The role of attention in emotional memory enhancement in pathological and healthy aging.

After short delays between encoding and retrieval, healthy young participants have better memory performance for emotional stimuli than for neutral stimuli. Divided-attention paradigms suggest that this emotional enhancement of memory (EEM) is due to different attention mechanisms involved during encoding: automatic processing for negative stimuli, and controlled processing for positive stimuli. As far as we know, no study on the influence of these factors on EEM in Alzheimer's disease (AD) and mild cognitive impairment (MCI) patients, as compared to healthy young and older controls, has been conducted. Thus, the goal of our study was to ascertain whether the EEM in these populations depends on the attention resources available at encoding. Participants completed two encoding phases: full attention (FA) and divided attention (DA), followed by two retrieval phases (recognition tasks). There was no EEM on the discrimination accuracy, independently of group and encoding condition. Nevertheless, all participants used a more liberal response criterion for the negative and positive stimuli than for neutral ones. In AD patients, larger numbers of false recognitions for negative and positive stimuli than for neutral ones were observed after both encoding conditions. In MCI patients and in healthy older and younger controls this effect was observed only for negative stimuli, and it depended on the encoding condition. Thus, this effect was observed in young controls after both encoding conditions, in older controls after the DA encoding, and in MCI patients after the FA encoding. In conclusion, our results suggest that emotional valence does not always enhance discrimination accuracy. Nevertheless, in certain conditions related to the attention resources available at encoding, emotional valence, especially the negative one, enhances the subjective feeling of familiarity and, consequently, engenders changes in response bias. This effect seems to be sensitive to the age and the pathology of participants.