The term is defined by the
Oxford Concise Dictionary as '
Medicine: a condition in which the sufferer responds to visual stimuli without consciously perceiving them.' It is associated with damage to human primary visual cortex (V1), which causes 'blindness' in corresponding parts of the affected visual fields, the field defect having a size and shape to be expected from the classical retino-cortical maps (Holmes 1918) (see
visual system: organization). Nevertheless, when subjects are required to 'guess' about stimuli in their blind fields, they may be able to discriminate them from each other or locate them in space, even though they say they do not see them and have no awareness of them. When a stimulus contains rapid transient onsets and offsets, or moves very rapidly, blindsight subjects may say they are 'aware' that something is happening, they may 'feel' it, although they do not 'see' it as such. This has been called 'Blindsight Type 2' in contrast to 'Blindsight Type 1', when discriminations occur in the total absence of any acknowledged awareness.
The historical origins stem from animal research and neuroanatomy. The primate retina, including that of humans, sends its major nerve tract (after a relay in the thalamus) to the visual cortex ('V1', 'striate cortex'). When the striate cortex is damaged or removed or blocked in monkeys, the animals can still carry out visual discriminations although their capacity changes in certain ways. That they can still make such discriminations is, in itself, not surprising because the retinal output also reaches a number of other brain targets, located mainly in the midbrain and thalamus. These routes remain intact even when V1 is entirely removed or damaged, and from these non-striate targets information can be relayed widely to a number of other regions in the brain. The 'extra-striate' tracts from the eye contain fewer nerve fibres than those in the pathway reaching V1, only about one-fifth as many, but this smaller number is not trivial — it amounts to about five times as many fibres as are in the whole auditory nerve. And so the surprise is not that animals can make visual discriminations in the absence of V1: it is that human subjects with damage to V1 claim they are blind. (Because damage to V1 is rarely complete, and is typically confined to one cerebral hemisphere, the region of blindness in most patients is confined to no more than one-half of the visual field, situated contralateral to the damaged hemisphere, although with bilateral cortical damage the blindness covers the entire visual field.)
The surprising discrepancy reflects the fact that the method of testing for visual capacity is usually deeply different in humans and in other animals, rather than there being differences in their brains as such. Humans are typically asked to give verbal descriptions or to comment on visual appearances and differences, whereas animals are trained to make alternative choices for which they are usually rewarded, devoid of any commentary. Even when a human subject is asked to make a discrimination between, say, two colours, he is usually explicitly instructed verbally as to what attribute he should be responding to, and more importantly there is an important implicit assumption that he will be aware of that attribute, or will tell us if he is not. But when the human subject is tested in a manner that is closer to animal methodology, being asked simply to make a forced-choice 'guess' about the visual stimuli whether or not he cannot 'see' them, e.g. whether a visual event is located at
A or
B, or whether it is colour
A or colour
B, or falls in one temporal interval or a second temporal interval, or whether its shape, or colour, or brightness is different in one or the other interval — in other words, tested in the same forced-choice discriminative way as animals are (but without peanut rewards!) — human subjects can match the performance of animals with visual cortex damage even though they may lack any acknowledged awareness of them. Hence, the term 'blindsight'. It is one example of a dissociation in brain-damage patients between an intact capacity and the commentaries about the contents of the capacity (see
brain function and awareness).
Given the counter-intuitive nature of blindsight, early scepticism abounded and questions have been raised about its validity (as was true for earlier examples of implicit processing in neuropsychology). It has been suggested, for example, that there may be stray light falling in the intact visual field, or that the cortical lesion in particular cases is incomplete and patchy, or that subjects really 'see' but deny this, perhaps because of a very conservative criterion, or that their vision is really essentially normal but the percepts are rendered very faint because of the brain damage. All of these alternatives have been roundly and directly addressed in various reviews and experimental analyses; the subject continues to provoke lively discussion not only among neuroscientists but also among philosophers and others interested in the nature of conscious awareness.
Attributes of vision that can be discriminated by blindsight subjects in the absence of their experience of the stimuli include colour, different orientation of lines or gratings, simple shapes, motion, onset and termination of visual events. Interestingly, the emotional expression of unseen faces in the blind field can be 'guessed' at better than chance levels. There are, however, changes in relation to normal vision. By altering the spacing of bars on a sine wave until it can no longer be discriminated from a homogeneous patch, one can measure the subjects' acuity. It is reduced, relative to their normal seeing fields, but is still creditable. Motion perception is retained for simple displacement of a bar or a spot, but more complex motion patterns ('third order motion') seem to be seriously affected. Good colour discrimination remains (again, in the absence of any experience of colour
per se,) but there is a shift towards a relative increase in sensitivity of long wavelengths (red) and a decrease of middle wavelengths (green). Otherwise the spectral sensitivity curve, and its change under dark adaptation, is relatively normal.
Because it is rare that damage to the visual cortex occurs in isolation in clinical cases, relatively 'pure' examples of blindsight may be rare. It is known from animal work that, if the damage extends very far outside the visual cortex, the residual visual capacity is reduced. It is for this reason that most human blindsight research has concentrated on a small number of well-chosen subjects. However, it now appears that this self-imposed restriction may be too conservative. Residual visual function has been reported to occur in the majority of cases of visual cortical damage if additional damage is only moderate.
The distinction between blindsight Type 1 and Type 2 has allowed us to carry out functional brain imaging contrasting states
with awareness and
without awareness, in both conditions using simple movement discriminations, which can be carried out at a high level of success — 90 per cent or better. In the unawareness condition, but not the aware condition, activity is seen in the superior colliculus of the midbrain. This also is active for red stimuli but not for equiluminant green. In contrast, in the aware state, dorsal cortical areas, including foci in the right prefrontal cortex, are active. Such research reflects one of the strong interests of neuroscientists in blindsight and other perceptual phenomena in helping to unravel the neural mechanisms that may underlie conscious awareness.
Because of the difficulty subjects may have in being asked to discriminate stimuli they cannot see, other methods of assessing residual function commend themselves, especially for screening of brain-damaged subjects for possible rehabilitation. Some of these methods depend upon asking the subject to discriminate stimuli lying entirely in their intact, seeing hemifields, but showing that their performance can be altered by the presentation of stimuli in their blind fields, which may enhance or interfere. By far the most quantitatively sensitive method depends upon changes in the diameter of the pupil, which constricts not only to increase in light energy, but to a wide variety of stimuli without any energy change. These include sine-wave gratings, movement, and colour. The acuity of the blind field can be accurately measured by pupillometry, as well as the sensitivity to colour and colour after-images, which appear to mirror the blindsight capacity as measured by forced-choice guessing. The pupil can also be used to measure similar capacities in animals or in human infants, where verbal report of course is impossible.
Finally, given that the existence of residual visual capacity was first demonstrated in animals with visual cortex lesions, the question arises as to whether they too show blindsight for the discriminations they can perform. Ingenious experiments appear to show a positive answer. They can detect and locate light stimuli with impressive sensitivity. They can also readily be trained to make differential responses in their normal visual fields for lights versus non-lights ('blanks'). But when the same lights are projected into their affected field, the monkeys reliably treat them as 'blanks', that is, the very stimuli that they can detect with impressive sensitivity are classified by them as being 'blanks', as 'non-lights' — just as a human blindsight subject does. Thus, the contribution made by the visual cortex to visual awareness appears to be similar in humans and other primates.
(Published 2004)— L. Weiskrantz
Bibliography- Cowey, A., and Stoerig, P. (1995). 'Blindsight in monkeys'. Nature, 373.
- Holmes, G. (1918). 'Disturbances of vision by cerebral lesions'. British Journal of Ophthalmology. 2.
- Pöppel, E., Held, R., and Frost, D. (1974). 'Residual visual function after brain wounds involving the central visual pathways in man'. Nature, 243.
- Sahraie, A., Weiskrantz, L., Barbur, J. L., Simmons, A., Williams, S. C. R., and Brammer, M. L. (1997). 'Pattern of neuronal activity associated with conscious and unconscious processing of visual signals'. Proceedings of the National Academy of Sciences of the USA, 94.
- Weiskrantz, L. (1990). 'Outlooks for blindsight: explicit methodologies for implicit processes. The Ferrier Lecture'. Proceedings of the Royal Society Biological Sciences, B series, 239.
- — — (1997). Consciousness Lost and Found: A Neuropsychological Exploration.
- — — (1986). A Case Study and Implications (new edn. 1998).