Abstract
Background
Alzheimer's disease (AD) is generally understood as a neurodegenerative condition where synaptic problems, network failure, and neuron loss lead to lasting cognitive decline [5,6]. Still, short lucid moments, especially terminal lucidity (a short return of clear thinking near death), have been reported many times in advanced dementia, including AD [1,2]. Also, clear moments in serious dementia have been seen in living patients [3,4]. This paper looks at lucidity both in terminal situations and as random, non-terminal moments reported in advanced dementia, treating both as possible state-dependent events. These observations question strict interpretations of decline and encourage a new look at whether serious amnesia means memory representations are fully erased.
Objective
This paper puts forward a way of thinking about network disconnection in Alzheimer's disease (AD). The idea is that memory loss in AD is largely due to problems retrieving memories because of disruptions in brain circuits and widespread communication between different brain areas; the actual memory traces might still be there, at least in part.
Lucid episodes in severe dementia, both terminal and spontaneous, can be understood as temporary shifts in brain activity that allow access to memories that are usually out of reach. Terminal lucidity is viewed as a specific, more extreme instance of this. We consider terminal lucidity a subtype of lucid episodes. ‘Reconnection’ can be seen as something we can measure and test that goes along with a change in brain state, but it is not necessarily the cause of lucidity.
This framework generates falsifiable predictions that distinguish genuine cognitive restoration from non-specific arousal, positioning lucidity as a naturally occurring perturbation experiment with immediate translational relevance. The framework specifies quantitative effect sizes and pre-registered analysis plans to enable rigorous falsification.
Methods
We combined evidence from several areas: (i) studies on engram reactivation that show memories that seem lost can be retrieved through targeted activation, even in Alzheimer's mouse models [11-13]; (ii) brain imaging studies in humans that link memory problems and disease progression to problems in how the hippocampus and cortex communicate, along with issues in the default mode network, beyond just brain shrinkage [14-24]; (iii) studies that point to damage to white matter and problems with brain tracts (like the fornix) as reasons for memory access issues [25-36]; and (iv) case studies of clarity in people with severe dementia [1-4]. From this, we made predictions that can be tested about how brain state changes can be measured using EEG and simple brain imaging techniques [40-52].
Results
Research on engrams backs up the idea that memory can stick around even when someone has amnesia. This lines up with the idea that problems in brain circuits can block memory retrieval [11-13]. Imaging studies on people show that when the hippocampus and cortex aren't talking to each other as well, and when there are issues in the default mode network, it's linked to problems with recalling events and keeping track of how a disease is progressing. Sometimes, this gives us clues beyond what we see from brain shrinkage [14-24]. When white matter breaks down, including in the fornix, it could explain why certain memory networks have trouble being reached [25-36]. Clear episodes (both before and at the very end of life) can be seen as changes in state that might briefly boost how well the brain is connected and how well the network works. This lets people briefly recall things from memories that are still somewhat intact [1-4]. Terminal lucidity is an extreme example of this [1,2]. This idea suggests that if we take into account things like alertness, artifacts, and (when someone is dying) oxygen levels, ventilation (including how CO₂ affects things), and how long someone has left to live, clear episodes should happen at the same time as some improvement in how the brain's electrical activity is organized and how well it's connected, even though the brain structure isn't recovering [40-49]. These expected changes in brain activity would be seen as linked to a more connected state, which allows for better access to memories. But, this doesn't necessarily mean that network connections are the only thing needed to recall autobiographical memories.
Conclusions
In disconnection theory, AD amnesia is viewed as a problem with accessing memories, not just permanent memory loss. This explains why some people have moments of clarity and fits with research showing memories can still exist even if they can't be used [11-13].
To test if lucidity depends on brain state, we could use wearable EEG and imaging to capture these lucid moments [40-52]. Treatments that change brain network states and boost brain connectivity, like neuromodulation or metabolic support, might help with symptoms even if there's a lot of damage [53-60].
Here, we see neuromodulation as a way to test if brain state affects memory access, not just as a way to make people lucid. Intervention studies still need to prove that changing brain network states really does lead to better memory retrieval.
Keywords: terminal lucidity; non-terminal lucidity; paradoxical lucidity; Alzheimer’s disease; memory engram; disconnection syndrome; functional connectivity; default mode network; EEG; white matter