Dorsomedial striatum involvement in regulating conflict between current and presumed outcomes Anna Mestres-Missé a,b, ⁎, Pierre-Louis Bazin b , Robert Trampel b , Robert Turner b , Sonja A. Kotz a,b a School of Psychological Sciences, The University of Manchester, Manchester, United Kingdom b Max Planck Institute for Human Cognitive and Brain Sciences, Leipzig, Germany abstract article info Article history: Accepted 3 May 2014 Available online 11 May 2014 Keywords: Cognitive control Conflict 7 T-fMRI Ambiguity Striatum The balance between automatic and controlled processing is essential to human flexible but optimal behavior. On the one hand, the automation of habitual behavior and processing is indispensable, and, on the other hand, stra- tegic processing is needed in light of unexpected, conflicting, or new situations. Using ultra-high-field high- resolution functional magnetic resonance imaging (7 T-fMRI), the present study examined the role of subcortical structures in mediating this balance. Participants were asked to judge the congruency of sentences containing a semantically ambiguous or unambiguous word. Ambiguous sentences had three possible resolutions: dominant meaning, subordinate meaning, and incongruent. The dominant interpretation represents the most habitual re- sponse, whereas both the subordinate and incongruent options clash with this automatic response, and, hence, require cognitive control. Moreover, the subordinate resolution entails a less expected but correct outcome, while the incongruent condition is simply wrong. The current results reveal the involvement of the anterior dorsomedial striatum in modulating and resolving conflict between actual and expected outcomes, and highlight the importance of cortical and subcortical cooperation in this process. © 2014 Elsevier Inc. All rights reserved. Introduction It is well-established that the prefrontal cortex supports executive functions, and that it has a compelling function in cognitive control (e.g., Badre, 2008; Koechlin et al., 2003; Miller and Cohen, 2001). Furthermore, topographically and functionally organized projections from different cortical regions to the striatum are well described and established both in human and nonhuman primates (Draganski et al., 2008; Haber, 2003; Haber et al., 2006; Kemp and Powell, 1970; Middleton and Strick, 2000; Parent and Hazrati, 1995; Selemon and Goldman-Rakic, 1985; Yeterian and Pandya, 1991; Yeterian and Van Hoesen, 1978). Cortical and subcortical regions interact with each other through these projections, which give rise to many parallel cortico-striatal-thalamo-cortical loops (Haber, 2003). Hence, owing to these extensive inputs from almost every cortical region to the striatum, the basal ganglia are considered to have a modulatory function, which complements that from the cortical regions it receives projections from, particularly by modulating, selecting, gating, and controlling the information flow (Bar-Gad et al., 2003; Frank et al., 2001; Houk and Wise, 1995). Consequently, one could hypothesize that subcortical regions, which receive inputs and form processing loops with lateral prefrontal cortex are involved in implementing cognitive control mechanisms, that is, aid cortex in light of a mismatch between what is expected and actual incoming information. The principal aim of this research is to directly test the involvement of the basal ganglia in modulating this aforementioned form of mis- match and its resolution. We advocate a general mechanism based on probabilistic inference and probability distributions within a Bayesian framework. In short, given the evidence (in our particular case prior knowledge of relative frequency) a probability for each outcome (i.e. in- terpretation) is computed and the different probabilities of occurrence are ranked. Cognitive control mechanisms are required when upcoming information clashes with a so far favored (high-ranking) interpretation. In a similar vein, more than a decade ago, Jurafsky (1996) formalized a probabilistic model of sentence processing; and crucially, Pouget et al. (2013) recently proposed that a probabilistic mechanism is at the core of neural computation, and this general probabilistic approach charac- terizes all levels of sensory and cognitive processing. The fundamental working hypothesis is that the basal ganglia play a critical role when stimulus incompatibility with probabilistic expectations creates a conflict, which in turn, requires the engagement of cognitive control mechanisms to: inhibit a prevalent response, implement retrospective reevaluation in search of the origin of conflict and a solution, and if NeuroImage 98 (2014) 159–167 ⁎ Corresponding author at: School of Psychological Sciences, The University of Manchester, Zochonis Building, Brunswick Street, M13 9PL Manchester, United Kingdom. E-mail address: anna.mestres-misse@manchester.ac.uk (A. Mestres-Missé). http://dx.doi.org/10.1016/j.neuroimage.2014.05.002 1053-8119/© 2014 Elsevier Inc. All rights reserved. Contents lists available at ScienceDirect NeuroImage journal homepage: www.elsevier.com/locate/ynimg