159 / 218
157
Controlling dorsolateral striatal function via anterior frontal cortex stimulation
congruous with a role for dopamine in mediating the interaction between reward and task
switching, which was evidenced by showing that the interaction between reward and task
switching depends on inter-individual differences in dopamine signaling (Aarts et al., 2010;
Aarts et al., 2011; Aarts et al., 2014a). In line with this, it is well established that TMS over the
cortex can alter dopamine release in the striatum (Strafella et al., 2001; Strafella et al., 2003; Ko
et al., 2008). TMS in the current design may therefore have inhibited dopamine release in the
striatum, eliciting indirectly a decrease in the striatal BOLD response (Knutson and Gibbs,
2007). In addition, the SNS account is congruous with the evidence from tracing work in
non-human primates, which has shown that the striatal-midbrain projections that originate
in the ventral striatum cover a wide range of dopamine neurons in the midbrain, including
neurons of the midbrain that project to the caudate nucleus. Also, cortical regions in the
monkey, involved in reward processing (areas 32, 25, 24) and cognitive control (area 46, 9)
do not directly innervate the posterior parts of the putamen (Haber 2000). Combined, these
results suggest that, at least anatomically, it is unlikely that direct corticostriatal connections
mediated the putaminal effects observed after aPFC stimulation, but rather that striatal-
nigral-striatal connections, mediated these effects.
A number of limitations require mentioning. First, aPFC stimulation modulated a region in
the caudate nucleus more dorsal than the region in the ventral striatum that was activated
in anticipation of a reward across all (TMS and baseline) sessions. One may expect that
stimulation of the cortical region involved in reward processing would modulate reward-
related signaling in the ventral striatum, rather than in the caudate nucleus. In fact, if we had
stimulated the orbitofrontal or ventromedial prefrontal cortex (OFC/vmPFC), this is indeed
what we would expect. However, given the corticostriatal connectivity pattern (Draganski et
al., 2008; Choi et al., 2012), the more dorsal aPFC is likely to modulate a region more dorsal
than the ventral striatum. Unfortunately, it is difficult to target the vmPFC/OFC with TMS,
especially due to the sensation of prefrontal TMS, which is increasingly uncomfortable when
moving to more ventral and anterior parts of the prefrontal cortex. In addition, in our previous
(Aarts et al., 2010) and current work, we observed activity in the aPFC when assessing the
main effect of reward, which is in line with others suggesting a role for the aPFC and the
caudate nucleus in reward-related processes (Kawagoe et al., 1998; Pochon et al., 2002; Locke
and Braver, 2008). Second, effects of aPFC stimulation may be experienced as less pleasant
than stimulation of e.g. the motor cortex and one may thus argue that the effects observed
in the current study may be due to the sensation of aPFC stimulation. However, given the
current pattern of results, such an explanation of the current results is highly unlikely. We
observed clearly distinct effects of aPFC stimulation as a function of task-related processing.
The effects of Reward and those of the interaction between Reward, Task switching and
Response switching were assessed in the same aPFC session. Any effects of the sensation of
TMS would have resulted in similar neural effects, irrespective of the condition, and such an
explanation cannot account for distinct effects of aPFC stimulation. Third, the current study
was designed to modulate processing in the striatum. We indeed show that stimulation of