Decision making, which pervades our daily lives, is ultimately a
computation carried out by the collective dynamics of millions of neurons
in the brain. Can we understand the neural circuit mechanism of decision
making, at the biophysical level, as we have so successively done in
studies of much simpler systems such as signal processing in the retina?
Recently, neurophysiologists have recorded from nerve cells in the brains
of monkeys performing simple decision tasks. These studies have begun to
reveal how neurons accumulate information in favor or against choice
options in a deliberate decision process, eventually leading to a
categorical response. In this talk, I will summarize experimental data and
present a biophysically-based recurrent neural network model of decision
making. I will show that this model accounts for a range of observations
from two sets of monkey experiments: one on perceptual decisions in the
visual system, the other on reward-based economic choice behavior. In
particular, I will discuss the highly stochastic nature of neural
activity, how it is generated and may be computationally harnessed. This
model suggests a circuit mechanism for decision making that can be
described theoretically in terms of stochastic attractor dynamical
systems.