Exp Brain Res (2002) 143:249–256
DOI 10.1007/s00221-001-0989-1

R E S E A R C H A RT I C L E

A. Sahraie · L. Weiskrantz · C.T. Trevethan
R. Cruce · A.D. Murray

Psychophysical and pupillometric study of spatial channels
of visual processing in blindsight
Received: 20 September 2001 / Accepted: 26 November 2001 / Published online: 2 February 2002
© Springer-Verlag 2002

Abstract To date no systematic method has been used
for characterising the residual capacity of blindsight subjects that would allow comparison and generalisation
across all subjects. The detection of isoluminant gratings
of varying spatial and temporal frequencies commends
itself for detailed between-subject comparison, and for
mapping results onto physiological properties in relation
to neuronal circuitry. We report the ability of a blindsight
subject (CS) to detect suprathreshold sine-wave gratings
over a range of spatial and temporal frequencies using
psychophysical techniques. A band-pass spatial channel
with an upper cutoff below 3.5 cycles/deg is specified.
The data also have been analysed to compare differences
between two types of blindsight performances, type I
and type II. Spatial gratings were also used to elicit a
pupillary grating response, offering an objective method
that is free of verbal nuances and response bias, and the
resulting band-pass channel can be used both for clinical
screening and for prediction and comparisons with
psychophysical profiles. Finally, we have compared our
results with those reported in studies of a well-known
subject, GY, which demonstrate remarkable similarities.
Implications are discussed in relation to blindsight
research.
Keywords Blindsight · Residual vision · Cortical
blindness · Spatial vision · Pupil

A. Sahraie (✉) · C.T. Trevethan · R. Cruce
Vision Research Laboratories, Department of Psychology,
University of Aberdeen, Aberdeen AB24 2UB, UK
e-mail: a.sahraie@abdn.ac.uk
Tel.: +44-1224-273919, Fax: +44-1224-273426
L. Weiskrantz
Department of Experimental Psychology, University of Oxford,
Oxford OX1 3UD, UK
A.D. Murray
Department of Radiology, Lilian Sutton Building, Foresterhill,
Aberdeen AB25 2ZD, UK

Introduction
Blindsight is defined as a condition in which a subject
“responds to visual stimuli without consciously perceiving them” (Oxford Concise Dictionary). The condition is
associated with lesions of the geniculostriate pathways,
resulting in perimetrically blind areas in the contralateral
visual field. Spatial responses without awareness were
first demonstrated in gunshot cases by Pöppel et al.
(1973) and detection and discrimination were studied in
a surgical patient DB (Weiskrantz et al. 1974), and extensively followed up over a 1-year period (Weiskrantz
1986, 1998). More recently some studies have used a
“commentary key paradigm” to obtain a record of the
subject’s conscious reports during testing. After every
experimental trial, the subject is required not only to
make a forced-choice response based on the presentation
of one of two visual events (target vs non-target) but also
to report whether or not there was any awareness of the
event. Awareness may be reported on a binary or multipoint scale (Weiskrantz et al. 1995; Sahraie et al. 1998).
Commentary responses allow a classification of responses as blindsight type I – visual capacity in the complete
absence of reported awareness – or type II – visual discrimination ability within the field defect with awareness
of the event, but in the absence of acknowledged “seeing” (Weiskrantz 1998). It is not assumed that the difference between type I and type II is necessarily sharp, but
some term is necessary to contrast responses of which
the subject has some awareness and those for which
there is none whatever. The effect of awareness appears
to be counterintuitive in that, although it enhances the
discrimination performance in some subjects, it hinders
discrimination in other blindsight cases. DB appeared to
perform better in the absence of any visual awareness
and so the experimental parameters were often chosen
such that he performed in type I mode. It was demonstrated that the overall performance of a well studied
subject GY was higher in type II mode of processing
compared to the type I, although given the right combination of the stimulus parameters discrimination perfor-

250
Fig. 1 T2 weighted transverse
MRIs of CS and GY, both
showing well defined left
occipital lesion. Lesion extends
more posteriorly in GY than in
CS. The binocular visual field
plots of both patients are shown
for the central 30 deg, both
showing a right hemianopia
with a greater macular sparing
in CS

mance could be matched between the two modes of processing.
To our knowledge, the spatial channels characterising
blindsight have been explored only in GY (Barbur et al.
1994), with the finding of spatial tuning curve with
band-pass characteristics and an upper cutoff of below
3 cycles/deg. Subcortical structures are reported to have
a lower frequency response than cortical areas; therefore,
a shift of sensitivity to lower spatial frequencies in cortically blind may reflect processing in subcortical areas
(Bisti and Sireteanu 1974).
Motion sensitivity in blindsight was first reported by
Weiskrantz in DB and subsequently by Barbur et al.
(1980) in subject GY. GY has since been tested on motion discrimination tasks by many other investigators
(Weiskrantz et al. 1995; Sahraie et al. 1998; Azzopardi
and Cowey 2001; Benson et al. 1998). In addition a
small number of other reports of motion discrimination
using first order apparent motion stimuli have been reported in blindsight. Both blindsight type I and type II
for motion discriminations have been reported in GY
using continuous motion of a single dot target.
Minute transient pupillary responses are elicited by
the onset of basic stimulus attributes such as structure,
motion and colour. Pupil grating responses appear to
have a similar sensitivity envelope to those measured
using contrast sensitivity functions in normal observers
and in non-human primates (Barbur et al. 1987). In both
human observers and monkeys, such responses may be
used as an objective measure of visual acuity. Pupil grating responses to static sine-wave gratings have been
shown to exist in GY and in two cases of destriated monkeys, with spatial tuning characteristics mirroring those
found in the behavioural studies. Such findings have

given rise to the suggestion that pupillometric techniques
may be used as an objective method to screen for blindsight.
It is interesting to note that in our latest literature
search on experimental human blindsight papers we have
found that out of 74 publications on this topic within the
past 30 years, 32 have either entirely or partially relied
on the results from a well studied subject GY. We have
compared our findings to those reported for GY in an
attempt to establish whether the blindsight characteristics reported in the literature could be generalised and attributed to our subject. In order to do this, we have used
identical paradigms and where possible even identical
apparatus in the study of a new subject CS. Here, we report on our findings of spatial processing for both blindsight types I and II. Motion sensitivity was also investigated using the same apparatus and experimental conditions as those used in previous studies on GY. Finally,
we have also measured pupil grating responses in CS using the same experimental procedures used in GY.

Materials and methods
Subject
CS is a 37-year-old female who suffered an ischaemic stroke as a
result of posterior cerebral artery embolisation during a cardiac
surgery at the age of 3 years. T2 weighted magnetic resonance images (MRI) show evidence of an old infarct in the medial aspect of
the left occipital lobe. This involves the primary visual cortex and
the adjacent white matter between the cortex and the trigone of the
left lateral ventricle, involving the terminal fibres of the optic radiation on the left side. No other intracranial abnormality was demonstrated (Fig. 1). The lesion resulted in a homonymous field defect covering the upper and to a lesser extent the lower visual

251
fields (see Fig. 1). GY has a complete right homonymous hemianopia with ≅3° macular sparing as a result of an occipital brain
lesion caused by a road traffic accident. A detailed report of his
lesion was first described by Barbur et al. (1980). Figure 1 shows
a comparison of MRI scans and the visual fields in CS and GY.
Both the extent of the brain lesions and the resultant visual field
losses are comparable.
The study reported was carried out in accordance to the
standards laid down in the 1964 Declaration of Helsinki and was
approved by the Grampian Research Ethics Committee, Scotland.
Spatial processing
Stimuli were programmed on an IBM compatible PC and were
generated using an SVGA graphics card and presented on a 21”
monitor (Sony Multiscan G520) at a 60 Hz refresh rate. The monitor gamma corrections were carried out using a Minolta Luminance-Meter (LS100) at 256 linear steps. The monitor was enclosed in a cubicle with a chin-head rest mounted at the entrance
to the cubical at a viewing distance of 760 mm. The room illuminance was 80 lux and the inside of the cubical and all the exposed
surfaces were covered with a non-reflecting mat-black felt to
avoid any possible scattered light from the surround. The screen
background luminance was 37 cd/m2 at the x,y chromaticity of
(0.309, 0.353) and subtended 26.6×20.6 deg. Spatial gratings were
either sine-wave or square-wave gratings with spaced averaged luminance equal to the background to avoid any possible artefacts
due to light flux changes. The subject’s eye movements were monitored using an ASL 5000 pupillometer (Applied Science Laboratories, MA), and trials with eye movements towards the target
were discarded from analysis. These formed ≅1% of the total
number of trials. The fixation was a high contrast black cross-hair
subtending 0.5 deg. The circular spatial gratings were 5 deg in
radius and were presented centred at 12 deg to the right and 7 deg
above the fixation. This insured that the edges of the stimuli were
at least 6 deg outside the edge of the blindfield. Where the gratings were spatially modulated by a gaussian profile, the diameter
of the stimuli was set to 4 times the spatial standard deviation
(σs). The stimuli onset and offset were either a temporal squarewave or modulated by a temporal gaussian with a minimum duration of 2 σt each, where σt was the standard deviation of the temporal gaussian. In addition, the spatial gratings were either static
or their contrast was temporally modulated at 10 Hz with an offset
cosine wave. The algorithms implemented to achieve the above
spatial and temporal gratings were the same as those reported by
Barbur et al. (1994) in the study of spatial processing in GY. Experiments were carried out using the two-alternative forced-choice
technique (2AFC). Each trial contained two time intervals separated by audio beeps. During one of the intervals, a spatial grating
was presented for 2016 ms, flanked by two 500-ms pre- and poststimulus blank presentations. Since the grating had the same
space-averaged luminance as that of the background, the remaining interval contained a blank presentation of the same duration.
The subject’s task was to indicate, if necessary by guessing, the interval containing the spatial target by pressing one of the two buttons on a response box. The commentary key paradigm was used
throughout the study. The subject was also asked to indicate by
pressing one of the two further response keys whether she was
aware or not aware of any visual presentation. She was instructed
to press “unaware” if she had no awareness of any visual stimulus
whatsoever; otherwise she was to press the aware button. Unless
otherwise stated, a minimum of 100 trials were presented for each
stimulus condition. A log file of each trial and the subject’s response was automatically generated.
Motion discrimination
The servo-driven computer-controlled optical projection system
used to generate the moving single dot target was the same apparatus that was used in a similar study of motion processing in GY

and reported previously (Weiskrantz et al. 1995; Sahraie et al.
1997, 1998). The stimulus target was generated by a semiconductor source (ACMT08/2092, 15 mW, 635 nm; Power Technology,
Little Rock, AR). A circular dot target with 600 cd/m2 luminance,
subtending 0.26°, was optically projected onto a wide screen
(40°×40°) at a viewing distance of 1.1 m via a servo-driven mirror
scanner system (model DMC 1510; Galil Motion Control, Sunnyvale, CA).
Motion discrimination capacity was tested in both lower and
upper visual fields for a speed range (0.5–20)°/s, and a displacement of 20° using a forced response procedure for horizontal motion away and towards the vertical meridians. The target was centred at 15 deg to the right and 5 deg below the fixation for lower
field testing, and centred at 15 deg to the right and +7 deg above
the vertical meridian when testing the upper field. The blindfield
and sighted field luminances were 180 and 5 cd/m2 respectively.
The luminances were achieved using two Kodak Carousel projectors fitted with appropriate neutral density filters. A minimum of
100 trials were presented for each stimulus speed tested. Target
speeds were selected in a pseudo-random order and multiple testing was carried out to avoid the order effect. Identical experimental procedures, similar to those reported earlier for GY, were carried out. That is, the target motion interval was indicated by audio
beeps. In every trial the target moved either away or towards the
vertical meridian. The subject’s task was to indicate, by pressing
one of the response keys, the direction of stimulus motion. As described above, the subject was instructed to report whether or not
she was aware of any visual target. Again, the subject was instructed to report “aware” if she had any feeling or awareness of
the stimulus presentation, and press “unaware” if she had no
awareness whatsoever.
Pupillometric study
An ASL5000 pupillometer with headrest mounted optics was used
to monitor eye movements. The pupil diameter was determined by
fitting a circle to the pupil margin and recorded at the sampling rate
of 60 Hz. The start and end of the recording were synchronised
with the refresh video signal of the stimulus monitor through the
parallel port of the PC via multiplexer electronics. The extracted
parameter data were transmitted from the pupillometer to a second
PC via an RS232 link. The data were analysed offline using P-trace
software developed at Vision Research Laboratories, University of
Aberdeen. The software allows examination of each pupil trace
segment and removal of those containing artefacts such as blinks;
sorting the pseudo-random order of segments, and assigning them
to relevant stimulus files; and averaging and extracting the pupil
response parameters such as latency, diameter change and percentage change in the pupil diameter. The pupillary responses elicited
by a circular patch of sine-wave grating (radius 5°) at a range of
spatial frequencies (0.25, 0.5, 1.0, 2.0, 3.49, 4.66 and 6.98 cycles/
deg) were recorded. The grating patch was presented at the same
eccentric location as those for the psychophysical testing of spatial
vision (12° to the right and 7° above the fixation). All the spatial
frequencies tested were presented in pseudo-random order with
pre-, post- and stimulus durations of 500 ms, 2500 ms and 1000 ms
respectively. A minimum of 60 pupil traces were recorded for each
stimulus condition.

Results
Spatial processing
Figure 2A shows the percentage of correct discrimination
scores in a 2AFC paradigm for detection of temporally
modulated (10 Hz) sine-wave gratings as a function of
spatial frequency (filled circles). Percentage of trials
where some awareness of the stimulus presentation was

252
Fig. 2 Discrimination (filled
circles) and awareness (filled
squares) performance for an
equiluminant sine-wave spatial
grating as a function of grating
spatial frequency using a two
alternative forced-choice technique are shown in A. A spatial
grating was presented in either
the first or the second interval
with a uniform patch being presented in the remaining interval. The subject had to indicate
(by guessing if necessary) the
interval containing the spatial
grating by pressing the appropriate response key (chance
level 50%). The 95% confidence interval for detection is
also shown, indicating that, except for the two highest spatial
frequencies, detection performance is significant at least at
P<0.05. Discrimination scores
are subdivided into type I and
II performance based on the reported awareness on each trial
and plotted in B (see “Materials
and methods”). A comparison
of CS and GY performance in
the same discrimination task is
shown in C

reported is shown by filled squares. CS’s performance in
detecting the temporally modulated spatial gratings is well
above chance for spatial frequencies below 3.5 cycles/deg.
Subsequently, the discrimination scores are subdivided
based on subject’s reported awareness. As described
above, blindsight type I performance refers to correct discrimination in the absence of any reported awareness,
whereas type II performance refers to correct discrimination with awareness of the visual event. Figure 2B shows
the spatial frequency tuning characteristics for both type I
and II performance in CS. Here, the type I performance
(correct when aware) is shown in filled circles and type II
performance (correct when aware) is shown in open circles. It appears that above chance performance for spatial
frequencies below 3.5 cycles/deg exists for both type I and
II. Although discrimination performance is better when
the subject reports some awareness of the visual event,
both modes of processing appear to have similar tuning
characteristics.
Figure 2C shows a comparison of the psychophysically determined spatial channel for CS and GY. GY’s data
was obtained using 6 Hz temporally modulated spatial
gratings; however, it appears that the spatial channel mediating blindsight in CS has similar tuning characteristics
to those reported for GY previously (Barbur et al. 1994).
Effect of grating contrast on correct discrimination and
reported awareness for 10 Hz temporally modulated
1 cycle/deg sine-wave gratings is shown in Fig. 3A. Significant discrimination performance is obtained for contrast gratings of 20% and above. Both subject’s reported
awareness and discrimination performance are higher

with increasing stimulus contrast. Discrimination scores
are subdivided into type I and II performance and are
shown in Fig. 3B. A monotonic relationship between discrimination and contrast exists for both type I and II. A
comparison of the data for CS and GY is shown in
Fig. 3C. Since chance performance in the discrimination
task was 50%, the 95% confidence interval for a discrimination score after 100 trials is 50±9.8% (binomial distribution). Therefore, for high contrast stimuli, it appears
that both CS and GY perform similarly. The available
data on GY’s contrast response are not sufficient to draw
clear conclusions on the differences between GY and CS.
Nevertheless, CS appears to respond to a larger contrast
range than GY.
We have attempted to isolate the contributions of both
temporal modulation and onset, to discrimination and
awareness of spatial targets in CS. An elegant way of
demonstrating the effect of temporal modulation on detection of spatial targets is to determine the temporal frequency characteristics of the mechanisms mediating
blindsight as demonstrated by Barbur et al. (1994). Due
to technical limitations, such a wide range of temporal
frequencies were not available to us. We have therefore
compared the differences between static gratings and
10-Hz modulated gratings only. The bar chart in Fig. 3D
shows the discrimination scores and reported awareness
for abrupt onset of both stimulus types. As demonstrated, although the discrimination performance remains
high for static gratings, there is a marked drop in awareness of the visual event. This indicates that the temporal
content of the visual stimulus is an important factor. We

253
Fig. 3 Discrimination and reported awareness for detecting
a 1 cycle/deg sine-wave grating
as a function of grating contrast
using the same 2AFC paradigm
as in Fig. 2 are shown in A.
B and C show the subdivisions
of the detection score into type
I and II performance and the
comparison of CS and GY in
the same task, respectively. The
effect of changes in temporal
onset of 10 Hz modulated and
static gratings on correct discrimination and reported
awareness are shown in D

have investigated this further by changing the temporal
onset of both types of stimuli. Changing the temporal
onset from a temporal squarewave to a slow temporal
gaussian [rise and fall times of 300 ms each (2×σt),
steady state = 1416 ms] again resulted in a similar lowering of awareness level compared to abrupt onset, but the
discrimination remained well above chance level. CS
was neither aware of nor was she able to discriminate the
presence of slow-onset static gratings. A similar drop in
the level of awareness for a linearly ramped static grating (rise, fall and steady state times of 500 ms each) was
also reported by Weiskrantz et al. (1998) in GY. However, DB was able to detect stimuli correctly even with
very slow temporal onsets (Weiskrantz 1986, Chap. 9),
whereas CS was at chance level for targets with even
faster temporal onset.

upper field. CS’ performance was subdivided into type I
and II performance as indicated in Fig. 4B. The percentage of trials where the direction of motion was correctly
identified was higher for type II than type I mode of processing. The number of unaware trials in faster speed
conditions (15 and 20) deg/s is small; therefore no firm
conclusions can be drawn on the tuning properties of the
motion processing channels for the two modes of processing.
Figure 4C shows a comparison of CS and GY’s performance for upper hemifield stimulus presentations.
The forced-response discrimination scores are not significantly different for any of the speeds tested. Also both
subjects appear to be unaware of the majority of presentations at slow speeds and their reported awareness increases monotonically with increasing speed.

Motion discrimination

Pupillometric study

Figure 4A shows the percentage of trials in which CS
correctly detected the direction of target motion as a
function of the stimulus speed. Results are plotted for
both lower and upper field stimulations (open symbols).
Although the upper and lower fields appear to have
different sensitivities, significantly above chance performance was observed for a wide range of speeds in both
conditions. The discrimination performance improves
monotonically with increasing stimulus speed. In addition, percentage of aware trials is shown for the same trials (filled symbols). The subject is more aware of the
stimulus presentation in the lower field than those in the

The amplitude of pupillary responses elicited by the onset of static sine-wave gratings (PGR) is plotted as a
function of grating spatial frequency for both sighted and
blindfield presentations and is shown in Fig. 5A. The
sighted field data show tuning characteristics similar to
those reported in normal observers by Barbur et al.
(1987). The sighted and blindfield presentations were in
different hemifields but the same eccentric distance from
the fixation point. The peak sensitivity is therefore approximately 2 cycles/deg for sighted field condition. The
normalised PGR and psychophysical detection performance for the same target size and visual field locus are

254
Fig. 4 Direction discrimination
and reported awareness of a
single-dot moving target are
plotted as a function of stimulus speed for upper and lower
visual fields in A. B and C plot
the type I and II performance
and comparison with GY’s data
respectively

Fig. 5 Pupil grating responses
elicited by the onset of suprathreshold static sine-wave
gratings presented in sighted
and blind hemifields are plotted
in A. A comparison of pupillometric and psychophysical
channels in CS are shown in B.
C plots a comparison of pupil
grating responses in CS and
GY

255

plotted in Fig. 5B. The normalised responses show that
although the spatial channels mediating the psychophysical performance have band-pass properties, the pupillometric responses have low-pass properties with a higher,
high cutoff frequency. It is important to note that the
psychophysical data are obtained using temporally modulated spatial patterns, whereas PGRs are elicited by
static gratings (see discussions) similar to those reported
in other studies (Weiskrantz et al. 1998). A comparison
of the normalised pupillary responses of GY and CS is
shown in Fig. 5C. The tuning curves for PGRs in the two
subjects appear to be different. GY’s pupillary responses
closely match the tuning characteristics of the psychophysical channels. CS’s pupillary responses on the other
hand extend to higher spatial frequencies and are low
pass in shape.

Discussion
We have investigated the residual visual capacities of a
hemianopic subject, CS, and found them to be closely
similar to those of the often studied blindsight case GY.
In particular, there was a narrowly tuned spatial channel
as assessed by forced-choice detection of sine-wave gratings of various spatial and temporal profiles. Directional
discrimination of single dot targets moving at a range of
speeds was also found to be similar, with a peak sensitivity at roughly 10 deg/s. Correct discrimination could be
found when she was not aware of any visual presentation
(blindsight type I), and also under other conditions when
she had some awareness of the visual target, but without
“seeing” as such (type II).
Pupillary responses to the onset of sine-wave gratings
also showed a similarly shaped spatial profile with that
of GY, and also with CS’s psychophysically determined
profile to the gratings. While the shapes were qualitatively similar, there were some differences in detail: e.g.
CS showed a maximal PGR to a higher spatial frequency
than was true of GY. GY’s pupillary responses closely
matched the tuning characteristics of the psychophysically determined channels. However, it is important to note
that the psychophysical data extracted from the earlier
study of Barbur et al. (1994) for GY as well as those
used here for CS were temporally modulated, whereas
the pupillary responses were necessarily obtained with
static stimuli because temporally modulated stimuli
would generate noisy pupil responses. It is worth noting
that with DB it was possible to use precisely the same
static gratings for both the psychophysical and the pupillometric determinations, with a close correspondence
in the shapes and peaks of the normalised functions
(Weiskrantz and Cowey, unpublished). Notwithstanding
the differences in details, the existence of the pupil grating response coexists with a psychophysical capacity for
spatially defined stimuli, and offers the exciting possibility of its use as an objective technique for residual capacity that would be reflected in discrimination performance measures.

CS is also capable, as was GY, in discriminating the
direction of a moving target. Above chance discrimination was observed for lower and upper field stimulation
presentations. The improved discrimination and increased percentage of “aware” trials with increasing
stimulus speed reported here are similar to those reported previously for GY.
Measurements of reported awareness showed that
both blindsight type I and II have similar peak sensitivities. It appears that the mechanisms mediating type II
performance have a larger bandwidth than those involved in type I, but the range of frequencies tested and
the small number of “aware” trials are not sufficient to
obtain a precise quantitative measure of the channel
characteristics for type II in CS. A comparison of CS’s
data with the type I and II performance in GY was not
possible since the commentary key paradigm was not
applied in the study for GY’s spatial channel. The spatial processing in blindsight indicates that in the absence
of a functioning striate cortex, whatever the remaining
functional pathways may be, the higher spatial frequency response is lost even for high contrast stimuli. The
experimental findings shown in Fig. 3 indicate that fast
transients play an important role, favouring magnocellular involvement. On the other hand, the overall contrast
response for both type I and II are closer to a parvocellular response in that they do not show a rapid contrast
gain expected for the magnocellular system (Fig. 3).
Blindsight performance, therefore, appears not to indicate the dominance of either a parvo- or a magnosystem.
It is an open question whether these findings can be
generalised to all blindsight subjects irrespective of the
age at injury or the aetiology of brain damage. CS suffered an ischaemic stroke following a cardiac operation
at the age of 3 years, whereas GY’s lesion occurred as a
result of a road traffic accident. DB had a surgical removal of a non-malignant tumour at the age of 33 years,
but pathological symptoms first came to light clinically
when he was a teenager and may have been present
much earlier. And so all three subjects had long-standing
lesions, and have been tested over 30 years following the
injury. Animal work strongly suggests that early lesions
give rise to greater recovery than later ones (Payne et al.
1996). On the other hand, the close similarity between
the pupillometric responses to gratings in mature monkeys whose lesions were not made at an early age, and
those of GY, as well as the similarity to GY’s psychophysical findings (Weiskrantz et al. 1998), suggests that
there may be a general correspondence that overrides the
importance of age at which lesion occurred. We are
currently investigating the relationship between psychophysically and pupillometrically determined spatial
channels in a large group of blindsight subjects of varying age and aetiology. The hope is that a common metric
may emerge that allows all subjects to be compared in
the same measurement domain.

256
Acknowledgements We would like to thank CS for many hours
of experimentation, Mr. Jim Urquhart for technical support, and
Dr. A. Purdie for help with visual field measurements. We are
grateful to The Royal Society for financial support (Grant 21510
to A.S. and Royal Society/NATO Postdoctoral Fellowship to
R.C.). The award of a Network grant from McDonnell-Pew to
A.S. and L.W. is gratefully acknowledged. Current blindsight research at VRL is supported by the Chief Scientist Office, Scottish
Executive (CZB/4/30).

References
Azzopardi P, Cowey A (2001) Motion discrimination in cortically
blind patients. Brain 124:30–46
Barbur JL, Ruddock KH, Waterfield VA (1980) Human visual responses in the absence of the geniculo-calcarine projection.
Brain 103:905–928
Barbur JL, Keenleyside MS, Thomson WE (1987) Investigation of
central visual processing by means of pupillometry. In: Seeing
contour and colour. Proceedings of the 3rd International Symposium of the Northern Eye Institute, Manchester, UK
Barbur JL, Harlow AJ, Weiskrantz L (1994) Spatial and temporal
response properties of residual vision in a case of hemianopia.
Philos Trans R Soc Lond B Biol Sci 343:157–166
Benson PJ, Guo K, Blakemore C (1998) Direction discrimination
of moving gratings and plaids and coherence in dot displays
without primary visual cortex (V1). Eur J Neurosci 10:3767–
3772

Bisti S, Sireteanu RC (1976) Sensitivity to spatial frequency and
contrast of visual cells in the cat superior colliculus. Vision
Res 16:247–251
Payne BR, Lomber SG, Macneil MA, Cornwell P (1996) Evidence
for greater sight in blindsight following damage of primary
visual cortex early in life. Neuropsychologia 34:741–774
Pöppel E, Held R, Frost D (1973) Residual visual function after
brain wounds involving the central pathways in man. Nature
203:295–296
Sahraie A, Weiskrantz L, Barbur JL, Simmons A, Williams SC,
Brammer MJ (1997) Pattern of neuronal activity associated
with conscious and unconscious processing of visual signals.
Proc Natl Acad Sci U S A 94:9406–9411
Sahraie A, Weiskrantz L, Barbur JL (1998) Awareness and confidence ratings in motion perception without geniculo-striate
projection. Behav Brain Res 96:71–77
Weiskrantz L (1986) Blindsight: a case study and implications.
Oxford University Press, Oxford
Weiskrantz (1998) Blindsight: a case study and implications, 2nd
edn. Oxford University Press, Oxford
Weiskrantz L, Warrington EK, Sanders MD, Marshall J (1974)
Visual capacity in the hemianopic field following a restricted
occipital ablation. Brain 97:709–728
Weiskrantz L, Barbur JL, Sahraie A (1995) Parameters affecting
conscious versus unconscious visual discrimination with damage to the visual cortex (V1). Proc Natl Acad Sci USA
92:6122–6126
Weiskrantz L, Cowey A, Le Mare C (1998) Learning from the
pupil: a spatial visual channel in the absence of V1 in monkey
and human. Brain 121:1065–1072