Chaotic versus stochastic dynamics: a critical look at the evidence for nonlinear sequence dependent structure in dopamine neurons.

The firing pattern of midbrain dopamine neurons is thought to have important behavioral consequences. Although these neurons fire regularly in vitro when deprived of their afferent inputs, they usually fire irregularly in vivo. It is not known whether the irregularity is functionally important and whether it derives from the intrinsic properties of dopamine neurons or network interactions. It is also not known whether the irregular firing pattern is fundamentally stochastic or deterministic in nature. Distinguishing between the deterministic nonlinear structure associated with chaos and other sources of structure including correlated noise is an inherently nontrivial problem. Here we explain the geometric tools provided by the field of nonlinear dynamics and their application to the analysis of interspike interval (ISI) data from midbrain dopamine neurons. One study failed to find strong evidence of nonlinear determinism, but others have identified such a structure and correlated it with network interactions.