Proefschrift_Holstein

Reward modulation of cognitive function: the nucleus accumbens

served to form an association between the task cue (noise and house light) and the relevant task (auditory or visual discrimination), although these associations did not become relevant until test. Trials were separated by inter-trial-intervals of 30s – 90s, during which the levers were retracted. For each ‘correct’ lever press a reward became available on average every 15 seconds (RI15 schedule), with the restriction that a reward was never delivered during the first 10s of a trial. All reward contexts, stimuli, and task cues were counterbalanced across animals. Table 6.1 shows the design and conditions for all animals. Cued task-switching paradigm with reward manipulation On the day immediately following the last training day, animals were tested. During test ( table 6.2, figure 6.1 ), the task cue was presented for 60s, followed by a 60s presentation of a response-incongruent compound stimulus (i.e. A1-V2 or A2-V1, for which one stimulus signals left lever presses will be rewarded, but the other stimulus signals right lever presses will be rewarded). Animals had to disambiguate these compound stimuli (i.e. whether to respond to the auditory or visual modality) by taking into account the task cue. For example, a presentation of A1-V2 preceded by the auditory task cue indicated that the animal should attend to the stimulus of the auditory modality, A1, and choose the associated lever press (left lever). The same A1-V2 compound preceded by the visual task cue, however, indicated that the animal should attend to the visual modality, and press the right lever. Again, ‘correct’ lever presses were rewarded according to an RI15 schedule, and never during the first 10s of each compound presentation. The design of this task afforded a unique opportunity to examine the animals’ performance on a trial-by-trial basis. Crucially, the task cue (i.e. AUD or VIS) could either remain the same (i.e. repeat: AUD -> AUD or VIS -> VIS) or change (i.e. switch: AUD -> VIS or VIS -> AUD) unexpectedly from trial to trial, allowing the assessment of task-switch performance (i.e. performance on task-switch versus task-repeat trials) ( figure 6.1 ). One test session consisted of 17 trials; one initiation trial (this was discarded, because is it not a repeat nor a switch trial), followed by a random alternation of eight visual (four A1-V2 and four A2-V1) and eight auditory trials (four A1-V2 and four A2-V1), with half the trials being task repetitions and the other half being switch trials. All 24 animals received the test twice on one day, once in each reward context (in counterbalanced order: AB or BA). After surgery all animals completed an additional two days of testing in an ABBA design. Surgery In the next phase of the experiment, half the animals received excitotoxic lesions of the AcbC and half underwent sham surgery. Next animals received an additional five days of discrimination training, followed immediately by two days of testing (again two tests on each day in each reward context), allowing the assessment of the effect of lesions of the AcbC on

123