Blindsight is a condition where patients display latent visual abilities in areas of their visual field
in which they report "seeing
" nothing. For instance, patients can look accurately towards stimuli presented in their field defects, localise stimuli by pointing, and detect and discriminate movement. One patient, reported by ter Braak, Schenk, and van Vlite (1971
), could follow a large moving striped display despite complete cortical blindness
. Among the abilities of patients are the ability to detect and discriminate flicker, orientation, and wavelength, and their pupils
respond to changes in light
, and contrast. Patients have been found to be able to reach for stimuli and will sometimes even adjust their grasp
to match an object's shape and size. Patients can also use the meaning of unseen words to select between pairs of words presented in their intact visual field.
Blindsight is associated with damage or destruction of the striate cortex, the cortex which contains the primary visual cortex (V1). However, blindsight is generally attributed to input to higher visual areas that bypass V1. There are many pathways leading from the retina. The main path is the dorsal lateral geniculate nucleus (dLGN) and this leads to V1 and through the main striate cortex. Another pathway is the superior colliculus (SC), which is likely to be the main pathway associated with blindsight. It is the second largest sub-component of the optic nerve, consisting of around 150,000 fibres, or 10% of the total retinal output.
Other pathways conducting visual information are the ventral lateral geniculate nucleus (vLGN), which appears to be used in reflexive, non-cognitive responses, such as detecting light levels and pupil dilation. There is also the Nucleus Optic Tract, which contains the Dorsal, Medial, and Lateral Terminal Accessory Optic Nucleus. These seem to be involved with self-motion, posture adjustments, and flow fields (it would be interesting to see if blindsight patients are affected by David Lee's Swinging Room). There is also the retinocellular projection which reaches extrastriate cortical visual areas.
All these pathways can be identified in monkeys with traceable chemicals, like HRP (horseradish peroxidase).
Much of the information about blindsight has been obtained from monkey studies. Mohler and Wurtz have found that monkeys with a small part of striate cortex removed could direct their eyes to a spot of light presented in their visual defect. This ability was lost when the corresponding part of the superior colliculus was removed - they were rendered totally insensitive in that part of the visual field. The SC was directly implicated in the recovery of monkeys with practice.
In many studies, Pasiks and collaborators found that monkeys' ability to detect which of two targets was illuminated survived total bilateral removal of striate cortex. This ability, however, was impaired by disruption of the accessory optic system and was further impaired by the removal of extrastriate visual cortical areas, which suggests that non-striate input from the SC and LGN to these areas is important.
There are many visual areas that have been found not to function after the removal of striate cortex. Girard and Butler (1989) found that area V2 has no functioning when the striate cortex in monkeys is cooled, and Rocha-Mirande et al. (1975) found that striate cortex ablations caused the inferotemporal cortex to be visually unresponsive. They also found no function in area MT (an area of the brain that responds to moving stimuli).
However, more recent work by Redman, Gross, and Albright (1989) say that 50% of cells in MT retain their function after the removal of striate cortex. Upon the removal of the superior colliculus, MT is no longer activated by visual stimuli. Other areas, such as V4 (dealing with colour information), have not been studied in this way.
Problems with Monkey Studies:
The main problem with monkey studies is that although the monkey brain is physiologically similar to the human brain, there are some differences. For instance, the location of the striate cortex is buried in the calcarine fissure in humans, but is on the occipital surface in monkeys. Experimenters have to assume that the functional similarities overcome the anatomical differences.
Striate cortex damage in humans is always clinical and damages surrounding tissue. Surgical lesions in monkeys are far more precise. Monkeys need to undergo lengthy training to describe what, if anything, they see. They cannot give the same responses to complex stimuli like a human could. Most significantly, there has been uncertainty as to whether monkeys experience the same dissociation between awareness and residual visual function that is seen in human patients.
A study by Cowey and Stoerig (1995) has shed some light on this issue. They showed that monkeys with unilateral striate cortical ablations classify stimuli briefly flashed in their hemiantopic field (i.e., the "blind" half-field) as blank trials when given this option, even though they can detect and localise these same or even briefer and dimmer stimuli in experiments lacking this option. This shows a direct similarity to human blindsight patients, so striate cortical lesions are likely to be the same in both species.
Role of Superior Colliculus:
Bruce Bridgeman has written that the SC fields have five characteristics which correspond to the properties of blindsight:
In blindsight, patients are better at detecting oscillating targets than stationary ones. SC receptive fields are selectively sensitive for moving stimuli.
Flickering stimuli are located better, reflecting the transient nature of SC responses.
Blindsight patients do have an inaccuracy in their spatial estimates, which could be accounted for by the large recpetive fields of the SC.
Blindsight patients have a low critical flicker fusion point (threshold of detecting flickering lights of the same brightness changing colour). This corresponds to the low temporal band pass of the Y-cells and W-cells which innervate the retina.
Patients also have a lack of psychophysical colour discrimination, which reflects the lack of colour opponency in Y-cells
Weiskrantz has shown that in GY, a blindsight patient, there is increased activity during the presentation of stimuli he is "unaware" of.
It is possible that, due to plasticity, the SC takes on other functions due the absence of striate cortex. One primary function the SC is known to perform is controlling eye movements. It mediates inhibition of return, that is the action of continuing to make saccadic eye movements over an object else the pigment in the visual receptors in the retina will become bleached and they will eventually lose all sensitivity. This ability is lost following SC damage and leads to the condition known as Supra-Nuclear Palsy.
Residual Striate Cortex:
Campion et al. (1983) suggest that blindsight is best explained by residual striate cortex mediating degraded vision. Alan Cowey, however, disagrees. Cowey (1967) has found in monkeys that within a field defect produced by complete removal of striate cortex, there is residual visual sensitivity. Therefore, blindsight is not explained by light scatter, as stimuli are not detected when they fall on a retinal lesion or the optic disc (blind spot). The light scatter idea is that light stimuli that would fall on the area of the retina that sends signals to the lesioned part of striate cortex is still detected because the light scatters in the eye and falls on a good part of the retina. Campion et al. argue that in normal vision we sometimes see things we are not consciously aware of. Cowey agrees, but points out that this is over a very small range of intensities and blindsight patients are unaware of visual stimuli even when the stimuli are at very high superthreshold intensities.
Campion et al. place too much emphasis on light scatter. Patient DB, for instance, can discern between striped and plain targets of matched luminance. This cannot be explained by scattered light. Appeals to residual striate cortex in hemispherectomised patients is incorrect.
Is Awareness Completely Gone?
In strict blindsight, awareness is completely gone, however, in Blindsight Type 2 there is awareness of some stimuli. Such patients report "awareness", "feeling", or a "knowing". One patient has said; "I know it's an 'X' because it's 'rougher'". This Blindsight Type 2 is evident in patient DB, but only sometimes in GY. Awareness of this sort may be reported, but subjects never report that they actually "see" anything. GY can mimic the movement of a laser beam with his arm - he "knew" something was moving, but he did not "see" anything.
Crick and Koch (1995) say that visual awareness is unlikely to arise directly within the striate cortex. It happens in higher areas, but striate cortex is needed for awareness in any significant sense. To be aware of an event, the brain constructs a multilevel, symbolic interpretation of the event. Only higher visual areas and prefrontal cortex, to which they have direct projections, have the required inputs for this. Physiological experiments in monkeys show a better correlation for higher visual areas than for striate cortex between neuronal response properties and what the trained monkey reports seeing. Some striate cortex neurons respond to things we are not aware of but can still influence other aspects of visual perception, finely spaced lines for instance. Many striate cortex cells respond only to visual stimulation of one eye, yet we are unaware of which eye is being stimulated when both are open and a stimulus only reaches one. No such cells are present beyond striate cortex. When all cortex besides visual cortex is removed, a monkey behaves as if blind.
Spectral sensitivity profiles of blindsight patients show the same peaks and troughs associated with colour opponency. Stoerig and Cowey (1992) found that wavelength discrimination was possible, even for some closely spaced wavelengths, in some subjects using forced-choice guessing methods (subjects have to be made to choose, since they do not believe that they can see the stimuli). Patients can sometimes tell colour from matched luminance grey stimuli, but they never report seeing colour, just as they never report seeing anything else. This is similar to some cases of achromatopsia, where patients can discern some colour but do not experience any colour sensation (such patients see only a monochrome world in drab shades of grey).
Colour is dealt with by the parvocellular system, but the magnocellular levels of the dLGN contain cells that have a weak, "hidden", chromatic opponency. So if the parvocellular system is destroyed, but the magnocellular system is spared it could explain this aspect of blindsight.
Pöppel et al. (1973) found that eye saccades to the supposed locus of a stimulus, a briefly flashed spot for example, could be made by blindsight patients. They could also sometimes point to, or touch, a stimulus on a screen. However, none of these localisation abilities are as accurate in their anopic fields as they are in their sighted visual fields.
Visual acuity in a scotoma is always less than in the corresponding part of an intact field.
DB can discriminate a difference in orientation of 10° between two diffraction gratings, even at 45° eccentricity. However, DB is better than average. But sensitivity to orientation is always better in an intact field.
DB can discriminate an "X" from an "O", however Warrington and Weiskrantz have reassessed the evidence and concluded that DB was using his considerable orientation skills to distinguish forms. They found that forms with roughly equivalent orientational components were poorly discriminated. Marcel (1998) and Perenin and Rossetti (1996) have found that when reaching for 3D objects, blindsight sufferers appropriately adjust their hand grasp before they grasp an object. Marcel (1998) has also found that words flashed into the blind field can influence the interpretation of the meanings of words subsequently shown in the intact field. For instance, the word "money" presented in the blind field biases the interpretation of the ambiguous word "bank". Maybe form perceptions are more advanced than were previously thought.
If a fear-evoking stimulus, like a strange doll, is placed into the blind hemifield of a monkey, the animal ignores it, even though it makes loud shrieks of fear when the doll is in its normal hemifield. GY could discriminate between different facial expressions presented in his blind field, but he failed to "see" the faces. He was at chance in his discrimination of inverted faces.
Indirect Testing Methods:
Pupillometry, that is, measuring the responses of the pupil to various stimuli, is a good testing method as it does not rely subject responses. This is particularly useful when testing blindsight patients, since it can be most frustrating being made to answer questions and give responses on things that they are totally unaware of.
The pupil automatically responds differently depending on the stimulus, so, changing the stimulus while monitoring pupil response can provide useful information. To test acuity, one can measure the sensitivity to gratings in the blind field. One can also measure in response to colour and movement. Pupillometry can be used to measure the extent of the blind field - this is the best way to do it for children and animals, as no response is needed.
GY reported being "aware" of a stimulus when a pupillometric response was elicited. It has been found that pupillometric acuity changes upon V1 removal, so there must be a direct pathway reaching the midbrain that controls pupil response. Similar results have been found in both humans and monkeys and it has shown that human blindsight functions do not depend on residual cortex.
Interactions between Intact and Blind Hemifields:
Torjussen (1976, 1978) has demonstrated the completion of shapes shown to both the blind and intact hemifields of blindsight sufferers. So a square presented half in the blind field, and half in the intact field would be seen by a blindsight patient as a complete square, and not as a rectangle (with one half of the square missing). It has also been noticed that there is the latency of saccades to stimuli in the intact field is increased if an "unseen" stimulus is presented just before in the blind field. Marzi et al. (1986) have found that subjects respond faster to two simultaneously presented stimuli. They found that the reaction time was faster when two stimuli are presented, one in the intact field, one in the blind field.
Blindsight is not just a quantitatively different thing - it is qualitatively different.
It probably is residual function that normally works in conjunction with normal vision, but we never notice it as it is not part of conscious awareness.
The superior colliculus appears to the crucial conduit in blindsight, and the condition provides us a window onto what the functions of the other visual pathways are actually doing. It seems reasonable to say that the responses we see in blindsight are what we use when fast responses are required, such as alerting us to immediate danger. For avoidance and "fight or flight" actions, we want instant responses. We do not want to wait the 300 milliseconds it takes to respond to these things consciously. Another possible function of these "hidden" visual abilities is to direct attention to alerted stimuli.
Cowey A, Stoerig P (1991) The neurobiology of blindsight. TINS 14(4): 140-145
Campion et al (1983) Open Peer Commentary on Blindsight. Behavioural and Brain Sciences 1983 pps. 448-483
Cowey A & Stoerig P (1995) Blindsight in monkeys. Nature 373: 247-249
Gazzaniga et al (1994) Blindsight reconsidered. Current Directions in Psychological Science. 3(3): 93-6
Meeres, Graves (1990) Localization of unseen visual stimuli by humans with normal vision. Neuropsychologia 28: 1231-1237