actually pressed it, and to choose it from a selection
of three that appeared after the button press.
Soon and colleagues found that by studying activ-
ity in the prefrontal and parietal cortex, they could
predict the outcome of a left-or-right decision up to
ten seconds before participants themselves became
aware of their choices. They conclude that ‘This delay
presumably reflects the operation of a network of
high-level control areas that begin to prepare an upcoming decision long before
it enters awareness’ (p. 543). The prediction rate, however, is relatively low (only
slightly above chance). This raises the question of whether these early predictive
cues are the precursors of intention or rather biasing signals that influence the
choice process without constituting it (see also Haynes, 2011). Interestingly, two
of the brain areas that carry early predictive information are the central nodes of
the default mode network (see
Chapter 8), which means that in
the context of decision-making
they might play an evaluative
role, providing background
information that biases deci-
sion-making (Brass et al., 2013,
p. 7) long before the moment
when we become aware of
making a choice.
But what is this moment? What
is the ‘time of awareness’, or the
time when a decision ‘enters
awareness’? The most radical
critique is given by Dennett,
who asks us to join him in the
following ‘all-too-natural vision’
of Libet’s wrist-flexing task
(Dennett, 1991, p. 165).
Unconscious intentions start
out somewhere deep in the
brain and then, gradually
becoming more definite and
powerful, make their way up to where ‘I’ am. At this point, they ‘enter conscious-
ness’ and ‘I’ have the experience of deciding to act. Meanwhile, representations of
spots on a clock face have been streaming up from the retina, gradually becom-
ing more definite in brightness and location, until they too reach consciousness
and ‘I’ can watch them parading past. So at the very moment when the intention
appears in consciousness, ‘I’ can say where the spot was.
As Dennett points out, this is so easy to visualise. Isn’t that how it has to be when
two things happen together in consciousness? No. Indeed, he says it cannot be.
There is no place or system in the brain where all the things currently ‘in con-
sciousness’ meet together, there is no time at which things ‘enter consciousness’,
and there is no self watching the display in that non-existent place. To try toKeypress Tone250 ms
Time on clock face
KeypressKeypress450 ms650 ms
–16 ms shiftTone–40 ms shiftTone–103 ms shift(a)(b)(c)JudgedJudgedJudgedActual ActualActual ActualActual Actualof intentional action requires an appropriate predictive
link between intentions and effects, rather than a retro-
spective inference that “I” caused the effect’ (Haggard
and Clark, 2003, p. 695).FIGURE 9.6 • Haggard and colleagues (2002)
report that the judged time of
a tone changes as a function of
the delay between the tone and
a previously executed voluntary
act. As the delay is lengthened
(a–c), the time misestimation
is reduced. Mean judged time is
represented by thought bubbles. In
the experiment, time judgements
are always retrospective, which is
why they can appear to precede
the actual times of occurrence
on the timelines (Eagleman and
Holcombe, 2002).