Talk:Attention and consciousness
... top-down attention cannot play much of a role (because gist is a property associated with the entire image, any process that locally enhances features is going to be only of limited use) ...
<review> although attention could pick some objects or scene aspects from the icon, which would last around 1/4 sec, and the subject could then infer gist from those</review>
Author Tsuchiya :
Could you let us know the reference that shows this point? We are not aware of a study that shows the reviewer's point.
Further Comments on Attention and Consciousness
(JG Taylor Dept of Maths KCL: firstname.lastname@example.org)
- I have numerous problems with this whole article, basically because it does not take account of the much greater complexity and subtlety of attention than the authors account for (let alone the even greater subtlety of consciousness). Attention has several important features: it is composed of a sensory and a motor component (handled mainly by the right and left brains respectively, following the work of Rushworth and colleagues); it has multiple foci in vision (from the work of McMains and Somers and many others); it can be used to follow several targets at once (related to its multi-focal character, as discussed by Cavanagh and Alvarez); it occurs either as top-down or bottom-up. These latter components are somewhat distinguishable in the brain, following the work of Maurizio Corbetta et al and many others by brain imaging and single unit studies (in monkeys), although share much of each others networks. The brain mechanisms of attention as a controller (in both the endogenous and exogenous cases) is being uncovered by the most recent work of Desimone, Bressler and numerous colleagues as to the temporal flow of activity in the attention control network. Moreover the endogenous and exogenous forms of attention both use a similar brain network, although there may be difference in timings between the two forms of attention, as would be expected.
- This more complete view of attention indicates the dubious character of some of the claims of the present article. To claim (as do the authors) "... current evidence suggests that top-down attention and perceptual processes are two distinct but often allied processes with distinct neurobiological processes" is on poor grounds. Where are the distinct neurobiological processes occurring as observed in all the brain imaging of attention and of conscious processing that has been proceeding apace? What is observed is instead a seamless control network in which attention feedback signals go back to posterior cortical sites to amplify the lower level representations, and to provide (I would suggest) the detailed content of consciousness when this attention-amplified activity attains some short-term site for general report (as in Baars' Global Workspace). The main difference between the endogenous and exogenous forms of attention seem to be in the different initial sources of bias to the attention movement signal generator observed by numerous methods to be in the superior parietal lobe and the intraparietal sulcus.
- Instead of assuming that consciousness appears independently in this process, I would suggest it may be more useful to look more carefully at the evidence presented by the authors for this independence. The important feature in the author's claim for such independence is in the upper left corner of their table 1: top-down attention is not required, the authors claim, for consciousness in the paradigms they list there. The paradigms proposed to support this claim are: pop-out/ gist/ animal and gender dual tasks/ partial reportability. Let us consider these briefly in turn.
- Pop-out: universally accepted that occurs by exogenous attention, so provided we take account of such attention as part of the overall complex of attention (using similar neural structures as top-down attention, as shown experimentally and mentioned earlier) it presents no problem.
- Iconic memory: This is holding sensory input over a short time for further processing so is likely pre-attentive, although may feed into pop-out. So no problem from the attention point of view.
- Gist: This may occur with a fleeting (30msecs) view but the processing time for the resulting gist to be extracted from a scene is much longer (although not well measured, as far as I know). It could well fit in with the exogenous/endogenous attention control system).
- Dual tasks: Can occur by means of multiple attention foci (but may require trained observers, as noted in some of the papers on this phenomenon). This can be given as an attention-based explanation of the Paustukh and Braun 2007 experimental data cited specifically by the authors.
- Partial reportability: A similar multi-focal account is possible for this, with a lower level of discriminability than normal if the foci are too widely spread.
- The above explanations are given rather briefly due to lack of space, but can be pursued fully and carefully, without change of the interpretation (and even at the level of detailed simulation, as in Taylor and Fragopanagos, 2007). More experimentation needs to be done, for example in the interesting Paustukh and Braun dual task paradigm, to observe multiple (or spread-out) foci at work in the brain. But that was not mentioned in the Koch and Tsuchiya paper, but should be considered seriously instead of claiming that "such findings are difficult to understand within a framework that aligns top-down attention with consciousness", as the present authors suggest. In any case they should expand 'top-down attention' to 'attention', thus including both the top-down form as well as that which is exogenous.
- In all of this discussion there is still the crucial question: where is the sense of 'I' - the inner self? Without such a component there can be no experience of the content being nursed carefully into place by attention in the short-term memory sites. My CODAM model, for example, based on a more complex attention-based model of processing, allows a glimpse of how more than short-term memory is activated through attention processing: so as to create what I proposed as a neural basis for the experience of the inner self. Even if the details of CODAM are incorrect at least it provides a possible basis for attacking the problem of consciousness, and one nowhere hinted at in other models of consciousness, such as of those of Baars, Edelman, Tononi, Rosenthal, etc. As I mentioned, without such a view of the further complexity of consciousness (composed at least of inner self + content) it will never be easy to see how attention could 'contain' consciousness. I would still say, as do many in neurosciences, that attention is necessary for consciousness, but is clearly not sufficient for it. Such a view is very different from that expressed by Koch and Tsuchiya.
- Bressler SL, Tang W, Sylvester CM, Shulman GL and Corbetta M (2008) Top-Down Control of Human Visual Cortex by Frontal and Parietal Cortex in Anticipatory Visual Spatial Attention. Journal of Neuroscience 28(40):10056-10061
- Cavanagh P & Alvarez GA (2005) Tracking multiple targets with multifocal attention. Trends in Cognitive Sciences 9(7):349-354
- Corebtta M, Patel G and Shulman GL (2008) The Reorienting System of the Human Brain: From Environment to Theory of Mind. Neuron 58:306-324
- Gregoriou GG, Gotts SJ, Zhou H and Desimone R (2009) High Frequency Long-Range Coupling Between Prefrontal and Visual Cortex During Attention. Science 324:1207-1210
- McMains SA & Somers DC (2005) Processing Efficiency of Divided Spatial Attention Mechanisms in Human Visual Cortex. Journal of Neuroscience 25(4):9444-9448
- Rushworth MFS, Paus T and Sipila PK (2001) Attention Systems and the Organization of the Human Parietal Cortex Journal of Neuroscience 21(14):5262-5271
- Taylor JG (2009/10) A Neural Model of the Loss of Self in Schizophrenia. Schizophrenia Bulletin (in press)
- Taylor JG (2009/10) The I's Eye View of Consciousness. Journal of Consciousness Studies (in press)
- Taylor JG and Fragopanagos N (2007) Resolving some confusions over attention and consciousness. Neural Networks 20(9):993-1003