One of the classic controversies in motor control is the extent to which discrete goal-directed movements unfold as planned or are modified based on response-produced feedback or changing environmental demands. An approach for examining the flexibility of movement planning and online control is to introduce an unexpected perturbation either during or some time after movement initiation. Of interest is how the perceptual–motor system is able to adapt to change in the task or environmental circumstance. In this chapter, we review research examining the effect of perturbations on the movement execution, movement goal, perceived goal, and visual context in which the action unfolds. Overall this research indicates that the visuomotor system is extremely flexible in adapting movement trajectories to rapidly changing task demands.
In general, the visual control of movement is studied through measurements of goal-directed behavior that describe the timing and spatial characteristics of a movement as it unfolds (e.g., Khan et al., 2006). Performance measures such as RT, MT, and end-point error describe overall performance or outcome. However, measuring and comparing the characteristics of the trajectory to the terminal outcome of the movement provide valuable additional information for researchers about the underlying visuomotor control processes responsible for the behavior observed after a perturbation (Elliott, Helsen, & Chua, 2001).
In the two-component model of goal-directed aiming (Woodworth, 1899), the kinematic events during the initial portions of the movement reflect movement planning toward the target, while later kinematic events reflect the processing of sensory feedback during the corrective process. The location of peak velocity has been associated with the separation point between the initial, planned portion of the movement and the latter, feedback-based portion of the movement (e.g., Elliott, Binsted, & Heath, 1999). Subsequently, decreased trial-to-trial spatial variability in the location of peak acceleration and peak velocity indicates increased planning efficiency (Khan & Franks, 2003), and decreased spatial variability at peak deceleration and movement end point indicates more efficient use of sensory feedback (Khan, Elliott, Coull, Chua, & Lyons, 2002).
Various other measures have been identified during the quest to discover the processes underlying visuomotor control. For example, Crossman and Goodeve (1963/1983) counted the number of discontinuities in the acceleration profile to quantify corrections to the initially executed movement (see also Elliott, Hansen, Mendoza, & Tremblay, 2004). Elliott and colleagues (1999) interpreted the average time elapsed between peak velocity and movement termination as an indication of the processes associated with vision gathered during movement execution. Khan, Franks, and Goodman (1998) employed an index of error correction effectiveness that quantified the ability of participants to correct for errors in the executed movement by normalizing the movement error before and after a correction. The variability in the locations of various kinematic markers throughout the movement trajectory over a series of trials has been used to infer the presence of corrective processes associated with vision (Darling & Cooke, 1987; Khan et al., 2002; Messier & Kalaska, 1999). More recently, several researchers have correlated the distance traveled at movement termination with the distance traveled at earlier kinematic markers such as peak velocity, or the primary submovement end point, to quantify the effectiveness of the corrective processes (Desmurget et al., 2005; Elliott et al., 1999; Heath, Westwood, & Binsted, 2004; Messier & Kalaska, 1999). However, the principal evidence for visuomotor processes occurring during a movement comes from trajectory corrections (Crossman & Goodeve, 1963/1983; see chapter 2 for a more detailed description of these various kinematic indexes). Perturbations induce changes in the kinematic measures, and theoretical inferences are made based on the direction and magnitude of these changes.
Although most models of the visual control of rapid action associate the initial portions of a movement with preplanning and the latter portions with online control, not all models concur with this view. Models explaining the control of rapid movements range from those positing complete preprogramming and execution without subsequent adjustment (e.g., Plamondon, 1995) to two-component models positing an initial movement to acquire the target area and subsequent adjustments to the trajectory near the target (Meyer, Abrams, Kornblum, Wright, & Smith, 1988; Woodworth, 1899). Finally, there are other models suggesting that sensory information gathering and concurrent adjustments to the movement trajectory occur continuously throughout the movement (Crossman & Goodeve, 1963/1983; Elliott, Carson, Goodman, & Chua, 1991; Elliot et al., 2001, 2004). In other words, models of the visual control of rapid movement range from completely open loop to entirely closed loop.
Discrete aiming movements are typically used to investigate visual control processes because they have an identifiable point of initiation and termination. Aiming movements are also relatively easy to observe and quantify with modern techniques. Researchers have employed multiple paradigms to investigate the processes underlying rapid and efficient action. These have included manipulations of practice (Elliott, Chua, Pollock, & Lyons, 1995; Proteau, 1995), vision of the limb (Elliott, Lyons, & Dyson, 1997; Pélisson, Prablanc, Goodale, & Jeannerod, 1986; Prablanc, Echallier, Komilis, & Jeannerod, 1979), vision of the target (Carlton, 1981b), and vision of both the limb and the target (Keele & Posner, 1968; Woodworth, 1899). In addition, researchers have changed the target location or size upon movement initiation (Elliott, Lyons, Chua, Goodman, & Carson, 1995; Heath, Hodges, Chua, & Elliott, 1998), manipulated the environmental context (Proteau & Masson, 1997), or altered the temporal and spatial association between the effector and the movement environment (Hansen, Elliott, & Tremblay, 2007; Redding & Wallace, 2001, 2002; Smith & Bowen, 1980). The important methodological questions have included how, when, and where to introduce manipulations within perturbation protocols investigating visual control processes.