Downloaded from rspb.royalsocietypublishing.org on July 20, 2010
6 A. E. Welchman et al. Reaction beats intention
Figure 4. The reactive advantage (expressed as a percentage)
for phase 1 and total execution times in experiment 3.
Data illustrate the between-subjects mean response with
error bars showing s.e.m. Data from the three different
movement conditions are shown: human opponent, virtual
opponent (in which the player faces a computer but is told
that this interfaces with another human) and computer
opponent (participants are told that they are playing with a
computer).
not suggest a pattern of results consistent with the
hypothesis that the anthropomorphic nature of the
opponent modulates the effect. In particular, if this
hypothesis underlies the effect, we would expect the reac-
tive advantage to be lowest for conditions in which the
player believed they were competing with a computer.
Rather, the mean reactive advantage is lowest when par-
ticipants believed they were competing against another
human interfaced through the computer.
Despite this null result, the findings from this experi-
ment are useful in addressing concerns that might arise
from the paradigm we have developed. In particular,
under conditions in which participants compete against
a computer, there is no visual motion and no auditory
cues from the opponent, nor any cues of movement prep-
aration. A previous report suggests that movements are
faster when participants are able to see motion (Smeets &
Brenner 1995), suggesting an alternative interpretation
for a benefit of viewing the opponent. However, this
cannot explain our findings as the reactive advantage per-
sists when there is no motion in the display (just three
simple illuminated squares). Moreover, auditory cues
from the human opponent hitting their buttons could
increase arousal, or provide a movement-timing signal.
However, such cues are not available when competing
against a computer, as there was no physical movement
or button pressing. Nevertheless, our findings were
unchanged. Finally, in the context of bimanual reaching,
it is known that participants slow down the easier of two
movements when reaching for two different targets, such
that both movements end simultaneously (Kelso et al.
1979). Similar effects of movement coordination have
been reported for social (two-person) movements
(Georgiou et al. 2007). It could be argued that inten-
tional movements are deliberately slowed to ensure
synchronous termination with an opponent’s reactive
movement. The competition between participants and
the considerable delay imposed by reaction make this
suggestion unlikely. Moreover, the sparse display in com-
puter-opponent conditions effectively rules out this
possibility.
4. GENERAL DISCUSSION
Here we consider the production of the same movement
sequence under conditions in which participants initiate
the movement or react to an opponent. We demonstrate
that reactive movements are associated with faster
execution times, and that this quickening of movement
does not appear to relate to having another human as a
model for one’s own action. We suggest different cortical
processing routes for the control reactive versus inten-
tional movements, and argue that faster movement
dynamics may constitute a basic property of reactive
movement production.
The suggestion of a distinction between reactive versus
intentional movements is consistent with a range of pre-
vious studies that report changes in the balance of the
involvement of a number of cortical and subcortical
areas during the production of different classes of action
(Laplane et al. 1977; Kurata & Tanji 1985; Romo &
Schultz 1987; Mushiake et al. 1991; Halsband et al.
1993; Cunnington et al. 1995, 2002; Deiber et al. 1999;
Jenkins et al. 2000; Maimon & Assad 2006; Sumner
et al. 2007). Previous behavioural work also supports
this distinction. For instance, countermanding the pro-
duction of an intended movement to react to an
external trigger can have a cost (Obhi & Haggard
2004), suggesting a delay imposed by switching between
different modes of movement triggering (Obhi et al.
2009b). Under our paradigm, participants could be pro-
voked to move sooner than they intended by seeing
their opponent’s actions. Based on Obhi and colleagues’
findings, the initiation of such reactive movements may
be slower than the initiation of internally generated move-
ments. (This suggestion is, of course, untestable as we
have no access to the timing of participant’s movement
triggering decisions.) Here, we assess a different aspect
of movement production, demonstrating that reactive
movements can be advantageous in producing faster
execution times (albeit with increased error rates).
It is interesting to speculate about the neural circuits
that might be involved in the production of the movements
we have studied. One candidate region of importance is the
pre-supplementary motor area (SMA) region of the medial
frontal cortex that is implicated in the control of intentional
actions. Moreover, it is thought to play a key role in
switching between different tasks (Rushworth et al. 2002)
and selecting an intentional action over a reactive one
(Isoda & Hikosaka 2007). It is possible that the pre-
SMA functions to remove the inhibition of potential
actions—a function carried out by the SMA (Sumner
et al. 2007)—thereby giving rise to the production of the
intentional movement sequence. The production of reac-
tive movements may involve an alternative route that
disinhibits the planned movement sequence via the parietal
cortex (Cunnington et al. 2006). Our observation that the
reactive advantage is focused on the initial, ballistic phase
of movement is suggestive of an effect limited to movement
onset rather than being general to the production of a
sequence of arm movements. This suggests different
types of disinhibition for reactive and intentional move-
ments. In particular, disinhibition designed to prevent
early movement (Sinclair & Hammond 2009; see Obhi
et al. 2009a for an excellent discussion) may be faster via
the parietal route, resulting in increased acceleration and
reduced movement execution times.
Proc. R. Soc. B