COGNITIVE NEUROSCIENCE Selective processing of buildings and faces during working memory: the role of the ventral striatum Alexa Haeger, 1,2 Hweeling Lee, 2 Juergen Fell 1 and Nikolai Axmacher 2,3 1 Department of Epileptology, University of Bonn, Sigmund-Freud-Str. 25, 53127 Bonn, Germany 2 German Center for Neurodegenerative Diseases (DZNE), Bonn, Germany 3 Department of Neuropsychology, Institute of Cognitive Neuroscience, Ruhr-University Bochum, Universitatsstrasse 150, 44801 Bochum, Germany Keywords: distraction, fMRI, interference resolution, striatum, working memory Abstract The ventral striatum seems to play an important role during working memory (WM) tasks when irrelevant information needs to be filtered out. However, the concrete neural mechanisms underlying this process are still unknown. In this study, we investigated these mechanisms in detail. Eighteen healthy human participants were presented with multiple items consisting of faces or build- ings. They either had to maintain two or four items from one category (low- and high-memory-load condition), or two from one category and suppress (filter out) two items from the other category (distraction condition). Striatal activity was increased in the distraction as compared with the high-load condition. Activity in category-specific regions in the inferior temporal cortex [fusiform face area (FFA) and parahippocampal place area (PPA)] was reduced when items from the other category needed to be selec- tively maintained. Furthermore, functional connectivity analysis showed significant reduction of striatalPPA correlations during selective maintenance of faces. However, striatalFFA connectivity was not reduced during maintenance of buildings vs. faces, possibly because face stimuli are more salient. Taken together, our results suggest that the ventral striatum supports selective WM maintenance by reduced gating of task-irrelevant activity via attenuating functional connectivity without increasing task-rele- vant activity correspondingly. Introduction Working memory (WM) depends both on regions exerting control functions and on areas supporting the perceptual representation of the information that needs to be maintained (Baddeley et al., 1974). During real-world situations, limited WM capacity requires one to assign transient storage space to relevant pieces of information while ltering out irrelevant ones. The basal ganglia, consisting of the stri- atum (caudate nucleus and putamen), globus pallidus, nucleus ac- cumbens and subthalamic nucleus, are implicated in such WM processes (Owen et al., 1998; Packard & Knowlton, 2002; Cools, 2005; Frank, 2005; Baier et al., 2010). The striatum in particular is hypothesized to act as a gateway for resolving interfering informa- tion by increasing the likelihood of processing relevant information whilst minimizing the inuence of irrelevant information (Cools et al., 2006; Dahlin et al., 2008; McNab & Klingberg, 2008; Frank & Fossella, 2011; Badre & Frank, 2012; Scimeca & Badre, 2012). This especially applies for novel items (e.g. Bunzeck & Duzel, 2006; Guitart-Masip et al., 2010). However, the exact mechanisms by which the striatum supports interference resolution during WM maintenance are still unknown. Here, we investigated the role of the ventral striatum in ltering of task-relevant and task-irrelevant complex representations of faces and buildings during visual WM. Participants performed a modied Sternberg WM task that manipulated target category (faces vs. buildings) and WM load (low, distraction, high) while being scanned. Faces predominantly induce activation in the fusiform gyri [fusiform face area (FFA); Kanwisher et al., 1997, 1999; Maguire et al., 2001], while buildings elicit selective responses in the para- hippocampal gyri [parahippocampal place area (PPA); Epstein & Kanwisher, 1998; Maguire et al., 2001]. Thus, we hypothesize that during selective maintenance of one of the two target categories (i.e. during the distraction condition), the striatum acts as a gateway to resolve interfering information by modulating activity in the FFA and PPA. These hypotheses extend previous work in two aspects. First, we investigated an issue that yielded controversial results in the existing literature, namely whether the striatum facilitates activity in task-rel- evant regions (Egner & Hirsch, 2005; Gruber et al., 2006), sup- presses activity in task-irrelevant regions (Maier et al., 2008), or both (Gazzaley et al., 2005; OReilly & Frank, 2006). We addressed this open question by using a novel design. In contrast to previous studies, we chose an experimental task that allowed us to compare maintenance in the presence of distraction with maintenance when only items from a single category were presented, which is impor- tant to isolate distraction-related effects. Second, we examined the Correspondence: Dr N. Axmacher, 3 Department of Neuropsychology, as above. E-mail: nikolai.axmacher@ukb.uni-bonn.de Received 8 August 2014, accepted 19 November 2014 © 2014 Federation of European Neuroscience Societies and John Wiley & Sons Ltd European Journal of Neuroscience, Vol. 41, pp. 505513, 2015 doi:10.1111/ejn.12808