Saccade detection during smooth tracking

Saccade detection in an eye-movement trace provides a starting point for analyses ranging from the investigation of low-level oculomotor mechanisms to high-level cognitive processes. When the eye tracks the motion of the object of current interest (smooth pursuit), of the visual background (OKN), or of the resultant visual motion from a head movement (tVOR, rVOR), the smooth tracking movement is generally intermixed with rapid-phase saccadic eye movements, which must be excised to analyze the smooth components of tracking behavior properly. We describe a simple method to detect saccades on a background trace of variable velocity, compare our saccade-detection algorithm with the performance of an expert human observer, and present an ideal-observer analysis to benchmark its detection performance.

► We describe a simple method to detect saccades on a background trace of variable eye velocity.
► We present an ideal-observer analysis to benchmark the detection performance of our algorithm.
► We compare the performance of the saccade-detection algorithm to that of an expert human observer.