Differential remoteness and emotional tone modulate the neural correlates of autobiographical memory

Autobiographical memory relies on complex interactions between episodic memory contents, associated emotions and a sense of self-continuity along the time axis of one's life history. The neural correlates underlying autobiographical memory are known to primarily comprise areas of prefrontal cortex, medial and lateral temporal cortex, as well as posterior cingulate and retrosplenial cortex. By contrast, the effect of encoding and/or storage parameters such as the emotional tone of the memories retrieved or the length of the time-interval between the initial encoding of information and retrieval remains to be clarified. Using blocked design functional MRI and statistical parametric mapping, we investigated the impact of remoteness (factor 1: recent, remote) and emotional valence (factor 2: positive, negative) on the neural correlates of autobiographical memory retrieval. Changes in neural activity (P < 0.05, corrected) related to autobiographical memory retrieval (irrespective of remoteness and emotional tone) relative to baseline were observed bilaterally in medial and lateral temporal, temporal-occipital, posterior cingulate and frontal cortices. Recent (relative to remote) memories were associated with differentially increased neural activity bilaterally in the retrosplenial cortex and the hippocampal region, whereas remote (relative to recent) memories did not show any statistically significant differential neural activations. Positive (relative to negative) memories bilaterally activated the orbitofrontal cortex, the temporal pole, as well as medial temporal areas, with the activation peak being in the entorhinal region. By contrast, negative (relative to positive) memories differentially increased neural activity in the right middle temporal gyrus only. The data suggest differential functional roles for temporal, prefrontal and retrosplenial regions during autobiographical memory retrieval depending on the remoteness and the emotional valence of the memories retrieved. In particular, our findings support the 'classic' model of long-term memory processing, which suggests a time-limited differential involvement of the hippocampus in memory consolidation. Interestingly, the observation of such a time-dependent involvement of the hippocampal region in memory consolidation corresponds to the course of retrograde amnesia observed in demented patients, with the loss of recent memories appearing during early stages of the disease when conspicuous neurofibrillary changes are restricted mainly to the hippocampal and parahippocampal regions. Only during later stages, as the neurofibrillary changes spread out to neocortical association areas, do remote memories also become impaired. We conclude that the brain regions involved in autobiographical memory retrieval are influenced by the triggered memories' emotional significance and their relationship to the individual time axis.