173 / 218
171
General discussion
from the prefrontal cortex. Reward motivation may serve to bias the gating mechanism of
the striatum to a more flexible state, which is beneficial in some cases (e.g. in the case of task
switching), but not others (e.g. on a Stroop task or on repeat trials) (
future research
).
To date not many studies have directly manipulated the dopamine system during motivated
cognitive control. The work in this thesis fills this gap and combines experiments using
pharmacology, excitotoxic lesions, and non-invasive brain stimulation to show that (prefrontal
input to) the striatum is involved in motivated cognitive control and that integration across
corticostriatal circuits occurs when integration across goals is required. These results may
even extend beyond motivation-cognition integration to other goals. For example, a role for
a prefrontal-striatal gating mechanism and integration across corticostriatal circuits has been
suggested when subjects need to integrate information across hierarchical rules (Badre and
Frank, 2012).
What can we learn from cross-species evidence?
The ability to use information about upcoming rewards to adapt their cognitive control
strategy changed with increasing age in healthy human subjects and in rats with lesions of the
ventral striatum (
chapter 5 and 6
). When their striatum was intact, the animals in
chapter
6
showed a reward-related improvement of flexible cognitive control, as we previously
showed in a subset of healthy young human subjects (e.g. the 9R carriers in
chapter 3
). When
comparing the results obtained in work with human subjects with those from the rodent
work in
chapter 6
, a number of things are worth considering. First, it is important to keep in
mind to what extent findings from the rodent literature generally translate to human work.
Secondly, it is important to take note of the parallels and differences between the human and
rodent rewarded task-switching paradigm.
Cross-species translation: Evidence from task switching in humans and rodents
Flexible behaviour and its underlying neural substrates have been extensively studied in both
human and rodent subjects. One paradigm that is commonly used to assess flexible cognitive
control in rodents is the strategy set-shifting paradigm (Ragozzino et al., 1999). In a T-maze
(Ragozzino et al., 1999) or operant (Floresco et al., 2008a) version of this task, animals learn
to obtain a food reward by making a left (turn or lever press) response, thereby ignoring
the illumination of a light (over either the left or right arm or lever). After the successful
acquisition of this response strategy, the previously correct strategy (e.g. a going left) is
no longer rewarded, and the animal has to learn that the rule has changed. Crucially, the
previously irrelevant stimulus (the light) now becomes relevant and the previously relevant
rule (‘go left’) needs to be suppressed. The number of trials the animals needs to reach a pre-
defined criterion (e.g. 8 consecutive correct responses) is taken as a measure for cognitive
flexibility. A major advantage of this paradigm is its ability to distinguish between several