131 / 218
129
Reward modulation of cognitive function: the nucleus accumbens
Histology and pre-surgery differences between groups
Figure 6.3
shows the extent (left) and peaks (right) of the lesions. Although lesions that
extended into the shell were excluded from analyses, a small portion of the striatum dorsal to
the core was sometimes affected. Seven animals were excluded after histology due to either
the presence of unilateral lesions and/or extension of damage into the NAc shell or peaks
located outside of the AcbC, resulting in 12 animals with AcbC lesions and 12 shams.
Prior to surgery, the groupswerewellmatchedondiscrimination training and test performance.
More specifically, on the final day of training, we observed no group-by-accuracy effect on
auditory (F(1,22) = 2.182, p > 0.1) or visual (F(1,22) = 2.644, p > 0.1) discrimination. In
addition, prior to surgery, the lesion group did not differ from the sham group in terms of
rewarded task-switching performance (F(1,22) < 1;
figure 6.4
), or in terms of overall accuracy
on test (F(1,22) < 1).
Retraining
The lesions did not affect discrimination training; by the end of five days of retraining, the
groups did not differ in terms of either auditory or visual discrimination (group x accuracy:
F(1,22) = 2.594, p > 0.1 and F(1,22) < 1, respectively). Also, the surgery did not differentially
affect retraining across 5 days (time x group x accuracy: all F(1,19) < 1). After correcting for
multiple (i.e. 10) tests, animals were able to reliably discriminate between the auditory stimuli
on all 5 days (all F(1,23) > 27.883, all p < 0.001, all η2 > 0.548) and between the visual stimuli
from day 3 onwards (all F(1,23) > 10.731, all p < 0.004, all η2 > 0.318). Animals were not able
to reliably discriminate between the visual stimuli on the first two days (the main effect of
accuracy was not significant after multiple comparison correction: F(1,23) = 8.273, p < 0.009
and F(1,23) = 9.192, p < 0.006, respectively).
Lesions of the nucleus accumbens core affect rewarded task-switching performance
No test-retest effect on rewarded task-switching performance
All animals performed the post-surgery test twice in each reward context. We first confirmed
that the effect of reward on task-switching performance did not vary as function of session
(F(1,22) < 1). Also, there were no overall differences in terms of accuracy between sessions
(F(1,22) = 2.537, p > 0.1) and we did not observe any group differences in accuracy between
the sessions (F(1,22 = 2.152, p > 0.1). In all subsequent analyses, the data were collapsed
across these two sessions.
Lesions of the nucleus accumbens core alter rewarded task-switching performance
Sham surgeries (i.e. pre-surgery test vs. post-surgery tests) had no effect on rewarded task-
switching performance. Surgery did not significantly affect performance in the sham group
(surgery x reward x switching x accuracy: F(1,11) = 2.806, p > 0.1), thus across all sessions,
task-switching performance in the sham group improved in the high- compared with low