Clinical Neuroscience and Neuropathology

neuroffpeport

Tue primate visual cortex is a mosaic of different areas
which are roughly organized into two major pathways,
a dorsal one along an occipito-parietal channel and a
ventral one along an occipito-temporal channel. It is
known that visual stimuli are processed differently within
these two systems and one might expect therefore that
behavioral deficits would differ according to the channel
damaged by cerebral lesions. Two patients with left
prestriate lesions presented perceptual deficits in the right
visual field. These deficits involved reading, recognition
of line drawings and colour perception and would be
compatible with a dysfunction of the ventral system.
Magnetic resonance imaging analysis confirmed that this
was the case.

Key words: Human visual cortex, Ventral and dorsal
system, Prestriate lesions, Form and colour perception
deficits, Neuropsychology

NeuroReport 2, 341 — 344 (1991)

Unmasking of visual
deficits following unilateral
prestriate lesions in man

Francois Michel,“ Marie-Anne Henaff
and Rémy Bruckert

U.280 INSERM, 151 Cours A. Thomas, 69003,
Lyon, France

©* Corresponding Author

Introduction

In monkeys physiological and anatomical studies
have uncovered a mosaic of as many as 20 separate
areas. On a larger scale the system can be thought
of as composed of two parallel channels of infor-
mation processing,'> the magno-cellular channel (the
‘motion’, ‘achromatic’ or ‘low-contrast’ channel) and
the parvo-cellular channel (the ‘form’, ‘colour’ channel).
The first channel projects to the parietal lobe and
connects areas in a dorsal system which processes
spatial properties, i.e. ‘where’ representations. The
second channel projects to the temporal lobe and
connects areas in a ventral system which processes
object properties, ic. ‘what’ representations. A
crucial station of the first channel is V5 while a crucial
station for the second channel is V4. Yet some
authors contest too strict a dichotomy between these
two major information-processing streams.**

The human visual system probably has a functional
organization similar to that of monkeys. However,
relevant anatomical data are lacking. The frontier
between area V1 and V2 is easily identified by the
stripe of Gennari.” Other borders of V2 are de-
lineated by cyto- and myelo-architectonic criteria and
possibly by a molecular marker." The boundaries
of other areas can be inferred from cyto- and myclo-
architectonic criteria combined with loci of callosal
connections.” The position of V4 remains approximate.
Area V5 has been tentatively localized on a flat map
as an heavily myelinated area in the parieto-occipital
sulcus.'°

PET studies confirm that the human visual cortex
is parcellated. Indeed blood flow changes show
activation of different discrete areas according to the
visual display attributes (shape, colour, movement,

‘Rapid Communications of Oxford Ltd

depth) attended to.'? Lueck et al.'"? have localized
a colour centre, homologous to V4, near the lingual
and fusiform gyri. V5 and MST seem to be located
on the dorso-lateral surface of the occipital lobe and
the inferior parietal lobule. Even specific loci activated
by passive word reading or visual memory tasks have
been localized in the prestriate cortex.'*'

One may expect that lesions would induce different
visual deficits according to the channel or the areas
damaged. Several cases of behavioral deficits re-
stricted to perception of colour (see Zeki’s compre-
hensive review!) or movement” have been reported.
In these cases lesions were mostly bilateral, in roughly
mirror locations, outside the striate areas. However,
such patients are so rare that it seems necessary to
resort to the more numerous patients with unilateral
lesions.” In such cases tachistoscopic visual tests
can be used in an effort to unmask deficits in one

hemifield.

Materials and Methods

We studied two patients with lesions damaging
ventral prestriate areas while sparing the striate areas
themselves.

Tachistoscopic hemifield presentation of stimuli
was monitored by a specific computer program.”
Patients sat 50-60 cm in front of a white CRT screen
driven by a Mac II computer. They were presented
with written material subtending from 2 to 4 degrees
of visual angle or with line drawings of objects
subtending 2 to 8 degrees. Before each presentation
patients were required to fixate a point appearing
for 1 s at the geometrical centre of the screen.

Vol 2 No6June 1991 341
neurofeport

F. Michel, M. - A. Henoff and R. Bruckert

Results

Case 1. Patient SK is a 55-year-old right-handed
male. He first had a cerebral infarct which damaged
the middle part of T4 and T5 in the right hemisphere.
A left upper quandrantanopia was then the only
noticeable deficit.

A year later SK had a second cerebral infarct in

letter was recognized. In the LVF + RVF pairs of
letters (n=160) were presented during various time
exposures between 100 and 250 ms, one letter in each
field, at 1° from the fovea. Right-hand side letters
were correctly read in only 30% of the pairs while
left-hand side letters were always correctly recog-
nized. Thus SK showed a clear extinction in the
right-hand side hemifield when tested using a right/

9g

FIG. 1. Magnetic resonance imaging of case 1. {a-e) Coronal cuts; (f) midline sagittal cut; (g) lateral sagittal cut. The lesion destroys the lateral part of the
lingual gyrus, the fusiform gyrus and the third occipitat gyrus, but spares the occipital pole and most of the calcarine area.

the left posterior temporo-occipital region (Fig. 1),
after which he presented a prosopagnosia and an
alexia. The alexia receded and SK showed no
cognitive disorder apart from a prosopagnosia.

On campimetric examination there was an incom-
plete left upper quadrantanopia and a normal RVF.
Tested with two sorts of grating patterns (IIIIIII vs III/III)
SK distinguished the string with the italic / in both
visual fields, but insisted that the pattern was fuzzier
in the RVF.

SK’s reading in natural conditions was slow but
not strikingly abnormal. However, when tested with
hemifield tachistoscopic exposure, SK showed an
obvious perceptual deficit in the RVF.

Letters and words. In the LVF nearly 100% of letters
(n=120) and 3-, 4- and 5-letter words (n=60) were
read easily with 100 or 250 ms time exposures. In
the RVF 92% of letters (n=130) were read at 1°
right to the fovea at various exposures, ranging from
50 to 250 ms. Almost no letters could be read
beyond 2°. No word could be read for exposures
between 100 and 300 ms, but most often the first

342 ~=Vol2 No 6 June 1991

left competition paradigm.

Line drawings. We presented object line drawings,
to the right or the left of the fovea, with a 250 ms
exposure time. SK correctly named all items in the
LVF and 17 out of 18 drawings in the RVF. This
excluded the hypothesis of a RVF scotoma.

With a bilateral presentation he recognized 24 out
of 25 items in the left hemifield, but only four out
of 25 in the right hemifield. Once again competition
induced an extinction phenomenon in the RVF.
Colour perception has not been tested.

Case 2. VM is a 25-year-old right-handed female
who had a surgical removal of an arterio-venous
angioma in the left occipital lobe. CT scans and
MRI showed the lesion located in the inferior aspect
of the left temporo-occipital junction (Fig. 2). Before
surgery she could read equally well in both visual
fields under bilateral tachistoscopic presentations.
Immediately after surgery she showed a word-finding
difficulty and a severe alexia without agraphia.
Goldman’s kinetic perimetry was normal while static
automatal perimetry proved a dense RVF hemianopia.
Visual deficits after prestriate lesions

neuroffpeport

FIG. 2. Magnetic resonance imaging of case 2. (a-d) Horizontal cuts. The lesion destroys the fusiform gyrus and the third occipital gyrus but spares the

lingual gyrus and the calcarine area.

Two months later VM had a normal neuropsychological
status and only complained of a slowness in reading.
The statokinetic dissociation had disappeared and
visual fields were normal. However, tachistoscopic
exposures revealed perceptual deficits in the RVF.
These deficits had partly resumed in the year following
surgery but remained severe. We report essentially
her latest, in fact her best, performances.
Contrast sensitivity. Three different shapes (square,
triangle and circle) were filled with gratings and
appeared grey on the white screen, subtending a 3°
visual angle. They were presented for 50 ms centred
at various positions between 0 and 4 degrees from
the fovea in the four quadrants. All were easily
recognized although VM noticed that shapes looked
different in the RVF. To lower the contrast the
same grey shapes were displayed on a grey background
square of a different texture, inducing a faint bright-
ness contrast. At 20 and 50 ms exposure times,
recognition remained perfect in the LVF but was
poor in the right inferior quadrant and null in the
right superior quadrant.

Words. VM was presented with a series of 40 8-letter
words, astride the fovea, for 200 ms. She read only
70% of the words, essentially because she could not
read the right-sided letters.

She was presented with 40 5-letter words in each
hemifield, at 1° lateral from the fixation point. At
100 ms VM read 75% of the words and 90% of
the letters in the LVF, 3% of the words and 30%
of the letters in the RVF. In the RVF the percentage
of recognized letters decreased along the successive
five positions within the word (respectively, 72, 45,
17, 12 and 15%),

VM was presented with 20 pairs of 4-, 5- and 6-
letter words, one word in each hemifield, during 200
ms. In the LVF she read 95% of the words. In
the RVF she read 70% of the 4-letter words, 10%
of the 5-letter words, and 0% of the 6-letter words.
With the same series presented for 100 ms she read
100% of the words in the LVF but 0% in the RVF.

She was also presented for 333 ms with three
randomly mixed series of 25 pairs of words, one
series on the horizontal meridian, one series 2° under
and one series 2° above. In the LVF she read 95%
of the words in the upper quadrant, 88% on the
horizontal meridian and 85% in the inferior quad-
rant. In the RVF she read, respectively, 18, 23 and
45% of the words, i.e. a lower performance in the
right superior quadrant.

Line drawings. In three different sessions VM was
presented with 75 pairs of line drawings for 200 ms.
She recognized 98% of pictures in the LVF and 62%
in the RVF.

Colours. In three different sessions, 60 pairs of
isoluminant coloured disks (blue, red, green or yellow)
were presented on a black screen, one disk in each
hemifield, for 100, 200 and 500 ms. With 100 and
200 ms exposures VM correctly named 95% of
colours in the LVF, 42% in the RVF. At 500 ms
all pairs of coloured disks were correctly named.

To eliminate a possible deficit of naming of RVF
stimuli, a same-different judgment task was used.
Twenty-seven pairs of isoluminant coloured disks (12
pairs of the same colour and 15 of different colours)
were presented for 200 ms. In the first session VM
judged all pairs as different. The same series of 27
pairs was presented for 200 ms for a three level
evaluation: same, slightly different, very different.
Responses were at chance level. These results reflect
a strong difference in colour perception in the two
hemifields that VM only became aware of during
these tachistoscopic testings. She stated that disks
appeared greyish on the right side.

A coloured video image was digitalized with two
different levels of saturation in order to obtain one
highly and one slightly coloured picture. A series
of 48 pairs of the two pictures, 12 of each kind
(highly-highly, highly-slightly, slightly-highly, and
slightly-slightly coloured) was randomly presented
for 200 ms, one picture in each hemifield, subtending
a 4° visual angle close to the fovea. The patient

Vol 2 No 6 June 1991 343
neurofpeport

F. Michel, M. - A. Henoff and R. Bruckert

was told to judge each picture of the pair as ‘bright’
or ‘pale’. All pictures presented in the left hemifield
were correctly judged, while all but four pictures
presented in the right hemifield were judged ‘pale’,
ie. 20 out of 24 ‘bright’ were reported ‘pale’.

Discussion

The first patient complained only of a slowness
in reading. The second patient considered that she
had completely recovered except that reading was
not as fast as prior to surgery.

Despite apparently normal visual fields both patients
showed perceptual deficits in their RVF when tachisto-
scopic presentations were used, They had no behavioral
difficulties which could be attributed to a dysfunction
of the dorsal system (perception of movement, pointing
or reaching, use of peripersonal space, spatial attention,
visual search, etc.). On the other hand, they showed
severe deficits which could be attributed to a dysfunction
of the ventral system (reading, colour and form
perception).

Radiology confirmed these clinical hypotheses. In
both cases MRI showed that the lesions involved
only prestriate areas. We cannot precisely identify
which particular visual areas were damaged but we
can state that in both patients the ventral system,
but not the dorsal system, was damaged.

In Case 1 the capacity of visual recognition in the
RVF was reduced. The lesion apparently involved
the ventral system extensively. The second patient
was less impaired and presented essentially difficul-
ties in reading and perceiving colours in the RVF.
Area V5 was probably spared and this may explain
the carly campimetric stato-kinetic dissociation. The
lesion probably involved V4 and V3 and temporal
areas in the close vicinity. This patient had her worst
performances in the.superior quadrant, which sug-
gests a lesion of areas topographically organized in
quadrants since a possible lesion of the inferior part
of the optic radiation is not consistent with the
normal campimetric examination.

In both patients the reading deficit was prominent.
Acuity and contrast sensitivity were sufficient to

344 Vol 2 No6 June 1991

allow recognition of line drawings. Why were letters
and words not recognized? Some procedures specific
to reading (parallel processing, spatial attention, word
form building, etc) were probably hampered.

In all the tests bilateral mirror presentations of
stimuli were very disabling. This may suggest that
during unilateral presentations the intact hemisphere
could palliate the failure of the damaged hemisphere
via callosal pathways.

Conclusion

Deficits consequent upon unilateral prestriate le-
sions may pass unnoticed. However, our results
show that it is possible to unmask them by com-
paring perceptual performances in each hemifield.
Although some authors have questioned the validity
of too strict a dichotomy, the present results seem
to validate the concept of dual pathways since the
specific deficits observed in these two patients reflect
damage to the ventral system.

Both patients exhibited a detectable recovery. This
suggests that the profuse interconnections within the
visual system, particularly the callosal ones, may be
an efficient anatomical substratum for a functional
recovery.

References

1 Ungerleger 1G, Misnkin M- Anaiysis of Visual Behavior Cambridge MIT Press.
1982

2 Livingstone MS, Hubet DH J Neuroscr 7. 3416-3468 1987

3 Morel A, Buller J VisuatNeurosc: 4. 555 578 11990)

4 7esS, Shipp S Nature 335, 311-317 11988)

5 Ettinger G Cortes 26, 319 341 11990)

6 Schiller PH, Logothetis NK Trends Neurosct ¥3, 392. 398 (19901

7 Stensaas SS, Eddington OK, Dobelle WH J Neurosurg 40. 747-755 (1974)

8 Hockheld S, Toote: RB, ZarembaS Proc Nat) Acad Sc: USA 87. 3027-3031 11990)
9 Clarke S, Milossy J J Comp Neurol 298, 188-274 11990)

10 Sereno MI, McDonaid CT, Aliman JM Sov Neurosc: Abstr 451, 3 (1988)

11 Corbetta M, Miezin FM, Oobmeyer, ef af Science 248 1566-1559 (1990)

12. Gulyas B, Roland PE, Stone f lander S, et al Soc Neurosc: Abstr 18. 621 (1990)
13° Lueck CJ, Zeki S, Friston KJ, ef af Nature 30, 386-389 (1990)

14 Petersen SE, Fox PT, Posse MI. ef at J Cog Neuroser 1, 153-170 11989}

15 Roland PE. Gulyas B. Seitz RU, ef al NeuroReport 1, 53 96 (1990)

16 Zeki S Bran 1913, 1721-1777 11990}

17 2inl J, Von Cramon D, Mar N. Bram 106, 313 340 (1983)

18 Albert ML, Reches A, Siverierg RJ Neuro! Newusurg Psych 38, 546.549 (1975)
19 Castro-Caldas A, Sa'gado V Arch Neuro! 41, 84-87 (1984)
20 Bub D, Gum T Phychiab Manuai Vol 85, Mortreal (19901

Received 22 March 1991;
accepted 24 April 1991