Information-limiting correlations

Computational strategies used by the brain depend strongly on
the amount of information that can be stored in population activity,
which in turn depends strongly on the pattern of noise correlations. In
vivo, noise correlations tend to be positive and proportional to the
similarity in tuning properties. Such correlations are thought to limit
information, which has led to the suggestion that decorrelation
increases information. In this talk, I will show analytically and
numerically that, in contrast, decorrelation does not imply an increase
in information. Instead, the only information-limiting correlations are
what we call differential correlations: correlations proportional to the
product of the derivatives of the tuning curves. Differential
correlations are likely to be very small, and buried under correlations
that do not limit information, making them particularly difficult to
detect. We show, however, that the impact of differential correlations
on information can be detected with relatively simple decoders.