The primary visual cortex (V1) codes fundamental attributes of visual
stimuli, representing them in the coordinated and dynamic activity of
populations of neurons. Does this representation follow simple
mathematical rules? Are these rules stable or do they change according
to stimulus history, strength, or configuration? We addressed these
questions with experiments performed in anesthetized cats, where we
recorded from populations of V1 neurons using multielectrode arrays.
Stimuli were rapid sequences of gratings, drifting gratings, and plaids obtained by summing two such gratings. In response to sequences
of gratings, the cortex adopts a very simple coding scheme: responses
to subsequent stimuli simply sum to each other, thus forming a
representation that is easily decoded. This basic linear
representation, however, is only the scaffolding for more complex,
nonlinear operations that make the cortex extremely adaptive. First,
when stimuli are varied in strength, the cortical responses are scaled
in a nonlinear way, without disrupting the ratios of firing rate
across neurons that participate in the response. Second, in response
to sums of stimuli of different strength, the cortex no longer
performs simple summation but rather switches to winner-take-all
competition, as explained by a simple model based on divisive
normalization. Third, the cortex shows a marked ability to adapt to
the statistics of the stimuli: it changes the selectivity and
responsivity of neurons just as needed to counteract any biases in the
recent history of stimulation. It is likely that the rules that we
have uncovered are not specific to area V1 but are rather general
rules of operation of cortical populations. They may act as guide to
research in the underlying mechanisms and circuits and to research in
the fundamental neural computations that lead to perception and
behavior. Website: http://www.cortexlab.net/
|