Alzheimer's disease [AD] is characterized by neuronal loss, neurofibrillary tangles, and senile plaques, which first develop in the entorhinal cortex and the hippocampus (Braak and Braak, 1991 and Delacourte et al., 1999). At the cognitive level, memory impairments come first and affect both the anterograde and retrograde components of episodic memory (Desgranges et al., 2002, Piolino et al., 2003 and Eustache et al., 2004). Such impairments are classically explained by changes in the medial temporal lobe (including the hippocampus), a critical brain area for memory processing (Squire and Alvarez, 1995, Morris, 1996 and Nadel and Moscovitch, 1997).
Recently, it has been suggested that memory decline in AD could also be the consequence of impairment in sleep-dependent memory consolidation (Rauchs et al., 2010). Indeed, cumulative evidence supports the existence of off-line reprocessing of recently acquired memories within hippocampal networks during sleep (for reviews, see Rauchs et al., 2005 and Diekelmann and Born, 2010). Sleep deprivation and neuroimaging studies have mainly demonstrated the crucial role of slow wave sleep ([SWS] = stage 3 + 4) for consolidation of declarative, episodic memories (e.g., Plihal and Born, 1997 and Peigneux et al., 2004). These works are in agreement with the “hippocampo-neocortical dialogue” hypothesis proposed by Buzsaki (1996). According to this model, recently acquired memories are reactivated, during SWS, within hippocampal networks. These reactivations occur in a coordinated manner with slow oscillations and sleep spindles, and will favor a transfer of memory traces towards neocortical areas where memories will be stored for the long term. Recent functional magnetic resonance imaging studies have provided evidence of such a transfer of memory traces during post-learning sleep (Gais et al., 2007 and Rauchs et al., 2011). Moreover, electroencephalographic [EEG] recordings have shown that the theta rhythm (brain activity between 4 Hz and 7.5 Hz) seems to be sensitive to the elicitation of memory consolidation processes that depend on hippocampus activity (Fogel et al., 2009, Klimesch, 1996 and Cantero et al., 2003). In AD, sleep disturbances have been investigated in several studies (see Petit et al., 2004 for review), mainly for diagnostic purposes or clinical viewpoints. They reported both an accentuation of sleep changes occurring during normal ageing (Bombois et al., 2010) and a reduction of the amount of REM sleep specific to AD patients (Petit et al., 2005, Christos, 1993 and Dykierek et al., 1998) but at a late stage of the disease (Gagnon et al., 2006). Cellular studies report disorders of the cortical cholinergic activity, involved in the regulation of sleep (Coyle et al., 1983 and Montplaisir et al., 1998). The cholinergic system has been identified as one of the major neurotransmitter systems involved in memory (Blokland, 1996 and Everitt and Robbins, 1997). This finding is particularly relevant, considering data from animal and human studies supporting the suggestion that AD is related to cholinergic system dysfunction (see White and Ruske, 2002, for a review). To date, the contribution of sleep disturbances to memory impairments in AD remains largely unknown (Rauchs et al., 2010). Mizuno et al. (2004) reported that increasing the amount of REM sleep by administration of the cholinesterase inhibitor donepezil in moderate AD patients improves cognitive functioning. However, this study did not clearly establish the link between declarative memory impairment and changes in sleep architecture, particularly because patients were relatively advanced in the disease and presented a global decline of sleep architecture.
Considering that unique compensatory brain processes have been reported in AD patients at the first stage of the disease (Dickerson et al., 2005), it can be assumed that (i) early changes could occur in post-learning EEG sleep before the alteration of sleep organization reported in mild to moderate AD patients, and (ii) these changes should be related to episodic memory decline. A recent sleep study with amnestic mild cognitive impairment [aMCI] patients, a prodromal stage of AD, confirms that inadequate memory consolidation is related to declines in subjective sleep indices (Westerberg et al., 2010). In our group, we also found that early AD patients with the faster spindles had the better performance in the episodic memory task (Rauchs et al., 2008). Converging studies point out that spindles are related to memory performance, but the involvement of brain structures sustaining the elicitation of the spindles has to be unraveled. These well-designed studies did not yet define what kind of brain processes are modulated in AD during sleep.
The aim of the present study is to further investigate the relationships between episodic memory deficits and sleep changes by focusing on changes during sleep stages in theta activity, a brain index of declarative memory processes. As reported in our previous study on spindles, we predict that changes in theta activity occur before the massive sleep disorganization occurring during the spread of AD. To examine subtle changes in brain activity during sleep, we focus on early AD patients with no clear sleep architecture disorganization.
