Proefschrift_Holstein

Controlling dorsolateral striatal function via anterior frontal cortex stimulation

Introduction The ability to adapt flexibly to our constantly changing environment requires our actions to be goal-directed, and goals to be hierarchically organized (Koechlin and Summerfield, 2007). Accordingly, when defining goals at different levels, we can distinguish between motivational goals (e.g. a reward), cognitive goals (e.g. a task-set), and action goals (e.g. a stimulus-response mapping). Reward-predictive signals engage cognitive control processes that implement and update abstract cognitive goal representations, which in turn direct action selection. Thus flexible behavior depends on a hierarchy of top-down selection processes, and requires the transformation of information about reward into abstract cognitive decisions, which in turn need to be translated into specific actions. The brain region most commonly implicated in such flexible, goal-directed behavior is the frontal cortex (Miller and Cohen, 2001; Clark et al., 2004; Jimura et al., 2010). However, the cortex does not act in isolation and is connected with subcortical regions, such as the striatum, which is also involved in (reward-guided) cognitive- and motor control (Eslinger and Grattan, 1993; Groenewegen, 2003; Cools, 2011). Studies combining non-invasive brain stimulation (i.e. transcranial magnetic stimulation; TMS) with brain imaging have shown that stimulation of the frontal cortex can alter signaling in the striatum (Strafella et al., 2001; Strafella et al., 2003; Kanno et al., 2004; Ko et al., 2008; van Schouwenburg et al., 2012). More specifically, stimulating the human motor cortex altered signaling in the motor part of the striatum (i.e. the putamen) (Strafella et al., 2003), whereas stimulating the dorsolateral prefrontal cortex (dlPFC) altered signaling in the cognitive part of the striatum (i.e. the caudate nucleus) (Strafella et al., 2001; Ko et al., 2008). These observations are in line with evidence from anatomical work showing that the cortex and striatum are organized in parallel circuits, linking distinct parts of the cortex with specific regions of the striatum in a topographically and functionally specific way (Alexander et al., 1986). More recent anatomical work has challenged the idea that the corticostriatal circuits are strictly parallel (Haber et al., 2000; Haber et al., 2006; Draganski et al., 2008; Haber and Knutson, 2010) and several researchers have shown that there is a unidirectional information flow between corticostriatal circuits, i.e. from anterior/ventromedial to posterior/dorsolateral parts of the striatum and/or cortex (Haber et al., 2000; Haber, 2003; Koechlin and Summerfield, 2007; Haber and Knutson, 2010; Badre and Frank, 2012). Such a cascade of information processing might be well suited to subserve integration across the distinct functional domains associated with different frontostriatal circuits. Thus reward motivational processing, associated with a circuit connecting the anterior/ventral prefrontal cortex with the nucleus accumbens and anterior caudate nucleus, might influence cognitive control, by altering processing in a circuit connecting the dlPFC and medial caudate nucleus, to ultimately guide action selection, by altering processing in circuits connecting motor cortices (e.g. the premotor cortex; PMC) and the putamen ( figure 7.1 ) (Cromwell and Schultz, 2003; Draganski et al., 2008; Seger, 2008; Haber and Knutson, 2010).

139