CODAM model/consciousness and attention

What's the Difference?

On the face of it there is a qualitative difference between the two: consciousness is all about experience - of being aware of the redness of a rose, the yellowness of daffodils or the sharpness in the throat from a shot of whisky, as well as of feelings and emotions - whilst attention is a process which allows you to focus your glance (and analysis power) onto an object or action out in the environment. Attention allows you to analyse a given stimulus more closely, while on the other hand being conscious of the stimulus gives you an extra subjective 'sense' or 'feel' about the stimulus. It is this latter subjective component that seems most crucial as differentiating the two. But is consciousness then just attention plus the subjective feel? Or are there different components in each of them beyond the absence or presence of subjectivity?

It is arguable that attention is an important gateway to consciousness. Without attention it would seem that one cannot be aware of an object in the external world. The phenomenon of the attentional blink (Vogel et al 1998), inattentional blindness (Rock and Mack 1998) and related phenomena indicate how crucial is attention to allow awareness of specific components of the outside world to arise. Yet there is still the possibility that there are other routes for stimulus activity in the brain to reach consciousness. It has been seriously suggested recently that attention is not so all-embracing and that attention and consciousness are two separate faculties (Lamme, 2003, 2006; Koch and Tsuchiya, 2006), instead of one (attention) leading on to the other (consciousness). Such a possibility of independence must therefore be considered seriously before proceeding further.

In this article we wish to consider briefly the evidence and related arguments for attention and consciousness being independent faculties in the brain. This is important to clarify, since especially the experiments of interest are those that probe most effectively into both of these facculties and refine their understanding very effectively.

Attention

To discuss differences we need some knowledge of the originals whose differences are being discussed. Attention has long been studied by psychologists by behavioural paradigms and more recently by neuroscientists using brain imaging methods (both globally and at single cell level). The original appreciation of attention as causing the speed-up of processing of an attended stimulus has now been put under the brain imaging microscope. This has used especially the paradigm of Posner, in which a target is presented after that of a cue on one side or the other of a visual display. The target may either agree in position, when it appears, with direction implied by the cue, or it may not - so the cue is valid or invalid - resulting in speeded response in the valid cue case compared with the invalid one. Furthermore the cue can either be a central arrow predicting on which side the target will appear, or a peripheral one, such as the brightening of one of two target boxes on the screen. The former of these cases is termed endogenous or top-down, corresponding to higher-level processes to evaluate the rule implied by the central cue and apply attention. The other method - the cue acting by brightening of a target box - is termed exogenous or bottom-up, since it involves attention breakthrough by the spatial position of the cue. We should also note that visual attention can either by spatially-based (using mainly the dorsal route to prefrontal cortex) or feature-based (using the ventral route down to object representations in the temporal lobe).

These two forms of attention - exogenous and endogenous - are known to be subserved by largely overlapping networks of fronto-parietal networks, although some studies have found a certain degree of specialisation for endogenous or exogenous attention within the overall network (Corbetta & Shulman, 2002; Yantis et l, 2002; Kincaide et al, 2005). Simlarily spatial attention and feature-based attention are observed to engage common frontoparietal areas, with spatial attention possibly activating the more dorsal cortical brain areas more strongly than feature-based attention (Giesebrecht et al, 2003)

There is also the important discovery that attention can be focussed on more than one place at once - the focus of attention can be split. This can happen in several foci, as reported in (McMains & Somers, 2004). The experiment used fMRI to detect retinotopically enhanced activations in visual areas of cortex when two separately positioned target stimuli were to be attended to simultaneously. This occurred in a rapid serial visual presentation stream of stimuli composed of two targets and three distracters or one target and four distracters, all positioned at the corners and centre of a square. It was possible to detect significant activations at the target positions compared to the distracter positions in all the cortical areas corresponding to V1, V2, V3/VP, V3A, V4v and V8, as well as increased separate activations in the two target versus the single target case. The authors concluded that 'This provides direct evidence that spatial attention can select, in parallel, multiple low-level perceptual representations' (McMains & Somers, 2004, p677). This result has been duplicated and extended in (McMains & Somers, 2005). and is related to earlier work on the ability of subjects simultaneously to track multiple moving targets (four or more) (Cavanagh & Alvarez, 2005).

So far we have discussed purely sensory attention (mainly visual) but we cannot neglect also the existence of motor intention, or as it is termed by some researchers, motor attention. This arises as if it were possible to focus attention on a specific motor response, thereby singling it out from all other motor responses. The behavioural effects of this can be studied by a motor response version of the Posner paradigm mentioned earlier: cue a subject to prepare a certain motor response (out of a range of them) and then use a GO signal to tell the subject to respond as rapidly as possible. A slower reaction time is observed when the cue is invalid, as in the visual case. Such a motor 'Posner benefit' (the difference between the reaction time to the invalid as compared to the valid cue) is found reduced in subjects with loss of certain left parietal sites, so these must play a crucial role in motor attention. and to be compared with right parietal sites being simlarily important for space-based attention but not for motor attention (Rushworth et al, 2001).

One of the basic mechanisms observed in the brain during attention that gives a strong clue to its action is that regions of cortex involved in relevant attended stimulus analysis, at either global level as seen by brain imaging or single cell level seen by implanted electrodes, have increased activity under attention. This has been observed most closely by (Mehta et al, 2000) using multi-unit recordings, which showed that in vision the initial stimulus input wave moves up the hierarchy of V1, V2, V3, V4 etc and then there is a return signal coming back from beyond V4 succesively to lower regions (as observed by the layer in a given cortical site to which it feeds). A similar increase in activity is observed under brain imaging in motor cortex when attention is drawn to a specific motor act (Binkofski et al, 2002,).

Thus attention acts in general as a feedback signal from higher cortex to increase neural activity representing an attended stimulus, and at the same time reduce that of distracters; this mechanism is now a well-established feature of attention. It is still controversial as to exactly how attention acts at a micro level in the cortex: does it feed back by contrast gain (so increasing synaptic weights of axons carrying input from lower areas associated with the attended stimulus) or is it purely an additive component to the activity of a relevant neuron, so giving an increase in background activity for all the neurons which receive input from an attended stimulus? A variety of approaches were discussed and analysed by simulation in (Taylor N et al, 2006); the contrast gain model seemed to fit single cell monkey data best (Reynolds & Desimone, 1999), although the threshold changing model is also very popular (see also Taylor JG et al, 2006). It is likely that attention acts through a mixture of contrast gain and additive feedback, although the proportions and learning processes for these two components are still to be explored both experimentally and by simulation.

To summarise, there has been an important advance in our knowledge of attention by use of the new tools of brain imaging (with ever more data becoming available day-by-day), although some of the details of the way that attention functions are still controversial. Yet the generally accepted thesis is that attention acts by boosting the neural activity in lower cortical sites of an attended stimulus representation, with associated inhibition of distracters. There is higher cortical guidance of the sites of attention feedback, this being either pre-set (in the endogenous case) or arising by breakthrough of lower-level salient inputs in the exogenous case.

Consciousness

There is not nearly as much understood about consciousness as there is about attention. This is partly due to the more subtle feature of consciousness as possessing the subjective experiential component mentioned earlier. There have been important advances, using brain imaging, to determine brain regions in which the presence of activity is correlated with consciousness of a related stimulus having a representation coded there. These brain regions are denoted 'neural correlates of consciousness' and have been detected in increasing numbers of paradigms and in a variety of sites in the brain. This has immeasurably advanced this 'topographic' study of consciousness, but has been slower in granting any understanding of consciousness itself, especially in terms of its subjective character. Furthermore the problem of constructing an acceptable model of how consciousness might be created by brain activity, at the same level of detail as for the action of attention, appears to be very far from being resolved.

Important advances in the analysis of brain sites coding for 'self' have also been made using brain imaging, which in general have emphasised the importance of cortical mid-line sites (especially congulate and posterior parietal) for the coding of representations of the traits of the self

However the aspects of self being explored by these studies have correctly to be regarded as those of the 'reflective' self - that which possesses attributes that are available to all and sundry, as lists of character traits ('is short-tempered', 'is rather brusque'), of special marks of appearance ('has a beard') and so on. These are still not allowing us to track down something much deeper - the 'pre-reflective self' or the 'inner self'. It is this component of the self that stands to one side as the ever present 'I' in one's own experience, and has so far resisted being detected at all by any of the methods now being employed so successfully in analysing the brain by neuroscience. It is the pre-reflective self of the Western phenomeonologists (Zahavi, 1999) that is at issue here. This is also a component emphasised very effectively by Block (Block, 1995), as a companion to what he termed 'access consciousness' (which involves representations of stimuli able to be broadcast round a brain and used for reasoning and other purposes). Indeed the distinction which can be drawn with general support about components of consciousness is that there is a 'content' component, Block's A-consciousness, and an inner self or 'I', Block's P-consciousness and answering to Nagel's question of 'What is it like to be?' (Nagel, 1974).

Whilst it is acceptable that A-consciousness is created by some competitive or threshold neural process involved in some way with attention, and with recurrence and other dynamic processes contributing, as suggested by (Lamme, Pollen and Grossberg), there is no similar set of models that could lead to such an understanding of the pre-reflective self or P-consciousness. It appears as if the 'I' is opaque to all the methods of science. It is possible however that in attempting to disentangle the pre-reflective self at the heart of consciousness from attention (as a possible gateway to consciousness) that real progress could be made. This could allow a better understanding of how the inner and contentful selves (in Block's terms P and A consciousness) interact and function together. It is therefore important to turn to analyse those cases where attention and consciousness seem to come apart: in some of those case the inner self may be hived off or accessed altogether separately from that of content.

Is Attention Really Crucial for Consciousness?

In general the experimental results from various paradigms seem to indicte that attention and consciousness are at least to a certain degree independent. As pointed out in the important discussion in (Koch & Tsuchiya, 2006) numbers of effects of stimuli outside awareness can be observed when attention is directed to them, as in the aftereffects from peripheral oriented gratings, masked so as to be outside awareness, or of the priming of stimuli caused by invisible words, and of attention dragged to them by suitably salient stimuli presented subliminally. On the other hand awareness can occur with little attention, as in the case of the gist being detected of pictures in a dual task (koch & Tsuchiya, 2006).

These results seem to imply that in the 2x2 table of conscious v non-consious against attended v non-attended all of the four possible entries are filled, especially the cases of the off-diagonal elements. Thse are the two important cases of

1) Non-attended but aware stimuli;
2) Attended but non-aware stimili.

That these two cases have non-trivial entries indicates that attention and consciousness really are independent. However the situation is muddied by the equating, in (Koch & Tsuchya, 2006), of attention with top-down attention. However, as we noted in section 2, the bottom-up form of attention is not only importantly present but is processed by similar brain circuits. Thus the exercise of filling in paradigms to the two case 1) and 2) above, in particular, must be performed for the broader sense of attention. In that case the paradigms remaining in the analysis of (Koch & Tsuchiya are: 1) Dual-task paradigms (especially that of Koch& Tsuchiya, 2006) (gist can now be handled by broad-based bottom-up attention, which can be fast); 2) The cases mentioned above under this heading involve the manner in which consciousness can be manipulated by earlier unaware but attended exposure; this can be envisaged in general as occurring by no access to buffer working memory sites of the unaware stimuli, as observed, for exaple very clearly in the attentional blink (Voel et al, 1998), with later modulation of access of related or the same stimulus to a buffer working memory caused by the unaware stimulus. Thus the awareness process is not outide attention but is affected by the effects of nonconscious stimuli being processed earier possibly up to semantic level.

In all we seem to be left with paradigms fitting the entries 1) above. We should note, however, that on the basis of various short-term memory experiments it has been claimed (Lamme, 2003, 2006) that consciousness is independent of attention and can arise from low level recurrent processing in, for example, cerebral cortex, so fitting entry 1) above. The paradigms claimed to justify this, especially a form of change blindeness, have been analysed in detail (Taylor, 2006a), where it was shown that they can be explained by attention processes under the CODAM model of attention control (Taylor, 2000). The specific paradigm did not use any outdoor scenes (which have been used in many studies of change blindness, for example), but involved eight objects presented simultaneously place equally round a circle, with a uniform grey mask presented for a range of times from 200 - 1500 msecs (the stimuli were on for 500 msecs). After the mask the stimuli were re-presented, although with one of them possibly having had its orientation or shape changed (but not its position). The subject then had to respond as to the change of orientation of an object at a cued position. The cue could have been presented at the first presentation of the objects, during the mask or during the second presentation of the objects.

The paradigm was modelled by inclusion in the model's attention control circuitry of buffer working memory sites (supposedly similar to such a mechanism in the human brain). The model is thus able to hold activity for several seconds in such buffer sites, as in a change-blindness type of paradigm in which an initial stimulus is turned off and then another presented to a subject a second or so later for comparison with the earlier stimulus. Such buffer working memories were proposed in the original model of distributed working memory of Baddeley and Hitch (Baddeley, 1986); the existence of such buffer sites has been given ample justification by numerous brain imaging experiments observing continued posterior activity after stimulus offset. They allowed an initial explanation of the data of Lamme by a particular form of the 'attention leads to consciousness' processing type of model. It was possible, using this extended attention model, to fit the data on success in this paradigm for the various cue types and timings (Taylor, 2006a). However this is only tentative as yet, and more experiments and modelling, for example of subject error levels in such change blindness task experiments for different stimulus timing parameters, needs to be followed up.

More recent claims that attention and consciousness are independent under certain circumstances (Koch & Tsuchiya, 2006) depend specifically on a paradigm also fitting case 1) above. This paradigm involves 10 hours of training of subjects to learn to detect animal features in peripheral visual scenes presented for 27 msecs at the same time as a foveal set of letters requiring recognition; both tasks could be performed together effectively without interference. However there is considerable knowledge of the ability of humans to learn to develop an automatic route for response after suitable training, and some of the brain regions involved have been observed at work (Pollman & Maertens, 2005). At the same time attention is known to have the ability to focus on up to four separate regions in space (McMains & Somers, 2005) as well as being able to track about four objects (Cavanagh & Alvarez, 2005). Thus either or both of these mechanisms (automatic processing or splitting the focus of attention) can help explain these results. There was apparently no questioning of the subjects for awareness of the details of the peripheral scene reported in (Koch & Tsuchiya, 2006) so there is no indication as to which one of the routes (split attention or automatic responding) was being used. There are the alternate and simpler explanations of automaticity or split attention, so the experimental results do not negate the claims that attention is a critical gateway to consciousness. However yet again the experimental paradigm of (Koch & Tsuchiya, 2006) and of related experiments needs to be investigated further, especially in association with the experience of the subjects of the peripheral stimulus, to determine their level of awareness.

We thus have the situation that currently attention does appear to be an important gateway to consciousness. There may be phenomena to be discovered in the future or certain paradigms to be analysed further which indicate that consciousness can arise without attention. Until these are properly validated, however, we are left without any strong evidence or any clear mechanism as to consciousness arising outside attention. The large majority of cases where stimulus awareness occurs is from attention being focussed on the stimulus; only in very special cases (such as after 10 hours of training or suitable change blindness experiences) could consciousness possibly arise outside attention.

These preliminary hints of a possible relation of interdependence of attention and consciousness do not at first seem to help us in our attempt to discover the 'inner' or pre-reflective self supposedly at the centre of our conscious experience. We turn to an alternative approach, which has made considerable progress on understanding of breakdown in schizophrenia, for example, by a global view of how the brain functions obtained by using ideas of motor control and self-monitoring (Feinberg, 1978; Frith, 1992). In particular this approach tackles the difficult problem of how the self is distinguished from the 'other' in motor action by some form of monitoring of self action as compared to passive movement caused by an external agent. The manner of any involvement of 'I' will be crucial here.

The Self as Monitor?

There have been important developments in the use of internal models to help understand the manner in which the brain controls motor responses. In particular considerable evidence has been brought forward from a range of experimental paradigms for the existence of forward and inverse models in motor control (Desmurget & Grafton, 2000). These models have been developed to help understand an extensive range of motor control aspects (such as in the MOSAIC ensemble of internal models) and quantitative and qualitative successes obtained in various applications. A similar approach has been suggested in attention, culminating in the CODAM model of sensory attention control (Taylor, 2000, 2003, 2006). This latter model has also been extended to the arena of motor attention or intention (Taylor & Fragopanagos, 2003) with explanation of a number of detailed motor control features beyond those obtainable by direct use of the original internal models. Thus this model has been applied in particular to the Posner motor paradigm, in which a precue indicates which motor response (of two possible responses) should be prepared, and a speeded response then made when a final cue is presented to a valid precue. The data arising from both normals and those with left or right parietal damage were able to be fit by the simulation based on a suitably coupled motor and visual attention control system (Taylor & Fragopanagos, 2003).

In many of these internal model approaches, either to motor control, to sensory attention control or in motor attention control, an important component is a monitor module, which checks for errors and brings about corrections so as to prevent further errors. Use of a copy of the motor or visual attention movement control signal allows the resulting control system to be able to provide an early error signal, using a forward model estimator of the next state updated by use of the efference copy of the control signal.

There have been numerous suggestions of how the experience of self can arise from some form of self-monitoring. In particular there has been an important recent discussion in (Vignemont & Fourneret, 2004), continuing earlier work by Feinberg and Frith (Feinberg, 1978; Frith, 1992) as to the manner in which the sense of agency of actions can be ascribed to a comparison (through a monitor) of a motor target state, a predicted state and an estimate of the actual state. They conclude that there are two forms of bodily movement agency: a) of initiating an action, by comparison of the target with the predicted state; b) of continued movement, by comparison of the predicted state with the estimated state. These are important results (still to be universally accepted), but our purpose to discuss them here is not only to look at the 'motoric' components of the self but also to ask if these analyses can also be applied to the movement of attention.

We might expect that in the movement of attention, it is the sense of initiating the movement that is crucial; that of continuing the movement of attention would be expected to apply only to holding attention to a given focus. However that might not be so: one can consider the first movement of attention as arising from a signal guided by the target to be attended to. Any error here would arise from the comparison between the target and the predicted state (from a forward model for attention state prediction using a corollary discharge of the attention movement control signal, as occurs in case a) above). It is exactly this signal that is used as an early error signal in CODAM to enable fast correction to be made to the attention control signal. However there is also the later possibility of comparing the predicted or the target state with the estimated state (as would only arise after attention movement had begun). The former of these corresponds to case b) above, as proposed by (Vignemont & Fourneret, 2004) for being behind the experience of continued motor response. The third case we have introduced for attention, that of comparing the estimated state against the target one, is an important quantity equal to the real error, not that predicted. Thus this error is important in making final error corrections: it could be related to the ERN (error related negativity) noted as arising in EEG studies of motor response. There is still the comparison of predicted and actual attention state results; this could be used in giving a sense of movement of attention after it has started.

It is to be expected that these comparisons will be calculated by the brain in order to improve the efficiency of attention movement. How they are thereby used to create experience of the two types of agency of attention movement (that of initiation and that of continued movement) is unknown. In any case how they are related to the notion of self is also unknown. CODAM uses the first of these comparisons (that of initiation) as the experience of the 'owner' of the resulting attended stimulus activity once it has reached its buffer working memory for report. The similar ownership experience expected to be created by the movement agency would only be appropriate for that of continued attention movement, but is difficult to relate to experience (awareness) of any external stimulus activity that reaches a suitable buffer site. The predicted attended state is that of the stimulus being attended to. The actual state is thus expected to be the same, unless there is a serious error in the attention movement control circuitry, in which case error-related learning is needed to take place to modify the predictor error to become zero. But this does not seem to involve any component of the self. Provided the forward attention model is working correctly there will be no error for movement. Thus an efficient attention movement control system should not have room for more than an experience of initiation, one already used for that purpose, it was suggested in CODAM. However more detailed development and analysis of relevant experiments to attempt to disern the presence of the attention error signal is relevant here.

Thus the use of monitors to generate self experiences under the attention control movement system is seen to involve at least two components and fitting those expected in consciousness in our earlier discussion: the content and the onner or pre-reflective self. A possible third componnt is hinted at but may only expectd to be involved in creating an effcient attention movement system; this aspect needs to be analysed further.

The Neural Correlates of Consciousness (NCC)

The components of brain activity crucially involved in the creation of the conscious experience contain at least those neural activities that are involved in the crucial ability to be used in making the buffer working memory activity reportable. What these are for a particular stimulus may depend on exactly what purpose the activity is to be used for as part of report. This could involve extensive detail, as in the awareness of fine and detailed structures in a visual stimulus. On the other hand the NNC could be quite limited owing to there being little demand on the attended stimulus and its lower cortical level components. However these levels of detail could change in time, so the nature of the requisite binding of higher and lower level components of NCC could vary in a highly dynamic manner through time.

The NCC are thus observable aspects of consciousness at work. But they do not seem to be the totality of that dynamic: there will be the ownership components which may also change in time with the attended stimulus, although not necessarily with the various contextual variations mentioned in the previous paragraph.

We thus arrive at the crucial question to be answered: is it possible to detect the inner self component of consciousness in and amongst the NCC? This is not only important to understand this second component of consciousness suggested earlier but even more importantly to determine if it exists. There are many proposed models of consciousness which do not have this component. Thus they have the content component of consciousness in abundance, together possibly with complex dynamics involving recurrent neural networks with much re-afference, complexity and frequency binding. However they do not seem to contain any 'I' as owner: there is no-one 'at home', it would seem in these models. Thus we desire to find the owner if that is possible. It may be that the owner is not present in the NCC signals; if so we should consider where else it might be found in the brain and what its signature might be. Thus if the owner is always 'home' when the subject is conscious then activity switched off in sleep (or only in slow-wave sleep) may be more relevant than the NCCs associated with specific psychological paradigms.

We may search for the ownership signal across the spatial extent of the brain or during its temporal flow of activity to the creation of the experience of consciousness content, supposedly in the relevant buffer working memory sites. But we should also look carefully at the complex dynamic flow of activity across these various brain sites. It is there, we suspect, that the pre-reflective self might be hidden. The parietal lobe is beginning to give up some of its secrets as to how the information flows between its various components in a range of attention paradigms. This investigation has especially been moved forward by the use of trans-cranial magnetic stimulation or TMS (Chambers & Mattingley, 2005). The results of this are that there is early activation of the supramarginal gyrus in two stretches: initially at about 110-130 msecs post stimulus onset, and then at about 230-250 msecs. Other parietal regions are also known to be sensitive to TMS at slightly different times (Chambers & Mattingley, 2005). The picture is still far from clear as to this delicate flow of activity back and forth across the many regions of the parietal lobe. It is expected also to vary for diffrerent attention paradigms. However careful analysis should gradually make clear what the overall flow of information is, and how activtiy at a given time is crucial to the two components of consciousness which have been suggested as basic to its composition. It is expected that in the near future subjects will be able, through TMS, to have various of the content components of consciousness modified or removed; hopfeully it will also be possible to analyse when and where the ownership experience arises.

Conclusions

In this article we have covered the nature of attention and consciousness, the present controversy over the nature of the dependency of one on the other, and the manner in which time-sensitive experimental study of the intermediate stages of information processing, especially in the parietal areas, should be able to begin to uncover at the same time the early stage of attention and the beginnings of the creation of the 'ownership' experience. We have also discussed if there is any light cast on consciousness by the motor-monitor approach to self and other, crucial in ascribing to oneself movement of one's body. It was proposed that TMS may prove a crucial tool in being able to reduce or remove this ownership experience, so that subjects begin to feel that (or lose the feeling that) what they are experiencing is their own. A more detailed analyhsis of the relation of such experiences to some of those reported by schizophrenic and by meditators would also be enlightening.

References

• Baddeley A (1986) Working Memory. Oxford: Oxford University Press
• Binkofski F, Fink GR, Geyer S, Buccino G, Gruber O, Shah NJ, Taylor JG, Seitz J, Zilles K & Freund H-J (2002) Neural Activity in Human Primary Motor Cortex: Areas 4a and 4p are Modulated Differentially by Attention to Action. J Neurophysiology 88:514-519
• Block NJ(1995) On a Confusion About the Function of Consciousness. Behavioural and Brain Sciences18:227-247
• Cavanagh P & Alvarez GA (2005) Tracking multiple targets with multifocal attention. Trends in Cognitive Sciences 9(7):349-354
• Chambers CD & Mattingley JB (2005) Neurodisruption of selective attention: insights and implications. Tresnds in Cognitive Neurosciences 9:542-550
• Corbetta M & Shulman GL (2002) Control of goal-directed and stimulus-driven attention in the brain. Nat Revs Neuroscience 3:201-215
• Desmurget M & Grafton S (2000) Forward modelling allows feedback control for fast reaching movements. Trends in Cognitive Sciences 11:423-431
• Feinberg I (1978) Efference copy and corollary discharge: Implications for thinking and its disorder. Schizophrenia Bulletin 4:636-640
• Frith C (1992) The cognitive neuropyschology of schizophrenia. Hove: Erlbaum
• Giesbrecht B, Woldorff MG, Song AW & Mangun GR (2003) Neural mechanisms of top-down control during spatial and feature attention. NeuroImage 19:496-512
• Koch & Tsuchiya (2006) Attention and Consciousness: two distinct brain processes. Trends in Cognitive Sciences 11(1):16-22
• Kincaide JM, Abrams RA Astafiev SV, Shulman G & Corbetta M (2005) An event-related functional magnetic resonance imaging tudy of voluntary and stimulus-driven orienting of attention. J Neuroscience 25:4593-4604
• Lamme V(2003) Why visual attention and awareness are different. Trends in Cognitive Sciences 7:12-18
• Lamme V (2006) Towards a true neural stance on consciousness. Trends in Cognitive Sciences 10(11):494-501
• McMains SA & Somers DC (2004) Multiple Spotlights of Attentional Selection in Human Visual Cortex. Neuron 42:677-688
• McMains SA & Somers DC (2005) Processing Efficiency of Divided Spatial Attention Mechanisms in Human Visual Cortex. Journal of Neuroscience 25(41):9444-9448
• Mehta AD, Ulbert I & Schroeder CE (2000) Intermodal Selective Attention in Monkeys I & II. Cerebral Cortex 10:343-358 & 359-370
• Reynolds J, Chelazzi L & Desimone R (1999) Competitive mechanisms subserve attention in macaque area V2 and V4. J Neuroscience 19(5):1736-1753
• Rushworth M, Krams M & Passingham RE (2001) The attentional role of the left parietal cortex: the distinct lateralization and localization of motor attention in the human brain. J Cog Neuroscience 13:698-710
• Taylor JG & Fragopanagos N (2003) Modelling the fusion of attention and motor control. Proc ICANN2003
• Taylor, JG Shapiro K & Nobre CA (editors)(2006) Special Issue on 'Brain and Attention', Neural Networks 19(9):1321-1462
• Taylor JG (2000) Attentionl movement: The control basis for consciousness. Soc for Neuroscience Abstracts #839.3
• Taylor JG (2003) Paying Attention to Consciousness. Prog in Neurobiology 71:305-335

& Taylor JG (2006) On the Neurodynamics of the Creation of Consciousness. Cognitive Neurodynamics (in press)

• Taylor JG (2005) Mind from Matter: An Answer to the Final Question? Physics of Life Reviews
• Taylor N, Hartley M & Taylor JG (2006) The micro-structure of attention. Neural Networks 19(9):1347-1370
• de Vignemont F & Fourneret P (2004) The sense of agency: A philosophical and empirical review of the "Who" system. Consciousness and Cognition 13:1-19
• Vogel EK, Luck SJ & Shapiro KL (1998) Electrophysiologicl evidence for a postperceptual locusof suppression during the attentional blink. J Exp Pschol Hum Percept Perform 24:1656-1674
• Wang X-J (1999) Synaptic reverbrations underlying mnemonic persistent memory. Trends in Neuroscience 24:455-463
• Yantis S, Schwarzbach J, Serences JT, Carlsen RI, Stemmer MA Pekar JJ & Courtney SM(2002) Transient neural activity in human parietal cortex during spatial attention shifts. Nat Neuroscience 5:995-1002
• Zahavi D (1999) Self-awareness and Alterity. Evanston, Ill: Northwestern University Press