From Towards a Science of Consciousness 3         Section 4: Vision and Consciousness -- Introduction       CogNet Proceedings

Attending, Seeing and Knowing in Blindsight

Robert W. Kentridge, Charles A. Heywood, and Lawrence Weiskrantz

Blindsight is the term coined by Weiskrantz (see Weiskrantz et al. 1974) to describe the condition in which subjects with damage to their primary visual cortex are able to perform simple visual tasks in the area of visual space corresponding to their brain-damage while maintaining that they have no visual experience there. In other words, they retain some visual abilities in an area where they report that they are phenomenally blind. This dissociation between conscious experience and visual performance is usually revealed in forced-choice tasks involving discriminations of simple stimulus properties such as location, contrast, orientation, color and so on. However, while conducting an experiment mapping the area of one subject's residual vision (Kentridge et al. 1997), we recently chanced upon an exception. In this experiment the subject had to decide which of two tones was accompanied by a flashing spot of light. By examining how his performance varied as the spot of light was presented in different positions we could map the area over which his blindsight extended. Quite by chance, during one of the breaks in testing the subject (known as GY) remarked that he had just realised that the stimuli were sometimes being presented well above the horizontal and so now he was trying to pay attention higher up in his blind visual field. This is an extraordinary remark since one's intuition is that it is attention that gives rise to consciousness. Our subjective experience is that we are most conscious of that part of the world to which we are attending. This apparently close relationship between attention and consciousness was remarked upon from the birth of modern psychology (see, for example, James [1890], Wundt [1912] and still influences many modern theories of consciousness. We followed the observation up in a series of experiments designed to establish exactly which aspects of attention continue to function in the blind region of GY's visual field. These experiments are reported in detail elsewhere (Kentridge et al., submitted), in the present chapter we will briefly summarize their results and then consider their implications for theories of consciousness and the central bases of visual attention.

Types of Attention


Attention has been studied, subdivided and categorized extensively (see, e.g., Pashler 1998). Two aspects of attention may influence awareness. First, the nature of the stimulus to which we are attending-are we simply attending to a specific location within the visual field or are we looking out for objects with a particular property, for example red things? Second, how was our attention initially directed-did we decide to attend voluntarily or did something unexpected or sudden catch our attention by itself? Both of these aspects of attention, the nature of its objective and the nature of its control, have featured in modern theories linking attention and consciousness.

Milner and Goodale (1996) suggested that we are conscious of stimuli that match a property to which we are attending, in other words, attention in the service of object identification gives rise to awareness. In contrast, stimuli appearing at an attended spatial location do not necessarily give rise to awareness. In other words, attention in the service of object location need not give rise to awareness. On the basis of neuropsychological evidence Milner and Goodale identify processes leading to object identification with pathways leaving the primary visual cortex and passing ventrally toward inferotemporal cortex. In contrast, processes associated with directing actions toward objects, and hence with their spatial characteristics, with pathways leaving the visual cortex and passing dorsally toward the parietal cortex (see figure 13.1).

Posner (1994) has suggested that a dissociation between attention and awareness can also be made in terms of the nature of attentional control. The voluntary direction of attention, in which memories are invoked in order to guide attention, is associated with awareness, whereas automatic direction of attention, in which a sensory stimulus captures attention for the processing of subsequent stimuli, can take place without awareness. Using brain-imaging Posner identifies the voluntary control of attention with activity in anterior cingulate and dorsolateral prefrontal cortex and automatic direction of attention with parietal cortex (see figure 13.2).

It should be clear that Milner and Goodale's and Posner's proposals are not in conflict with one another, they are, however, not entirely orthogonal. The automatic direction of attention is almost inevitably spatial whereas attention can be voluntarily directed either on the basis of location or object features. Our experiments addressed the question of whether attention could be automatically or voluntarily directed to target locations within the scotoma of a blindsight patient. It has been found that blindsight patients can exhibit different modes of awareness within their scotomata. In addition to blindsight, the presence of residual visual function with no acknowledged awareness, these patients also sometimes report awareness of events within their scotoma. This awareness is not accompanied by report of normal visual experience, rather, it is described as a sense that "something happened," that a stimulus gives rise to "awareness, but you don't see it," that decisions are not being based purely on guesswork. This "aware" mode of perception can be produced using moving or transient stimuli of high contrast (see, e.g., Weiskrantz et al. 1995). The ability to manipulate awareness in a blindsight patient without producing any report of visual qualia allowed us to investigate the role of awareness in both the automatic and voluntary direction of attention.

Attention without awareness in blindsight


We used Posner's attentional cueing paradigm (Posner 1980) to assess the effects of various cues on GY's reaction time to stimuli presented within his scotoma. GY's task was to determine whether or not a tone was accompanied by a visual target presented within his scotoma. A cue preceded each trial that usually indicated the likely location of a subsequent visual target. One some trials, however, the cue signalled an incorrect location. If the cues were being used to direct attention then we would expect to find significantly quicker reaction times on trials where the cue indicated the correct location compared to those from trials where the cue was misleading. In addition to measuring the effects of cueing on his reaction time we could also assess GY's accuracy in discriminaing the presence of targets. The possibility that reaction-time advantages produced by cueing were due to changes in strategy, trading-off accuracy for speed rather than gaining a speed advantage with no loss of accuracy by virtue of attending to the target location, could be discounted if reaction-time advantages are not accompanied by decreases in discrimination accuracy.

First we established that GY could indeed direct attention within his scotoma using cues that we presented in an undamaged area of his visual field pointing toward target locations in his blind field (the sequence of stimuli used in this experiment and those about to be desribed are shown in figure 13.3). We then went on to assess the role of awareness in voluntary and automatic direction of spatial attention. All cues were now presented within GY's scotoma, as were the targets. We manipulated GY's awareness of cues by using two different cue contrasts. In one set of experiments cues (apart from the misleading ones) and targets were presented at the same locations. These cues could automatically direct attention to the target location. In a second set of experiments we instructed GY that when a cue appeared at one location it indicated that a target was likely to appear at a second specified location. Appropriate direction of attention relied on recall of this instruction and so was necessarily voluntary-automatic engagement of attention to the location of the cue would not aid detection of the target, which was at another location in these experiments. We found that GY could only direct his attention voluntarily with high contrast cues that produced reports of awareness on nearly all trials. GY's attention could, however, be directed automatically regardless of the cue-contrast. These results are summarised in figure 13.4.

Anatomical bases for blindsight


What implications do these results have for our understanding of the anatomy of attention, awareness and qualia? GY's scotoma resulted from a road accident he suffered aged 8. The accident produced a unilateral occipital lesion, apparently restricted to striate cortex (area V1) as revealed by neuroimaging (Blythe et al. 1987, Barbur et al. 1993). There are a number of possible bases for GY's residual vision given this lesion. One suggestion (Campion et al. 1983, Fendrich et al. 1995) is that the damage to GY's striate cortex is incomplete and his residual vision depends upon small patches of spared cortex within the volume of the apparent lesion. There are a number of reasons to doubt this explanation. First the area of GY's residual vision did not appear patchy when mapped using procedures to eliminate possible confounding effects of eye-movements (Kentridge et al. 1997). Second, functional magnetic resonance imaging of brain activity in another blindsight patient showed that stimuli presented in his blind field did not produce any detectable activation of primary visual cortex but did result in an increase in activity in extrastirate visual areas (Stoerig et al. 1998). This evidence shows that blindsight cannot be a purely subcortical phenomenon. Blindsight may, under some circumstances, be mediated by pathways that reach extrastriate cortical areas without passing through primary visual cortex. There are two possible sources, direct projections from the lateral geniculate nucleus to areas V2 and V4 (Cowey and Stoerig 1989) and projections, possibly via the superior colliculus, to the pulvinar and on to areas MT (Standage and Benvenuto 1983) and V4 (Ungerleider and Desimone 1986).

Anatomical bases for attention


How could these potential sources of residual function in blindsight link up with the anatomies of voluntary and automatic attention? Should we be surprised that attention can be directed in the absence of primary visual cortex? An elegant experiment by He et al. (1996) showed that the after-effects of looking at a pattern with a particular orientation in a particular location and the ability to attend to just that location could be dissociated. The range over which the after-effect occurred was much smaller than the distance over which distactors items could be suppressed in an attentional task. The scale over which the after-effect and the focus of attention act are likely to depend on the receptive field sizes of cells in the brain areas that mediate these two processes. As the after-effect is known to be due adaptation in primary visual cortex and the scale over which the focus of attention operates is different from the scale over which the after effect operates it is unlikely that attentional selection is occurring is primary visual cortex. This aspect of their conclusion is unsurprising since it has been known since the mid 1980s that attention modulates neuronal activity in extrastriate cortex but has little effect on striate cortex (Moran and Desimone 1985). The key conclusion of the He et al. paper, however, is that is "that the attentional filter acts in one or more higher visual cortical areas to restrict the availability of visual information to conscious awareness" (p. 335). We show that this cannot be the case. Attentional selection can take place without awareness, so selection per se, is not sufficient for awareness.

There is more to attention than selection. Consider the steps involving attention in our experimental designs using peripheral cues. Initially, as the subject's attention is not directed to either of the target locations, the appearance of a cue in the periphery must capture attention, disengaging it from its current focus. Once attention has been captured by the cue, attention must be redirected to its new focus. This redirection may be automatic or voluntary. The newly redirected focus of attention must now selectively enhance the processing of stimuli that match the focus of attention (in this case, location). The consequences of this enhanced processing may be an increased likelihood of perceiving the attended stimulus or a predisposition to act in response to it. The posterior parietal lobe is clearly implicated in the capture of attention. Hemispatial neglect is a neuropsychological disorder usually caused by unilateral damage to the parietal lobe. Although parietal syndrome may have many components, including somatosensory and oculomotor deficits (see, e.g., Cole et al. 1962, Ishiai et al. 1979) its most studied aspect is patients inability to redirect their attention to the visual hemifield contralateral to their lesion (Posner et al. 1984). This attentional component has been specifically associated with lesions of the inferior parietal lobule (Galletti et al. 1997). Electrophysiological studies in monkeys have shown that attending to a location at which subsequent targets are likely to appear reduces the sensitivity of a majority of parietal neurons (in area 7a) to stimuli (Steinmetz et al. 1994) at the attended location. The implication is that the sensitivity of neurons in nonattended locations is enhanced in order to allow peripheral events to capture attention. Posterior parietal cortex receives many projections from the pulvinar (see, e.g., Asanuma et al. 1985) in addition to those from striate and extrastriate cortex. It is reasonable to assume that it can be activated in blindsight in the absence of striate cortex.

The automatic redirection of attention to the location at which a cue appeared may not require control by areas beyond parietal cortex. Activity in parietal areas is revealed by PET studies in tasks both voluntary and automatic attention (e.g., Corbetta et al. 1993). Constantinidis and Steinmetz (1996) showed that during the interval between the presentation of cue and target, when no stimulus was therefore present, activity in a subpopulation of neurons within area 7a of parietal cortex of monkeys was elevated in the region corresponding to the cue's location. The relationship between the set of neurons whose sensitivity is suppressed once attention has been captured by a stimulus and the set of neurons whose activity is enhanced is not known. These studies indicate that the automatic redirection of attention may not require control from areas beyond the parietal lobe. This is not the case when attention must be voluntarily redirected. In this case the new focus of attention is depends on interpreting cues in the light of some remembered rule or meaning. When tasks in which attention is automatically engaged are compared to those requiring voluntary control differences in PET are found in frontal areas but not parietal ones. Corbetta et al. (1993) report activation of superior frontal cortex and anterior cingulate cortex in voluntary but not in automatic tasks.

In 1985 Moran and Desimone (1985) reported that cells in cortical area V4 and in the inferior temporal cortex of monkeys responded much more strongly to stimuli when the monkeys were attending to them than when the stimuli were being ignored. The enhancement of the response of a specific subset of cells is a very good candidate for the selective mechanism of attention. Since Moran and Desimone's (1985) paper similar selective enhancements have been reported in a number of visual areas, for example, MT and MST (Treue and Maunsell 1996), V1 (striate cortex), V2, V4 (Motter 1993). There is some controversy over evidence for attention modulation of responsivity in striate cortex. Motter (1993) reports attentional modulation when using large number of distractors whereas Moran and Desimone (1985) failed to find any. Although selection may take place in V1, it probably operates more strongly in extrastriate areas, particularly V4 (Desimone and Duncan 1995). Given the fact that, among the many thalamo-cortical connections that bypass striate cortex, there are projections from both the LGN and the pulvinar to V4 we should not be surprised that attentional selection can operate in blindsight.

Awareness, control and qualia


The evidence outlined above is compatible with the capture, redirection and selection mechanism of attention in blindsight. Remember, however, that all of this is occurring without normal visual experience. One of the most curious features of attention in blindsight is its voluntary control in response to cues within the blind field. Although this was only effective when GY reported awareness of stimuli, he still denied normal visual experience of them. Although he was aware of stimuli he did not "see" them. As we have seen, studies of voluntary attention suggest that it involves prefrontal areas, in particular dorsolateral prefrontal cortex and anterior cingulate cortex. Willed action (Frith et al. 1991), spatial memory (Ungerleider et al. 1998) and the suppression of habitual responses to remembered instructions (Jahanshahi et al. 1998) have also all been associated with activity in dorsolateral prefrontal cortex. Dorsolateral prefrontal cortex has also been shown to be active in GY when he reports awareness of a stimulus but not when he makes successful discriminations without awareness (Sahraie et al. 1997), although superior colliculus and medial and orbital prefrontal cortices are activated in this unaware blindsight mode. It therefore appears that while the involvement of dorsolateral prefrontal cortex in the voluntary direction of attention may give rise to awareness, on its own it does not give rise to visual qualia, but rather to a feeling of knowing free from qualia, what James (1890) referred to as fringe consciousness (see Mangan 1993 for a modern review of this concept). This fringe consciousness is sufficient to allow the retrieval of a rule from memory and a resultant attentional modulation of responses to targets, but, it does not give rise to visual qualia. This awareness is sufficient for control but not experience. Although awareness of visual cues defined by luminance contrast may be mediated by activity reaching dorsolateral prefrontal cortex by a route that bypasses striate cortex, the experience of visual qualia is not complete without striate cortex.

We can draw the following conclusions:

* Attention can be directed to locations within the scotoma of a blindsight patient.

* Attention can be directed by cues within the scotoma of a blindsight patient.

* Although the automatic direction of attention need not be associated with awareness, the voluntary direction of attention requires awareness.

* There is no conflict between current understanding of the anatomy of attention and both voluntary and automatic control of attention bypassing striate cortex.

* The activation of dosrolateral prefrontal cortex in GY's aware mode revealed with functional magnetic resonance imaging by Sahraie et al. (1997) and the selective involvement of dosrolateral prefrontal cortex in voluntary but not automatic control of attention indicate that GY's "aware" mode is sufficient for the engagement of rules from memory and the control of action via attention.

* 'Aware' mode is not seeing, activation of dosrolateral prefrontal cortex may coincide with consciousness in the form of a feeling of knowing, but, without striate cortex it is does not produce visual qualia.

References


J. Comp. Neurol. 241:357-81.

Barbur, J. L., J.D.G. Watson, R.S.J. Frakowiak, and S. Zeki. 1993. Conscious visual perception without V1. In Brain 116:1293-1302.

Blythe, I. M., C. Kennard, and K. H. Ruddock. 1987. Residual vision in patients with retrogeniculate lesions of the visual pathways. In Brain 110:887-894.

Cole, M., H. S. Shutta, and W. K. Warrington. 1962. Visual disorientation in homonymous halffields. In Neurology 12:257-263.

Constantinidis, C., and M.A. Steinmetz. 1996. Neuronal activity in posterior parietal area 7a during the delay periods of a spatial memory task. In J. Neurophysiol. 76:1352-1355.

Corbetta, M., F. M. Miezin, G. L. Shulman, and S. E. Petersen. 1993. A PET study of visuospatial attention. In J. Neurosci. 13:1202-1226.

Cowey, A., and P. Stoerig. 1989. Projection patterns of surviving neurons in the dorsal lateral geniculate nucleus following discrete lesions of striate cortex: implications for residual vision. In Exp. Brain Res. 75:631-638

Desimone, R., and J. Duncan. 1995. Neural mechanisms of selective visual attenton. In Ann. Rev. Neurosci. 18:193-222.

Frith, C. D., K. Friston, P.F. Liddle, and R. S. Frackowiak. 1991. Willed action and the prefrontal cortex in man: a study with PET. In Proc. R. Soc. Lond. B. Biol. Sci. 244:241-6.

Galletti, C., P. P. Battaglini, and P. Fattori. 1997. The posterior parietal cortex in humans and monkeys. In News in Physiol. Sci. 12:166-171.

He, S., P. Cavanagh, and J. Intriligator. 1996. Attentional resolution and the locus of visual awareness. In Nature 383:334-337.

Ishiai, S., T. Furukawa, and H. Tsukagoshi. 1987. Eye-fixation patterns in homonymous hemianopia and unilateral spatial neglect. In Neuropsychologia 25:675-679.

Jahanshahi, M., P. Profice, R.G. Brown, M.C. Ridding, G. Dirnberger, and J. C. Rothwell. 1998. The effects of transcranial magnetic stimulation on suppression of habitual counting during random number generation. In Brain 121:1533-1544.

James, W. 1890. Principles of psychology. London: Macmillan.

Kentridge, R. W., C. A. Heywood, and L. Weiskrantz. Attention without awareness in blindsight. Submitted to Nature Neuroscience.

Kentridge, R. W., C. A. Heywood, and L. Weiskrantz. 1997. Residual vision in multiple retinal locations within a scotoma: Implications for blindsight. In J. Cog. Neurosci. 9:191-202.

Mangan, B. 1993. Taking pheomenology seriously: The "fringe" and its implications for cognitive research. In Consciousness and Cognition 2:89-108.

Milner, A. D. and M. A. Goodale. 1996. The visual brain in action. Oxford: Oxford University Press.

Moran, J., and R. Desimone. 1985. Selective attention gates visual processing in the extrastriate cortex. In Science 229:782-884.

Motter, B. C. 1993. Focal attention produces spatially selective processing in visual cortical areas V1, V2, and V4 in the presence of competing stimuli. In J. Neurophysiol. 70:909-19

Pashler, H. E. 1998. The psychology of attention. Cambridge, Mass.: MIT Press.

Posner, M. I. 1980. Orienting of attention. In Q. J. Exp. Psychol. 32:3-25.

Posner, M. I. 1994. Attention: the mechanisms of consciousness. In Proc. Natl. Acad. Sci. USA 91:7398-8403.

Posner, M. I., J. A. Walker, F. J. Friedrich, and R. D. Rafal. Effects of parietal injury on covert orienting of attention. In J. Neurosci. 4:1863-84.

Sahraie, A., L. Weiskrantz, J. L. Barbur, A. Simmons, S. C. Williams, and M. J. Brammer. 1997. Pattern of neuronal activity associated with conscious and unconscious processing of visual signals. In Proc. Natl. Acad. Sci.USA 94:9406-11.

Standage, G. P., and L. A. BeneventoA. 1983. The organization of connections between the pulvinar and visual area MT in the macaque monkey. In Brain. Res. 262:288-94.

Steinmetz, M. A., C. E. Connor, C. Constantinidis, and J. R. McLaughlin. 1994. Covert attention suppresses neuronal responses in area 7a of the posterior parietal cortex. In J. Neurophysiol. 72:1020-3.

Stoerig, P., A. Kleinschmidt, and J. Frahm. 1998. No visual responses in denervated V1: high-resolution functional magnetic resonance imaging of a blindsight patient. In Neuroreport 9:21-5.

Treue, S., and J. H. Maunsell. 1996. Attentional modulation of visual motion processing in cortical areas MT and MST. In Nature 382:539-41.

Ungerleider, L. G., S. M. Courtney, and J. V. Haxby. 1998. A neural system for human visual working memory. In Proc. Natl. Acad. Sci. USA 95:883-90.

Ungerleider, L. G., and R. Desimone. 1986. Cortical connections of visual area MT in the macaque. In J. Comp. Neurol. 248:190-222.

Weiskrantz, L., J.L. Barbur, and A. Sahraie. 1995. Parameters affecting conscious versus unconscious visual discrimination with damage to the visual cortex (V1). In Proc. Natl. Acad. Sci. USA 92:6122-6126.

Weiskrantz, L., E.K. Warrington, M.D. Sanders, and J. Marshall. 1974. Visual capacity in the hemianopic field following a restricted occipital ablation. In Brain 97:709-28.

Wundt, W. 1912. Introduction to psychology (Trans. R. Pintner). London: George Allen.