The effects of intracerebral infusions of the non-specific muscarinic receptor antagonist, scopolamine, into the anteroventral (AV) thalamic nucleus on spatial working memory performance in delay and standard 12-arm radial maze tasks were assessed. The first experiment provided evidence that implants of angled bilateral cannula guides, rather than vertical unilateral cannulae combined with a unilateral AV lesions, were optimal to assess the effects on spatial working memory of scopolamine infusions to the AV thalamic nucleus. Experiment 2 examined the effects of drug infusions (1, 2.51, 6.31, 10 and 15.85 µg/side) made at the beginning of a IO-minute delay period between the 6th and ]1h arm choices, in a two-phase 12-arm radial maze task (Part One, lower three doses, and Part Three, higher three doses. During Parts One and Three, scopolamine (6.31, 10 and 15.85 µg/side) impaired performance temporarily in a dose-dependent manner, but infusions of phosphate buffered saline vehicle (PBS) also caused temporary impairments in performance which was probably the effect of the internal cannula rather than infusions per se (Part Two, Experiment 2). Drug infusions (10 µg/side) were also administered in Part Four before the daily session in a standard 12-aiu1 radial maze task. In Part Four clear and significant differences in choice accuracy, but not choice pattern, between Scopolamine (10 µg/side) and PBS infusions were obtained; PBS infusions produced only minor changes relative to No cannulae. It was concluded that infusions of scopolamine produce more significant impairments in performance on spatial working memory task when infused before the daily session, rather than during the delay between the 6th and 7th arm choices in 12- arm radial maze tasks. The results of this research contribute novel findings on the involvement of both the AV thalamic nucleus, and especially of the brainstem cholinergic innervation to the AV thalamic nucleus, during spatial working memory.