Proefschrift_Holstein

Chapter 6

reward condition (reward x task switching x accuracy: F(1,11) = 7.651, p = 0.018, η2 = 0.410; figure 6.4 ). In addition, sham surgeries did not affect overall accuracy (surgery x accuracy: F(1,11) < 1; main effect of accuracy: F(1,11) = 4.997, p = 0.047, η2 = 0.312). Surgeries in the group with lesions of the AcbC on the other hand, did affect rewarded task- switching performance (surgery x reward x task switching x accuracy: F(1,11) = 6.782, p = 0.025, η2 = 0.381; figure 6.4 ), without affecting overall performance (surgery x accuracy: F(1,11) < 1; main effect of accuracy: F(1,11) = 13.189, p = 0.004, η2 = 0.545). More specifically, prior to surgery, the task-switching performance improved in the high compared with a low reward context (reward x task switching x accuracy: F(1,11) = 4.850, p = 0.05, η2 = 0.306), while this pattern was reversed after lesions of the core (reward x task switching x accuracy: F(1,11) = 6.133, p = 0.031, η2 = 0.358). The reversal of the effect was characterized by improved task-switching performance on low reward trials after surgery (F(1,11) = 6.788, p = 0.024, η2 = 0.382), in combination with a numerical impairment on task-switching performance during high reward trials (F(1,11) = 3.229, p = 0.1, η2 = 0.227). This effect was confirmed by directly comparing the groups, showing a difference between the sham and lesion group after surgery (group x reward x switch x accuracy: F(1,22) = 8.099, p = 0.009, η2 = 0.269), which was not present before surgery (group x reward x switch x accuracy: F(1,22) < 1; group x accuracy (F(1,22) < 1). This difference was due to a relative task-switch improvement in the lesion group in the low reward context (F(1,22) = 7.492, p = 0.012, η2 = 0.254), and a numerical impairment in this group in task-switching performance in the high reward context (F(1,22) = 1.188, p > 0.1), compared with the sham group. In summary, prior to surgery, task-switching performance was overall better in the high, compared with low reward context. Sham surgeries did not affect this beneficial effect of a high reward context on task-switching performance, whereas lesions of the AcbC did. Importantly, lesions of the AcbC did not affect overall accuracy on test, or performance during discrimination training. Discussion The current study assessed whether the core of the nucleus accumbens plays a crucial role in exerting motivational control over behaviour, thereby facilitating an agent to effectively pursue goals. Using a paradigm in rodents that parallels a well-established paradigm in human subjects (Aarts et al., 2010; van Holstein et al., 2011; Aarts et al., 2014a; Aarts et al., 2015), (and see Etzel et al., 2015; Fuentes-Claramonte et al., 2015) we showed that animals were able to use cues to prepare for an upcoming change in cognitive demands and that this task-switching ability was improved in a high reward context. This ability to improve cognitive control in a high-reward situation changed after excitotoxic lesions of the AcbC. More specifically, animals with an intact AcbC exhibited better cognitive control in a high relative to a low reward situation, whereas this effect was reversed after lesions of the AcbC. The reversal of this effect was characterized in particular by improved task-switching

130