Neurocase

2004, Vol. 10, No. 3, pp. 207214

Differential Impact of Parvocellular

and Magnocellular Pathways on Visual
Impairment in Apperceptive Agnosia?

Helmut Hildebrandt1,2, Cathleen Schutze1,3, Markus Ebke1 and Karoline Spang3

1
Municipal Hospital of Bremen, Neurology, Bremen, Germany, 2Institute for Psychology, University of Oldenburg,
Oldenburg, Germany and 3Institute for Human-Neurobiology, University of Bremen, Bremen, Germany

Abstract

The term visual form agnosia describes a disorder characterized by problems recognizing objects, poor copying,
and distinguishing between simple geometric shapes despite normal intellectual abilities. Visual agnosia has
been interpreted as a disorder of the magnocellular visual system, caused by an inability to separate figure from ground
by sampling information from extended regions of space and to integrate it with fine-grain local information. However,
this interpretation has hardly been tested with neuropsychological or functional brain imaging methods, mainly
because the magnocellular and parvocellular structures are highly interconnected in the visual system.
We studied a patient (AM) who had suffered a sudden heart arrest, causing hypoxic brain damage. He
was/is severely agnosic, as apparent in both the Birmingham Object Recognition Battery and the Visual Object and
Space Battery. First- and especially second-order motion perception was also impaired, but AM experienced
no problems in grasping and navigating through space. The patient revealed a normal P100 in visual
evoked potentials both with colored and fine-grained achromatic checkerboards. But the amplitude of the P100 was
clearly decreased if a coarse achromatic checkerboard was presented.
The physiological and neuropsychological findings indicate that AM experienced problems integrating
information over extended regions of space and in detecting second-order motion. This may be interpreted as a
disorder of the magnocellular system, with intact parvocellular system and therefore preserved ability to detect both
local features and colors.

Introduction
Benson and Greenberg (1969) described a syndrome which
they named visual form agnosia. The case they described
suffered from no serious intellectual impairment, but from
object recognition problems and poor copying abilities. As
shown by Efron (1969) the patient was unable to distinguish
between simple geometric shapes. Several other patients
with similar impairments have been described (Adler,
1944, 1950; Abadi et al., 1981; Campion and Latto, 1985).
Extensive testing of these patients showed multiple dissocia-
tions between figure-ground separation (Kartsounis and
Warrington, 1991), shape discrimination (Davidoff and
Warrington, 1993), and luminance and color discrimina-
tion (Heywood et al., 1994). These dissociations have been
taken as an argument for segregated visual pathways in the
human brain, which can be lesioned focally and therefore
produce different classes of symptoms. A subgroup of these
patients cannot distinguish between different shapes, sizes,
and shadings of grey. One popular explanation for these
symptoms proposes a selective damage of the magnocellular
visual pathway (Milner et al., 1991; Davidoff and Warrington,
1993).
The term magnocellular refers to a subdivision of the
visual system originating in the alpha ganglion cells of the
retina, projecting to layers 4B and 4C of the striate cortex
(V1), and thence to the thick stripes of the adjacent area
V2 (Livingstone and Hubel, 1988). The magnocellular path-
way is supposed to predominantly analyse depth and motion
cues, while it is color blind. On the other hand, because of
the larger receptive fields of its neurons, it has also been
related to the perception of size and length differences

Correspondence to: Helmut Hildebrandt, University of Oldenburg, Institute of Psychology, 2611 Oldenburg, Germany. e-mail: helmut.hildebrandt@
uni-oldenburg.de

10.1080/13554790490495168$16.00 # Taylor & Francis Ltd.

208 H. Hildebrandt et al.
(Kosslyn et al., 1989, 1990), to perception of low spatial
frequencies (i.e., coarse spatial information) and therefore of
surfaces or entities in figure-ground separation (Milner et al.,
1991).
Though many of the problems of patients suffering from
visual form agnosia can be explained by an isolated lesion
of the magnocellular system, not all form agnosic patients
have problems in recognizing motion and in localizing
objects for action. Based on the importance of the magno-
cellular pathway for motion processing, one would predict
that any lesion of this system should severely impair motion
perception (for an extended discussion of this aspect, see
Goodale and Milner, 1995). Motion perception was tested
extensively only in the case of Milner et al. (1991), D.F.
and her motion perception was indeed impaired. Neither
Bartolomeo et al. (1998) nor Davidoff and Warrington
(1993), Ferreira et al. (1998) or Kartsounis and Warrington
(1991) reported an impairment of motion perception in their
cases.
One possibility to reconcile these apparently contradictory
findings is that motion perception has not been investigated
extensively in visual form agnosic patients because the
impairment was so subtle that it did not produce obvious
clinical deficits. For example, only higher-order motion de-
tection may be affected while first-order motion detection,
relying on luminance changes, remains intact (Chubb and
Sperling, 1988; Cavanagh and Mather, 1989; Smith et al.,
1994; Braun et al., 1998).
Another possibility is to postulate different magnocellular
subsystems of visual information processing responsible for
low spatial frequencies, for motion perception and for object
localization. Low and high spatial frequencies are indeed
processed predominantly in different parts of the brain (Fink
et al., 1997, 1999) and may be disturbed differentially after
brain lesions (Robertson and Lamb, 1991), although the
hemispheric difference in global/local processing only be-
comes apparent at later stages of information processing
(Heinze et al., 1998). A strong argument for such a separation
of the magnocellular pathway comes from reading research.
Patients with developmental dyslexia show poor control of
saccades (Biscaldi et al., 1994, 1998) and impaired visual-
spatial attention (Facoetti et al., 2000). They show differences
in the structure of their magnocellular neurons at the lateral
geniculate nucleus (Livingstone et al., 1991) and in the
physiological activity of magno-recipient areas such as MT
(Eden et al., 1996; Sperling et al., 2003; Stein, 2003). But
clearly dyslexic patients are not as impaired in motion per-
ception and grasping (visuomotor integration) as they are in
reading.
Patients with visual form agnosia were also impaired in
reading (Kartsounis and Warrington, 1991; Milner et al.,
1991; Davidoff and Warrington, 1993; Bartolomeo et al.,
1998; Ferreira et al., 1998), but not necessarily in shape
perception (Kartsounis and Warrington, 1991), and in identi-
fying single letters. Recent investigations have shown that in
normal readers the separation of the letters in letter strings
involves primarily the magnocellular pathway, whereas single
letter reading depends on parvocellular activity (Omtzigt
et al., 2002). The high incidence of reading impairments in
visual form agnosia may therefore support the explanation
of their deficits in terms of damage to the magnocellular
pathway.
Visual evoked potentials (VEPs) provide useful neuro-
physiological information with respect to the functioning of
the two systems. Two predictions can be made in this
respect: (1) Selective damage of the magnocellular pathway,
leaving the parvocellular pathway intact, should result in an
intact P100 when using a checkerboard defined by color,
independent of the square size since the parvocellular-blob
pathway processes color-defined stimuli for all visible spa-
tial frequencies (Tobimatsu et al., 1995; Spinelli et al., 1996;
Arakawa et al., 1999). (2) If the magnocellular system
involves the sampling of low spatial frequencies, then
achromatic checkerboards consisting of small squares should
produce a normal VEP whereas a checkerboard with rela-
tively large squares should evoke clearly smaller VEP ampli-
tudes (Livingstone et al., 1991). In particular, the amplitude
of the N1/P1 complex reflects spatial frequency (Tobimatsu
et al., 1995; Arakawa et al., 1999). In achromatic checker-
boards this amplitude increases only with patterns of a
spatial frequency well above 2 cycles per degree, whereas
in chromatic checkerboards it decreases starting at 2 cycles
per degree.
Recently, we studied a patient with visual form agnosia
due to sudden heart arrest using a series of neuropsycho-
logical tests and investigating first and second order motion
perception. Moreover cortical sum potentials were evoked
by reversing achromatic and colored checkerboards of
different sizes.
Case description
AM, a technical drawer, suffered from sudden heart arrest
when he was 46 years-old. He was reanimated and awoke
from coma three days after the incident. Five weeks after
his acute medical treatment, AM was admitted to a neu-
rological rehabilitation unit where all the investigations
reported below took place. At this time, he had no obvious
motor problems, was aware of having visual and memory
problems, but was disoriented in time and place. No
confusional state was evident, although he left the hospital
several times without being allowed to do so. His mood
state alternated between calm and anxious, but he never
became really upset or aggressive. The patient was always
motivated to participate in the investigations and in train-
ing sessions, and became tired only after one hour of
testing.
A neurological examination showed intact motor, sensori-
motor and cranial nerve functions. There was no visual field
defect as tested with a Goldman Perimeter. A 1.5 Tesla MRI
scan, taken 2 months after the heart arrest, showed no focal
damage and no major brain atrophy.
 Differential impact of parvocellular and magnocellular pathways 209

Neuropsychological investigation
Visual fields during visual extinction
We used the TAP-Neglect test (see Hildebrandt et al., 2002,
for an extensive description) to analyze the visual fields of
AM during visual extinction. The test was performed in a
dimmed room, with maximal screen contrast and a 21-inch
screen. AM had no visual field deficits, but response times
especially to targets on the left side were increased (725 ms for
left-sided targets and 602 ms for right-sided targets), and he
also missed one item there.

Clinical testing of visual perception

Illusionary contours
A set of high frequency gratings were shown to AM and he
had to decide whether or not a simple geometric figure was
hidden in these gratings, defined by small line breaks. AM had Fig. 1. Copy of the Rey-Osterrieth Figure. Notice the piecemeal fashion of the
no problems identifying circles and squares in the gratings, copy.
although his reactions were slow and he gave the impression
of not always feeling totally sure.

Kanisza figures
A set of Kanisza figures was presented and AMs task was to
recognize the figures. To enhance the sensitivity to impair-
ment we started with distorted figures (rotating one or two of
the Pac-men by about 90 degrees and then gradually back
rotating these circles). AM was able to discriminate between
the conditions, although his reactions were again slow and
appeared sometimes uncertain.

Color perception
A set of colored squares was shown on a computer monitor
and AM was asked to name the colors. This turned out to be a
very easy task for him.

Copying and visuoconstructional abilities

When copying the complex Rey-Osterrieth figure, AM started
in a piecemeal fashion and ended up with just a partial
solution, documenting some visuoconstructional problems
(see Fig. 1).
In the Gailinger Abzeichentest (Gailinger copy test) geo-
metrical figures have to be copied on a sheet of paper with
dots, indicating specific parts of the original figure. Some of
these figures require three-dimensional perception. AM had
no major problems perceiving the three-dimensional figures
but in assessing the relative length of parts of the figures, and
made four errors during copying (see Fig. 2).
In the Block Design test of the WAIS AMs score was 5,
indicating a medium impairment.
Reading
AM could read single letters if the letters were spatially
separated (independent of font type). The ability to read
words depended on two aspects: First, he performed much
better in reading fonts like Arial than Times Roman, range of our control group, but in the normal range of the
Fig. 2. Item of the Gailinger Abzeichentest.
probably because the letters in Arial lack serifs. Second, his
reading ability depended on letter size: words written in small
letters (letter size < 16 pt.) led to confusions between letters.
Neuropsychological testing of visual
abilities of AM
Space perception was investigated with the Visual Object
and Space Perception Battery (VOSP, see Table 1) and the
Birmingham Object Recognition Test (BORB, see Table 2).
Because the control data used in BORB reflect a 65 age
group, we additionally collected data from 11 control subjects
(mean age of 49 years, minimum age of 43, maximum age of
54 years). AM scored in all spatial tasks BORB below the
210 H. Hildebrandt et al.

Table 1. Results in the visual object and space perception battery

Visual space perception Object recognition

VOSP Correct trials Cut-off score Correct trials Cut-off score

Shape detection (screening task) 18 15 Incomplete letters 8 17

Dot numerosity 10 8 Foreshortened Silhouettes 4 (only animal pictures) 16
Visual location 13 18 Object recognition 2 15
Spatial location 3 7

Performance in tests written in italics was impaired.

Table 2. Results in the Birmingham object recognition battery

Visual space perception Object recognition

BORB Correct trials Controlsa Range of controlsa Correct trials Controlsa Range of controlsa
(maximum) Mean (SD) & (cut-off score (maximum) Mean (SD) & (cut-off score
BORBb) BORBb)

Length match 18 (30) 26.6 (1.6) 2430 (<24) Minimal feature 16 (25) 24.8 (0.4) 2425 (< 19)
task match
Size match task 22 (30) 27.7 (2.3) 2430 (<25) Foreshortened 15 (25) 24.5 (0.5) 2425 (< 16)
match
Orientation 20 (30) 25.3 (1.4) 2328 (<20) Object decision 22 (32) 31.7 (0.6) 3032 (< 28)
match task (easy)
Position of gap 28 (40) 34.8 (2.4) 3039 (<27) Item match 25 (32) 32 (0) 3232 (< 26)
match task
Association match 25 (30) 29.6 (0.8) 2830 (< 22)
Picture naming 10 (15) 14.4 (0.8) 1315 (< 8)
(short version)
Animate drawings
Picture naming of 12 (20) 20 (0) 2020 (< 17)
inanimate
drawings from
Subtest 7
Picture naming 50 (76) 73.8 (2.4) 6976 (< 64)
(long version)

a
Eleven age-matched control subjects were tested with the BORB.
b
Scores of the BORB control group.
Performance in tests written in italics was below the cut-off scores of the BORB. But note that AM scored outside the range of our control group in all subtests.

original BORB control group for two tasks: the orientation
match task and the position of gap task, both requiring a
comparison of directions. Furthermore, AMs performance
was normal in the screening task and in dot counting of the
VOSP.
Identifying nonoverlapping figures, examined with the
subtest of the BORB, was possible for single letters (AM
misnamed 2 of 18 letters, but this may be due to his first
language being Iranian) and almost perfect for geometrical
shapes (3 misidentifications on 36 trials). Identification of
single line drawings (14 errors in 40 trials), of paired letters
and of triplets was impaired (13 errors on 36 presentations of
paired letters and 28 errors out of 36 presented triplets). The
same held for triplets of shapes (10 mistakes on 36 trials) but
less so for paired geometrical shapes (2 errors on 36 trials).
Overlapping letters, geometrical shapes and line drawings
could hardly be identified. In general, AM replied correctly on
less than half of the trials with the exception of geometrical
shapes, on which he performed slightly better (he misidenti-
fied only 8 out of 36 paired stimuli presented and 22 out of 54
triplets presented). However, this performance is far below the
normal range.
Object recognition was also investigated with the VOSP
and BORB (see Tables 1 and 2). AM scored in all of these
tasks below the range of our control group. He was also
impaired on all these tasks taking the results of the original
BORB control group, except for association match and
partly for picture naming of the BORB. In the case of
picture naming his identification of animals was in the normal
range, whereas the naming of manmade objects was highly
impaired. His errors in the association match task involved
only manmade objects, but not animals. It is worth noticing
that the only four correct answers AM gave in the Silhouette
task of the VOSP all concerned animals. We further tested this
difference by means of pictures of the Rivermead Behavioural
Memory Test using 16 animals and 16 manmade objects.
AM identified all 16 animals correctly, but only 13 of the
manmade objects.
 Differential impact of parvocellular and magnocellular pathways
Motion perception
First-order and second-order motion perception was inves-
tigated using sets of moving random dots (Braun et al.,
1998). For each kind of motion the subject had to identify
the direction of motion of a small rectangle, at 108 distance
on the right or on the left of a central fixation point, moving
upwards or downwards on the screen. The target was defined
as a rectangle area with dots moving preferentially in one
direction, while dots in the remainder of the screen moved in
random directions. The coherence of the dots in the target
was varied between 0% (random, no coherence) and 100%
(all dots in one direction, perfect coherence), and the mini-
mal amount of coherence required by the patient to identify
the motion direction of the target was measured. In the first
type of motion condition, the dots within the target area
moved coherently in the same direction as the target itself
( first-order motion). In the second condition, the dots
within the rectangle were moving in the opposite direction
as the target itself ( second-order motion). In this condi-
tion, the elementary movement detectors signal the direction
of the dots, not that of the target (cf. Braun et al., 1998),
and thus higher-order detectors are required to give correct
responses to upward or downward motion of the target
(Theta-motion).
Stimuli were produced by a Silicon Graphics computer
(O2) and presented on a LG Studioworks 221U Monitor
(1152  870 pixel, monitor size 40  30.2 cm), viewed from
an observation distance of 50 cm with frame frequency of
75 Hz. The screen contained overall 50% white and 50%
black pixels. The target for both conditions was 28 high and 48
wide, and moved for 2 s at a speed of 68/s from bottom
upwards or from the top down. AM and 21 age-matched
control subjects (mean age of 43 years, minimum age of 38,
maximum age of 54 years) were tested. Subjects were seated
in a dimly lt room and were instructed to fixate a central
small red arrow (1.38) that indicated whether the target would
appear on the left or right side of the screen. Subjects had
to indicate verbally whether the rectangle seemed to move
upwards or downwards (a binary forced choice task). The
examiner entered the responses on the keyboard without
seeing the stimulus. False answers were followed by acoustic
feedback. For both hemifields 50 trials were presented each
and coherence level was controlled using a staircase pro-
cedure (PEST; Taylor and Creelman, 1967). To determine
thresholds for motion perception, PEST varied the percentage
of dots within the stimulus area, by gradually decreasing the
percentages of dots within the rectangle that moved in the
defined direction, starting with a correlation of 100%. Hence,
with 0% correlation the movements of the dots within the
rectangle were absolutely identical to those in the surrounding
and the target was no longer present. Direction discrimination
thresholds, defined as percentage of signal dots that lead to
75% correct responses, were calculated using probit analysis
(Finney, 1962).
Motion detection did not differ significantly between hemi-
fields both in AM and in the control group, hence we averaged
211
Table 3. Thresholds of AM and controls in detecting first-order (Fourier) and
second-order (Theta) motion
Fourier motiona Theta motiona
Controls (n 21)
Mean 32.85 36.39
Standard Deviation 5.82 6.77
Minimum 25.20 26.6
Maximum 45.85 47.10
AM: t1 45.85 74.05
AM: t2 63.6
a
Numbers indicate the percentages of dots moving coherently in the target area,
necessary for discrimination of motion direction. Thresholds of the left and right
hemifield are averaged.
the thresholds for the left and right sides (see Table 3).
AM was severely impaired in second order motion perception,
as he scored 4 standard deviations above the threshold of
the healthy controls. His first order motion perception was
somewhat better and at least one of the control subjects
performed of the same level (see Table 3).
Visual evoked potentials
Cortical potentials (VEP) were evoked by presenting a
contrast-reversing checkerboard pattern, either achromatic
white/black or else chromatic blue/red. In the control condition
the size of each square element was 1.5 cm. In the experimental
condition element size was 10 cm. AM was positioned in a
darkened room at a distance of 1 m from the screen of a 14-
inch VGA monitor. The elements of the checkerboard there-
fore comprised 0.868 in the control condition and 5.78 in the
experimental condition. An additional stimulus consisted of a
black/white checkerboard with very small elements. The first
set of VEPs (t1) started with these small elements (0.438),
continued with the control condition (0.868), and ended with
the experimental condition (5.78). In the achromatic series
Fig. 3. VEP for different, chromatic and achromatic spatial layouts in AM.
Left side: VEP for left eye stimulation. Right side: VEP for right eye
stimulation. Each stimulation was replicated two times producing two graphs
per panel. Note the reduced or almost absent VEPs for both eyes in the
achromatic condition with low spatial frequency patterns (lower panels).
212 H. Hildebrandt et al.

Table 4. Results of the VEP investigation

Black/white Blue/red

High spatial frequency Low spatial frequency High spatial frequency Low spatial frequency

Left eye Right eye Left eye Right eye Left eye Right eye Left eye Right eye

Control (n 17)s
Mean 9.35 8.76 6.61 6.12 6.78 6.48 6.25 6.78
SD 3.68 4.24 2.59 3.06 2.32 2.29 2.28 2.90
Maximum 18.32 20.50 13.23 13.20 10.70 11.50 9.68 9.68
Minimum 3.52 3.08 3.71 2.90 2.10 3.86 1.96 3.17
AM: t1 3.26 1.54 0.91 1.89
AM: t2 6.81 7.08 2.60 2.45 6.02 4.87 6.27 4.60
AM: t3 5.16 6.23 4.30 2.20

All figures: peak to peak mV differences of N1/P100.

AM evoked a normal VEP in the small-sized and in the control
conditions, but there was no clear P100 in the experimental
condition. Because of a possible order effect, we replicated
the experimental condition in reverse order after two days (see
Fig. 3 and t2 in Table 4) and again seven weeks later (t3 in
Table 4). At t1 there was no clear-cut N1/P100 complex. At t2 a
peak-to-peak amplitude (N1/P100) of 2.6 mV was measured
for the left eye and 2.45 mV for the right eye. At t3 the
corresponding amplitudes were 4.3 mV and 2.2 mV. With the
exception of the amplitude for the left eye at the last measure-
ment, all amplitudes were completely outside the range of an
age-matched control group (n 17, mean age of 47, minimum
age of 38, maximum age of 60 years). During stimulation by
chromatic reversals the size of the checkerboard elements did
not influence the results. In both conditions, AM showed a
pronounced P100, within the range of the control group (see
Table 4). Therefore, AMs reduction of the peak-to-peak am-
plitude was specific for the large sized black/white checker-
board.
The latencies of the P100 of AM were prolonged in each of
the black/white reversal conditions, but not in the color
reversals.
Discussion
Visual form agnosia has been documented in the past in
several case descriptions of patients suffering from hypoxia,
carbon monoxide poisoning, or bilateral posterior infarcts.
Our patient AM suffered from a sudden heart arrest. His
visual and memory impairment were therefore caused by
brain hypoxia although no morphological brain damage
could be traced on an MRI two months after the event.
The profile of his impairments was similar to that of the
case SMK of Davidoff and Warrington (1993), while he
was somewhat less impaired than the case of Milner et al.
(1991). The similarity of our case with SMK is striking.
Like SMK, AM was able to perceive subjective contours,
colors and individual geometrical objects, but he could not
give exact size estimations. In both cases the reading ability
was highly impaired and both had problems recognizing
objects in their environment, especially drawings and pic-
tures without color information. On the other hand, while
SMK was able to trace overlapping figures, AM was un-
able to do so. In this respect, AM resembles FGP more
(Kartsounis and Warrington, 1991), who also had problems
recognizing both overlapping and embedded figures. But
FGP was able to discriminate between different object
sizes, while AM had problems with this task. The visual
impairment of AM is an additional argument for the
existence of independent mechanisms for figure/ground
discrimination on one hand and the perception of the re-
lative size of surfaces on the other. FGP was able to give a
correct size estimation, while SMK and AM could not.
From a functional neuroanatomical viewpoint it can be
argued that V1 and V2 have to be intact in AM. In case of
a lesion of V1 one would expect visual field deficits and an
attenuation of the VEP potentials. This was not the case in
AM. V2 is concerned with higher order boundary percep-
tion. Peterhans and von der Heydt (1989, 1991) have shown
that neurons in this part of the visual system of monkeys
are activated by illusionary contours presented in their
receptive fields. Because AM was able to detect such
patterns and to discriminate between different Kanisza
figures, V2 presumably was not impaired by the hypoxia.
This assumption can also be illustrated by the fact of
AMs abilities to identify letters and short words if printed
in an appropriate font type. Zeki et al. (1990) and Barbur
et al. (1993) have shown that V4 is a major centre of
conscious color perception. Again, at least some parts of
V4 must have been spared from the hypoxia, because AM
was able to discriminate between colors easily and showed
a normal checkerboard VEP based on blue/red reversals.
The main functional problem for AM was to separate figure
from ground, to separate individual letters in reading, and to
estimate sizes. We would like to argue that his impaired
performance in the visual and spatial location tasks of the
VOSP can also be explained by the latter impairment, because
both tasks rely on locating stimuli relative to a background
 Differential impact of parvocellular and magnocellular pathways
figure (the squares). Kartsounis and Warrington (1991) and
Milner et al. (1991) assumed that the impairment of FGP and
DF might be due to a lesion of the magnocellular pathway.
Such a hypothesis would be too general for AM because he
(like DF) was able to walk, to grasp and to localize objects in
the environment. All these performances rely on precise
visual-spatial perception. Furthermore, AM was able to copy
reasonably well (although not in the normal range), which
also argues for an intact perception of spatial coordinates.
Motion perception is certainly one of the major functions of
the magnocellular system and recently Braddick et al. (2003)
proposed tests of global motion processing, i.e., the detection
of coherently moving dots, as an appropriate means to analyse
dorsal stream vulnerability. In their view the relative
threshold increase (how many dots have to move coherently
in a specific area for a predominant direction being seen)
may be a sensitive measurement of the integrity of the
magnocellular system.
We used two such tests one for first-order and one for
second-order motion perception to investigate AMs motion
perception abilities. In both tests AM scored more than two
standard deviations above i.e., worse than the control
group, and at least in the second-order motion task his
performance was completely outside the range of the control
group. This result points to an impairment of the magnocel-
lular system, especially in the processing of higher-order
motion signals, which presuppose a decoupling of local
directional selectivity at V1 from the direction of a globally
moving object.
On the other hand, one could argue that the deficit in motion
perception of AM is not necessarily the result of a defect of
magnocellular neurons but may also be a consequence of his
visual form agnosia. We did not notice any clinical impairment
in motion perception and, although AM could identify surfaces
of degraded stimuli such as the X in the screening test of the
VOSP, he was not perfect in this task. Detecting the predo-
minant motion direction of more or less coherently moving
dots involves a similar perceptual function as in the screening
test of the VOSP. Therefore, a subclinical impairment
obviously led to a significant deficit in the motion perception
task, with its more extended area to integrate and lower
coherence level which make detection much more difficult.
It has been argued that another function of the magnocel-
lular system is the selection of items within arrays of similar
items (Omtzigt et al., 2002). While the parvocellular system
identifies the feature in the focus of attention, the magnocel-
lular system identifies, according to this hypothesis, the global
structure in which this feature is embedded. Such a view of the
magnocellular system has been used to explain dyslexia as an
impairment of the magnocellular system (Livingstone et al.,
1991; Sperling et al., 2003; Stein, 2003). AM was not only
dyslexic, he also resembled dyslexic children in two other
respects, which are considered as a sign for a magnocellular
deficit (Stein, 2003). First, he had problems in higher-order
motion perception (see above) and second, he showed a left
minineglect (Hari et al., 2001) as he was about 120 ms
213
slower in responding to targets in the left hemifield compared
to the right hemifield.
If integration of different focal aspects into one global pic-
ture relies on magnocellular structures, the investigation of
the visual evoked potential is also of relevance. AMs
amplitudes in the black/white reversal VEPs for coarse
checker-boards were subnormal while normal for colored
checker- boards of all element-sizes. Hence, the deficit seems
to be specific for large receptive fields, low spatial frequency
stimulus, i.e., a coarse checkerboard, evokes a normal VEP
amplitude for a chromatic pattern while an impaired VEP for
an achromatic pattern. This dissociation could be explained
by the involvement, in the first case, of parvocellular struc-
tures, whereas this is not the case primarily in the achromatic
condition. The investigations of Tobimatsu et al. (1995) and
Arakawa et al. (1999) demonstrate that achromatic stimula-
tion of the parvocellular pathway can be dissociated from
chromatic stimulation using different spatial frequencies.
Higher achromatic spatial frequencies (more than 2 cycles
per degree) seem to stimulate primarily the parvocellular
interblob pathway, whereas low frequency spatial patterns
have a higher impact on magnocellular structures, as demon-
strated in research on reading and neglect (Spinelli et al.,
1996).
The comparatively better ability of AM to recognize draw-
ings of animals, compared to drawings of manmade objects,
may also be explained by such a difference between the
parvocellular and the magnocellular system, while, of course,
other explanations are possible. Animal pictures, as in the
Snodgrass and Vanderwart (1980) drawings, often possess
typical features, which allow identification irrespectively of
their global shape. Manmade objects like tables, chairs, etc. are
characterized as drawings by symmetrical and partly over-
lapping features. Recognition therefore presupposes a global
integration of local features. In summary we suggest that the
visual deficits of AM may be caused primarily by a defect of
the magnocellular system as a result of the anoxic episode.
Acknowledgement
We would like to thank Paul Eling (Nijmegen) and Manfred
Fahle (Bremen) for their comments on an earlier draft of this
article.
Supported by German Research Council, SFB 517.
References
Abadi RV, Kulikowski JJ, Meudell P. Visual performance in a case of visual
agnosia. In: Hoff MW, Hohn E editors. Functional recovery from brain
damage. Amsterdam: Elsevier, 1981; 27586.
Adler A. Disintegration and restoration of optic recognition in visual agnosia:
Analysis of a case. Archives of Neurology and Psychiatry 1944; 51: 24359.
Adler A. Course and outcome of visual agnosia. Journal of Nervous Mental
Diseases 1950; 111: 4151.
Arakawa K, Tobimatsu S, Tomoda H, Kira J, Kati M. The effect of spatial
frequency on chromatic and achromatic steady-state visual evoked
potentials. Clinical Neurophysiology 1999; 110: 195964.
Barbur JL, Watson JDG, Frackowiak RSJ, Zeki S. Conscious visual perception
without V1. Brain 1993; 116: 1293302.
214 H. Hildebrandt et al.
Bartolomeo P, Bachoud-Levi AC, De Gelder B, Denes G, Dalla Barba G,
Brugieres P, Degos JD. Multiple-domain dissociation between impaired
visual perception and preserved mental imagery in a patient with bilateral
extrastriate lesions. Neuropsychologia 1998; 36: 23949.
Benson DF, Greenberg JP. Visual form agnosia. Archives of Neurology 1969;
20: 829.
Biscaldi M, Fischer B, Aiple F. Saccadic eye movements of dyslexic and
normal reading children. Perception 1994; 23: 4564.
Biscaldi M, Gezeck St, Stuhr V. Poor saccadic control correlates with dyslexia.
Neuropsychologia 1998; 36: 1189202.
Braddick O, Atkinson J, Wattam-Bell J. Normal and anomalous develop-
ment of visual motion processing: Motion coherence and dorsal-stream
vulnerability. Neuropsychologia 2003; 41: 176984.
Braun D, Petersen D, Scho nle P, Fahle M. Deficits and recovery of first- and
second-order motion perception in patiens with unilateral cortical lesions.
European Journal of Neuroscience 1998; 10: 211728.
Campion J, Latto R. Apperceptive agnosia due to carbon monoxide poisoning.
An interpretation based on critical band masking from disseminated lesions.
Behavioral Brain Research 1985; 15: 22740.
Cavanagh P, Mather G. Motion: The long and the short of it. Spatial Vision
1989; 4: 10329.
Chubb C, Sperling G. Drift balanced stimuli: A general basis for studying non-
Fourier motgion perception. Journal of the Ophthamological Society of
America A. 1988; 5: 19862006.
Davidoff J, Warrington EK. A dissociation of shape discrimination and figure-
ground perception in a patient with normal acuity. Neuropsychologia 1993;
31: 8393.
Eden GF, Van Meter JW, Rumsey JM, Maisog JM, Woods RP, Zeffiro TA.
Abnormal processing of visual motion in dyslexia revealed by functional
brain imaging. Nature 1996; 382: 669.
Efron R. What is perception? Boston Studies of the Philosophy Sciences 1969;
4: 13773.
Facoetti A, Paganoni P, Turatto M, Marzola V, Mascetti GG. Visual-spatial
attention in developmental dyslexia. Cortex 2000; 36: 10923.
Ferreira CT, Ceccaldi M, Giusiano B, Poncet M. Separate visual pathways
for perception of actions and objects: Evidence from a case of apperceptive
agnosia. Journal of Neurology, Neurosurgery, and Psychiatry 1998; 65:
3825.
Fink GR, Halligan PW, Marshall JC, Frith ChD, Frackowiak RSJ. Neural
mechanisms involved in the processing of global and local aspects of
hierarchically organized visual stimuli. Brain 1997; 120: 177991.
Fink GR, Marshall JC, Halligan PW, Dolan RJ. Hemispheric asymmetries in
global/local processing are modulated by perceptual salience. Neuropsy-
chologia 1999; 37: 3140.
Finney DJ. Probit Analysis. Cambridge University Press, 1962.
Goodale MA, Milner AD. A. The Visual Brain in Action. Oxford: Oxford
University Press, 1995.
Hari R, Renvall H, Tanskanen T. Left minineglect in dyslexic adults. Brain
2001; 124: 137380.
Heinze HJ, Hinrichs H, Scholz M, Burchert W, Mangun GR. Neural
mechanisms of global and local processing: A combined PET and ERP
study. Journal of Cognitive Neuroscience 1998; 10: 48598.
Heywood CA, Cowey A, Newcombe F. On the role of parvocellar (P) and
magnocellular (M) pathways in cerebral achromatopsia. Brain 1994; 117:
24554.
Hildebrandt H, Spang K, Ebke M. Case report: Visuo-spatial hemi-inattention
following a cerebellar/brain stem bleeding. Neurocase 2002; 8: 32329.
Kartsounis LD, Warrington EK. Failure of object recognition due to a
breakdown of figure-ground discrimination in a patient with normal acuity.
Neuropsychologia 1991; 29: 96980.
Kosslyn SM, Flynn RA, Amsterdam JB, Wang G. Components of high-level
vision: A cognitive neuroscience analysis and accounts of neurological
syndromes. Cognition 1990; 34: 20377.
Kosslyn SM, Koenig O, Barrett A, Cave CB, Tang J, Gabrieli J. Evidence
for two types of spatial representations: Hemispheric specialization for
categorical and coordinate relations. Journal of Experimental Psychology:
Human Perception and Performance 1989; 15: 72335.
Livingstone MS, Hubel DH. Segregration of form, color, movement, and
depth: Anatomy, physiology, and perception. Science 1988; 240: 7409.
Livingstone MS, Rosen GD, Drislane FW, Galburda AM. Physiological and
anatomical evidence for a magnocellular defect in developmental dys-
lexia. Proceedings of the National Academy of Sciences (USA) 1991; 88:
79437.
Milner AD, Perrett DI, Johnston RS, Benson PJ, Jordan TR, Heeley DW,
Bettucci D, Mortara F, Mutani R, Terazzi E, Davidson DLW. Perception and
action in visual form agnosie. Brain 1991; 114: 40528.
Omtzigt D, Hendriks AW, Kolk HH. Evidence for magnocellular involvement in
the identification of flanked letters. Neuropsychologia 2002; 40(12): 188190.
Peterhans E, Heydt Rvd. Mechanisms of contour perception in monkey
visual cortex. II. Contours bridging gaps. Journal of Neuroscience 1989; 9:
174963.
Peterhans E, Heydt Rvd. Subjective contours bridging the gap between
psychophysics and physiology. Trends in Neurosciences 1991; 14: 1129.
Robertson LC, Lamb MR. Neuropsychological contributions to theories of
part/whole organization. Cognitive Psychology 1991; 23: 299330.
Smith AT, Hess RF, Baker CL. Direction identification thresholds for second
order motion in central and peripheral vision. Journal of the Ophthalmo-
logical Society of America A. 1994; 11: 50614.
Snodgrass JG, Vanderwart M. A standardized set of 260 pictures: Norms for
agreement, familiarity and visual complexity. Journal of Experimental
Psychology: Human Learning and Memory 1980; 6: 174215.
Sperling AJ, Lu Z, Manis FR, Seidenberg MS. Selective magnocellular deficits in
dyslexia: A phantom contour study. Neuropsychologia 2003; 41: 14229.
Spinelli D, Angelelli P, de Luca M, Burr DC. VEP in neglect patients have
longer latencies for luminance but not for chromatic patterns. Neuroreport
1996; 7: 8159.
Stein J. Visual motion sensitivity and reading. Neuropsychologia 2003; 41:
178593.
Taylor MM, Creelman CD. PEST: Efficient estimates on probability functions.
Journal of the Acoustical Society, Am. 1967; 41: 7827.
Tobimatsu S, Tomoda H, Kato M. Parvocellular and mognocellular contribu-
tions to visual evoked potentials: Stimulation with chromatic and achromatic
gratings. Journal of the Neurological Sciences 1995; 134: 7382.
Zeki S. A century of cerebral achromatopsia. Brain 1990; 113: 172177.