Automatic pilot: cognitive, attentional and neurological aspects of the online correction of manual aiming movements.
When the target of a reaching movement is displaced suddenly, people update their movement to take account of the jump, correcting their trajectory online to end the movement at the new target location. These corrections are initiated too rapidly to be conscious, and occur when they are uninstructed (Pisella et al., 2000) or the participant is unaware of the change in location (Goodale et al., 1986). These findings have been taken as evidence that fast corrections occur automatically, and the spatial updating of reach trajectories has become known as the ‘automatic pilot’ (Pisella et al., 2000). This thesis set out to investigate the cognitive, attentional and neurological aspects of the automatic pilot, in three series of related experiments, all employing a double-step reaching task. Experiments 1 - 4 investigated how strongly automatic reach corrections are, by manipulating the influence of conscious intention and cognitive load. These experiments confirmed that the automatic pilot is at most weakly automatic, as correction efficiency is enhanced by an explicit instruction to follow target jumps and, conversely, corrections can be overridden by an intention to resist them. However, voluntary inhibition of the automatic pilot can be disrupted by placing participants under heavy cognitive load, whilst voluntary enhancement is unaffected by this manipulation. Thus, voluntary suppression of the automatic pilot is effortful, but enhancement towards greater responsiveness is seemingly effortless. Experiments 5 - 8 explored the properties of the visual target displacement that drive the automatic pilot response in a double-step reaching task. These experiments demonstrate that correction efficiency is lawfully related to jump salience, but that the onset of the new target location drives correction responses more powerfully than the offset of the original target. However, the maximal correction rates obtained from a simultaneous onset and offset, were too great to be attributed simply to the additive influences of onsets and offsets. The onset and offset components of the target jump are thus synergistic. It is suggested that this reflects the contribution of an apparent motion signal induced by simultaneous onset and offsets, which strongly drives the automatic pilot system. Experiment 9 examined an asymmetry in correction efficiency, favouring rightward over leftward target jumps, evident throughout the earlier experiments. Correction efficiency was assessed for right- and left-handed participants responding to rightward and leftward target jumps. The pattern of results indicated that each hand is advantaged for responding to ipsilaterallydirected jumps, which may reflect biomechanical or hemispheric compatibility effects. However, there was also an overall differential advantage for rightward jumps, which was independent of handedness, or hand used. This suggests a left-hemispheric advantage for automatic correction behaviour, independent of handedness. Finally, Experiments 10 - 14 considered whether the automatic pilot deficit in optic ataxia is simply a manifestation of the more general misreaching deficit. Across several different target conditions, the pattern of online correction in optic ataxia refuted a simple misreaching explanation, suggesting that it is a specific functional consequence of dorsal stream damage.