Recollection and familiarity in the human thalamus.

Recollection and familiarity are two distinct forms of recognition memory that differ in terms of the associative richness of the memory experience. In recollection, exposure to a previously encountered item cues the recollection of a number of contextual, temporal and other associative information. In the case of familiarity, instead, the item is recognized as previously encountered, but it does not cue any associative information. According to the dual-process theory, the memory processes that underlie recollection and familiarity are qualitatively different and this distinction is reflected in the existence of different neural substrates underlying the two processes. Thus far, research has primarily focused on distinct regions of the medial temporal lobe as implicated mostly in recollection (hippocampus) or familiarity (perirhinal cortex). Aggleton and Brown (1999) suggested extending the neuroanatomical distinction to other cortical and subcortical areas of the brain, including the thalamus. In particular, they proposed the existence of two reciprocally independent neural circuits for recollection and familiarity. The former would include the hippocampus, the fornix, the mammillary bodies and the anterior thalamic nuclei. The second would involve the mesial magnocellular portion of the mediodorsal nucleus connected to the perirhinal cortex through the ventroamygdalofugal pathway. Here we review neuropsychological evidence in experimental animals and brain-damaged individuals and functional neuroimaging evidence in healthy humans that supports Aggleton and Brown's model at the level of the thalamus. The evidence substantially supports the functional relationship between recollection processes and integrity of the thalamic anterior nuclei. Additional evidence, not predicted by the model, has been provided in favour of the reliance of recollection on the integrity of the lateral portion (parvocellular) of the mediodoral nucleus. Finally, there is sparse and controversial evidence in support of the reliance of familiarity on the integrity of the mesial portion of the mediodorsal nucleus, possibly due to neuroimaging methodological limits which did not satisfactorily distinguish between the medial and lateral portions of the mediodorsal nucleus.