177 / 218
175
General discussion
ADHDwho take methylphenidate on a daily basis and who refrained from taking their normal
medication prior to the experimental session. When comparing this to drug administration
studies in non-medicated subjects, the long term effect of methylphenidate should be taken
into account. For example, it has been suggested that long-term use of methylphenidate may
– among other things – decrease excessive reward-related signalling in the striatum (Seeman
and Madras, 1998; Robbins, 2002). It remains unclear from
chapter 4
to what extent the
effects in the ADHD patients after medication withdrawal reflect deficits related to ADHD, or
to long term effect of methylphenidate use. Another limitation of the work in
chapter 4
is that
it failed to reveal clear behavioural effects, i.e. it remains unclear how the change in striatal
signalling would affect behaviour.
The results in
chapter 5
speak to opposite effects of dopamine on cognitive focusing (task
repetition) and cognitive flexibility (task switching). However, cognitive flexibility and stability
were not formally manipulated in this study. Moreover given the age-related degeneration of
dopamine neurons, and the opposite effects of reward on task switching and repeat trials, it
is reasonable to speculate that dopamine mediated these opposite reward-related effects on
cognitive control. However, future work should test this hypothesis (
future research
).
In
chapter 6
I present a novel rodent paradigm allowing the independent manipulation of
reward motivation and cognitive control, enabling the assessment of the interaction between
reward and task switching. Although this paradigm parallels the paradigm used in human
subjects (
box 2.3
), at least two differences should be considered when comparing the results
between species. First, whereas the size of the reward varied on a trial-by-trial basis in the
human version, the rodents performed the task-switching paradigm in separate high and
low reward blocks. However, similar block-designs have been used in humans to reveal
motivation–cognition interactions (Locke and Braver, 2008; Kouneiher et al., 2009), although
not with the rewarded task-switching paradigm. It is therefore unlikely that these differences in
trial-by-trial fluctuations in reward expectancy will have dissociable effects on task-switching
behaviour. Second, the trial duration in the rodent paradigm is much larger than that in the
human paradigm: The animals were presented with the task cue for one minute and were then
given another minute to press the levers during the presentation of the target, whereas these
events lasted merely a second in the human paradigm. As a consequence, the time between
the task cue and the target was much shorter in the human version of the paradigm than it
was in the rodent paradigm (i.e. 400ms instead of 1 minute). Longer cue-target intervals in
humans are known to decrease the switch cost (Meiran et al., 2000). Moreover, situations in
real-life often require instantaneous decisions and the need for lengthy preparation times
can be detrimental for survival. In the future it is thus advisable to reduce the the cue-target
interval. The large cue-target interval in chapter 6 may potentially account for a switch benefit
in the high reward condition, and a non-significant switch cost in the low reward condition:
If increasing the cue-target interval will increase the difficulty of switching, animals may well
show the same pattern of results as that observed in the 9R subjects in the healthy population.
One additional question that is raised when comparing the results in rodents to those in