus possible that it can represent and integrate the outputs of all of the functional areas of both hemispheres of the brain: sensory, cognitive, limbic, and motor (cf. Edelman, 2004). Although this relatively simple picture has changed signicantly during the intervening years (e.g., Jones, 2007), the overall structure of the more modern understanding remains similar to that portrayed in Fig. 1. I will describe some of the important additions to this picture in Section 6. The thalamic nuclei (excluding TRN) contain neurons of several types. About 70% of them are excitatory glutamatergic neurons that interact with cortex and TRN. The other 30% are GABA-ergic inhibitory interneurons that form reciprocal connections with the excitatory neurons within the separate nuclei. The anatomy of the dorsal thalamic neurons is very different from that of the pyramidal neurons, spiny stellate excitatory interneurons, and aspiny stellate inhibitory interneurons found in the various layers of the cortex (Fig. 2). The thalamic neurons have larger cell bodies and more extensive dendritic proliferation, where, along with cholinergic, noradrenergic, and seratonergic inputs from the brainstem, they receive the excitatory axonal synapses from layer 6 of their associated cortical area and from layer 5 of an earlier cortical area, as well as inhibitory synapses from nearby interneurons and from the associated area of the TRN to which they in turn also project (Jones, 2002). Note that the neurons of the TRN (Fig. 2) resemble typical cortical interneurons, and indeed they are all GABA-ergic inhibitory neurons, both in their projections to nearby neurons in the TRN and in their back projections to the thalamic neurons. 3.2. Cortico-cortical connectivity Functional cortical areas interact with each other in a small world fashion via an extensive system of reciprocal axonal connections comprising the white matter of the brain. Some of this connectivity is illustrated in Fig. 3. The corpus callosum, which connects the two hemispheres (not shown in Fig. 3), is the most obvious of these pathways, but there are many others. Fig. 3 also shows that, whereas the primary sensory projection areas of the cortex are relatively small, their unimodal association areas, where higher order computations are carried out and modality-specic information is integrated, are very large, in particular for vision. The multi-modal association areas, where information from all sensory systems is integrated as well as memory storage and retrieval and other kinds of computations are carried out, are also very large. The prefrontal cortex, which receives a massive ow of information from much of the rest of the cortex, comprises the largest single associative region except for the visual association areas. There is, in the cortex, a general ow of information from back to front, but again, all connections are reciprocal so that there are numerous intra-cortical feedback loops as well. Each of the cortical areas (and there are many subdivisions not represented in Fig. 3) is also reciprocally connected to a particular thalamic
Fig. 2. Detail of neurons and their connections in cortex and thalamus. Modied from Steriade (1999).
469
Fig. 3. Cortico-cortical connectivity. Primary sensory projection areas in yellow; unimodal sensory association areas in light blue; multi-modal association areas in pink; motor areas in gray. Short-range pathways in black arrows, longer-range pathways in pink arrows. PFC: prefrontal cortex; MA: motor association; M1: primary motor; S1: primary somatosensory; SA: somatosensory association; TP: tempor-polar; A1: primary auditory; AA: auditory association; PTHA: parieto-temporal heteromodal association; VA: visual association; V1: primary visual. All pathways are reciprocal. Based on Nieuwenhuys et al. (2007). (For interpretation of the references to color in this gure legend, the reader is referred to the web version of this article.)
nucleus as depicted in Fig. 1. Layer VI of each area sends thin bers with many small synaptic boutons to the specic thalamic nucleus from which that area received thalamic projections, whereas layer V projects to the thalamic nucleus associated with a later cortical area, and also to subcortical areas such as the superior colliculus, with large bers having fewer but larger boutons. 3.3. Mumfords proposal One of the ideas that stimulated me to think about the special role of the thalamus in consciousness was that of Mumford. Mumford (1991) proposed that the thalamic nuclei might act as an active blackboard for the cortical areas to which they are reciprocally connected. As mentioned in Section 3.1, functional cortical areas, especially the association areas, send massive descending projections to specic thalamic nuclei, and receive back from them much less massive ascending projections (Fig. 1). Mumford (1991) pointed out that most thalamic nuclei, called higher order by Sherman and Guillery (2006) to distinguish them from the relay nuclei that innervate the primary sensory projection areas, receive the vast majority of their input from the cortical area(s) to which they are reciprocally connected (the remainder being from other cortical and subcortical areas). He asked: What information could these nuclei be processing, since they receive no direct sensory input? To quote Mumford (1991, p. 139): . . . the thalamo-cortical bers convey to the cortex the current picture of those aspects of the world with which that area of the cortex is concerned... (whereas)... The cortico-thalamic bers convey to the thalamus proposed additions and revisions to this picture arrived at by many computations carried out in the cortex, which are integrated in the thalamus via the dendritic arbors of the thalamic neurons. Essentially, these thalamic nuclei could be displaying the results of cortical computations from moment to moment, acting as an active blackboard that is constantly being updated as new information arrives and new computations are performed. To quote Mumford (1991, p. 140) again: . . .the thalamus (is) an active blackboard, i.e., a blackboard which is volatile and continually represents the latest ideas, synthesized from multiple cortical sources. Mumford (1992) later proposed that reciprocally-connected cortical areas (Fig. 3) are usually of the higherlower variety, in the sense that the lower area passes data to the higher area, which in turn passes back to the lower area hypotheses as to the properties, objects, and/or events represented by the data. These exchanges typically settle into a form of
470
adaptive resonance (Grossberg, 2000) whereby any error signals are nulled. This proposal is consistent with the now routine idea that the cortex computes the contents of consciousness (cf. Rees et al., 2002). It is also consistent with the proposal that the cerebral cortex operates as a hierarchical, Bayesian belief propagation network that is constantly updating a current construct (based on sensory, emotional and cognitive input) that functions as a model of the organisms circumstances in the world (e.g., Friston, 2009; George & Hawkins, 2009; Hinton & Sejnowski, 1983). Thus, Mumford proposed specic computational roles for both thalamus and cortex consistent with the thalamo-cortical anatomy and physiology known in 1985 or so. This is a suggestive picture and I will modify and extend it in what follows based on the four evidential pillars I mentioned earlier, one of which is comprised of some new research results concerning the physiology of the thalamus. 4. We are conscious of results not processes The rst evidential pillar of the thalamic dynamic core hypothesis is a foundational idea from the cognitive realm that is remarkable, puzzling, and critical. It is the idea that the contents of consciousness are the results of cortical computations, not the computations themselves. (I use the word computation to mean formally speciable information processing operations, not necessarily mathematical calculations and certainly not necessarily the Boolean bit manipulations done by digital computers. Indeed it is possible that the computations performed by neural systems are best characterized as analog.) This idea has been around for a while in psychology (e.g., Baars, 1995; Newman, 1995; Nisbett & Wilson, 1977) and seems to be generally accepted there. It might seem obvious to philosophers of mind because it is closely related to the idea of phenomenal transparency, in which we can access introspectively the content properties of a conscious mental state but not the carrier (vehicle properties) of that content (e.g., Metzinger, 2003; Moore, 1903). It also seems to be generally accepted by neuroscientists (e.g., Rees et al., 2002). It is consistent with Mumfords (1991, 1992) proposal for the computational roles of the corticocortical and thalamo-cortical loops, and with the idea that many cognitive and most perceptual modules are not cognitively penetrable (Fodor, 1983; Ward, 1992). The evidence for this idea is multifarious, coming both from analyses of experience and from laboratory data. Nonetheless, its implications for the NCC of the conscious state have not yet been fully explored. Perhaps the most dramatic example of the segregation of these functional domains arises from various forms of perceptual failures such as saccadic blindness (e.g., Matin, 1974), inattention blindness (e.g., Mack & Rock, 1998), and its close relative change blindness (e.g., Simons & Rensink, 2005). It is well known that visual consciousness seems to be suppressed during saccadic eye movements greater than 24 of visual angle. Indeed we cannot perceive our own saccades in a mirror! The majority of changes introduced into the visual eld during a saccade are not reported (e.g., Bridgeman, Hendry, & Stark, 1975; Grimes, 1996), including even large obvious ones such as an exchange of the heads of two cowboys in a photograph. We are simply blind to visual stimuli that only occur during a saccade and even to most obvious persisting changes that occur then as well. Visual awareness is actually a series of snapshots of the world that is integrated seamlessly across the 10100 ms durations of the more than 100,000 saccades that occur each day. If this didnt happen we would inhabit a wildly swinging and jerking visual world as the swings of the visual image across the retina during saccades would generate dramatic movement signals (push gently on the outside of your open eye to see this). It was suggested by Helmholtz (1867,1962) that the efference copy of the saccadic eye movement motor signal, or for head or body movements the efference copies of those signals, cancels the perceptual movement information, yielding a stable world except for changes within that world. Gibson (e.g., 1966) pointed out that movement within the image signies movement in the world, and that movement of the image signies movement of the perceiver, and others have proposed other explanations over the years. But it seems incontrovertible that what seems to be a continuous and stable visual world is in fact a construct made from snapshots, with the intervening eye-movement moments edited out. Inattention blindness and change blindness refer to the situation in which people fail to be aware of large changes in the sensory array that, when pointed out to them or when they later notice them, are clear and obvious, regardless of whether or not they occurred during a saccade. In inattention blindness, experimental subjects attention is directed to a particular place in the sensory array and they often fail to notice conspicuous stimuli occurring elsewhere than the attended location. In some change blindness studies subjects are exposed to a dynamic scene and often fail to notice large changes in that scene, such as when the person they are talking to is replaced by a different person (Simons & Levin, 1998). In others, changes are introduced into photographs or other displays in brief temporal gaps between exposures and subjects actively search for the change (e.g., Bayless & Ward, 2009; Rensink, ORegan & Clark, 1997), a process that often lasts for many seconds without success (Fig. 4). In this paradigm the blank gap between exposures of successive frames introduces massive transients that mask the transients that would have been generated by the sought-after change, forcing subjects to rely on their memory of the images from one frame to the next. In order for the change to be noticed the subject must attend to the changed portion of the image in each of two or more successive frames (Rensink, ORegan, & Clark, 1997). In all of these phenomena, information that is clearly present in the sensory array, and is thus processed by sensory and perceptual systems at least, does not seem to make it into consciousness. Consciousness appears to be up-to-date only for the parts of the sensory array that are currently in attention. Other parts of the conscious eld either are updated only when attention rests there or, in the absence of recent attention focus, only reect global features. This implies that our experience of the world is a construct that does not always contain all available information, even when it seems to do so. This is often called the grand illusion of complete perception. Our conscious experience of the world seems complete but it is demonstrably not so. This means that consciousness is a synthetic construct that is somewhat removed from the elementary information processing operations carried out by sensory and perceptual systems. It is a separate type from the processes that construct it, rather than being a
471
560ms
100ms
560ms
100ms
Fig. 4. The icker paradigm for change blindness experiments (based on Bayless and Ward (2009)). Gray panels induce 100 ms-duration icker or gap. After the second icker the sequence repeats. This is an easy example that takes on average less than 2 repeats to detect the change. More difcult cases, for example with 64 rectangles, can take over 11 repeats. Changes in naturalistic scenes can require many seconds for detection.
blend of their activities. If it were the latter, it would have to contain all of the information that is being processed, and it clearly does not do so. Another convincing demonstration of this idea involves asking people to do Sternbergs (e.g., 1970) short-term memory scanning task. In this task subjects are required to say whether a digit, say 6, is a member of a set of previously presented, to-be-remembered digits, say 3, 4, 7, 9. Sternberg (1970) observed (p. 22): Perhaps because of its high speed, the scanning process seems not to have any obvious correlate in conscious experience. Subjects generally say either that they engage in a self-terminating search, or that they know immediately, with no search at all, whether the test stimulus is contained in the memorized list. When asked to do the task as quickly as possible as in Sternbergs experimental situation, most people report an experience of simply inquiring whether the target digit is among those in the memory set, and experiencing the answer immediately. They do not have the experience of checking the target digit against each of the memory set digits in turn, and, regardless of whether a match is found, continuing this process until all digits have been checked (serial, exhaustive search, the model advocated by Sternberg, 1970), or of serially decoding the memory set digits and then comparing the representations of the decoded digits to that of the target in parallel (e.g., Corballis, 1979; Ratcliff, 1978). Yet, according to the common experimental determination of the identical linear dependence of the time to make either yes or no answers on the number of digits in the memory set (both relations have slope approximately 2535 ms per item), something like this is what their brains must be doing in the standard paradigm. When trying to answer as quickly as possible, people simply do not have access consciousness for, and apparently do not experience at all, the cortical computations that result in their experience of knowing the answer, although of course they do have both for the answer itself. Interestingly, if they wish to, people can consciously image each member of the memory set individually and then compare it to the target, stopping when and if a match is found, in effect simulating what they think they are doing in the speeded case (cf. Hesslow, 2002; Kahneman & Tversky, 1982). However, this results in slowing the process down to the speed of conscious operations, about 4 Hz (250 ms per item; see Crovitz, 1970), 10 times slower than the usual speed of about 3040 Hz. More examples can be piled up. For example, we do not experience and cannot report the memory codes used to retrieve material from long term memory, we only know whether we can or cannot retrieve the material. When asked whether we know someones telephone number, we do not consciously compare the target name to the names of all those in our memorial phone book, we simply either produce the number (or sometimes know we know it but cannot nd it) or we know we do not know it. Do you, for example, know Elizabeth Taylors telephone number? Do you know your mothers telephone number? We have no access at all to the lower levels of perceptual processing, including those that give rise to dramatic illusions such as those involving converging lines (e.g., Mller-Lyer illusion), the moon illusion, or motion aftereffects. Problem solutions just pop into our heads, emotions come from nowhere, complex verbal utterances pour forth with no effort and certainly with no consciousness of the complex processes that must be occurring to create the surface structure from the deep structure, not to mention the details of phoneme coarticulation. Emotions, feelings and moods arise with no insider information about how the brain might be giving rise to them. We can often identify external causes for these states, but we have no experience of how those external causes get translated into our feelings. We get dumped by a spouse and feel sad, we get insulted by a co-worker and feel mad, we win a prize and feel happy. But we dont experience the brain processing giving rise to these feelings we simply feel them. It can even be argued that our experience of free will, the feeling that we can act arbitrarily whenever and however we wish, is an illusion arising from the fact that we have no conscious awareness of the neural computations that are giving rise to our often apparently whimsical decisions about what to do at a given moment (e.g., Crick, 1994; Wegner, 2002 and cf. Libet, 2004). It would seem that if consciousness arises directly from cortical activity, we would be aware of more of the details of the computations, since it is the cortex that is doing them, or at least that perceptual and cognitive modules would all be penetrable by consciousness. For example, consciously adjusting memory retrieval codes would seem to be a useful ability to have. Yet, short of thinking of different things we seem to have no mechanism to do this, even though consciously intended explicit memory may be a uniquely human skill (Donald, 1995). The functional segregation between computing the
472
preliminaries of conscious experience and the actual experiencing of the higher-order synthetic construct based on those preliminaries suggests that there might be a similar segregation between brain loci or circuits where the preliminaries happen and those where the synthetic construct based on them, the experiencing, happens. Sections 58 of the paper argue that the two processes happen in different loci, although the possibility of different circuits in the same general locus, the cortex (e.g., Laberge & Kasevich, 2007), cannot be ruled out on the basis of available research. Why should such a functional segregation evolve? As mentioned in Section 1.1, conscious processing is integrative, involving the interaction of information from several sources, including sensory systems, memorial systems, emotional systems, linguistic systems, reasoning, decision-making, and so forth, although of course not necessarily all of these at once. This interaction gives rise to what appears to be a seamless stream of consciousness (James, 1890) that portrays an environment with affordances for action at a causal level appropriate to those actions, at least for humans and by extension for any animal with a similar primary consciousness. Thus, both the computational operations involved in producing such representations, and any environmental causal levels below or above those at which an organism acts, are irrelevant to moment-to-moment adaptive action in the organisms niche (although they might be relevant to long-term projects or constructions of artifacts by humans). It seems reasonable that evolution would select a level of environmental representation appropriate to the niche occupied by an organism, because that is the level at which the organisms actions will operate to produce tness-relevant outcomes. 5. The thalamus is a common locus in vegetative state and general anesthetic action In searching for a brain locus critical for experiencing it is useful to focus on two states in which experiencing seems not to occur. The vegetative state resulting from brain damage and the state induced by sufcient doses of general anesthetics have in common that their subjects appear to lack primary consciousness. In this section, the second pillar, I argue that they also share a brain locus of inaction: the thalamus. The two states are different, however, in that general anesthetics also induce deep sleep, whereas people in the vegetative state display a sleep-wake cycle both in the EEG and in the sense of eyeopening and movement. Fig. 5 shows where these two states, and some others, fall in the space composed of sleep/arousal and primary consciousness axes. I have separated the sleep/arousal continuum from that of consciousness, because, among other evidence, the vegetative state clearly differentiates them, viz. the comment of Jennett and Plum (1972) in the article that rst named the vegetative state: It seems there is wakefulness without awareness. (p. 734). This means that vegetative patients show no evidence whatsoever of responsiveness to external stimuli, for example eye-tracking moving objects or responding to speech. Note that REM sleep/dreaming is paradoxical in that some brain regions, including the thalamus, are quite highly activated in this state whereas others, including motor cortex, are suppressed. REM sleep has been argued to be both a primary conscious state in adults and a protoconscious state that functions to prepare the brains of developing mammals and birds to encounter the world (Hobson, 2009).
Fig. 5. Some states of consciousness displayed as loci in a two-dimensional space with sleep/arousal and primary consciousness as the axes. The continuum of normal states runs from normal deep sleep to the normal waking state (with a detour through dreaming), whereas the minimally conscious state and the vegetative state remain at the low end of the consciousness axis while ranging from one end to the other of the arousal/sleep axis.
473
Those in a minimally conscious state (Giacino & et al., 2002) do display some, inconsistent, evidence of consciousness, such as responding occasionally to external stimuli, although they often also seem vegetative. Of course they too display a sleep-wake cycle. Those in a coma do not respond in any way at all except for some spinal or brain stem reexes and show no evidence of a sleep-wake cycle, even on EEG. The vegetative state is of particular interest for the present argument because of the apparent lack of consciousness in an awake body. The evidence points to the thalamus as the common locus of severe brain damage in these patients, and thus to the thalamus as a critical locus for conscious awareness. Similarly, various general anesthetics have a variety of effects on the brain, including a general reduction in brain metabolism. But all examples studied to date dramatically reduce activity in the thalamus and in the brain stem reticular activating system, thus inducing both unconsciousness and deep sleep, indeed a state similar to coma, while leaving cortical activity reduced but resembling that seen in deep sleep. 5.1. Brain damage and consciousness Possibly the most well-studied vegetative patient has been Karen Ann Quinlan, who lost consciousness after a cardiopulmonary arrest. An autopsy of her brain revealed extensive scarring in the thalamus but relatively little damage to the cortex (Kinney, Korein, Panigrahy, Dikkes, & Goode, 1994). Jellinger (1994) described a similar case and stated I can conrm their (Kinney et al.s) suggestion that the thalamus is critical for cognition and awareness but less essential for arousal. . . (p. 1378). More recently several similar studies of the brains of vegetative patients have been published. In one of these, Adams, Graham, and Jennett (2000) compared vegetative patients with traumatic brain injuries (e.g., trafc accidents, falls, assaults) to those whose injuries were non-traumatic (e.g., hypoxia from cardiac arrest or other causes, infection). Of those with a nontraumatic injury there was damage to thalamus (especially the anterior and dorsomedial nuclei and with relative sparing of lateral and ventral nuclei) in 100% of cases. Of those with traumatic injury there was damage to either subcortical white matter or thalamus, or both, in 100% of cases. Traumatic injury often results in diffuse axonal injury (DAI) which can lead to extensive damage to lateral and ventral thalamic nuclei from retrograde degeneration, sometimes with relative sparing of anterior and dorsomedial nuclei, pulvinar, centromedian nuclei and lateral geniculate nuclei. This usually takes about three months to occur, whereas the thalamic damage arising from hypoxia is usually apparent shortly after the insult. Interestingly, the thalami of traumatically-injured vegetative patients show much lower neuronal integrity, as indicated by the thalamic N-acetyl aspartate:creatine ratio measured by proton magnetic resonance spectroscopy, than do those of severely or mildly disabled patients who had regained consciousness after being in a vegetative state (Uzan et al., 2003). In yet another study, Jennett, Adams, Murray, and Graham (2001) compared the brains of patients who in life had been in vegetative state to the brains of patients who had been severely cognitively disabled, but conscious. Vegetative state was associated with either severe DAI (7/35 cases) or thalamic damage (28/35 cases) and/or both (18/35 cases), whereas severely disabled patients brains often displayed neither type of damage (15/30). From the data reported by Jennett et al. (2001) I computed the v2 for the simple comparison of vegetative state vs. severe disability and thalamic damage vs no thalamic damage to be 16.5, df = 1, p < 0.0001. Finally, damage to the dorso-medial nucleus in particular is predictive of the eventual outcome of traumatic brain injury (Maxwell et al., 2004). In this latter study, loss of neurons from the dorso-medial nucleus was associated with the worst outcomes, and vegetative patients had suffered the most loss of nucleus area and of neurons (approximately 30% smaller/fewer) relative to normal controls. Vegetative patients had also suffered signicant loss of area and neurons in the ventral posterior nucleus, which could explain their lack of response to tactile stimuli. Moderately disabled patients, on the other hand, had suffered only a minor loss (<5%) of dorsomedial and ventral posterior neurons. Thus it seems that vegetative state could arise in two important ways, either from direct extensive damage to the thalamus usually caused by hypoxia, infection, or infarct, or from extensive diffuse thalamic damage arising from degenerative changes caused by DAI resulting from traumatic cortical damage. Either way, the neuropathology of the vegetative state seems to point toward the thalamus as critical for consciousness, although not for a sleep-wake cycle. 5.2. Anesthetics and consciousness Another body of direct evidence for a key role of the thalamus in primary consciousness comes from studies of the loci of action of general anesthetics, which cause loss of such awareness when applied in sufcient dosage. In one of the most telling of these, Alkire, Haier, and Fallon (2000) used PET to ascertain which brain areas were suppressed by both of two very different but commonly used general anesthetics. They found that only the thalamus and the midbrain reticular formation were suppressed by both agents. Because of the convergence between these results and studies of the role of the thalamus in sleep (e.g., Steriade, McCormick, & Sejnowski, 1993), Alkire and colleagues (2000) proposed that a hyperpolarization block of thalamocortical neurons (those in the relay nuclei) is the common mechanism of unconsciousness in general anesthesia (Fig. 6). This implies that activity of these thalamic neurons is critical for normal consciousness. Such a hyperpolarization block would not simply suppress ring in these neurons, however, it also would disrupt their ability to re synchronously with each other and with cortical neurons. In a more recent review of the literature on anesthetics and consciousness, Alkire and Miller (2005) provided even more evidence that the thalamus is a critical and common site of anesthetic action (Fig. 7). Eight different general anesthetics agents have in common mainly that they suppress metabolism in the thalamus and brainstem reticular formation. Moreover, the sensory cortex has been shown to be responsive to stimuli even under large doses of anesthetics such as desurane, although sustained responding is impaired, presumably because of interruption
474
Fig. 6. The thalamic switch hypothesis of Alkire et al. (2000). (a) Active cortico-thalamic, thalamo-cortical, reticulo-thalamic and reticular activating system connections during normal consciousness. (b) Dotted lines represent connections inactivated by a general inhalational anesthetic rendering a person unconscious. Elsevier. Reprinted from Alkire et al. (2000) with permission from Elsevier.
Fig. 7. The thalamus is the common locus of action of eight different general anesthetic agents. Elsevier. Reprinted from Alkire & Miller, 2005 with permission from Elsevier.
of thalamo-cortical and cortico-cortical loops (e.g., Hudetz, Vizuete, & Imas, 2009). Interestingly, and closely related to the role of midline nuclei in vegetative and minimally conscious states, blockade of potassium channels in the central medial thalamic nucleus of rats reverses desurane anesthesia (Alkire, Asher, Franciscus, & Hahn, 2009). Finally, another recent summary of the effects of anesthesia on consciousness, although arguing against the hypothesis of a thalamic switch, nonetheless concluded that the effects of anesthetics on the thalamus imply that thalamic activity might
475
be comprised of a readout of global cortical activity (Alkire, Hudetz, & Tononi, 2008). This idea is consistent with the present proposal that the cortex computes the contents of consciousness, including those related to self, whereas the thalamus is the critical locus for the experiencing of those contents. Alkire et al. (2008) also review data that conrm that the role of the thalamus in cortical arousal is not unique, with much cortical arousal arising from intra-cortical sources and being in some cases decoupled from activity in the thalamus. 6. Computational roles of cortex and thalamus It is clear from the data reviewed in Section 5 that the thalamus must be one of the critical substrates involved in the generation of the conscious state. An even more intriguing possibility is that the thalamus itself, or some particular nuclei in it, is the critical site of a dynamic core of neural activity that gives rise directly to the state of primary phenomenal consciousness. That is the possibility I wish to pursue here. Two further evidential pillars support this notion. The rst, described in this section, extends the proposals by Mumford (1991, 1992; cf. Harth, 1983) described in Section 3.3 regarding the computations performed by cortico-cortical and thalamo-cortical circuits. 6.1. The core and the matrix An important modication of the earlier picture of cortico-thalamic interaction arises from Jones discovery of two types of thalamic neurons (Jones, 2009) and the resulting improved characterizations of the non-sensory thalamic nuclei and their targets and inputs. The two major types are characterized by their expression of calcium-binding proteins, the core neurons expressing parvalbumin and the matrix neurons expressing 28-kDa or 29-kDa calbindin, and by their very different types of projections to their targets. All dorsal thalamic nuclei contain a diffuse matrix of calbindin-expressing cells, which is overlaid in some nuclei by a core of specically-projecting parvalbumin-expressing neurons. These core neurons occur mostly in the sensory and motor relay nuclei, such as the lateral geniculate nucleus (LGN) that relays visual information from the retina, the medial geniculate nucleus (MGN) that relays auditory information from the cochlea, and the ventral posterior nuclei (VPL and VPM) that relay tactile information from the skin (Fig. 8). They project to layer IV stellate neurons and layer III, V, and VI pyramidal neurons in sensory- or motor-specic cortical areas, as in the classical picture. The matrix neurons, on the other hand, project diffusely to stellate neurons in the supercial layers of several related cortical areas, primarily in
Fig. 8. Relative densities of core (right side) and matrix (left side) relay neurons in the thalamus indicated by densities of dotted textures. Core neurons are more numerous in the primary sensory nuclei (VPM and VPL, ventral posterior lateral and medial, LGN, lateral geniculate nucleus, and MGN, medial geniculate nucleus not shown). Matrix neurons are more numerous in nuclei that interact with frontal areas. John Wiley & Sons. Reprinted from Jones (2009) with permission from Wiley.
476
non-sensory nuclei and especially in frontal areas. Both types of thalamic neurons receive back projections from layer V pyramidal neurons and the core neurons also receive back projections from layer VI pyramidal neurons (Fig. 9). Jones (2001, 2002, 2009) proposed that the core neurons serve the function of relaying information within specic sensory and motor pathways, whereas the matrix neurons serve the function of binding together the activities of thalamus and cortex, thus promoting thalamo-cortical synchrony. He emphasized the two major modes of action in the thalamo-cortical circuitry: burst mode promoting drowsiness and sleep and tonic mode promoting wakeful consciousness and action. The former occurs when the brainstem arousal system is quiescent and the core and matrix neurons are suppressed by inhibition from the TRN. The latter occurs when the brainstem arousal system is active and reticular inhibition is at its weakest (although not absent). During the tonic mode, thalamo-cortical synchronization, as well as corticocortical synchronization, are promoted by the binding inuence of the matrix neurons. In this scheme the TRN, in combination with the brainstem arousal system, plays the role of determining whether the thalamus will promote thalamo-cortical synchronization at 40 Hz (consciousness) or at much lower frequencies, in the delta (23 Hz) range (sleep). Some recent work is beginning to address directly the functional properties of the core and matrix thalamic neurons. For example, Zikopoulos and Barbas (2007) studied pathways linking prefrontal cortex and the ventral anterior thalamic nucleus in rhesus monkeys. The ventral anterior thalamic nucleus in monkeys is involved in linking prefrontal, premotor, and motor cortices. Zikopoulos and Barbas (2007) characterized both the paralbumin-calbindin expressing properties of the cortical and thalamic neurons involved as well as the driving (large, ionotropic) or modulatory (small, metabotropic) character of their synapses. They found one very dense circuit in which thalamic parvalbumin (core) expressing neurons terminated in large driving boutons in the middle layers of the prefrontal target regions, and were in turn innervated by small metabotropic boutons originating in cortical layer VI. A separate circuit originated from thalamic calbindin-expressing (matrix) neurons and projected more diffusely to supercial cortical layers in several cortical columns, being in turn innervated by layer V ionotropic synapses. One interpretation of these circuits would be that the thalamic core neurons drive the middle layers of prefrontal cortex and are modulated by the layer VI projections, whereas the thalamic matrix neurons modulate the supercial layers of the prefrontal cortex and are driven by layer V projections. Other recent work amplies the differences between the corticothalamo-cortical circuitry in sensory-motor regions and those in the higher association areas and their associated thalamic nuclei. In one study, for example, Zikopoulos and Barbas (2006) described projections from prefrontal cortex to the thalamic reticular nucleus (TRN) that differ greatly from those arising from primary sensory areas to their thalamic associates. In the latter case, layer VI cortical neurons project specically to the TRN and the thalamic relay nuclei, and layer V cortical neurons project only to the next-higher thalamic nucleus and also to motor centers (Sherman & Guillery, 2006). From prefrontal cortex, however, both layer VI and layer V cortical neurons project to the TRN as well as to neurons in their associated thalamic nuclei, and do so with both large, ionotropic and small metabotropic synapses. Thus it does seem that frontal cortex, and its associated thalamic nuclei, are ideally suited both to direct attention to salient events and thoughts (e.g., Zikopoulos & Barbas, 2006), and to construct and display a synthesis of current circumstances (cf. Pico, 2002). In the absence of denitive information about the exact corticothalamo-cortical projections for all thalamic nuclei, and in the light of the newer data available, it seems reasonable to attempt an interpretation that integrates what we do know so
Fig. 9. Jones model of the core and matrix thalamo-cortical circuits. Roman numerals refer to cortical layers; RTN = thalamic reticular nucleus. Layers II and IV neurons are stellate cells whereas layers III, V, and VI neurons are pyramidal cells. Arrows indicate direction of action potential movement. John Wiley & Sons. Reprinted from Jones (2009) with permission from Wiley.
477
far. In this interpretation, the thalamus could still function as an active blackboard, but it would echo back to the cortex an integrated impression of more diverse inputs, given that each higher-order thalamic nucleus receives inputs from layer VI of its own associated cortical area and layer V of perhaps several other cortical areas. Thus, for example, the thalamic neurons associated with modality-specic association areas of the cortex would be integrating where information from one, lower cortical area with what information from another, higher cortical area, and coincidentally sending that integrated information back to their own area. This is consistent with Kanwishers (2001) argument that both dorsal and ventral stream information must be integrated for consciousness to take place. Moreover, as mentioned earlier, it is unlikely that this scheme has the same function in multi-modal sensory and in non-sensory domains. For example, there may be no strictly ascending hierarchy of cortical and thalamic areas once we proceed beyond the rst several levels of each sensory system. Certainly this is the case in the visual system (e.g., van Essen & Deyoe, 1995), where the visual pathways form an ever more tangled web as we ascend from the retina to the associative cortices. Thus, in addition to providing a synthetic perceptual construct, the diverse information integrated in thalamic nuclei associated with multi-modal association areas, and in particular in those associated with frontal and prefrontal cortex, would include that of a non-sensory nature as well, encompassing different facets of information computed by frontal or associative circuits, for example, a perceptual construct of a face accompanied by a memory of having seen the face before and the associated persons name, occupation, social standing, and personal relationship to oneself. It is possible that the integration of neural activity giving rise to primary awareness takes place mostly in the dendrites of the matrix neurons, as these are more common in the non-sensory and non-motor nuclei, rather than in those of the core neurons. This would help to explain why the detailed activities of primary and secondary sensory cortical areas do not enter primary awareness, although their outputs are certainly necessary for awareness of specic sensory content. Juxtaposing Mumfords (1991, 1992) proposals about the role of the thalamus with some others made about consciousness also is suggestive. For example, Baars (e.g., 1997, 2002) and Dehaene and colleagues (e.g., Dehaene et al., 1998; Gaillard et al., 2009) global workspace theory of consciousness emphasizes the role of consciousness as providing a blackboard where various parallel processes occurring in the brain can be integrated and can affect each other. Navon (1989) proposed that attention was essentially a process whereby various perceptual and cognitive processes could be coupled or decoupled to each other in a kind of active blackboard. It seems a short step to the proposal that active-blackboard-like activity in the thalamus is conscious awareness. 6.2. The thalamus and attention Attention and consciousness are very closely related (e.g., James, 1890) although not necessarily the same process (e.g., Koch & Tsuchiya, 2006; Lamme, 2003). A critical role of the thalamus in attention focusing has also been suggested by many authors (e.g., Newman, 1995). For example, Laberge (1995) implicated the pulvinar nucleus of the thalamus in the ltering action of attention. Crick (1984) suggested that the thalamic reticular nucleus (TRN) is the locus of the attention spotlight. He proposed a mechanism involving short-term synaptic plasticity in which the TRN created temporary bindings of the contents of diverse cortical areas constituting the contents of the focus of attention. Although the exact mechanism by which this was supposed to happen has not been conrmed, the proposal that the TRN is involved in modulating the ring of thalamic neurons is now widely accepted (e.g., Sherman & Guillery, 2006) and could be involved in the mechanism implementing temporary binding. As Fig. 1 shows, the neurons of the TRN also project directly with inhibitory connections (GABA-ergic) to all of the nuclei of the dorsal thalamus. Both the cortex and the thalamic nuclei project to the TRN with excitatory (glutamatergic) connections, and the TRN neurons themselves interact locally in an inhibitory manner via GABA-ergic connections. Exploiting this latter fact, Suffczynski, Kalitzin, Pfurtscheller, and Lopes da Silva (2001) proposed a center-surround model of focal attention mediated by the lateral inhibitory interactions of neurons in the TRN, and Min (2010) proposed that the TRN produces primary consciousness by modulating synchronization in thalamo-cortical networks. The TRN is organized in a mosaic of sectors that connect with one or more functionally-related cortical areas and their associated thalamic nuclei (e.g., Crabtree, 1999). The sensory sectors contain topographic maps that mirror the organization of the associated cortical and thalamic regions. They form open loops with their associated thalamic nuclei and appear to receive modulatory inputs from associated cortical regions. Because of the specicity of their connections it is thought that the sensory sectors of the TRN exert only local effects, perhaps enhancing transmission of salient information as postulated by Suffczynski et al. (2001). The higher-order sectors have been less studied, but it is known that they lack the detailed topographic organization of the sensory sectors, and so probably only exert global effects on activity in these nuclei. Attention gating involving the TRN has been conrmed both during classical conditioning (McAlonan, Brown, & Bowman, 2000) and in visual perception of simple patterns (McAlonan, Cavanaugh, & Wurtz, 2008). In the latter study, LGN activity was greater, and TRN activity in its associated visual sector was less, when a stimulus appeared in the attended receptive eld relative to when it appeared out of that receptive eld, and this within the rst 2575 ms after stimulus onset. So attention appears to inhibit TRN response, in order to enhance LGN response, to an attended stimulus eld. Zhang and Jones (2004) pointed out that reciprocal inhibition between TRN neurons is important in promoting spindlefrequency oscillations in the thalamo-cortical loops, and thus the sleep state. They also suggested that although reciprocal inhibitory connections within the TRN are mostly local, the thalamo-cortical system can recruit the entire TRN though these connections through either cortical or thalamic input. Nonetheless, although broad stimuli elicit widespread inhibition in the TRN, focal stimuli initiate oscillations that persist for some time, indicating that intra-reticular inhibition could . . .restrict
478
intrathalamic activity to particular spatiotemporal patterns to allow focal recurrent activity that may be relevant for normal thalamocortical function while preventing widespread synchronization as occurs in seizures. (p. 1735, Sohal, Huntsman, & Huguenard, 2000). Moreover, interactions can occur between modality-related thalamic relay nuclei via their mutual connections to the TRN (Crabtree, Collingridge, & Isaac, 1998). Such interactions could be particularly important in allowing thalamic sensory relay neurons responding to a salient stimulus to inuence the activity of higher order neurons in the same modality, and even to inuence a multi-modal salience map (see later). Indeed, Newman, Baars, and Cho (1997) proposed that attentional gating of information ow from specialized processors into the global workspace of consciousness was accomplished by just such a TRN-centered system in the thalamus. Although the TRN is the origin of most of the inhibitory inputs to the thalamic neurons, recent work has discovered two additional sources of such input, at least in rats. One of these is the zona incerta, a nearby region of the diencephalon that receives collateral cortico-thalamic inputs from layer V but not thalamo-cortical ones (Barth, Freund, & Acsdy, 2002). These inputs target zones of the thalamus that are dense with matrix cells, especially in higher-order zones. The zone incerta also projects to brainstem centers, such as the superior colliculus, involved in sensory-motor integration. Unlike the TRN, with its excitatory-inhibitory loops, the zona incerta exerts its inhibitory inuence without thalamic feedback and with the possibility of affecting widespread thalamic and cortical areas because of its targeting of diffusely-projecting matrix neurons. It has been shown that the zona incerta gates sensory inputs to the thalamus under the inuence of cholinergic inputs from the brainstem that reect behavioral state (Trageser & Keller, 2004; Trageser et al., 2006). Yet another GABA-ergic inuence, restricted to higher-order thalamic neurons, is exerted by the neurons of the anterior pretectal nucleus (APT; Bokor et al., 2005). Interestingly, and in contrast to those from the zona incerta, the APT projects to higher-order thalamic neurons in a focal manner, suggesting that it exerts control over specic neural groups. It is reciprocally connected to the zona incerta and also receives input from layer V of the cortex and projects to brainstem centers such as the superior colliculus. In addition the APT seems to have been evolutionarily conserved, appearing homologously even in reptiles. Together the two centers form a gating system that acts in parallel to the TRN system in controlling the activity in the higher-order thalamic neurons but without thalamic feedback. This system, if it exists in some form in humans, could form a mechanism whereby mood or emotion acts to focus attention on particular kinds of state-consistent sensory input. The pulvinar nucleus appears to play a special role in attention orienting and selection (e.g., Laberge, 1995). It can be divided into several subnuclei that interact differently with cortical and subcortical regions, viz. the inferior, lateral, oral, and medial subnuclei. The more ventro-lateral parts (inferior, and ventral part of lateral) are dedicated to vision, with extensive reciprocal connections to the visual cortices and input from supercial superior colliculus, whereas the more dorso-medial parts (medial, and dorsal part of lateral) seem to be more related to attentional focusing, with connections to orbito-frontal, parietal, temporal and cingulate cortex and to the amygdala, and input from intermediate layers of the superior colliculus (e.g., Grieve, Acua, & Cudeiro, 2000). The oral pulvinar is a polysensory section, with connections to parietal and temporal cortices as well as to visual cortices. Shipp (2004) suggested that, on the basis of such recent anatomical and physiological evidence, the pulvinar nucleus contains a salience map that could coordinate neural activity directed toward locations in space within multiple cortical visual spatial maps. Clearly, given the polysensory connections that also exist, the pulvinar could coordinate attention focusing across sensory domains as well. Moreover, it is ideally situated to integrate bottom-up orienting, driven either by sensory intensity or by subcortical inputs such as from the amygdala that would signal danger, with top-down orienting, driven by goals and context and computed by frontal and other association cortices. Although no generally-agreed-upon, complete, brain-based model of attention exists at this time, it seems possible that both TRN and pulvinar nuclei could be involved in implementing attention, in concert with several other important parts of the brain, including superior colliculus (Lovejoy & Krauzlis, 2010), posterior parietal cortex, and frontal cortex. It is likely that all of these areas cooperate in this task, integrating information from limbic, action, and perceptual-cognitive systems to focus processing on the most important aspects of the current situation. Attention thus critically affects what is represented in consciousness but is not the same as consciousness. We are conscious of more than we are attending to at the moment and also have access to that content for report. But the contents of currently-unattended consciousness are likely not to be up-to-date and also tend to be somewhat coarse. Attention essentially functions to update the conscious construct with new information from moment to moment, making sure it contains the most relevant information available. Inattention or change blindness thus arises because parts of the construct get stale, and do not reect recent changes because attention has not been directed to them recently.
7. Consciousness and neural synchrony: the dynamic core In Section 6 the role of the thalamic matrix neurons in promoting cortico-thalamic and cortico-cortical synchronization at different oscillation frequencies was mentioned. This role is important because there is considerable evidence that synchronization at the higher frequencies is present for neurons representing perceptual-cognitive and emotional contents that are in consciousness, and absent for those representing contents not in consciousness. Thus, the nal evidential pillar on which the thalamic dynamic core proposal rests is the relationship between consciousness and synchronous neural activity (e.g., Edelman & Tononi, 2000; Engel & Singer, 2001; Melloni et al., 2007; Rodriguez et al., 1999; Srinivasan, Russell, Edelman, & Tononi, 1999; Tononi & Edelman, 1998a, 1998b; Ward, 2002, 2003).
479
7.1. Neural synchrony and consciousness Signicant direct support for a dynamic core of synchronous neural activity as an important neural correlate of phenomenal awareness comes from several MEG imaging studies of the binocular rivalry paradigm using ickering, frequencytagged stimuli (Cosmelli et al., 2004; Srinivasan et al., 1999; Tononi, Srivinivasan, Russell, & Edelman, 1998. Tononi et al. (1998) found that the spectral power at the driving frequencies was modulated over the entire brain, not just in the occipital cortex, according to whether the stimulus ickering at a particular frequency was seen or not seen; the peak spectral power was 5085% lower when the relevant stimulus was not seen than when it was seen. Srinivasan et al. (1999) not only replicated these results but also showed that there was signicantly greater coherence between the signals from widely separated MEG sensors when the stimulus ickering at a given frequency was consciously seen than when it was suppressed, and thus not seen. Finally, Cosmelli et al. (2004) found that the inter-regional phase synchrony between those (locally coherent) brain regions responding to a 5-Hz ickering stimulus was dramatically greater when that stimulus was being seen than when it was suppressed (and a rivaling, non-ickering, stimulus was being seen; Fig. 10). I want to emphasize that the locally coherent responses did not go away when the ickering rings werent in phenomenal awareness. The local cortical areas that responded to the ickering rings still did so coherently even when the face was in phenomenal awareness. What did go away when the face appeared in consciousness, and returned when the ickering rings appeared there, was the inter-regional synchrony at the icker frequency of the rings. Thus, inter-regional phase synchrony of neural activity in diverse cortical areas is closely associated with the appearance in phenomenal consciousness of the perceptual objects that neural activity represents. Most recently, studies in my own laboratory have demonstrated that endogenously-generated gamma-band synchronization is a signature of primary perceptual consciousness. Following a demonstration by Doesburg, Kitajo, and Ward (2005) that a burst of brain-wide gamma-band synchronization precedes switches of percepts in binocular rivalry of even stable
Fig. 10. A composite of gures from Cosmelli et al. (2004) binocular rivalry experiment. The face stimulus rivaled with the 5 Hz expanding-ring stimulus, giving rise to uctuating phase locking between widespread cortical areas that oscillated at 5 Hz with the ring stimulus. Phase-locking among these areas at 5 Hz was greater when the 5 Hz rings were being seen than when they were suppressed (and the face was seen; graph). Reddish splotches on the brain images represent cortical areas that oscillated at 5 Hz; black lines between these areas represent statistically reliable phase-locking relative to baseline. Horizontal series of brain images represent individual time points (black dots) in Section 3 of the graph, illustrating a transition from face to rings and back to face over the series. Vertical series of brain images are different sections of brain of one subject with optimal phase-locking (while rings were seen) portrayed as black lines. Elsevier. Figure parts from Cosmelli et al. (2004) used with permission from Elsevier. (For interpretation of the references to color in this gure legend, the reader is referred to the web version of this article.)
480
visual images, Doesburg, Green, McDonald, and Ward (2009) discovered that such synchronization occurs between specic areas in a consciousness-associated network in the cortex (Fig. 11). This network contains both frontal and parietal sources, and includes task-specic regions as well when the percept switches. Moreover, during perceptual rivalry and even in periods of stable dominance by one or the other of the rivaling percepts, both gamma-band amplitude in these brain regions and gamma-band synchronization between them are modulated by the phase of ongoing theta-band oscillations. The theta oscillations in turn are phase locked to each other, although most are apparently generated separately in each region. Thus, the cortical areas that are active during and between switches of conscious percepts in binocular rivalry participate in a thetamodulated gamma-synchronous network, a synchronous rhythm of consciousness. Finally, similar results have been found in studies in which the brain activity resulting from presentation of masked stimuli is compared between trials when the masked stimulus was visible and those when it was invisible (e.g., Melloni et al., 2007). In the study by Melloni et al., briey-presented words were masked either by a weak mask (visible condition) or by an intense mask (invisible condition) in a delayed match-to-sample task, as well as in two unmasked control conditions, while the EEG was recorded. P3a event-related potentials occurred to the weakly-masked but visible words but not to the invisible words. But more relevant here, signicantly increased gamma-band phase-locking occurred widely over the scalp from around 40 ms to 182 ms after sample word presentation, and then lateralized to the left side in accordance with the language-related task from 182 ms to 333 ms, only for the visible condition. Thus, perceptual awareness of a masked word in this study was closely associated with widespread synchronization in the EEG. 7.2. Mechanisms of neural synchrony Neural synchrony probably plays many roles in implementing cognitive processes in the brain in addition to implementing the dynamic core, and accomplishes this through several mechanisms (e.g., Ward, 2003). Suggested additional roles include: sensory binding and integration (e.g., Singer, 1999); memory encoding, consolidation and retrieval (e.g., Buzsaki & Draguhn, 2004); temporal coordination (Bressler & Kelso, 2001); temporal coding of novel feature relations (Singer,
Fig. 11. The cortical consciousness network at 3545 Hz during binocular rivalry. DLPFC = dorsolateral prefrontal cortex; SFG = superior frontal gyrus; PreCG = precentral gyrus; PreC = precuneus; ITG = inferior temporal gyrus. Reprinted with permission from Doesburg et al. (2009).
481
1999). These functions are accomplished by one or more of several mechanisms, including the fact that groups of neurons ring synchronously at around 40 Hz have greater downstream effect on their targets (e.g., Singer, 1999), and that groups of neurons ring asynchronously at 10 Hz avoid spike adaptation in their targets (Fries, Reynolds, Rorie, & Desimone, 2001). Most importantly, in the present view conscious contents are computed by both sustained and transient communication among various and ever-changing groups of neurons. Recently Fries (2005) pointed out that such neuronal communication is best accomplished through neuronal coherence that allows groups of neurons to inuence each other when their excitability peaks are aligned (phase locked) but not to do so when they are not aligned. Interestingly, neuronal communication is not best accomplished through simple rate modulation, as traditionally thought, because direct experiments show that the impact of increasing the rate of ring in presynaptic neurons can be a reduction in their postsynaptic effect unless the sending group of neurons sends output that arrives at the receiving groups excitability peak (Fries, 2005). Although there are several mechanisms through which such coherent interaction might come about (e.g., Fox, Jayaprakash, Wang, & Campbell, 2001), one of the most suggestive in the present context involves subcortical, especially thalamic, regions of the brain (cf. Newman, 1995). First among the several reasons for focusing on this mechanism is that much neural synchrony is characterized by near-zero phase-lags, in spite of conduction and synaptic delays between the cortical areas involved (Chawla, Friston, & Lumer, 2001; Fox et al., 2001; John, 2001; Roelfsema, Engel, Knig, & Singer, 1997). This imposes signicant constraints on the mechanism, since conduction delays between excitatorily-connected cortical neurons usually result in substantial phase lag (although there is a mechanism whereby such networks can attain zero-phase-lag synchrony, see Bose, Kopell, & Terman, 2000; Fox et al., 2001; Knig & Schillen, 1991). Second, computer simulation of cortico-thalamic circuits stimulated by sine-wave gratings gives rise to spontaneous synchronous oscillations in the gamma range (Lumer, Edelman, & Tononi, 1997a, 1997b). Third, direct stimulation of some thalamic neurons can evoke synchronous, gamma-range activity in cortical neurons (e.g., Barth & MacDonald, 1996). Fourth, a prime function of many thalamic neurons, especially those in the TRN, may be that of modulating the ring of primary and secondary relay neurons in the dorsal thalamic nuclei (e.g., Sherman & Guillery, 2006). Fifth, as was mentioned earlier, Jones (2001, 2002, 2009) proposed that the diffusely-projecting matrix neurons, found in most nuclei of the thalamus, including especially in the intralaminar and midline nuclei, play the role of propagating synchronous oscillations throughout the thalamic nuclei and their associated cortical areas, thus playing a critical role in creating the coherent neural activity that characterizes cognition. Finally, it has been shown that synaptic depression, a common form of short-term plasticity in neurons, can induce synchronous oscillations in neural excitatoryinhibitory networks like those found in the TRN and relay nuclei of the thalamus (Bose, Manor, & Nadim, 2000). Recent modeling efforts have studied how information transmission in thalamo-cortical circuits is related to their rhythmic uctuations. Mazzoni, Panzeri, Logothetis, and Brunel (2008) created a model of V1 consisting of sparsely-connected pyramidal cells and inhibitory interneurons. The simulated V1 network used a synchronous theta-range (47 Hz) channel to encode slow dynamic changes in input from thalamus (LGN), whereas it used a synchronous gamma-range (3050 Hz) channel to encode static input spike rates, for both articial and natural stimuli. Koepsell et al. (2009) showed that thalamic relay neurons (LGN) in both cat and model generated output spike trains consisted of two multiplexed synchronous neural oscillatory channels: a low frequency channel (<30 Hz) carried information about local changes in the stimulus over time, and a high-frequency channel (4080 Hz) carried information about global features of the visual scene. 8. A thalamic dynamic core? Put simply, the hypothesis of this paper is that the cortex computes the contents of consciousness whereas the thalamus displays, and thus experiences, the results of those computations. In line with Mumfords (1991) proposal that the higher-order thalamic nuclei form an active blackboard for the cortical areas to which they are reciprocally connected, I propose that the primary neural correlate of consciousness is a dynamic core of coordinated (synchronous) neural activity within the dendrites of a subset of higher-order thalamic nuclei, particularly those serving frontal areas (such as the medial dorsal nucleus), modulated and/ or gated by the TRN, the pulvinar nucleus, various subcortical inputs, and diffuse cortico-thalamic projections, a thalamic dynamic core. Although the dendritic trees of all neurons perform an integrative function relative to their synaptic input, the extensive dendritic proliferation of the dorsal thalamic neurons, plus their central location in the brain, makes them seem ideally suited to integrate information from cortical computations along with modulations sent from subcortical sites and to generate the neural activity most closely associated with conscious experience. The idea that the dendritic tree is the site of such information integration is of course not new (Cook, 2008; Yuste & Tank, 1996). Most relevant to consciousness, Romijn (2002) and Vadakkan (2010) made it a central aspect of their theories of primary consciousness. In contrast, Laberge and Kasevich (2007) and Cook (2008) differentiated the neural activity involved in cognitive information processing from that involved in generating primary consciousness. The former takes place in the cortical shell circuits for Laberge and Kasevich and is related to dendritic currents for Cook, whereas the latter arises around the apical dendrites of pyramidal neurons within the input-stayput axis circuits (cortico-thalamic) for Laberge and Kasevich, and from cortical action potentials for Cook. All four of these diverge from the present proposal in that they argue that it is in or around cortical neurons where primary consciousness is generated. I would argue instead that the cortical dendritic currents and action potentials reect mainly neural activity specic to the function of the cortical area in which they are found, prior to the level of integration represented by consciousness. In addition, cortical-level integration, especially in interaction with the hippocampus, seems ideal for the formation of memories (Merker, 2004).
482
Each of the theories just mentioned, as well as many others, propose a specic aspect of neural activity as the generator of subjectivity per se. These range from a eld of virtual photons generated in dendrites (Romijn) to an electro-magnetic eld generated near apical dendrites in axis circuits (Laberge & Kasevich) to free ion ow during action potentials (Cook). I will refrain from adding to this list here, as the purpose of the present proposal is to focus attention on the role of synchronous thalamic activity in generating primary consciousness rather than to propose exactly what it is about this activity that comprises subjectivity. I do believe, however, that none of the previous proposals has been able to show that subjectivity is entailed by the particular aspect of neural activity they spotlight (cf. Chalmers, 1996). In this proposal, attention is the mechanism by which information is selected for special processing by the higher cortical areas and is perforce integrated into the thalamic dynamic core by this selection process. As pointed out earlier, we are blind to (not aware of) most information that is not in attention (e.g., Mack & Rock, 1998), although the dynamic core may contain representations of information that was in attention in the past, or may contain representations, from templates or default frames, of information at a (hypothesized) illusorily high level of detail, presumably resulting from activity of higher cortical association and memory areas. Cortical activity concerned with this selected information is also coordinated and synchronized at one or more of theta, alpha, beta, or gamma frequency bands because of the coordination of the reciprocally-connected-with-cortex thalamic nuclei (cf. Ward, 2003). This means that even cortical areas separated by several synaptic delays can have (near) zero-phase-lag synchrony because activity in these areas is modulated by input from the synchronously ring neurons in their respective thalamic nuclei. When a general anesthetic or a large lesion blocks activity in the thalamus, it blocks the intricate connectivity both within the thalamus (TRN modulation) and between cortex and thalamus upon which consciousness depends (cf. Laberge & Kasevich, 2007; Llins et al., 1998). Deep sleep is a similar state to general anesthesia, in which thalamic activity is very different from the waking state (burst ring instead of tonic ring see Steriade et al., 1993). However, in REM sleep, the thalamus is functioning similarly to the way it does in a conscious state except that input to the sensory relay nuclei from the peripheral sensory systems (but not from cortical areas) is blocked. Dreams thus have the feel of consciousness but are disconnected from the sensory input that, during waking, keeps higher cortical activity calibrated with the current environment and helps to distinguish imagination from perception. To put the proposal another way, cortical computations are necessary but not sufcient for consciousness (cf. Kanwisher, 2001; Rees et al., 2002). They are necessary in the sense that particular contents of consciousness will not be possible unless the relevant cortical computations are performed. Thus, a lesion in a particular cortical area will eliminate the particular experiences computed by that cortical area. For example, a lesion in area V5/MT in the right temporal cortex would eliminate the experience of visual motion in the left visual eld, while leaving other visual experiences from stimuli in that visual eld intact (although possibly different in some ways, e.g., no shape from motion). Moreover, activity in V1 seems to determine whether activity in V5/MT reaches visual awareness at all (Silvanto, Cowey, Lavie, & Walsh, 2005), so early cortical activity is clearly necessary to support later cortical computations. Finally, a lesion that destroys V1 bilaterally curtails visual awareness (but not awareness itself) because it eliminates input to both dorsal and ventral visual streams (but not to the tecto-pulvinar system, which seems to be sufcient to guide blindsight; e.g., Weiskrantz, 1986), which compute the contents of visual awareness. Such observations were previously taken to indicate that cortical activity in the lesioned area constitutes the experience. In the present proposal such cortical activity supports the experience but the experience actually occurs as a result of informing the activity of the associated thalamic nuclei of the results of the cortical computations and integrating those results into the thalamic dynamic core. Theoretically, it would be possible to reconstitute the experience, e.g. of visual motion, by performing the relevant computations externally and then stimulating the thalamus in the same way it would have been stimulated by the associated disabled cortical area, e.g. V5/MT, and producing the conditions under which this activity would be included in the dynamic core, e.g. by telling the subject to pay attention to the moving stimulus. This would be a potentially falsifying test of the present theory. Grosser tests, such as stimulating the cortex and recording the associated experiences, as carried out by Peneld (e.g., 1958, 1975) have already shown that experiences can be elicited by such means. Interestingly, Peneld (1975) argued that during such stimulation the stimulated cortical area was disabled and the areas to which the stimulated area projected were the source of the experience. Of course, all cortical areas project to the thalamic nuclei associated with those areas, and thus Peneld might indeed have been observing the induction of experiences directly in the thalamus. 9. The future There are, inevitably, unanswered questions and problems with the proposal of the thalamic dynamic core. One conceptual problem arises when we consider the consequences of cutting the corpus callosum. Does a split-brain patient have two consciousnesses, each conned to one hemisphere of the brain, as would be predicted by any purely cortical or even corticothalamic theory? Or is there only one consciousness residing in integrated thalamic activity, or perhaps even in only one thalamus, the one connected to the (usually left) language-understanding hemisphere? This is a very difcult problem to deal with, and how it is dealt with might eventually either conrm or deny the thalamic dynamic core theory. Because each hemisphere of a split-brain patient can be seen to operate independent of the other, and as each responds in complex ways to certain stimuli, it would seem that at least the appearance of two consciousnesses is sustained. There are signicant limits to the abilities of the non-linguistic (usually right) hemisphere, however, and the split-brain patient acts as if there is only one integrative consciousness directing behavior. As Gazzaniga (e.g., 2000) reports, split-brain patients feel the same regarding their consciousness before and after surgery: they claim to experience no trace of a dual consciousness. Gazzaniga
483
(e.g., 2000) believes this feeling arises because only one hemisphere, usually the left one, is capable of integrating separate stimuli into a coherent story; the left hemisphere is, in his terms, the interpreter, and the only one that generates consciousness. Indeed the left hemisphere in split-brain patients easily and routinely generates explanations for the sometimes unexpected behavior initiated by the right hemisphere (see Gazzaniga, 2000, for examples). Another problem of course is falsiability. The problem of empirically testing the necessity and sufciency of neural synchrony for conscious awareness is just as difcult as that for any other dynamic proposal (cf., Rees et al., 2002). Moreover, establishing that the thalamus, rather than the cortex or cortico-thalamic loops, is the site of the necessary and sufcient synchronous neural activity will be difcult, since the cortex is so intimately linked to the thalamus. Most exclusion tests that reduce or interrupt activity in one part would also, inevitably, interrupt it in the other (cf. Frith, Perry, & Lumer, 1999). There is some promise, however, that phenomena like blindsight, neglect, and extinction, which dissociate behavior from experience, can also dissociate experience supported by the cortex from that arising from thalamic activity alone. Syndromes in which communication between particular cortical areas and their associated thalamic nuclei is interrupted, e.g., thalamocortical dysrhythmias (Llins, Ribary, Jeanmonod, Kronberg, & Mitra, 1999), would be good candidates for studies of the thalamic dynamic core. More subtle, computer controlled, direct stimulation studies than those conducted by Peneld and his successors (e.g., Ojeman, 1983) are possible, perhaps also on epileptic or other patients during the course of operations and coupled with the intra-cortical EEG so often used to map cortical function. This is already being done in the course of thalamic deep brain stimulation therapy in clinical situations involving epilepsy (e.g., Loddenkemper et al., 2001) and Parkinsonism and essential tremor (e.g., Benabid, 2003). A particularly fascinating possibility is the direct and unique stimulation of thalamic neurons using laser light targeting cells that express channelrhodopsin, created by injecting a viral vector (optogenetics; e.g., Boyden, Zhang, Bamberg, Nagel, & Deisseroth, 2005). Recently this has been accomplished in rats by Lee et al. (2010), and there is also a method available for selectively suppressing populations of neurons using a similar optogenetic technique (Chow et al., 2010). It would be necessary to emulate cortical input to the thalamus in this experiment and this already has been accomplished in a rudimentary way, albeit digitally in model systems involving modulation of the LGN by input from V1 (e.g., Bickle, Bernstein, Heatley, Worley, & Stiehl, 1999). Brain imaging using PET, fMRI or MEG, which can record subcortical as well as cortical activity, coupled with interventions such as trans-cranial magnetic stimulation or specic anesthetics, could possibly localize the source(s) of necessary neural activity for particular conscious experiences. One example of this is the very telling work of Alkire and colleagues (e.g., 2000, 2005) and John (e.g., 2001) on the action of general anesthetics. Another is the ongoing work using a combination of brain imaging and binocular rivalry (e.g., Doesburg et al., 2005, 2009). The promise is that through the accumulation of creative and sound empirical insights we will eventually begin to understand how the brain gives rise to the conscious mind. Acknowledgments This research was supported by a Discovery Grant from the Natural Sciences and Engineering Research Council of Canada, and by the Peter Wall Institute of Advanced Studies at the University of British Columbia, neither of which had any other role in any phase of it. I thank the members of my lab, particularly Michael Baumann and Sam Doesburg, and David Mumford, Larry Roberts, Olaf Sporns, Eugene Izhikevich and Giulio Tononi for useful observations on earlier versions. I particularly wish to thank Bjorn Merker for generously reading several versions of the paper and for providing perspicuous comments that have enhanced its evidential and conceptual bases as well as its structure. References
Adams, J. H., Graham, D. I., & Jennett, B. (2000). The neuropathology of the vegetative state after an acute brain insult. Brain, 123, 13271338. Alkire, M. T., Asher, C. D., Franciscus, A. M., & Hahn, E. L. (2009). Thalamic microinfusion of antibody to a voltage-gated potassium channel restores consciousness during anesthesia. Anesthesiology, 110, 766773. Alkire, M. T., Haier, R. J., & Fallon, J. H. (2000). Toward a unied theory of narcosis: Brain imaging evidence for a thalamocortical switch as the neurophysiologic basis of anesthetic-induced unconsciousness. Consciousness and Cognition, 9, 370386. Alkire, M. T., Hudetz, A. G., & Tononi, G. (2008). Consciousness and anesthesia. Science, 322, 876880. Alkire, M. T., & Miller, J. (2005). General anesthesia and the neural correlates of consciousness. Progress in Brain Research, 150, 227242. Baars, B. J. (1995). Tutorial commentary: Surprisingly small subcortical structures are needed for the state of waking consciousness, while cortical projection areas seem to provide perceptual contents of consciousness. Consciousness and Cognition, 4, 159162. Baars, B. J. (1997). In the theatre of consciousness: Global workspace theory, a rigorous scientic theory of consciousness. Journal of Consciousness Studies, 4, 292309. Baars, B. J. (2002). The conscious access hypothesis: Origins and recent evidence. Trends in Cognitive Sciences, 6, 4752. Barth, D. S., & MacDonald, K. D. (1996). Thalamic modulation of high-frequency oscillating potentials in auditory cortex. Nature, 383, 7881. Barth, P., Freund, F. T., & Acsdy, L. (2002). Selective GABAergic innervation of thalamic nuclei from zona incerta. European Journal of Neuroscience, 16, 9991014. Bayless, S. & Ward, L.M. (2009). Information processing in a change blindness paradigm. Unpublished ms., University of British Columbia. Benabid, A. L. (2003). Deep brain stimulation for Parkinsons disease. Current Opinion in Neurobiology, 13, 696706. Bickle, J., Bernstein, M., Heatley, M., Worley, C., & Stiehl, S. (1999). A functional hypothesis for LGN-V1-TRN connectivities suggested by computer simulation. Journal of Computational Neuroscience, 6, 251261. Bogen, J. E. (1995a). On the neurophysiology of consciousness: I. An overview. Consciousness and Cognition, 4, 5262. Bogen, J. E. (1995b). On the neurophysiology of consciousness: Part II. Constraining the semantic problem. Consciousness and Cognition, 4, 137158. Bokor, H., Frre, S. G. A., Eyre, M. D., Slzia, A., Ulbert, I., Lthi, A., et al (2005). Selective GABAergic control of higher-order thalamic relays. Neuron, 45, 929940. Bose, A., Kopell, N., & Terman, D. (2000). Almost-synchronous solutions for mutually coupled excitatory neurons. Physica D, 140, 6994.
484
Bose, A., Manor, Y., & Nadim, F. (2000). Bistable oscillations arising from synaptic depression. SIAM Journal of Applied Mathematics, 62, 706727. Boyden, E. S., Zhang, F., Bamberg, E., Nagel, G., & Deisseroth, K. (2005). Millisecond-timescale, genetically targeted optical control of neural activity. Nature Neuroscience, 8, 12631268. Bressler, S. L., & Kelso, J. A. S. (2001). Cortical coordination dynamics and cognition. Trends In Cognitive Sciences, 5, 2636. Bridgeman, B., Hendry, D., & Stark, L. (1975). Failure to detect displacement of the visual world during saccadic eye movements. Vision Research, 15, 719722. Buzsaki, G., & Draguhn, A. (2004). Neuronal oscillations in cortical networks. Science, 304, 19261929. Chalmers, D. J. (1996). The conscious mind: In search of a fundamental theory. New York: Oxford University Press. Chawla, D., Friston, K. J., & Lumer, E. D. (2001). Zero-lag synchronous dynamics in triplets of interconnected cortical areas. Neural Networks, 14, 727735. Chow, B. Y., Han, X., Dobry, A. S., Qian, X., Chuong, A. S., Li, M., et al (2010). High-performance genetically targetable optical neural silencing by light-driven proton pumps. Nature, 463, 98102. Cook, N. M. (2008). The neuron-level phenomena underlying cognition and consciousness: Synaptic activity and the action potential. Neuroscience, 153, 556570. Corballis, M. C. (1979). Memory retrieval and the problem of scanning. Psychological Review, 86, 154157. Cosmelli, D., David, O., Lachaux, J.-P., Martinerie, J., Garnero, L., Renault, B., et al (2004). Waves of consciousness: Ongoing cortical patterns during binocular rivalry. NeuroImage, 23, 128140. Crabtree, J. W. (1999). Intrathalamic sensory connections mediated by the thalamic reticular nucleus. Cellular and Molecular Life Sciences, 56, 683700. Crabtree, J. W., Collingridge, G. L., & Isaac, J. T. R. (1998). A new intrathalamic pathway linking modality-related nuclei in the dorsal thalamus. Nature Neuroscience, 1, 389394. Crick, F. (1984). Function of the thalamic reticular complex: The searchlight hypothesis. Proceedings of the National Academy of Sciences of the USA, 81, 45864590. Crick, F. (1994). The astonishing hypothesis: The scientic search for the soul. London: Simon & Schuster. Crick, F., & Koch, C. (1990). Some reections on visual awareness. Cold Spring Harbor Symposia in Quantitative Biology, 55, 953962. Crick, F., & Koch, C. (2003). A framework for consciousness. Nature Neuroscience, 6, 119126. Crovitz, H. F. (1970). Galtons walk: Methods for the analysis of thinking, intelligence and creativity. New York: Harper and Row. Damasio, A. (1999). The feeling of what happens. London: William Heinemann. Dehaene, S., Kerszberg, M., & Changuex, J.-P. (1998). A neuronal model of a global workspace in effortful cognitive tasks. Proceedings of the National Academy of Sciences of the USA, 95, 1452914534. Dennett, D. C. (1991). Consciousness explained. Boston: Little, Brown and Company. Doesburg, S. M., Green, J. J., McDonald, J. J., & Ward, L. M. (2009). Rhythms of consciousness: Binocular rivalry reveals large-scale oscillatory network dynamics mediating visual perception. PLoS One, 7. e6142 (114). Doesburg, S., Kitajo, K., & Ward, L. M. (2005). Increased gamma-band synchrony precedes switching of conscious perceptual objects in binocular rivalry. NeuroReport, 16, 11391142. Donald, M. (1995). The neurobiology of human consciousness: An evolutionary approach. Neuropsychologia, 33, 10871102. Edelman, G. M. (2004). Wider than the sky. New Haven, CN: Yale University Press. Edelman, G. M., & Tononi, G. (2000). A universe of consciousness. New York: Basic Books. Engel, A. K., & Singer, W. (2001). Temporal binding and the neural correlates of sensory awareness. Trends in Cognitive Sciences, 5, 1625. Fechner, G. T. (1966). Elements of psychophysics (Vol. 1; H. E. Adler, Trans.). New York: Holt, Rinehart & Winston (Original work published 1860). Fodor, J. A. (1983). The modularity of mind. Cambridge, MA: MIT Press. Fox, J. J., Jayaprakash, C., Wang, D., & Campbell, S. R. (2001). Synchronization in relaxation oscillator networks with conduction delays. Neural Computation, 13, 10031021. Freud, S. (1984). The unconscious. In A. Richards (Ed.), The Penguin Freud library. On metapsychology (Vol. 2, pp. 159222). London: Penguin Books (Original work published 1915). Fries, P. (2005). A mechanism for cognitive dynamics: Neuronal communication through neuronal coherence. Trends in Cognitive Sciences, 9, 474480. Fries, P., Reynolds, J. H., Rorie, A. E., & Desimone, R. (2001). Modulation of oscillatory neuronal synchronization by selective visual attention. Science, 291, 15601563. Friston, K. (2009). The free-energy principle: A rough guide to the brain? Trends in Cognitive Sciences, 13, 293301. Frith, C., Perry, R., & Lumer, E. (1999). The neural correlates of conscious experience: An experimental framework. Trends in Cognitive Sciences, 3, 105114. Gaillard, R., Dehaene, S., Adam, C., Clmenceau, S., Hasboun, D., Baulac, M., et al (2009). Converging intracranial markers of conscious access. PLoS Biology, 7, e1000061. Gazzaniga, M. S. (2000). Cerebral specialization and interhemispheric communication: Does the corpus callosum enable the human condition? Brain, 123, 12931326. George, D., & Hawkins, J. (2009). Towards a mathematical theory of cortical micro-circuits. PLoS Computational Biology, 5, e1000532. Giacino, J. T. et al (2002). The minimally conscious state: Denition and diagnostic criteria. Neurology, 58, 349353. Gibson, J. J. (1966). The senses considered as perceptual systems. Oxford, England: Houghton Mifin. Grieve, K. L., Acua, C., & Cudeiro, J. (2000). The primate pulvinar nuclei: Vision and action. Trends in Neurosciences, 23, 3539. Grimes, J. (1996). On the failure to detect changes in scenes across saccades. In K. Akins (Ed.), Perception (Vancouver Studies in Cognitive Science, Vol. 5, pp. 89110), NY: Oxford University Press. Grossberg, S. (2000). The complementary brain: Unifying brain dynamics and modularity. Trends in Cognitive Sciences, 4, 233246. Harth, E. (1983). Windows on the mind. New York: Quill. Helmholtz, H. von (1867). Treatise on physiological optics. Trans. J.P.C. Southall (1962). New York: Dover Books. Herbart, J. F. (1824). Psychologie als Wissenschaft, neu gegrndet auf Erfahrung, Metaphysik und Mathematik, Knigsberg. Reissued in English translation in 1961 as Psychology as a science, newly founded on experience, metaphysics and mathematics, (Ed. T. Shipley), New York: Philosophical Library (Classics in Psychology). Hesslow, G. (2002). Conscious thought as simulation of behaviour and perception. Trends in Cognitive Sciences, 6, 242247. Hinton, G. E. & Sejnowski, T. J. (1983). Optimal perceptual inference. In Proceedings of the IEEE conference on computer vision and pattern recognition (pp. 448 453). Washington, DC. Hobson, J. A. (2009). REM sleep and dreaming: Towards a theory of protoconsciousness. Nature Reviews: Neuroscience, 10, 803813. Hudetz, A. G., Vizuete, J. A., & Imas, O. A. (2009). Desurane selectively suppresses long-latency cortical neuronal response to ash in the rat. Anesthesiology, 111, 231239. Huntsman, M. M., Porcello, D. M., Homanics, G. E., DeLorey, T. M., & Huguenard, J. R. (1999). Reciprocal inhibitory connections and network synchrony in the mammalian thalamus. Science, 283, 541543. James, W. (1890). The Principles of Psychology. New York: Holt. Jellinger, K. A. (1994). Letter to the Editor. The New England Journal of Medicine, 331, 13781379. Jennett, B., Adams, J. H., Murray, L. S., & Graham, D. I. (2001). Neuropathology in vegetative and severely disabled patients after head injury. Neurology, 59, 486490. Jennett, B., & Plum, F. (1972). Persistent vegetative state after brain damage. A syndrome in search of a name. Lancet, 1, 734737. John, E. R. (2001). A eld theory of consciousness. Consciousness and Cognition, 10, 184213. John, E. R. (2002). The neurophysics of consciousness. Brain Research Reviews, 39, 128.
485
Jones, E. G. (1985). The thalamus. New York: Plenum Press. Jones, E. G. (2001). The thalamic matrix and thalamocortical synchrony. Trends in Neurosciences, 24, 595601. Jones, E. G. (2002). Thalamic circuitry and thalamocortical synchrony. Philosophical Transactions of the Royal Society of London Series B, 357, 16591673. Jones, E. G. (2007). The thalamus (2nd ed.). New York: Cambridge University Press. Jones, E. G. (2009). Synchrony in the interconnected circuitry of the thalamus and cerebral cortex. In N. D. Schiff & S. Laureys (Eds.), Disorders of Consciousness. Annals of the New York Academy of Sciences (Vol. 1157, pp. 1023). Kahneman, D., & Tversky, A. (1982). The simulation heuristic. In D. Kahneman, P. Slovic, & A. Tversky (Eds.), Judgment under uncertainty: Heuristics and biases (pp. 201208). Cambridge, UK: Cambridge University press. Kanwisher, N. (2001). Neural events and perceptual awareness. Cognition, 79, 89113. Kinney, H. C., Korein, J., Panigrahy, A., Dikkes, P., & Goode, R. (1994). Neuropathological ndings in the brain of Karen Ann Quinlan The role of the thalamus in the persistent vegetative state. The New England Journal of Medicine, 330, 14601475. Kinsbourne, M. (1988). Integrated eld theory of consciousness. In A. J. Marcel & E. Bisiach (Eds.), Consciousness in contemporary science (pp. 239256). New York: Oxford University Press. Koch, C., & Tsuchiya, N. (2006). Attention and consciousness: two distinct brain processes. Trends in Cognitive Sciences, 11, 1622. Koepsell, K., Wang, X., Vaingankar, V., Wei, Y., Wang, G., Rathbun, D., et al (2009). Retinal oscillations carry visual information to cortex. Frontiers in Systems Neuroscience, 3(4), 118. Knig, P., & Schillen, T. B. (1991). Stimulus-dependent assembly formation of oscillatory responses: I. Synchronization. Neural Computation, 3, 155166. Laberge, D. (1995). Attentional processing: The brains art of mindfulness. Cambridge, MA: Harvard University Press. Laberge, D., & Kasevich, R. (2007). The apical dendrite theory of consciousness. Neural Networks, 20, 10041020. Lamme, V. A. F. (2003). Why visual attention and awareness are different. Trends in Cognitive Sciences, 7, 1218. Lee, J. H., Durand, R., Gradinaru, V., Zhang, F., Goshen, I., Kim, D.-S., et al (2010). Global and local fMRI signals driven by neurons dened optogenetically by type and wiring. Nature, 465, 788792. Leibniz, G. W. (1973). The monadology. In G. H. R. Parkinson (Ed.), Philosophical writings. London: Dent (Original published 1714). Libet, B. (2004). Mind time. The temporal factor of consciousness. Cambridge, MA: Harvard University Press. Llins, R. R., & Par, D. (1991). Commentary of dreaming and wakefulness. Neuroscience, 44(3), 521535. Llins, R., Ribary, U., Contreras, D., & Pedroarena, C. (1998). The neuronal basis for consciousness. Philosophical Transactions of the Royal Society of London, Series B, 353, 18411849. Llins, R., Ribary, U., Jeanmonod, D., Kronberg, E., & Mitra, P. P. (1999). Thalamocortical dysrhythmia: A neurological and neuropsychiatric syndrome characterized by magnetoencephalography. Proceedings of the National Academy of Sciences of the USA, 96, 1522215227. Loddenkemper, T., Pan, A., Neme, S., Baker, K. B., Rezai, A. R., Dinner, D. S., et al (2001). Deep brain stimulation in epilepsy. Journal of Clinical Neurophysiology, 18, 514532. Lovejoy, L. P., & Krauzlis, R. J. (2010). Inactivation of primate superior colliculus impairs covert selection of signals for perceptual judgments. Nature Neuroscience, 13, 261267. Lumer, E. D., Edelman, G. M., & Tononi, G. (1997a). Neural dynamics in a model of the thalamocortical system. I. Layers, loops and the emergence of fast synchronous rhythms. Cerebral Cortex, 7, 207227. Lumer, E. D., Edelman, G. M., & Tononi, G. (1997b). Neural dynamics in a model of the thalamocortical system. II. The role of neural synchrony tested through perturbations of spike timing. Cerebral Cortex, 7, 228236. Mack, A., & Rock, I. (1998). Inattentional blindness: Perception without attention. Cambridge, MA: The MIT Press. Matin, E. (1974). Saccadic suppression: A review and an analysis. Psychological Bulletin, 81, 899917. Maxwell, W. L., Pennington, K., MacKinnon, M. A., Smith, D. H., McIntosh, T. K., Wilson, J. T. L., et al (2004). Differential responses in three thalamic nuclei in moderately disabled, severely disabled and vegetative patients after blunt head injury. Brain, 127, 24702478. Mazzoni, A., Panzeri, S., Logothetis, N. K., & Brunel, N. (2008). Encoding of naturalistic stimuli by local eld potential spectra in networks of excitatory and inhibitory neurons. PLoS Computational Biology, 4. e1000239 (120). McAlonan, K., Brown, V. J., & Bowman, E. M. (2000). Thalamic reticular nucleus activation reects attentional gating during classical conditioning. The Journal of Neuroscience, 20, 88978901. McAlonan, K., Cavanaugh, J., & Wurtz, R. H. (2008). Guarding the gateway to cortex with attention in visual thalamus. Nature, 456, 391394. Melloni, L., Molina, C., Pena, M., Torres, D., Singer, W., & Rodriguez, E. (2007). Synchronization of neural activity across cortical areas correlates with conscious perception. The Journal of Neuroscience, 27, 28582865. Merker, B. (2004). Cortex, countercurrent context, and dimensional integration of lifetime memory. Cortex, 40, 559576. Merker, B. (2007). Consciousness without a cerebral cortex: A challenge for neuroscience and medicine. Behavioral and Brain Sciences, 30, 63134. Metzinger, T. (2003). Phenomenal transparency and cognitive self-reference. Phenomenology and the Cognitive Sciences, 2, 353393. Min, B.-K. (2010). A thalamic reticular networking model of consciousness. Theoretical Biology and Medical Modelling, 7(10), 118. Minsky, M. (1986). The society of mind. New York: Simon and Schuster. Moore, G. E. (1903). The refutation of idealism. Mind, 12, 433453. Mumford, D. (1991). On the computational architecture of the neocortex. I. The role of the thalamo-cortical loop. Biological Cybernetics, 65, 135145. Mumford, D. (1992). On the computational architecture of the neocortex. II. The role of cortico-cortical loops. Biological Cybernetics, 66, 241251. Navon, D. (1989). The importance of being visible: On the role of attention in a mind viewed as an anarchic intelligence system: I. Basic tenets. European Journal of Cognitive Psychology, 1, 191213. Newman, J. (1995). Thalamic contributions to attention and consciousness. Consciousness and Cognition, 4, 172193. Newman, J., Baars, B. J., & Cho, S.-B. (1997). A neural global workspace model for conscious attention. Neural Networks, 10, 11951206. Nieuwenhuys, R., Voogd, J., & van Huijzen, C. (2007). The human central nervous system (4th ed.). New York: Springer. Nisbett, R. E., & Wilson, T. D. (1977). Telling more than we can know: Verbal reports on mental processes. Psychological Review, 84, 231259. Ojeman, G. A. (1983). Brain organization for language from the perspective of electrical stimulation mapping. The Behavioral and Brain Sciences, 2, 189230. Peneld, W. (1958). The excitable cortex in conscious man. Atlantic Highlands, NJ: Liverpool University Press. Peneld, W. (1975). The mystery of the mind. Princeton, NJ: Princeton University Press. Peneld, W., & Jasper, H. H. (1954). Epilepsy and the functional anatomy of the human brain. Boston: Little, Brown and Company. Pico, R. (2002). Consciousness in four dimensions: Biological relativity and the origins of thought. New York: McGraw-Hill. Ratcliff, R. (1978). A theory of memory retrieval. Psychological Review, 85, 59108. Rees, G., Kreiman, G., & Koch, C. (2002). Neural correlates of consciousness in humans. Nature Reviews: Neuroscience, 3, 261270. Rensink, R. A., ORegan, J. K., & Clark, J. J. (1997). To see or not to see: The need for attention to perceive changes in scenes. Psychological Science, 8, 368373. Rodriguez, E., George, N., Lachaux, J.-P., Martinerie, J., Renault, B., & Varela, F. J. (1999). Perceptions shadow: Long-distance synchronization of human brain activity. Nature, 397, 430433. Roelfsema, P. R., Engel, A. K., Knig, P., & Singer, W. (1997). Visuomotor integration is associated with zero time-lag synchronization among cortical areas. Nature, 385, 157161. Romijn, H. (2002). Are virtual photons the elementary carriers of consciousness? Journal of Consciousness Studies, 9, 6181. Searle, J. R. (2000). Consciousness. Annual Review of Neuroscience, 23, 557578. Seth, A. K., & Baars, B. J. (2005). Neural Darwinism and consciousness. Consciousness and Cognition, 14, 140168. Sherman, S. M., & Guillery, R. W. (2006). Exploring the thalamus and its role in cortical function. Cambridge, MA: MIT Press. Shipp, S. (2004). The brain circuitry of attention. Trends in Cognitive Sciences, 8, 223230.
486
Silvanto, J., Cowey, A., Lavie, N., & Walsh, V. (2005). Striate cortex (V1) activity gates awareness of motion. Nature Neuroscience, 8, 143144. Simons, D. J., & Levin, D. T. (1998). Failure to detect changes to people in a real-world interaction. Psychonomic Bulletin and Review, 5, 644649. Simons, D. J., & Rensink, R. A. (2005). Change blindness: Past, present, and future. Trends in Cognitive Sciences, 9, 1620. Singer, W. (1999). Neuronal synchrony: A versatile code for the denition of relations. Neuron, 24, 4965. Sohal, V. S., Huntsman, M. M., & Huguenard, J. R. (2000). Reciprocal inhibitory connections regulate the spatiotemporal properties of intrathalamic oscillations. The Journal of Neuroscience, 20, 17351745. Srinivasan, R., Russell, D. P., Edelman, G. M., & Tononi, G. (1999). Increased synchronization of neuromagnetic responses during conscious perception. Journal of Neuroscience, 19, 54355448. Steriade, M. (1999). Coherent oscillations and short-term plasticity in corticothalamic networks. Trends in Neurosciences, 22, 337345. Steriade, M. (2005). Sleep, epilepsy and thalamic reticular inhibitory neurons. Trends in Neurosciences, 28, 317324. Steriade, M., McCormick, D. A., & Sejnowski, T. J. (1993). Thalamocortical oscillations in the sleeping and aroused brain. Science, 262, 679685. Sternberg, S. (1970). Memory-scanning: Mental processes revealed by reaction-time experiments. In J. S. Antrobus (Ed.), Cognition and affect (pp. 1358). Boston: Little, Brown and Company. Suffczynski, P., Kalitzin, A., Pfurtscheller, G., & Lopes da Silva, F. H. (2001). Computational model of thalamo-cortical networks: Dynamical control of alpha rhythms in relation to focal attention. International Journal of Psychophysiology, 43, 2540. Tononi, G., & Edelman, G. M. (1998a). Consciousness and the integration of information in the brain. In H. H. Jasper, L. Descarries, V. F. Castellucci, & S. Rossignol (Eds.), Consciousness: At the frontiers of neuroscience (pp. 245279). Philadelphia: Lippencott-Raven. Tononi, G., & Edelman, G. M. (1998b). Consciousness and complexity. Science, 282, 18461851. Tononi, G., Srivinivasan, R., Russell, D. P., & Edelman, G. M. (1998). Investigating neural correlates of conscious perception by frequency-tagged neuromagnetic responses. Proceedings of the National Academy of Sciences of the USA, 95, 31983203. Trageser, J. C., Burke, K. A., Masri, R., Li, Y., Sellers, L., & Keller, A. (2006). State-dependent gating of sensory inputs by zona incerta. Journal of Neurophysiology, 96, 14561463. Trageser, J. C., & Keller, A. (2004). Reducing the uncertainty: Gating of peripheral inputs by zona incerta. The Journal of Neuroscience, 24, 89118915. Uzan, M., Albayram, S., Dashti, S. G. R., Aydin, S., Hanci, M., & Kuday, C. (2003). Thalamic proton magnetic resonance spectroscopy in vegetative state induced by traumatic brain injury. Journal of Neurology and Neurological Psychiatry, 74, 3338. Vadakkan, K. (2010). Framework of consciousness from semblance of activity at functionally linked postsynaptic membranes. Frontiers in Psychology(168), 112. van Essen, D. C., & Deyoe, E. A. (1995). Concurrent processing in the primate visual cortex. In M. S. Gazzaniga (Ed.), The cognitive neurosciences (pp. 383340). Cambridge, MA: MIT Press. Ward, L. M. (1992). Mind in psychophysics. In D. Algom (Ed.), Psychophysical approaches to cognition (pp. 187249). New York: North-Holland. Ward, L. M. (2002). Dynamical cognitive science. Cambridge, MA: MIT Press. Ward, L. M. (2003). Synchronous neural oscillations and cognitive processes. Trends in Cognitive Sciences, 7, 553559. Ward, L. M. (2004). Oscillations and synchrony in cognition. In V. K. Jirsa & J. A. S. Kelso (Eds.), Coordination dynamics: Issues and trends (pp. 217242). New York: Springer. Wegner, D. (2002). The illusion of conscious will. Cambridge, MA: MIT Press. Weiskrantz, L. (1986). Blindsight: A case study and implications. Oxford: Oxford University Press. Yuste, R., & Tank, D. W. (1996). Dendritic integration in mammalian neurons, a century after Cajal. Neuron, 16, 701716. Zelazo, P. D. (2004). The development of conscious control in childhood. Trends in Cognitive Sciences, 8, 1217. Zhang, L., & Jones, E. G. (2004). Corticothalamic inhibition in the thalamic reticular nucleus. Journal of Neurophysiology, 91, 759766. Zikopoulos, B., & Barbas, H. (2006). Prefrontal projections to the thalamic reticular nucleus form a unique circuit for attentional mechanisms. Journal of Neuroscience, 26, 73487361. Zikopoulos, B., & Barbas, H. (2007). Parallel driving and modulatory pathways link the prefrontal cortex and thalamus. PLoS One, 9, e848.

Ward 2011 The Thalamic Dynamic Core Theory of Conscious Experience

Diunggah oleh

Informasi Dokumen

Deskripsi Asli:

Judul Asli

Hak Cipta

Format Tersedia

Bagikan dokumen Ini

Bagikan atau Tanam Dokumen

Opsi Berbagi

Apakah menurut Anda dokumen ini bermanfaat?

Apakah konten ini tidak pantas?

Hak Cipta:

Format Tersedia

Ward 2011 The Thalamic Dynamic Core Theory of Conscious Experience

Diunggah oleh

Hak Cipta:

Format Tersedia

Consciousness and Cognition 20 (2011) 464486

Contents lists available at ScienceDirect

Consciousness and Cognition

The thalamic dynamic core theory of conscious experience

Fax: +1 604 822 6923.

L.M. Ward / Consciousness and Cognition 20 (2011) 464486

L.M. Ward / Consciousness and Cognition 20 (2011) 464486

L.M. Ward / Consciousness and Cognition 20 (2011) 464486

L.M. Ward / Consciousness and Cognition 20 (2011) 464486

L.M. Ward / Consciousness and Cognition 20 (2011) 464486

L.M. Ward / Consciousness and Cognition 20 (2011) 464486

L.M. Ward / Consciousness and Cognition 20 (2011) 464486

L.M. Ward / Consciousness and Cognition 20 (2011) 464486

L.M. Ward / Consciousness and Cognition 20 (2011) 464486

L.M. Ward / Consciousness and Cognition 20 (2011) 464486

L.M. Ward / Consciousness and Cognition 20 (2011) 464486

L.M. Ward / Consciousness and Cognition 20 (2011) 464486

L.M. Ward / Consciousness and Cognition 20 (2011) 464486

L.M. Ward / Consciousness and Cognition 20 (2011) 464486

L.M. Ward / Consciousness and Cognition 20 (2011) 464486

L.M. Ward / Consciousness and Cognition 20 (2011) 464486

L.M. Ward / Consciousness and Cognition 20 (2011) 464486

L.M. Ward / Consciousness and Cognition 20 (2011) 464486

L.M. Ward / Consciousness and Cognition 20 (2011) 464486

L.M. Ward / Consciousness and Cognition 20 (2011) 464486

L.M. Ward / Consciousness and Cognition 20 (2011) 464486

L.M. Ward / Consciousness and Cognition 20 (2011) 464486

Anda mungkin juga menyukai