In this paper, we study a simple discrete-time neural oscillator model that
, in certain parameter regimes, exhibits periodic or chaotic dynamics. The
present model with intrinsically chaotic dynamics is capable of spatiotempo
ral information processing: in response to constant external stimulation, t
he oscillator can switch into different chaotic states restricted to distin
ct parts of the phase space. Of particular interest is the processing of ti
me-dependent input in a master-slave configuration of two coupled oscillato
rs. Here, the response of an oscillator is studied by driving it with the s
ignal of the other. Following the input, the response system adapts to the
state of the drive. For a chaotic drive, we can observe generalized synchro
nization. The onset of adaptation to the drive state by the response is acc
ompanied by on-off intermittency resulting in irregular bursting behavior.