“…These quick pulselike movements are not captured by the more slowly changing input reconstruction model or the variable time delay model, and they may need to be modeled as a second, possibly open-loop, mode of tracking. Such an episodic model would be more complex than those presented in this article, in that both the structure of multiple motion generators and the rules for starting and stopping them must be specified (e.g., see Jagacinski, Plamondon, & Miller, 1987). Repperger and Vickmanis (1979) have previously suggested that subjects may switch to a different tracking mode when the target velocity and acceleration become large.…”