The neural correlates of personal semantic memory across the lifecourse

Even though progressive memory loss is a defining feature of Alzheimer's disease (AD), there is currently no universally recognised diagnostic test. One major challenge for research and clinical practice is that it is difficult to differentiate healthy ageing from AD, both being associated with decline in memory function. It is thus critical to develop more specific tests of memory functions to differentiate memory profiles associated with healthy ageing and those associated with risk for AD.

Our long-term memories are generally divided into two categories. On the one hand, semantic memory is our general knowledge of the world (for instance: "1984" is the title of a book written by George Orwell). This knowledge is considered to be detached from its context of acquisition (i.e., we would not remember when or where we have learned it) and to be culturally shared. On the other hand, episodic memory concerns our memories of context-specific personal events (I remember reading "1984" at Hyde Park yesterday). The distinction between these 2 types of long-term memory, semantic and episodic, is now well-established. We know that these memory systems are associated with different brain regions and are impaired differently in healthy ageing and AD: semantic memory tends to be well-preserved in ageing and in early AD, while episodic memory performance tends to decrease with advancing age and to be impaired in AD.

However, recent data suggest that there may be additional forms of memory between the categories of semantic and episodic. These additional types, collectively referred to as personal semantics concerns knowledge of one's past (for instance: "I was born on the 25th of June like Orwell"; " I used to study Orwell's work at secondary school", or "'I am a fast reader"). Personal semantics share features with both semantic and episodic memory but appear to be distinguishable from them. Personal semantic memories are personal (like episodic memory), yet detached from a specific context of acquisition (like semantic memory). However, these forms of memories have not been comprehensively studied and we do not know how they are represented in the brain.

What is crucially needed to better understand memory are experiments that compare different types of personal semantics against semantic and episodic memory processes. Moreover, since these distinct types of memory are known to evolve differently through ageing and AD, it is therefore important to study personal semantics in young and older adults and to identify the specific memory profiles of individuals at risk for AD. This is what we are proposing to do in the present research project. We will study the distinct types of personal semantics that have been described so far and compare their neural bases in groups of young and older adults, taking into account genetic risk for AD. For this, we will recruit 60 young (18-30) and 100 older (65-80) participants, half of which are at genetic risk for AD. Using electroencephalography (EEG) and magnetic resonance imaging (MRI), we will be able to: 1) Uncover how personal semantics, as a new component of declarative memory, is represented in the brain; 2) Identify how personal semantics evolve with age and in those associated with risk for AD; and 3) Propose new tools for the clinical diagnosis of AD by monitoring early departure from memory profiles observed in healthy older adults. Crucially, as no psychological test currently available allows the evaluation of the 3 types of memory together (i.e., semantic, episodic and personal semantics), our results will directly contribute to the development of new memory tests.

Taken together, the present project will provide unique opportunities to better understand understudied forms of memory in younger and older adults, but also to improve the evaluation of memory functions in healthy ageing and in those at risk for AD.

Declarative memory is typically defined as consisting of two independent systems: episodic and semantic. However, since this dichotomy was established, additional aspects of declarative memory have been described, which appear to share some features with episodic and semantic memory, but are also dissociable from them. These aspects of memory are commonly referred to as personal semantics (or semantic autobiographical memory). Despite offering high potential for more accurate models of memory and new diagnostic and rehabilitation pathways, personal semantics is not integrated into current models of declarative memory and knowledge, as it remains poorly studied. I have recently introduced a new framework to operationalise personal semantics according to 4 main types and their distinct neural correlates (Renoult et al., 2012). In the proposed research project, we will study these 4 types of personal semantics and examine how they relate to semantic and episodic memory in 60 younger (18-30) and 100 older (65-80) adults and in relation with genetic risk (APOE4 carriers) for Alzheimer's disease (AD). We will measure their corresponding neural correlates using event-related potentials (ERPs), multi-variate statistics and dipole source modelling using each participant's MRI. It is well established that semantic memory tends to be well-preserved in ageing and in early AD, while episodic memory performance tends to decrease with advancing age and to be impaired in AD. However, very little is known about personal semantics in ageing and AD. It is thus still unclear which types of personal semantics are modified in healthy ageing and in individuals at genetic risk for AD. By studying personal semantics alongside episodic and semantic memory in relation to age and genetic risk for AD, this project will contribute more accurate models of declarative memory, new memory tests, and will provide avenues for new rehabilitation approaches in memory-impaired individuals.

The present project fits under 2 priorities of the MRC 2014-2019 strategic plan: the "Living a long and healthy life / life course perspective" priority, as this project will allow to understand better the fundamental biological mechanisms of cognitive ageing; and the "Resilience repair and replacement" priority, as we will advance knowledge on the neurobiology of neurodegeneration by investigating the brain correlates of genetic risk for Azheimer's disease (AD). We summarise below the likely clinical, academic, and policy and societal impact to emerge from the research, which will be promoted to the general public through a dedicated website.

Clinical Impact: The memory tests that we propose to develop (including a new version of the "autobiographical interview") will constitute the first tests allowing the measurement of personal semantics alongside semantic and episodic memory. It will thus be possible to identify preserved and impaired memory types much more precisely with these new memory tests, opening the door to new diagnosis and rehabilitation approaches in memory-impaired individuals, as well as assessment of risk prognosis in healthy older adults.

Academic Impact: The incorporation of personal semantics into current models of memory will constitute a highly significant theoretical change in the cognitive neuroscience of declarative memory and will thus have a substantial impact in the field. Moreover, the availability of a large dataset on differentiation of memory types in older adults and in individuals at high risk for AD will allow progressing our understanding of the most fundamental changes associated with cognitive ageing.

Impact on policy and society:
Stakeholder steering group: We will create a stakeholder steering group at the start of the project, composed of health professionals, relevant 3rd sector organizations, and policy experts. In the 1st year, its purpose will be to refine the design of our experiments and help with participants' recruitment. In the 2nd and 3rd years, the goals of the steering group will be to raise awareness of the outcomes of our research among interested stakeholders and to identify future applications of the research via "memory in practice guidelines". The focus of the discussions will be on impact. After the last meeting of the steering group, we will produce a report on the outcomes, including recommendations for changes in policies and practice, not just nationally but also internationally.

Public workshop: The 2nd phase of dissemination of the outcomes of our research will be a 2 day public workshop at UEA on declarative memory in ageing and AD, bringing leading national and international academics, with a view to giving broader reach and dissemination that started locally with the steering group. We will produce a report on the outcomes of the workshop, which will be shared widely.

Further developments: The steering group and the workshop will be vehicles for knowledge exchange. They will enable strengthening of existing partnerships across all sectors as well as creating new ones. In the autumn of 2021, we will use the momentum of this engagement to work with these partners to feed research recommendations into practice. We will aim to give guidance and advice to improve health and social care by directly influencing the following guidelines from the National Institute for Health and Care Excellence (NICE) based on our research: "independence and mental wellbeing", "home care" and guidelines on "challenging behaviour and learning disabilities". These guidelines will greatly benefit from integrating new findings in memory research and specifically on personal memory and knowledge. Some of these changes could result in more accurate diagnostics of memory difficulties, more efficient rehabilitation approaches, less referrals and hospitalizations, more appropriate timing for move into care homes and better environment in these homes.