Exp Brain Res (1992) 88:292-302
Experimental
BrainResearch
© Springer-Verlag 1992
Effects of inferotemporal cortex lesions
on form-from-motion discrimination in monkeys
Kenneth H. Britten 1, William T. Newsome I, and Richard C. Saunders 2
l Department of Neurobiotogy, Stanford University School of Medicine, Stanford, CA 94305, USA
2 Clinical Brain Disorders Branch, NIMH Neuroscienccs Center, St. Elizabeth's Hospital, Washington, DC 20032, USA
Received June 6, 1991 / Accepted September 23, 1991
Summary. The inferotemporal cortex of primates plays
a prominent role in the learning and retention of visual
form discriminations. In this experiment we investigated
the role of inferotemporal (IT) cortex in the discrimina-
tion of two-dimensional forms defined by motion cues.
Six monkeys were trained to a criterion level of perfor-
mance on two form-from-motion problems. Three of
these animals received complete bilateral lesions of IT
cortex, while the other three served as unoperated con-
trols. All animals were then retrained to criterion to
evaluate the effects of IT lesions on the retention of
form-from-motion learning. Compared with the control
group, the lesion group was significantly impaired on
both problems. Following retention testing, we trained
both groups of monkeys on two new form-from-motion
problems to investigate the effects of IT lesions on ac-
quisition rates for new learning. The lesion group per-
formed well on the new problems; the learning rates of
the operated and control groups were not significantly
different. When forms were defined by luminance cues,
monkeys with IT lesions, like those in previous studies,
were impaired both for retention and for acquisition.
These findings indicate that the anterograde effects of IT
lesions on learning new form discriminations are less
severe for forms defined by motion cues than for forms
defined by luminance cues. However, the retrograde ef-
fects of IT lesions on retention are severe for forms
defined by either cue.
Key words: Extrastriate visual cortex - Inferotemporal
cortex - Motion processing - Cortical pathways - Lesion
effects - Structure-from-motion
Introduction
Behavioral and anatomical evidence indicates that extra-
striate visual cortex of primates consists of at least 2
major pathways with distinct functional roles (Unger-
Offprint requests to: K.H. Britten
leider and Mishkin 1982). A ventral pathway connecting
the striate and posterior extrastriate visual areas with
inferior temporal (IT) cortex is involved in the identifica-
tion of patterns and is often referred to as the "form
pathway". The second pathway" courses more dorsally
and connects striate and extrastriate areas with inferior
parietal cortex. This pathway, which is often referred to
as the "spatial pathway", contributes to the visual loc-
alization of objects and is involved in the control of
visual attention as well. Removal of IT cortex in mon-
keys results in striking deficits on form and object recog-
nition tasks, while the ability to localize objects in space
appears unaffected. The opposite pattern of effects oc-
curs after removal of the posterior parietal cortex (Mish-
kin et al. 1982; Mishkin and Ungerleider 1982; Pohl
1973), although the spatial localization deficits that fol-
low parietal lesions may result in part from an attentional
impairment (Lawler and Cowey 1987). These results are
consistent with clinical observations of human patients
with restricted cortical damage. Temporal lobe lesions
can cause various agnosias involving identification or
recognition of visual stimuli (Damasio et al. 1982; Kimu-
ra 1963; Meadows 1974), while parietal lesions often
result in mislocalization of visual stimuli (Cole et al.
1962; Holmes 1918; Ratcliff and Davies-Jones 1972) as
well as impaired spatial attention (Battersby et al. I956;
Brain 1941 ; Heilman et al. 1970; Heilman and Valenstein
1979).
Motion is a fundamental visual dimension that is used
both for the perception of form and for the analysis of
spatial relations between objects. Psychophysical studies
have provided ample evidence that relative motion cues
are important for segregating and identifying two-dimen-
sional forms and for identifying the shape of three-
dimensional surfaces (Nakayama and Loomis 1974;
Wallach and O'Connell 1953). In addition, motion in-
formation is of critical importance for establishing an
accurate representation of the spatial relationships be-
tween moving objects and for orienting visual attention.
Thus one might expect motion information to be promi-
nently represented at the single neuron level in both the