SPECIAL SECTION

2D BUT NOT 3D: PICTORIAL-DEPTH DEFICITS IN A CASE

OF VISUAL AGNOSIA
Oliver H. Turnbull1, Jon Driver2 and Rosaleen A. McCarthy3

(1School of Psychology, University of Wales, Bangor, UK; 2Institute of Cognitive Neuroscience and Department of
Psychology, University College London, UK; 3Department of Experimental Psychology, University of Cambridge, UK)

ABSTRACT

Patients with visual agnosia exhibit acquired impairments in visual object recognition, that may or may not involve
deficits in low-level perceptual abilities. Here we report a case (patient DM) who after head injury presented with object-
recognition deficits. He still appears able to extract 2D information from the visual world in a relatively intact manner; but
his ability to extract pictorial information about 3D object-structure is greatly compromised. His copying of line drawings
is relatively good, and he is accurate and shows apparently normal mental rotation when matching or judging objects tilted
in the picture-plane. But he performs poorly on a variety of tasks requiring 3D representations to be derived from 2D
stimuli, including: performing mental rotation in depth, rather than in the picture-plane; judging the relative depth of two
regions depicted in line-drawings of objects; and deciding whether a line-drawing represents an object that is ‘impossible’
in 3D. Interestingly, DM failed to show several visual illusions experienced by normals (Muller-Lyer and Ponzo), that
some authors have attributed to pictorial depth cues. Taken together, these findings indicate a deficit in achieving 3D
intepretations of objects from 2D pictorial cues, that may contribute to object-recognition problems in agnosia.

Key words: pictorial depth, agnosia, object recognition, 2D, 3D

INTRODUCTION three dimensions are extracted from two-

dimensional images. There is an extensive
The period spanned by Ennio De Renzi’s psychological literature on this topic (Gibson, 1966;
inspirational career has seen an enormous growth in Gregory, 1966; Julesz, 1971; Rock, 1973), but
our understanding of vision and its neurobiological rather less in the way of neuropsychology (see
basis, due in part to single-cell recording methods Carey et al., 1998). De Renzi’s (1982) classic text
and more recent technical developments such as covers the issue of depth perception in only two
functional imaging, but also due to the careful study brief sections (pp. 60-63 and pp. 141-145).
of neuropsychological deficits after brain injury in Nevertheless, in reviewing the literature two
neurological patients, as pioneered by De Renzi. As decades ago, De Renzi characteristically
regards theory, the recent decades also have seen an foreshadowed a number of issues that have since
increasing understanding of the importance of become more widely recognised (e.g. Turnbull,
regional specialization - in domains such as spatial 1999). For example, De Renzi warned of the
cognition, object recognition, and visually-guided potential risks of designating disorders as being of
action (Farah, 1990; Grusser and Landis, 1991; exclusively perceptual origin, when the findings
Milner and Goodale, 1995), together with an might also involve motor processes, as when a
understanding that small-scale ‘modules’ are ‘depth’ deficit is evident only in action (e.g. in
typically part of wider patterns of cortical distance errors when misreaching, De Renzi, 1982,
organization, and form interacting components of p.141).
widespread networks (Milner and Goodale, 1995; De Renzi (1982) also commented on the likely
Ungerleider and Mishkin, 1982; see also Cowey, anatomy of disorders in depth processing. He noted
1981). De Renzi has of course made seminal (pp.141-145) that the most common lesion sites
contributions to virtually every domain of implicated by these deficits seemed to involve the
neuropsychology alluded to above, across several occipital and/or parietal lobes (i.e. including dorsal
decades (e.g. De Renzi and Faglioni, 1967, to De rather than ventral pathways). The more recent
Renzi and Di Pellegrino, 1998) - though his work is literature, as reviewed by Carey et al. (1998),
perhaps most accessibly known from his classic contains only a limited number of
monograph (De Renzi, 1982) neuropsychological studies on depth, which mainly
The extent of progress in understanding the approached the problem from the ‘perception
neuropsychology of vision has been far from versus action’ perspective of Milner and Goodale
uniform, however, and there are several apparently (1995), and have focused primarily on the dorsal
important topics that may have been under- ‘visuo-motor’ stream. Thus, recent studies have
investigated to date. One such topic is visual depth tended to focus on the role of depth information in
perception: and in particular the question of how the processing of relative or absolute distance, as

Cortex, (2004) 40, 723-738

724 Oliver H. Turnbull and Others

evidenced in either perceptual judgements or else accurate copies of seen line-drawings – may be
in the biomechanics of reaching, grasping and inadequate.
catching, with binocular versus monocular cues It was in this context that Farah (1990, pp. 59-
being compared (see Haffenden and Goodale, 69) listed several lines of evidence which
1999; Mon-Williams and Dijkerman, 1999; Marotta suggested that many cases of so-called ‘associative’
et al., 1997; Mon-Williams et al., 2001; Servos et agnosics might in fact have some perceptual
al., 1992; Servos and Goodale, 1998). The main deficits in extracting object structure, despite
thrust of this literature has been in general preservation of elementary visual processes. While
agreement with De Renzi’s (1982) proposal that such patients may eventually produce relatively
binocular depth information may be useful in the accurate copies of line drawings, in fact these may
control of spatial action (see also Dijkerman et al., be produced in a pathologically ‘slavish’ and
1996; Gallese et al. 1995, Otto-de Haart et al., piecemeal manner (see also Humphreys and
1999, Marotta et al., 1998), by the dorsal stream. Riddoch, 1987). Moreover, one patient categorised
Any possible role for pictorial cues to depth as having associative visual agnosia (with
processing in the ventral (occipito-temporal) stream apparently ‘normal’ copying) case M.S. of Ratcliffe
appears to have been less studied by and Newcombe (1982), was unable to classify line-
neuropsychologists in recent years, at least within drawings of the type introduced by Penrose and
the literature inspired by the dorsal/ventral Penrose (1958) as ‘possible’ or ‘impossible’ in 3D.
dichotomy. However, in mainstream psychology it The ‘impossible’ versions among these stimuli
has often been argued that pictorial cues to depict perspective views which are very unlikely
distance might have major impacts on perceptual (Gregory, 1970) to correspond to a coherent three-
processes, as in size-constancy (e.g. Gibson, 1966) dimensional object, because of inconsistencies in
and in potentially related illusions of size (e.g. the pictorial depth information at various junctions
Gregory, 1966). on the shape (Cowan, 1974). We report a task
We turn now to consider the possible role of based on such stimuli later in this paper (and an
pictorial depth-cues in ‘visual agnosia’, the example can be observed in Figure 6).
generic term for acquired neurological disorders Distinguishing possible and impossible versions of
of object-recognition (see Farah, 1990; Grusser these stimuli is trivial for normal observers, but
and Landis, 1991; Humphreys and Riddoch, 1987 this distinction was problematic for patient MS
for reviews). Some forms of brain-damage (Ratcliffe and Newcombe, 1982). Moreover, MS
(typically affecting occipito-temporal cortex, in apparently took the same amount of time to copy
the ventral stream) can disrupt visual object both possible and impossible nonsense objects
recognition, while leaving ‘elementary’ vision (Farah, 1990, p.63), whereas normal subjects are
(brightness, acuity, colour, motion) fairly intact. considerably faster to copy the possible objects,
Recognition by touch, sound, or verbal description presumably as they can use the cues from coherent
may also be preserved (see De Renzi, 1982; 3D structure to aid them.
Farah, 1990; Grusser and Landis, 1991). Many Thus some patients with agnosic deficits,
patients diagnosed on this basis as suffering from including some of those who would traditionally
visual agnosia are described as apparently having be classified as ‘associative’ on the basis of
access to a fairly accurate description of the relatively preserved copying, may be restricted in
structural properties of the visual objects before their ability to obtain the appropriate 3D structure
them, which they cannot recognise. Lissauer from the 2D cues available in line-drawings.
(1890) originally distinguished between Perhaps patients such as MS are able to extract
‘apperceptive’ disorders (for patients in whom 2D edge-based information in the image while the
there was some fundamental perceptual deficit in extraction of 3D structure from pictorial cues
perceiving shape, as apparent for instance in their within the 2D image might be impaired. If so,
inability to match or copy simple shapes), versus this might demonstrate one form of spatial
those with ‘associative’ impairment, in whom low processing that might be specific to the
level perception was apparently intact. Teuber representation of objects, having little or nothing
(1968) elegantly described associative agnosics as to do with the various forms of ‘depth’ processing
having “a normal percept stripped of its that have been associated with the dorsal stream
meaning”. However, it has since been argued that in studies of visually-guided action (c.f. Carey et
many patients who would classically have been al., 1998).
diagnosed as associative agnosics may in fact In the remainder of this paper, we report the
have subtle ‘high-level’ perceptual impairments new case of a patient (DM) who suffers from
that disrupt the full extraction of object structure visual agnosia following head-injury. He still
(e.g. see Farah, 1990; Humphreys and Riddoch, appears able to extract stable 2D information from
1987; Farah and Feinberg, 1997). It has also been visual images; but his ability to extract 3D
noted by such authors that the traditional clinical structure from the pictorial depth-cues within 2D
test for intact structural perception in agnosic images (e.g. line-drawings) is greatly compromised.
patients – namely a preserved ability to make This deficit in deriving 3D structure from pictorial
 Pictorial depth deficits in visual agnosia 725

Fig. 1 – DM’s copy drawing.

cues may contribute directly to his object- contrast to his poor recognition abilities, his
recognition difficulties. copying of line drawings was good (see Figure 1),
even though he often could not recognise what the
original or his copy depicted. The copy shown here
CASE REPORT is representative in being fairly accurate. His
performance is creditable, especially considering
DM was a 19-year-old man who had suffered a that DM often could not recognise what the
closed-head injury in a car-accident around three original or the copy depicted, and his copies
years before testing began. He had no language include many areas with excellent reproduction of
deficits, no visual neglect, and full visual fields, on detail. However, portions of the copy also appear
bedside testing. In addition to his closed head injury, to indicate some difficulties in segmentation and in
he also suffered a spinal cord injury that left him extracting accurate 3D structure. For example, in
paraplegic, and wheelchair-bound. He had a his copy shown in Figure 1, DM did not segment
profound amnesia, with intact short-term or working the conical hat held in the person’s right hand from
memory (for example on Digit Span), but an inability the hand itself; he did not represent the right arm
to retain material from preceding minutes, hours and as a structure separate from the torso that it
days. On assessment he was disoriented for time, partially occludes; and he did not represent the
although not for place and person. He was alert and upper arm as a structure entering the sleeve.
had good insight into his condition. It is noteworthy that in everyday situations and
He and his family reported that his visual during assessment DM appeared entirely normal in
recognition was extremely poor. He was unable to his reaching for objects and other visuo-motor
recognise familiar faces (for example those of his activities, although this was not formally tested.
family), although he could recognise them by their For example, he was remarkably dexterous in the
voices. His family also reported a recognition operation of his wheelchair - directing it around
deficit in daily life for visual objects other than obstacles and into narrow spaces. He was also able
faces. For example, while travelling by car to one to accurately reach for objects in near space, for
testing session, he misrecognised a dog in the back example picking up a cup, or a packet of
of the car ahead as a ‘horse’, although he then cigarettes. He was also able to successfully control
realised that it should not be possible to fit a horse a computer mouse.
into a car. DM’s recognition abilities were
investigated clinically with a diverse range of
stimulus materials, and he was found to be CLINICAL ASSESSMENT
particularly poor (see below) at recognising line-
drawings, especially for animate objects rather than DM was assessed on subsets of the WAIS-R.
artefacts, possibly relating to the greater visual He scored within the average range on Digit Span
similarities that exist between animate categories (Scaled Score 10), but was poor on Vocabulary
(Damasio, 1990; Gaffan and Heywood, 1993; (Scaled Score 5) and Arithmetic (Scaled Score 4),
Sheridan and Humphreys, 1993; Riddoch et al., Picture Completion (Scaled Score 3) and Block
1988; though see Turnbull and Laws, 2000). In Design (Scaled Score 7).
726 Oliver H. Turnbull and Others

TABLE I
DM’s object recognition in various stimulus formats

Animate Artefacts Total (%) Test

Real Objects or Object Models 13/17 30/30 91% Not standardised

Photographs 3/20 13/19 68% Winslowe Press Black and White
Photographs
14/19 44% Winslowe Press Colour
Line Drawings 5/13 22/47 38% Snodgrass and Vanderwart (1980)
0/10 6/20 20% Object Synonym Matching
McCarthy and Warrington (1986)
0% Graded Naming Test
McKenna and Warrington (1980)

DM performed within normal limits on the descriptions of objects. He scored only just outside
Efron Squares/Rectangles task (Efron, 1968), a test normal limits (11/15) on the stringent Naming to
for low level impairments in form discrimination; Description task (Coughlan and Warrington, 1978),
and on the VOSP (Warrington and James, 1991) which asks questions such as “What is the name of
tasks of Shape Detection (19/20), Dot Centring the bird that flies at night and hoots?”. Moreover
(20/20), Dot Counting (10/10), Incomplete Letters when given a name only, and asked to provide a
(18/20), and Cube Counting (8/15). However, he verbal description of the named object, DM
was poor on the VOSP perceptual tasks of Object provided accurate and concise descriptions of
Decision (9/20), recognised only 1/10 of the objects for 10/10 artefacts and 8/10 animate
Warrington and Taylor Unusual Views items, and objects.
was also gravely impaired in recognising
conventional views of the same items. He scored Comment: DM shows a severe deficit in the
only 34/60 on the Benton et al (1983) Face recognition of visual objects, in the presence of
Matching Test. His object recognition was also relatively preserved copying ability - and might
tested by using a variety of stimuli in different well be classified as an ‘associative’ visual agnosic
formats (see Table I). On these tasks DM was best in the traditional scheme (Farah, 1990; Grusser
in recognizing real objects (or 3D models of and Landis, 1991; McCarthy and Warrington,
objects), poorer with photographs, and poorest with 1990). Moreover, while DM is clearly able to
line drawings. For example, he recognised none of derive accurate 2D information from the visual
the items on the Graded Naming Test, describing a world, there are also some preliminary grounds for
kangaroo as a ‘hedgehog’, a scarecrow as a ‘man’, suspecting that DM’s recognition errors may
a buoy as a ‘boat’, a thimble as a ‘sieve’, and follow from an inability to use pictorial depth
made no response to the handcuffs item before this cues. For example, he shows a ‘standard’ pattern
test was terminated. Also, he was poorer in seen in visual agnosia (Farah, 1990; Humphreys
recognizing animate objects than in recognizing and Riddoch, 1987; Farah and Feinberg, 1997) of
artifacts. Both the line drawings vs. real object and poorer performance in recognizing line drawings
the animate vs. inanimate trends are, of course, than photographs and real objects. Thus, like other
common in patients with visual agnosia (Farah, cases of early visual processing impairment
1990; Humphreys and Riddoch, 1987; Farah and (Chainay and Humphreys, 2001), DM may rely, to
Feinberg, 1997). a pathological extent, on cues that are absent in
In a task that we adapted from Biederman’s simple line-drawings, such as texture, shading,
(1987, p. 135) studies with normals, DM was stereo and/or motion parallax cues. Also, he
shown the five fragmented and complete line performs much more poorly in the recognition of
drawings of a cup, wine-glass, watering-can, pair fragmented line drawings that lack some of the
of scissors and stool. He was asked to name each information regarded as a central pictorial depth
on six occasions. He recognised only 11/30 of the cue in the formation of stable object structure
‘fragmented but geon-intact’ items, even though (Biederman, 1987). Finally, while DM’s copies are
normals are at ceiling in recognising these. accurate in many respects, he appears to show
However, he correctly named 29/30 of the some difficulties in deriving adequate 3D
‘complete’ items. (This unusually high success rate structural information from line-drawings objects
for him, on the complete items, may be a function (see Figure 1).
of our repeated presentation of these stimuli, and To investigate these issues further, DM was
their high visual familiarity.) given a series of experimental tasks that
In contrast to his typically poor performance investigated more formally his ability to extract 2D
when naming visually presented items, DM did and 3D information from line-drawings. Our
much better when presented with verbal general finding was that his performance was
 Pictorial depth deficits in visual agnosia 727

relatively intact on the 2D tasks, and by contrast recognition (see above). In a later session, we
was substantially impaired on the 3D tasks. asked him to identify the same drawn objects as
used in this experiment, in their upright orientation,
and he identified only 10/20 even when given
2D TASKS unlimited response time.
As was the case with normal subjects, DM’s
Picture-matching across rotations in the picture- errors were not evenly distributed across all
plane orientations (see Figure 2), but showed a linear
This task was performed in order to establish increase in error rate between 0° and 120° [r (4) =
whether DM was able to rapidly process drawn 0.92, p < 0.01]. DM also showed a significant
objects and compare them across transformations in linear increase in mean reaction time with
two-dimensional space. The task required matching misorientation [r (4) = 0.96, p < 0.01]. (As in
of line drawings for identical objects that could Jolicoeur (1985), this analysis of linearity excluded
undergo picture-plane misorientation. There is a the 180° items, which can show faster
substantial literature showing that reaction times are performance than neighbouring orientations.)
typically slowed with increasing misorientation
when neurologically normal individuals perform Comment: While DM’s reaction time
such tasks (e.g. Jolicoeur, 1985). Moreover, we have performance was somewhat slower than that of
also conducted several previous patient studies using normal subjects, and the slope of his RT function
such tasks (e.g. Turnbull and McCarthy, 1996a, against orientation somewhat steeper, his
Turnbull et al., 2002), and shown that reliably linear performance still retains the general form seen in
reaction-time performances can be produced by the sample of normal subjects - i.e. slower reaction
various patients with neurological lesions. Since only time with increasing misorientation from the
a matching response was required in the present task, canonical orientation (Jolicoeur, 1985). This
identifying the object was not a logical requirement. performance is remarkable, considering the
Moreover, since any transformation was in the profound nature of DM’s agnosic deficits. His
picture-plane only, DM was not logically required to performance on this task suggests that he performs
develop or manipulate a 3D description of the object the matching of pictures using the incremental
in order to successfully complete the task. transformation strategy that is also observed in
normal subjects.
Materials and Procedure: Line drawings of 20
common objects (taken from the Snodgrass and An interesting feature of DM’s reaction time
Vanderwart (1980) corpus, see Turnbull and performance was his rapid response to 180° items.
McCarthy, 1996a, for further details) were This finding is similar to that found by Jolicoeur
presented. As in Turnbull and McCarthy (1996a), (1985) in normal subjects for 180° items, although
the first object was presented offset by 10° to the the effect on reaction times for 180° items was
left of the screen center for 200ms, followed by a more marked in DM than was reported in Jolicoeur
blank screen for 300ms, after which the second (1985). DM’s error rate on the 180° items was also
object appeared in the center of the screen until lower than for the 120° and 240° items. Taken
DM responded. The first object was always in an together, these data suggest that DM, like normal
upright orientation, and the second object was subjects, might not employ an incremental
rotated from the upright: randomly presented in transformation strategy to match 180° misoriented
one of six orientations (0°, 60°, 120°, 180°, 240°, items (Jolicoeur, 1985).
300°). DM was asked to judge the items as ‘same’
(e.g. dog-dog - note that ‘same’ trials always
involved identical albeit rotated images) or SHEPARD-METZLER MENTAL ROTATION
‘different’ (e.g. horse-dog) in a speeded manner. He IN THE PICTURE PLANE
responded with a key on the right side of the
keyboard if the items were the same, and a key on The object-matching experiment described
the left side if the items were different. He above suggests that DM employed an incremental
completed three blocks of 240 trials in three transformation strategy in dealing with misoriented
different sessions. drawn objects. To test this more directly he was
given the classic mental-rotation task of Shepard
Results: DM made only 70 errors on the 720 and Metzler (1971), which requires mirror-image
items of the task (9.8%). This rate of errors is discrimination rather than object matching per se
somewhat higher than that found in a small group (see Turnbull et al., 2002). This task involves the
of university-educated undergraduate control presentation of a pair of perspective line drawings
participants (mean 4.4% errors, s.d. 1.0%), who of objects, each constructed of ten cubes (see Figure
were matched with DM for age but not education. 3). One of the objects can be rotated relative to the
However, it remains a creditable performance, other, but this only occurred within the picture plane
especially given his profound difficulties in object in the present experiment (i.e only 2D differences in
728 Oliver H. Turnbull and Others

900

850

800
Reaction time (ms)

750
D.M. reaction time
D.M. errors

550 normals reaction time

normals errors

500

450

20
15

Error (%)
10
5
0
0 60 120 180 240 300 360

Orientation (degrees)

Fig. 2 – Performance in recognizing misoriented line drawing of objects.

orientation could occur). One form was either performance within normal limits compared with
identical (but possibly rotated) with respect to the eight neurologically-normal control subjects, who
other in the same display, thus requiring a ‘same’ averaged 5.7% errors (s.d. 5.5%). The data points
response; or was a mirror-image of that shape which for stimulus pairs with orientation-differences
could also be rotated, requiring a ‘different’ greater than 180° were combined with the mirror-
response. image angle below 180° (as in Shepard and Metzler,
1971). Incorrect responses were excluded from
Materials and Procedure: The figures were reaction time analyses. The mean reaction times and
those used by Shepard and Metzler (1971), with a errors, in relation to angular disparity, for ‘same’
single model compared against a target presented in responses (again as in Shepard and Metzler, 1971)
one of seven different orientations: 0°, 40°, 80°, are presented in Figure 3. DM showed a significant
120°, 160°, 200°, and 240°, in either a ‘standard’ or increase in error rate with angular disparity for
‘reflected’ form (as in Turnbull et al., 2002). There orientations between 0° and 120° [r (4) = 0.81, p <
were 70 trials in one block, consisting of five 0.05]. DM’s reaction time performance, as a
versions of the 14 stimulus pairs (seven were ‘same’ function of orientation, likewise showed a linear
and seven ‘different’ responses). The order of items increase [r (4) = 0.97, p < 0.001].
in each block was randomised before each
presentation. DM completed three blocks of the task Comment: In spite of his substantial object
over different sessions. The task was also performed recognition deficit for line-drawings, as described
by a control group of eight subject; four male and above, DM appears able to perform the complex
four female, with age ranging from 21 to 30 years. anaglyph-discrimination required by the Shepard-
Metzler task across transformations in the picture-
Results: Across the three blocks, DM produced plane (i.e. across changes in 2D orientation). His
an average error rate of 10.5%. This places DM’s performance was within the normal range in terms
 Pictorial depth deficits in visual agnosia 729

D.M. reaction time

6000
normals reaction time

5000 D.M. errors

Normals errors
Reaction time (ms)

4000

3000

60
2000 50

Error (%)
40
30
1000
20
10

0 40 80 120 160

Orientation (degrees)

Fig. 3 – Performance in picture-plane mental rotation.

of accuracy, and showed the linear increase in Materials and Procedure: The figures were
reaction time with differences in orientation that is again drawn from Shepard and Metzler (1971, see
traditionally interpreted as evidence for ‘mental Figure 4). There were again 70 trials in one block,
rotation’ in normal subjects (Shepard and Metzler, consisting of five groups of the 14 stimulus pairs
1971). DM’s ability to compare unfamiliar complex (seven requiring ‘same’ and seven requiring
shapes across transformations in the picture-plane ‘different’ responses). The order of items in each
suggests that he has access to some useful block was randomised before presentation. DM
description of object structure, at least at a 2D completed three blocks of the task over different
(picture-plane) level. However, we noted earlier that sessions. The task was also performed by the same
DM appears to have some difficulty in deriving control group of eight subjects as before. Four were
adequate 3D structural information about objects male and four female, with age ranging from 21 to
from line-drawings. To examine this issue further, 30 years.
DM was given a variety of tasks assessing his
ability to extract 3D from line drawings, and to cope Results: The responses of normal subjects and
with 3D transformations. DM, for correct ‘same’ responses, are presented in
Figure 4. Across the three blocks DM’s accuracy
performance was not significantly different from
3D TASKS chance, with an average error rate of 46.7% [χ2
SHEPARD-METZLER MENTAL ROTATION IN DEPTH (209) = 112, p > 0.05]. This effect was stable over
the three successive blocks (25/70, 38/70 and 35/70
To further investigate the adequacy of DM’s errors). In contrast, normal subjects made an
derivation of 3D structure from line drawings, he average of 8.0% errors (s.d. 6.0%).
was given a depth-rotation version of the Shepard- As might be expected, given that the DM’s
Metzler task reported above. This experiment was accuracy on the task was close to chance, his reaction
identical to the previous one, except that the pair of time performance as a function of orientation no
stimuli that had to be compared for the anaglyph- longer showed a significant linear increase with
discrimination on each trial could now differ in angular disparity [r (4) = 0.44, p > 0.05].
angular rotations only in the depth plane, not in the
picture-plane. Thus 3D rather than 2D Comment: DM’s performance on this 3D mental-
transformations of the objects were now involved rotation task was at chance, which contrasts with his
for the first time. good performance for the picture-plane mental
730 Oliver H. Turnbull and Others

D.M. reaction time

Normals reaction time

D.M. errors

6000 Normals errors

5000
Reaction time (ms)

4000

3000

2000

60
1000 50

Error (%)
40
30
20
10

0 40 80 120 160

Orientation (degrees)

Fig. 4 - Performance in depth mental rotation.

rotation task in the previous experiment (compare appropriate 3D structural interpretations of the forms
Figures 2 and 3, against Figure 4), in which DM shown, consistent with some of the clinical evidence
had shown an error rate within the normal range we presented earlier. Alternatively (or in association)
(10.5%). A comparison between DM’s error scores it might also involve a deficit in implementing 3D
on the depth versus picture-plane mental rotation transformations of extracted forms. To investigate
tasks confirmed a significant difference in this further, a new task was devised in which DM’s
performance (t = 3.7, p < 0.05). ability to derive 3D interpretations of shape from line
The striking difference in DM’s performance drawings was tested, but now without requiring any
between the two mental rotation tasks (2D versus mental rotation, nor any comparison across
3D) may have some bearing on findings about transformed images.
mental rotation in normal subjects. The classic
cognitive psychology literature on normal
performance has reported that the accuracy and DEPTH JUDGMENTS WITHIN LINE-DRAWINGS
speed of performance might be equivalent for OF SHEPARD-METZLER FORMS
picture-plane and depth-plane mental rotation tasks
(R. Shepard and J. Metzler, 1971; S. Shepard and D. In this task DM’s ability to derive 3D
Metzler, 1988; but see Jolicoeur et al., 1985; Bauer information from line-drawings was assessed, by
and Jolicoeur, 1993). However, the dramatic asking him to judge which of two dots placed on the
difference in performance on the two versions of the surface of a form depicted in a line-drawing was
task in DM suggests that they can involve different ‘closer’ to him. The task employed the same type of
processes, which may be selectively disrupted. drawn Shepard-Metzler form that DM had
DM’s inability to perform the 3D mental-rotation previously seen in the ‘mental rotation’ task (see
task could directly reflect an inability to derive Figure 4). This new task was devised in order to
 Pictorial depth deficits in visual agnosia 731

Fig. 5 – Stimuli for dot-depth judgement;

establish whether DM’s failure on the 3D ‘mental interpretation. We stress the factor of elevation,
rotation’ task was due solely to an inability to because this can relate to depth as shown not only in
perform 3D transformations, or might involve a Gibson’s classic work on pictorial depth-cues
deficit in extracting the 3D structure of the depicted (Gibson, 1966) but also by several recent findings.
shape. For example, Gardner and Mon-Williams (2001)
have shown that elevation cues provided by gaze
Materials: Six examples of Shepard and Metzler angle can influence depth processing in visually-
(1971) figures were used. Each form now had two guided action. Similarly, Marotta and Goodale (1998)
dots placed on the surface, such that one dot would have suggested that height in the visual field can be
clearly be considered ‘closer’ to the viewer when the used to influence aiming, reaching and grasping
line drawing is given an appropriate 3D interpretation movements under monocular conditions.
(see Figure 5A and 5B). The placement of the dots in
terms of their relative elevation was balanced so that We employed six Shepard-Metzler shapes in the
for some forms (such as Figure 5A) elevation-in- task. Three variations in dot position were used for
plane was a correct predictor of the relative depth each form, which altered the vertical and horizontal
represented by the two dots: i.e. the lower dot would distance between the dots. In a ‘Large vertical
be ‘closer’ to the observer in the three-dimensional separation’ version (e.g. Figure 5B), the vertical
form represented by the drawing. In the other forms distance between the dots greatly exceeded the
(such as Figure 5B) the dot lower on the screen would horizontal distance (across all stimuli the average
be ‘further’ from the subject following 3D ratio was 1.9 :1). The distances were approximately
732 Oliver H. Turnbull and Others

equal in a ‘Balanced vertical separation’ version elevation differences between the two dots. DM
(average ratio 0.9 :1). In a ‘Small vertical separation’ made accurate judgments for the forms for which
version (e.g. Figure 5A) the vertical distance was elevation correctly predicted depth (2% errors). For
smaller than the horizontal distance (average ratio the items in which relative elevation did not predict
0.5 :1). These differences in vertical separation were the relative depth of parts of the form, he performed
included to establish whether there was any gradient much more poorly (51% errors). Although
of accuracy associated with task difficulty. horizontal distance between the dots also varied
In sum, here we employed 18 stimuli, consisting across the stimuli, this did not predict DM’s
of six Shepard-Metzler forms, each with three performance. In contrast, DM could accurately
possible differences in vertical separation between discriminate the relative two-dimensional position
the dots. A block of 90 items was devised, of the dots (0.6% errors).
consisting of 5 repetitions of each of the 18 items.
As a control task, to tap DM’s ability to derive two Taken together, these findings suggest some
dimensional information from the drawings, DM deficit in extracting correct 3D interpretations of
was asked to judge which of the two dots was line-drawings. His failure in the earlier 3D mental
‘higher’ in each image. It was anticipated that DM’s rotation task, reported above, cannot necessarily be
‘elevation’ judgment would be good, given that he attributed to a deficit in mental rotation in depth
seems able to extract 2D information with a per se, given that it might also or instead reflect a
reasonable degree of accuracy. However, we deficit in the more elementary ability of deriving
predicted that his 3D ‘depth’ judgments would be appropriate depth information from 2D
poor, since they require information about pictorial information.
depth to be extracted.

Procedure: In the 3D task, DM was asked to DEPTH JUDGMENTS FOR PARTS OF SHADED
push a key, identified on the right of the keyboard, SHEPARD-METZLER FORMS
if the dot on the right would be ‘closer’ to him in a
three dimensional version of the form, and a key on Our general clinical assessment of DM had
the left of the keyboard if the dot on that side was suggested that his object recognition performance
‘closer’. In the 2D control task, DM was asked just might be better when viewing real objects, or their
to ignore the Shepard-Metzler forms altogether and photographs, than for simple line-drawings with no
simply concentrate on the height of the dots, shading. We next examined whether he would
pressing the right-sided key if the dot on the right perform better at judging the relative depth of two
was higher on the screen, and the left sided key if dots placed on surfaces of a Shepard-Metzler block-
the left dot was higher. DM was given four blocks, form if these had additional depth cues (with
each with 90 trials. The 3D or 2D tasks were shading in particular). The new task was as in the
presented in a counter-balanced (ABBA) order, previous experiment, except that the stimuli were
beginning with the 3D task. DM was given a short now digital photographs (see Figures 6C and 6D),
break between blocks. converted to bit-map graphics.

Results: For the 3D depth task, DM made Procedure: The procedure was identical to the
48/180 errors (26.6%). Almost all of these errors previous experiment, apart from the change in
were on those items where elevation did not stimuli. DM again performed four blocks of 90
correctly predict depth. His error rate ranged from trials each, two for the 3D judgements and two for
0% to 7% for items where the vertical separation the 2D control task.
predicted depth (mean 2.2%), and from 33% to 87%
for items where the vertical location did not predict Results: For the 3D depth task, DM made 55/180
depth (mean 51.1%). There was no significant errors (30.5%). All of these errors were on those
difference in DM’s performance as a function of the items where elevation did not correctly predict
three classes of ratio between horizontal versus depth. His error rate ranged from 37% to 87% for
vertical distance between the dots. Reaction time items where the vertical separation did not predict
performance did not correlate with either the depth (mean 61.1%). As with the line drawing
(signed) vertical [r (17) = 0.09, p > 0.05] or stimuli in the previous experiment, there was no
horizontal distance between the dots [r (17) = 0.10, significant difference in DM’s performance as a
p > 0.05]. function of the three classes of ratio between
On the 2D control task, DM made only 1/180 horizontal and vertical distance between the dots.
errors (0.56%). There was a significant relationship Reaction time performance did not correlate with
between vertical separation and reaction time either vertical [r (17) = 0.4, p > 0.05] or horizontal
performance on this task [r (17) = 0.60, p > 0.05]. distance between the dots [r (17) = 0.13, p > 0.05].
In the control 2D task, DM performed perfectly
Comment: DM’s performance on the 3D task (0/180) errors. As in the previous task, there was a
appears strongly related to the variable of 2D significant relationship between elevation and
 Pictorial depth deficits in visual agnosia 733

Fig. 6 – Stimuli for impossible object and visual illusion tasks.

reaction time performance on the task [r (17) = neuropsychological case of visual agnosia, patient
0.62, p < 0.05]. MS of Ratcliffe and Newcombe (1982), who was
able to copy line drawings of objects accurately that
Comment: These results serve to replicate those he could not recognise. He was unable to
for line-drawings in the previous experiment. DM discriminate possible from impossible objects, and
performed well on items where the relative apparently also took as long to copy possible as
elevation of the dots was an accurate predictor of impossible objects (Farah, 1990, p.63), unlike
3D form, but was at chance, or worse, for the items normals who are typically faster to copy line-
in which relative elevation did not predict the drawings of objects that are ‘possible’ in 3D.
relative depth of parts of the form. However, he Given DM’s apparent deficit in processing the
could accurately discriminate the relative 2D 3D structure of objects depicted in line-drawings,
position of the dots. The fact that DM’s we anticipated that he would also be impaired at
performance with the ‘shaded’ stimuli did not judging whether they depicted possible or
improve over that with simple line drawings impossible objects. As a control to check that he
suggests that this did not provide a sufficient cue to was nevertheless able to resolve the fine 2D details
depth for him in this task. that distinguished possible and impossible versions,
in a separate task we asked him to judge whether
two versions were the same or different, regardless
DISCRIMINATING ‘POSSIBLE’ VERSUS ‘IMPOSSIBLE’ of their possible or impossible nature.
OBJECTS FROM LINE-DRAWINGS
Stimuli and Procedure
The ability to discriminate structurally ‘possible’
from ‘impossible’ objects (Penrose and Penrose, The stimuli were the ‘diamond’ and ‘rectangle’
1958) has been used as one index of normal drawings taken from the set of ‘possible’ and
observers seemingly effortless ability to derive 3D ‘impossible’ objects generated by Young and
structure from line drawings (e.g. see Gregory, Deregowski (1981, see Figure 6A). These stimuli
1970; Young and Deregowski, 1981). A test of this were presented in pairs, with each pair on one sheet
has also been applied to at least one of paper, for DM to view for as long as he wished.
734 Oliver H. Turnbull and Others

In the 3D task (see Figure 6A), one ‘possible’ and pictorial depth cues leading to the misapplication of
one ‘impossible’ drawing was presented to DM as a size-constancy mechanisms (Gregory, 1966).
pair. He was told that he would see two drawings, However, while there is some experimental support
“one of which is a ‘good’ drawing, while one is a for this hypothesis from studies of normal
‘bad’ drawing that doesn’t make sense if you look perception, it remains contested (e.g. De Lucia and
closely” (c.f. Young and Deregowski, 1981). In a Hochberg, 1991; Hochberg, 1987). We reasoned that
forced-choice paradigm DM was asked to select the if the Muller-Lyer and Ponzo illusions are indeed at
‘bad’ or ‘impossible’ drawing by pointing at it. The least in part due to pictorial depth-cues, and if
position of the correct item was pseudo-randomly patient DM fails to utilise such depth-cues in the
varied across blocks of ten trials, such that the normal manner, then he might fail to show these
correct response item was not biased towards the illusions to the normal extent. We note also that
left or right side. many recent investigators have suggested that the
In the 2D control task (Figure 6B), DM was conscious perception of visual illusions may depend
again presented with pairs of drawings, drawn from on ventral stream mechanisms (e.g. Aglioti et al.,
the same possible and impossible sets, but now 1995) that are classically damaged in object
paired such that both could be possible, both agnosics, although this perspective is not without its
impossible, or one possible and the other impossible critics (e.g. Franz et al. 2000; Jackson and Shaw,
(the latter applied to 50% of trials). The task was 2000; for review see Carey, 2001).
now to indicate whether the two drawings on any DM was now required to perform line-length
trial were the ‘same’ or ‘different’, regardless of judgments in situations that can induce the Muller-
whether they were possible or impossible. Lyer or Ponzo illusions. Note that according to the
hypothesis presented above, we derived the
Results: On the 3D possible-impossible paradoxical prediction that in these tasks DM would
discrimination task DM was correct on 5/10 of the actually be more accurate than normal subjects!
Diamond items, and 6/10 of the Rectangle items,
and thus no better than chance overall (11/20). On Stimuli and Procedure: This experiment was run
the same-different discrimination items (Task B) using ‘Eye Lines’ software (Beagley, 1990), which
DM was correct on 10/10 of the Diamond items, offers one standard experimental paradigm for
and 10/10 of the Rectangle items. This difference testing the Muller-Lyer (Figure 6C) and Ponzo
in performance between the 3D and the 2D tasks illusions (Figure 6E). The subject actively
was significant (χ2 = 11.6, p < 0.001). manipulated the length of a plain line (ii), by mouse
control, to match the length of a target line (i),
Comment: DM had great difficulty in deciding under ‘illusion’ or ‘control’ conditions. In the
on the 3D plausibility of the ‘impossible’ objects Muller-Lyer experiment ‘illusion’ condition, the
presented as drawings. This is similar to the target was the classic horizontal ‘Y’-junction
performance of the agnosic patient reported by Muller-Lyer form (Figure 6C). Below this target,
Ratcliff and Newcombe (1982) who also had good and offset relative to it, was a line (with no Y-
copying skills. However, DM was able to see the junctions) whose length was under DM’s control
2D features of the line-drawings clearly enough to through mouse-manipulation. DM was required to
discriminate between same and different items (an match the length of the two lines. In the control
important control that was not implemented by condition (Figure 6D) the target was a plain line, of
Ratcliff and Newcombe). A plausible interpretation the same length as in the ‘illusion’ condition, and
of these data is that the same-different task can be again the length of a plain second line was
completed without a requirement to extract an controlled by DM, who was asked to match the
appropriate 3D description of the drawn object, but length of the two lines.
the possible-impossible task requires the extraction In the Ponzo version of the experiment (see
of 3D structure. Figure 6E) the ‘illusion’ condition involved the
same two horizontal lines now placed between two
vertically angled lines providing the standard Ponzo
INSENSITIVITY TO MÜLLER-LYER AND PONZO illusion of depth - such that the target line is usually
VISUAL ILLUSIONS experienced as longer than its veridical length.
Again, the control condition consisted of the two
Normal observers show various illusions in lines alone (Figure 6D). The veridical length of the
which they fail to accurately judge the two- control line was 10.6 cm, viewed at a distance of
dimensional extent of lines. The best known approximately 40cm.
examples are the Muller-Lyer and Ponzo illusions The Muller-Lyer task was administered in one
(Goldstein, 1996; Gregory, 1966; Sekuler and session. It involved 15 ‘illusion’ and 15 ‘control’
Blake, 1990, see Figures 6C and 6E). Although the trials, presented in random sequence - again such
exact psychological and neural basis of these that with any illusion the target line should be
illusions remains controversial, on one influential experienced as longer than its veridical length. The
account they may result at least in part from Ponzo task was given in a separate session, also in a
 Pictorial depth deficits in visual agnosia 735

random sequence of 15 ‘illusion’ and 15 ‘control’ might be able to distinguish these possibilities. For
trials. Four neurologically-normal participants, now, the present results show one way in which
matched for age, also completed the tasks. DM’s vision is pathologically supranormal; his 2D
line-length judgements are not subject to illusions
Results: Muller-Lyer illusion: DM’s mean that lead to powerful distortions of perception in
chosen length under the ‘illusion’ condition was normals.
10.8 cm (s.d. 0.9cm). For the control condition his
chosen length was also a mean of 10.8 cm (s.d.
0.9cm); thus, DM did not exhibit any illusion. For GENERAL DISCUSSION
the neurologically-normal control participants, the
mean length under the illusion conditions was 12.2 DM, a patient with visual agnosia, is capable of
cm (s.d. 0.6 cm), which was significantly different successfully performing a variety of tasks involving
from DM’s illusion condition (Mann-Whitney U’ = the representation and manipulation of pictorial
206, p < 0.001). The neurologically-normal stimuli that require only two-dimensional coding.
participants gave a mean length of 10.8 cm (s.d. 0.3 Thus, he is accurate and shows an apparently
cm) in the no-illusion condition, which was normal strategy in matching picture-plane
significantly different from their illusion misoriented object stimuli, and when performing
performance (U’ = 152, p > 0.05). Every control picture-plane image rotations. Also, his copying of
subject showed the illusion. line drawings is relatively good - given that copying
need not rely on the use of pictorial depth cues.
Ponzo illusion: DM’s mean chosen length under However, he performs poorly on a variety of tasks
the ‘illusion’ condition was 10.9 cm (s.d. 0.6cm). that require him to derive three-dimensional
For the control condition his chosen length was also representations from line drawing, such as
a mean of 10.9 cm (s.d. 0.5cm); thus he showed no discriminating possible from ‘impossible’ objects,
Ponzo illusion. For the neurologically-normal performing mental rotation in the depth-plane for
participants, the mean under illusion conditions was 2D line-drawings of 3D objects, and judging the
11.7 cm (s.d. 0.3 cm), which was significantly depth depicted in line drawings and shaded
different from DM’s illusion condition (U’ = 199, photographs. These findings strongly suggest that he
p < 0.001). The neurologically-normal participants has a primary deficit in his ability to extract
showed a control condition performance of 10.8 cm pictorial depth cues. He appears to ignore cues to
(s.d. 0.3 cm), which was significantly different from perspective, such as junction information that
their illusion performance (U’ = 116, p > 0.05). signals occlusion and interposition. Moreover, he
Again, every control subject showed the illusion. does not show the classic Muller-Lyer and Ponzo-
illusions, possibly because of his deficits in pictorial
Comment: While neurologically-normal depth-cues (and/or because of pathological
participants showed a Muller-Lyer and a Ponzo segmentation of the inducing stimuli). There is
illusion, DM’s performance on the ‘illusion’ some evidence that DM might rely on the simple
condition did not differ from the control condition. cue of elevation when forced to make depth
Thus, he was exceptional in not showing either judgements where the only cues to depth are
illusion, making his performance more accurate than pictorial.
that of normal control subjects on the 2D length Our findings of deficits in extraction and use of
judgements. 3D object structure from pictorial depth cues begs
the question of whether DM’s pictorial-depth
One intepretation of this pathologically good deficits form the basis for his difficulties in
performance would be consistent with Gregory’s recognising visually presented objects. There is
(1966) intepretation of the Muller-Lyer and Ponzo some preliminary evidence to support this claim.
illusions in terms of pictorial depth cues leading to Notably, DM is severely impaired in his recognition
misapplication of size constancy. We have shown of line drawings of objects, but is far less impaired
earlier that DM has deficits in 3D interpretation of in recognizing real objects. One explanation of this
line-drawings which contain only pictorial cues to effect might be that the primary difference between
depth. A deficit in use of pictorial depth cues by line drawings and real objects is the quality of
DM might then eliminate the Muller-Lyer and surface information that they provide - with real
Ponzo phenomena on Gregory’s (1966) account of objects providing excellent surface cues, and line
these illusions, consistent with our findings. Another drawings being devoid of such information. Also,
possible way to formulate the finding may be that real objects can be parsed in depth by the use of
patient DM groups or segments the components of stereopsis, as well as visuo-motor (i.e. cephalo-
Muller-Lyer and Ponzo stimuli differently to normal ocular) exploration, yielding motion parallax and
observers. This could either provide an alternative changing patterns of occlusion. One interpretation
account to the pictorial-depth-cue hypothesis, or of the present findings would be that DM has lost a
merely a different way of stating the Gregory key subset of the edge-based recognition skills that
interpretation. Further research with agnosic patients most humans use for rapid object identification
736 Oliver H. Turnbull and Others

(Biederman, 1987). On this account, DM would be in processing pictorial cues to depth with the
left only with ‘surface-based’ cues, which would be remarkable sparing of his ability to use 3D cues for
adequate for the recognition of many real objects, action that is apparent in his daily life. With the
but would severely impair his recognition ability exception of De Renzi’s pioneering work, most
when the stimuli were line drawings of objects. previous investigations of depth perception in
These data may also be of interest in the context patients with cortical visual disorders have
of a recent debate on the necessity of three focussed almost exclusively on 3D depth cues in
dimensional information for object recognition. the context of visually-guided action. In such cases
There have, of course, been long-standing attempts the role of the dorsal visual system has been
to model object recognition with the requirement emphasised (see Carey et al., 1998; Dijkerman et
that early vision derive a full three dimensional al. 1999; Servos and Goodale, 1998). While DM’s
(structural) description of the object before ability to use depth for the purposes of action was
recognition (e.g. Marr, 1982; Biederman, 1987). never formally assessed by us, our impression was
However, a number of more recent attempts at that he could use three dimensional depth cues
modelling recognition have attempted to with great accuracy for the purposes of action. For
demonstrate that object recognition can be achieved example, he was able to manipulate his wheel-chair
on the basis of two dimensional information alone without apparent error through narrow spaces,
(e.g. Bulthoff et al., 1995; Tarr and Bulthoff, 1998). judging the distance to ramps, corners and the edge
In this respect, then, DM might conceivably of pavements. He could reach accurately for
represent a test case of the extent to which humans everyday objects at a range of distances, even
can recognise objects from line-drawings without accurately performing precise tasks such as lighting
pictorial depth information. On this interpretation, a cigarette, and at no stage showed errors of
the answer would be that DM’s recogniton visually-guided action that have been traditionally
performance appears to be greatly impaired. It classified as optic ataxia. Nevertheless, more
remains an open question as to whether DM’s visual formal testing of his abilities in these domains
agnosia is a secondary consequence of this primary would be useful.
depth-perception deficit. However, it is notable that, DM’s apparently good performance of such
like many visual agnosics, DM’s recognition is depth-related visuo-motor skills contrasts with his
enhanced when objects are presented with richer very poor performance when he was required to
depth cues - in photographs and with real objects. extract depth information from pictures. This
It has long been recognised that 2D information potential dissociation between depth processing for
is the more primitive class of cue available from pictorial perception versus depth perception for
pictorial input, and that depth information can be action agrees with many other lines of evidence
constructed from pictorial cues (amongst other types that depth information is not a unitary information
of depth cue) and may require construction on the resource, processed by a single ‘module’. Rather,
basis of a range of secondary cues (Marr, 1982). different aspects of depth processing appear to be
Does DM’s profile simply reflect ‘task difficulty’, represented separately, so that these can be
rather than a selective perceptual impairment? We selectively impaired. We might speculate that cues
would contend that his disorder should be such as motion parallax, optic flow, optic
considered alongside those of other cases whose convergence and possibly binocular disparity might
deficit appears by contrast to be more significant for be optimally employed by the (action-oriented)
2D cues – patients who are able to identify line dorsal visual system, while linear pictorial depth
drawings and complex forms – but who are quite cues (T-junctions, occlusion-interposition, and
unable to judge their location and orientation in perspective cues related to the vanishing point) are
space. For example, patients who cannot employed by the (recognition-oriented) ventral
discriminate between a picture and its mirror-image system. Additional flexibility is suggested by
(Gold et al., 1995; McCloskey et al., 1995; the fact that, under sub-optimal conditions,
McCloskey and Rapp, 2000, Turnbull and sensorimotor systems can ‘make do’ with
McCarthy, 1996), and patients who do not show monocular cues (e.g. Gardner and Mon-Williams,
stability in orientation either for the entire picture- 2001; Marotta and Goodale, 1998), and may use
plane (Solms et al., 1988), or for objects within it two-dimensional pictorial cues for obstacle
(Solms et al., 1998; Turnbull et al., 1995; 1997). avoidance (Haffenden and Goodale, 2001).
DM’s impairment in extracting 3D cues from 2D The fact that DM appeared unable to extract
stimuli may reflect yet another dissociation within pictorial depth, and also was not affected by
spatial processing systems. Unlike these other cases, common visual illusions, is of some interest with
in which selective aspects of 2D processing are regard the recent debate on the role of 2D vs 3D
impaired, DM shows a high level of competence in information in visual illusions of extent. The
dealing with 2D cues, but is significantly impaired striking Titchener Circles findings of Aglioti et al
in his ability to go beyond the second dimension (1995) suggested that such visual illusions might
and into the third based on pictorial cues. arise only for consciously mediated (ventral?)
We also wish to contrast DM’s severe problems estimates of size, but not for (dorsally-mediated?)
 Pictorial depth deficits in visual agnosia 737

visually guided reaching (see Carey, 2001 for prehension of objects oriented in the depth plane: 1. Effects of
visual form agnosia. Experimental Brain Research, 112: 442-
review of this controversial topic). However, as 451, 1996.
Carey points out (p.112), there are some tasks in DIJKERMAN HC, MILNER AD and CAREY DP. Motion parallax
which visually guided action appears to be affected enables depth processing for action in a visual form agnosic
when binocular vision is unavailable. Neuropsychologia, 37:
by purely two dimensional information - a point 1505-1510, 1999.
recently supported by the findings of Kwok and EFRON R. What is perception? Boston Studies in Philosophy of
Braddick (2003) on reaching to Titchener Circles. Science, 4: 137-173, 1968.
FARAH MJ. Visual Agnosia: Disorders of Object Recognition and
To conclude, across a number of tasks we have what they tell Us about Normal Perception. Cambridge: MIT
shown that the agnosic patient DM presents with a Press, 1990
deficit in his ability to extract pictorial depth FARAH MJ and FEINBERG TE. Visual object agnosia. In TE Feinberg
and MJ Farah. Behavioral Neurology and Neuropsychology.
information from 2D images, often relying on New York: McGraw Hill, 1997, pp. 239-244.
height in plane instead - a simple monocular cue to FRANZ VH, GEGENFURTNER KR, BULTHOFF HH and FAHLE M.
Grasping visual illusions: No evidence for a dissociation
depth. In contrast, his ability to extract accurate 2D between perception and action. Psychological. Science, 11:
information from the same images, and even to 20-25, 2000.
manipulate such images in two dimensional rotation GAFFAN D and HEYWOOD CA. A spurious category-specific visual
agnosia for living things in normal human and nonhuman
tasks, remains basically intact. He therefore primates. Journal of Cognitive Neuroscience, 5: 118-128,
represents a further subtype of the many selective 1993.
disorders of space exploration and cognition that GARDNER PL and MON-WILLIAMS M. Monocular and binocular
control of human interceptive movements. Experimental
were so clearly outlined in De Renzi’s pioneering Brain Research, 136: 379-385, 2001.
monograph. GIBSON JJ. The Senses Considered as a Perceptual System. Boston:
Houghton Mifflen, 1966.
GOLD M, ADAIR JC, JACOBS DH and HEILMAN KM. Right-left
Acknowledgements: Our thanks to DM for taking the confusion in Gerstmann’s syndrome: A model of body centred
time to complete the various tasks reported in this paper spatial orientation. Cortex, 31: 267-283, 1995.
so cheerfully, and to David Carey for encouragement and GOLDSTEIN EB. Sensation and Perception. Pacific Grove:
patience. Both JD and OT are supported (separately) by Brooks/Cole, 1996.
the Wellcome Trust, and JD holds a Royal Society- GREGORY RL. Eye and Brain. New York: McGraw Hill, 1966.
Wolfson Research Merit Award. GRUSSER O-J and LANDIS T. Visual Agnosias and other
Disturbances of Visual Perception and Cognition. London:
Macmillan Press, 1991.
REFERENCES HAFFENDEN AM and GOODALE MA. The effect of pictorial illusion
on prehension and perception. Journal of Cognitive
AGLIOTI S, DESOUZA JFX and GOODALE MA. Size-contrast Neuroscience, 10: 122-136, 1999.
illusions deceive the eye but not the hand. Current Biology, 5: HAFFENDEN AM and GOODALE MA. The dissociation between
679-685, 1995. perception and action in the Ebbinghaus illusion: Non-illusory
BAUER B and JOLICOEUR P. Stimulus dimensionality effects in effects of pictorial cues on grasp. Current Biology, 11: 171-
mental rotation. Canadian Society for Brain, Behaviour and 181, 2001.
Cognitive Science: 53-54,1993. HOCHBERG JE. Machines should see as people do, but must know
BEAGLEY WK. Eye-lines: A tool for studying visual perception of how people see. Computer Vision, Graphics and Image
size and angle. Behaviour Research Methods, Instruments and Processing, 39: 221-237, 1987.
Computers, 23: 183-186,1990. HUMPHREYS GW and RIDDOCH MJ. To see but not to see: A Case
BIEDERMAN I. Recognition-by-components: A theory of human Study of Visual Agnosia. London: Lawrence Earlbaum
image understanding. Psychological Review, 4: 115-147, Associates., 1987.
1987. JACKSON SR and SHAW A. The Ponzo illusion affects grip-force but
BULTHOFF HH, EDELMAN SY and TARR MJ. How are three- not grip-aperture scaling during prehension movements.
dimensional objects represented in the brain? Cerebral Cortex, Journal of Experimental Psychology: Human Perception and
3: 247-260, 1995. Performance, 26: 418-423, 2000.
CAREY DP. Do action systems resist visual illusions? Trends in JOLICOEUR P. The time to name disoriented natural objects.
Cognitive Sciences, 5: 109-113, 2001. Memory and Cognition, 13: 289-303, 1985.
CAREY DP, DIJKERMAN HC and MILNER AD. Perception and action JULESZ B. Foundation of Cyclopean Perception. Chicago:
in depth. Consciousness and Cognition, 7: 438-453, 1998. University of Chicago Press, 1971.
CHAINAY H and HUMPHREYS GW. The real-object advantage in KWOK RM and BRADDICK OJ. Does the Titchener Circles illusion
agnosia: Evidence for a role of surface and depth information exert an effect on grasping two or three-dimensional targets?
in object recognition. Cognitive Neuropsychology, 18: 175- Neuropsychologia, 41: 932-940, 2003.
191, 2001. LISSAUER H. Ein Fall von Seelenblindheit nebsteinen Betrag
COWAN TM. The theory of braids and the analysis of impossible zur Theorie derselben. Archiv fur Psychiatrie and
figures. Journal of Mathematical Psychology, 11: 190-212, Nervenkrankheiten, 21: 222-270, 1890.
1974. MAROTTA JJ, BEHRMANN M and GOODALE MA. The removal of
COWEY A. Why are there so many visual areas? In FO Schmidt, binocular cues disrupts the calibration of grasping in patients
FG Warden, G Adelman and SG Dennis (Eds), The with visual form agnosia. Experimental Brain Research, 116:
Organisation of the Cerebral Cortex. Cambridge, MA: MIT 113-121, 1997.
Press, 1981, pp. 395-413. MAROTTA JJ, DESOUZA, JFX, HAFFENDEN AM and GOODALE MA.
DAMASIO AR. Category-related recognition deficits as a clue to the Does a monocularly presented size-contrast illusion influence
neural status of knowledge. Trends in Neuroscience, 13: 95- grip aperture? Neuropsychologia, 36: 491-497, 1998.
98, 1990. MAROTTA JJ and GOODALE MA. The role of learned pictorial cues
DELUCIA PR and HOCHBERG J. Geometrical illusions in solid in the programming and control of grasping. Experimental
objects under ordinary viewing conditions. Perception and Brain Research, 121: 465-470, 1998.
Psychophysics, 50: 547-554, 1991. MCCARTHY RA and WARRINGTON EK. Visual associative agnosia:
DE RENZI E. Disorders of Space Exploration and Cognition. A clinico-anatomical study of a single case. Journal of
Chichester: Wiley, 1982. Neurology, Neurosurgery and Psychiatry, 49: 1233-1240,
DE RENZI E and DI PELLEGRINO G. Prosopagnosia and alexia 1986.
without object agnosia. Cortex, 34: 41-50, 1998. MCCARTHY RA and WARRINGTON EK. Cognitive neuropsychology:
DE RENZI E and FAGLIONI P. The relationship between visuo- A clinical approach. San Diego: Academic Press, 1990.
spatial impairment and constructional apraxia. Cortex, 3: 327- MCCLOSKEY M and RAPP B. A visually based developmental
342, 1967. reading deficit. Journal of Memory and Language, 43: 157-
DIJKERMAN HC, MILNER AD and CAREY DP. The perception and 181, 2000.
738 Oliver H. Turnbull and Others

MCCLOSKEY M, RAPP B, YANTIS S, RUBIN G, BACON WF, familiarity, and visual complexity. Journal of Experimental
DAGNELIE G, GORDON B, ALIMINOSA D, BOATMAN DF, Psychology: Human Perception and Performance, 6: 174-215,
BADECKER W, JOHNSON DN, TUSA RJ and PALMER E. A 1980.
developmental deficit in localizing objects from vision. SOLMS M, KAPLAN-SOLMS K, SALING M and MILLER P. Inverted
Psychological Science, 6: 112-117, 1995. vision after frontal lobe disease. Cortex, 24: 499-509, 1988.
MCKENNA P and WARRINGTON EK. Testing for nominal dysphasia. SOLMS M, TURNBULL OH, KAPLAN-SOLMS K and MILLER P.
Journal of Neurology, Neurosurgery and Psychiatry, 43: 781- Rotated drawing: The range of performance, and anatomical
788, 1980. correlates, in a series of 16 patients. Brain and Cognition, 38:
MARR D. Vision. San Francisco: Freeman, 1982. 358-368, 1998.
MILNER AD and GOODALE MA. The Visual Brain in Action. TARR MJ and BULTHOFF HH. Image based object recognition in
Oxford: Oxford University Press, 1995. man, monkey and machine. In MJ Tarr and HH Bulthoff.
MON-WILLIAMS M and DIJKERMAN HC. The use of vergence Image based Object Recognition in Man, Monkey and
information in the programming of prehension. Experimental Machine. Amsterdam: Elsevier Science Publishers, 1998, pp.
Brain Research, 128: 578-582, 1999. 1-20.
MON-WILLIAMS M, TRESILIAN JR, MCINTOSH RD and MILNER AD. TEUBER H-L. Alteration of perception and memory in man. In L
Monocular and binocular distance cues: insights from visual Weizkrantz. Analysis of Bevavioral Change. New York:
form agnosia I (of III). Experimental Brain Research, 139: Harper and Row, 1968.
127-136, 2001. TURNBULL OH. Of two minds about two visual systems. Psyche, 5:
OTTO-DE HAART EG, CAREY DP and MILNE AB. More thoughts on 1-5, 1999.
perceiving and grasping the Müller-Lyer illusion. TURNBULL OH, BESCHIN N and DELLA SALA S. Agnosia for object
Neuropsychologia, 37: 1437-1444, 1999. orientation: Implications for theories of object recognition.
PENROSE LS and PENROSE R. Impossible objects: A special type of Neuropsychologia, 35: 153-163, 1997.
illusion. British Journal of Psychology, 49: 31-33, 1958. TURNBULL OH, DELLA SALA S and BESCHIN N. Agnosia for object
RATCLIFFE G and NEWCOMBE F. Object recognition: Some orientation: Naming and mental rotation evidence. Neurocase,
deductions from the clinical evidence. In AW Ellis (Ed), 8: 296-305, 2002.
Normality and Pathology in Cognitive Function. London: TURNBULL OH and MCCARTHY RA. When is a view unusual? A
Academic Press, 1982. single case study of orientation-dependent visual agnosia.
RIDDOCH MJ, HUMPHREYS GW, COLTHEART M and FUNNELL E. Brain Research Bulletin, 40: 497-503, 1996a.
Semantic system or systems? Neuropsychological evidence TURNBULL OH and MCCARTHY RA. Failure to discriminate
re-examined. Cognitive Neuropsychology, 5: 3-25, 1988. between mirror-image objects: A case of viewpoint-
ROCK I. Orientation and Form. New York: Academic Press, 1973. independent object recognition? Neurocase, 2: 63-72, 1996b.
SEKULER R and BLAKE R. Perception. New York: McGraw Hill, TURNBULL OH, LAWS KR and MCCARTHY RA. Object recognition
1990. without knowledge of object orientation. Cortex, 31: 387-395,
SERVOS P and GOODALE MA. Monocular and binocular control of 1995.
human interceptive movements. Experimental Brain TURNBULL OH and LAWS KR. Loss of stored knowledge of object
Research, 119: 92-102, 1998. structure: Implications for ‘category-specific’ deficits.
SERVOS P, GOODALE MA and JAKOBSON LS. The role of binocular Cognitive Neuropsychology, 17: 365-389, 2000.
vision in prehension. Vision Research, 32: 1513-1521, 1992. UNGERLEIDER LG and MISHKIN M. Two cortical visual systems. In
SHEPARD RN and METZLER J. Mental rotation of three-dimensional DJ Ingle, MA Goodale and RJW Mansfield (Eds), Analysis of
objects. Science, 171: 701-703, 1971. Visual Behavior. Cambridge, MA: MIT Press, 1982, pp. 549-
SHEPARD S and MELTZER D. Mental rotation: Effects of 586.
dimensionality of objects and type of task. Journal of WARRINGTON EK and JAMES M. The Visual Object and Space
Experimental Psychology: Human Perception and Perception Battery. Thames Valley Test Company, 1991.
Performance, 14: 3-1, 1988. YOUNG AW and DEREGOWSKI JB. Learning to see the impossible.
SHERIDAN J and HUMPHREYS GW. A verbal-semantic category- Perception, 10: 91-105, 1981.
specific recognition impairment. Cognitive Neuropsychology,
10: 143-184, 1993.
Oliver Turnbull, Centre for Cognitive Neuroscience, School of Psychology,University
SNODGRASS JG and VANDERWART M. A standardized set of 260 of Wales, Bangor, Wales, LL57 2AS, UK
pictures: Norms for name agreement, image agreement, e-mail: [email protected]