To investigate the neural basis of visual behavior one
needs to compare neurophysiological events in the visual
cortex with the visual psychophysical performance of an
animal. It may be possible to determine which neural structures
are responsible for a particular aspect of perception
if one can readily demonstrate a correlation between a
neuron's activity and an animal's responses.
Dr. William Newsome has done this in the most direct
way-he performed concurrent psychophysical and electrophysiological
experiments in awake behaving macaque monkeys. Measuring
single cell properties in some cortical structures can
provide insight into the perceptual function of that structure.
Newsome studied the properties of neural cells in the
medial-temporal (MT) cortical area of the macaque visual
cortex while the animal performed a psychophysical task.
The MT area is organized into columns of cells selective
for the direction of movement of a visual target.
To stimulate these cells, a random-dot moving display
was used in which the fraction of coherentiv moving dots
could be systematically varied. At 0% coherence all dots
moved randomly, while at 100% coherence all dots moved
with uniform speed in the same direction. Thus at 50%
coherence half of the dots in the display moved at the
same velocity while the other half moved randomly.
After a cluster of cells in an MT column were found and
their preferred direction and receptive field determined,
the cells' percentage of coherence threshold was determined
by electrophysiologically recording the responses to moving-dot
displays of different coherence levels. Concurrently,
the monkey performed a two-alternative forced choice psychophysical
task whereby he responded to the perceived movement of
the dots in the receptive field of the recorded cells
by making saccades in the appropriate preferred or null
direction. The monkey's psychophysical threshold and psychophysical
curve corresponded well with that of the recorded cluster
of MT neurons, implying strongly that the performance
of these MT neurons underlies the psychological performance
of the animal. In order to further test the idea that
these neurons' activity provides the neural basis for
the animal's motion perception, Newsome went on to study
how altering the cells' activity affected the performance
of the whole animal.
Running the same psychophysical task, an identified cluster
of directional cells in an MT column was electrically
stimulated while the animal viewed the moving- dot stimulus.
It was found that the animal's two-choice responses were
biased towards the preferred direction of the stimulated
cortical neurons.
This result undoubtedly demonstrated that the activity
of a small number of cells does determine the animal's
performance. However, it is harder to tell from this work
whether the stimulation of these neurons caused an alteration
in the perception of motion of the display or in the process
of decision making independent of what direction was perceived.
Besides direction selectivity, many cells in area MT
also code for retinal disparity, a measure of the location
of an object in depth. Using similar psychophysical paradigms
adapted to test for performance on retinal disparity instead
of direction of movement, Newsome and his colleagues found
as before that the neural responses were at least as good
as the performance of the monkey, and that stimulating
a particular disparity column biased the animal's psychophysical
performance accordingly. These results demonstrate that
the neural basis of other aspects of visual perception
is similar to motion direction perception. Thus, we may
generalize this understanding to other related neuro-
perceptual questions.
