A common view of diencephalic amnesia is that a single diencephalic structure is responsible for the memory impairment, but an alternative view is that different diencephalic structures contribute to the memory impairment in subtly different ways. This study directly compared the effects of highly selective lesions to three thalamic aggregates (the AT, MT and LT) on an odour­ place paired-associate task and a spontaneous object recognition task and used a novel quantitative analysis to calculate the damage caused by these lesions. AT and LT, but not MT lesions, severely impaired performance on the odour-place paired-associate task. Spatial probe trials introduced at the end of the odour-place paired-associate task suggested that animals may use a combination of allocentric and egocentric strategies to solve the task. No group (including controls) showed clear detection of object or object-in-place changes in the spontaneous object recognition task. The impairment in odour-place paired-associate learning in the AT group is consistent with previous research (Sziklas and Petrides, 1999) and supports Aggleton and Brown's (1999) proposal that the AT is part of an 'extended hippocampal system'. The impairment in the LT group provided new insight into the potential role of the LT in pattern association. Findings from the spatial probe tasks and the spontaneous object recognition task highlight the need for future studies to control for factors that could potentially affect performance in these tasks, such as the use of egocentric response strategies and innate object preference. The results of this study provide new information regarding the role of the thalamic nuclei in pattern association processes, and suggest that traditional models of memory function (for example, Kesner, 1998; White & McDonald, 2002) may need to be revised to take into account the important role of the thalamic nuclei in memory.