157 / 218
155
Controlling dorsolateral striatal function via anterior frontal cortex stimulation
Integration across goals
Across sessions, participants exhibited a significant effect of reward on task switching in terms
of response times (Reward x Task switching: F(1,26) = 56.089, p < 0.001), but not in terms of
error rates (Reward x Task switching: F(1,26) < 1). Breaking down this effect in the response
times revealed that participants exhibited a switch benefit (i.e. repeat – switch performance)
on low reward trials (F(1,26) = 12.446, p = 0.002) and a switch cost on high reward trials
(F(1,26) = 21.376, p < 0.001). We also observed a Task switch x Response switch interaction
in terms of error rates (F(1,26) = 7.224, p = 0.012), but not in terms of response times (F(1,26)
1.130, p = 0.298). Breaking down this effect in the error rates revealed a larger task-switch cost
(task switch – task repeat) on response repeat trials (F(1,26) = 31.019, p < 0.001) compared
with response switch trials (F(1,26) = 4.224, p = 0.05). There was no Reward x Task switch x
Response switch interaction (F(1,26) < 1 for response times and error rates).
No effects of TMS on behavior
None of the main effects (i.e. of Reward anticipation, Task switching Response
switching) were modulated by TMS (i.e. TMS stimulation vs. baseline) for
any of the TMS sites (response times and error rates all F(1,26) < 3.612,
all p < 0.05).
We did not observe any significant main effects of stimulation irrespective of task conditions
(in response times and error rates: F(1,26 < 1) and no interactions effects of stimulation on the
Reward x Task switching, Task switching x Response switching or Reward x Task switching x
Response switching effects for any of the stimulation sites (in response times and error rates:
all F(1,26) < 2.316, all p > 0.1).
Discussion
In the current study, we aimed to provide evidence for a functionally cascading architecture of
corticostriatal circuits by assessing the consequence of manipulating distinct frontal regions
during the processing of reward, cognitive (task switch) and action (response switch) goals
while measuring signals in distinct striatal subregions. In support of our hypothesis, we
show that task-specific information can be integrated across corticostriatal circuits. More
specifically, manipulation of the prefrontal region involved in reward processing decreased
reward-related processing in an anterior part of the striatum (in the caudate nucleus: x = 6, y
= 16, z = 2), whereas stimulation of this same region decreased processing in a more posterior,
dorsal and lateral region of the striatum (in the putamen: x = -28, y = -8, z = 8) when assessing
the interaction between reward, task switching and response switching.
The anterior/ventromedial to posterior/dorsolateral task-related gradient we observed
during reward processing, task switching and response switching fits well with evidence
from anatomical studies, suggesting exactly this gradient in the striatum, forming parallel