Behavior of the saccadic system: Metrics of timing and accuracy

The behavior of saccades in response to a peripheral target is discussed. The saccade latency comprises sensory and motor processing delays of about 80ms, leaving on average more than 100ms for central processing. Many factors influence the latter. Yet, programming express saccades requires little to no central processing time. Typical saccades are hypometric by about 10%, which seems to be a deliberate strategy. A correction saccade requires only about 50ms of central processing. There is no strict dead zone for saccades, as they can be elicited by target jumps as small as 0.05deg. There seems to be no strict refractoriness in the system either, because saccade metrics can be continuously modified during the preparation interval by new target information. This suggests semi-independent processes for the "when" and "where" of saccades, which is incorporated into a neurophysiologically-inspired model. Saccades are not kept in retinotopic coordinates but are goal-directed by incorporating intervening changes in eye position. Although the updating mechanism is unclear, there is strong evidence that it involves the use of efference copy information (the outflow theory). Although the spatial percept of a target may be erroneous around saccades, the motor system seems to be more accurate. The chapter closes with a discussion on the potential function of microsaccades and slow drifts, when fixating a target.