Computational models of decision making: integration, stability, and noise

► In many settings, evidence must be accumulated over time before sound decisions can be made. ► This computation can often be implemented by simply integrating neural activity over time, and new experimental data provides increasing evidence that this computation occurs biologically. ► Precisely defined modifications to the computation of integration are needed in unpredictable and variable decision making environments. ► Neurally based circuit models are providing new insights into how these computations can be implemented.