Brain Signals May Reveal How Prospective Memory Falters in MCI, But Evidence Remains Preliminary

A PhD thesis combines EEG neurophysiology and machine learning to map memory-for-intentions across ageing and mild cognitive impairment, offering early-stage insights into neural markers that may distinguish healthy cognitive ageing from early decline.

Why This Matters

Prospective memory, the ability to remember to carry out a future intention, is one of the earliest cognitive domains affected in mild cognitive impairment and dementia, yet its neurophysiological basis remains poorly understood. Identifying reliable brain-based markers of prospective memory decline could eventually support earlier detection and more targeted intervention in populations at risk for Alzheimer’s disease. This thesis represents a first attempt to bridge that gap using EEG, making its findings scientifically credible as a starting point even though the evidence remains exploratory and has not been fully peer-reviewed.

Clinical Summary

Prospective memory (PM) refers to the capacity to remember and execute a planned action at the right moment, such as taking medication at a specific time. Unlike retrospective memory, PM depends heavily on executive control and attentional monitoring, making it particularly vulnerable to early-stage cognitive decline. This 2020 PhD thesis from Nottingham Trent University, conducted in collaboration with Auckland University of Technology, used electroencephalography (EEG) and behavioural testing across at least four empirical studies to examine PM in young adults, healthy older adults, and older adults with mild cognitive impairment (MCI). The mechanistic hypothesis centres on event-related potentials (ERPs), specifically the reorientation negativity and parietal positivity, as neural signatures of the attentional and metacognitive processes that support PM.

The thesis reports that healthy older adults did not show significant behavioural impairment on PM tasks, which the author attributes to possible compensatory neural strategies. Older adults with MCI, however, appeared impaired on certain PM tasks, with deficits potentially linked to reduced attentional reorientation and diminished “feelings of knowing,” a metacognitive process. Additionally, spiking neural network (SNN) modelling was applied as an exploratory computational tool to characterise spatiotemporal EEG patterns across groups. However, individual study sample sizes were not reported in the available text, full quantitative results and effect sizes remain unverified, and the thesis has not been independently peer-reviewed. The author notes that further research, including longitudinal designs with larger samples, is needed before any clinical applications can be considered.

Dr. Caplan’s Take

This thesis asks a genuinely important question: can we detect early prospective memory decline using EEG before it becomes clinically obvious? The identification of ERP markers like the reorientation negativity in PM paradigms is mechanistically plausible and aligns with what we know about attentional monitoring deficits in MCI. However, patients and family members frequently ask whether a brain scan or test could catch memory problems early, and this work is not yet at the stage where it informs that conversation. The sample sizes are unclear, the findings have not been replicated, and the computational modelling approach, while innovative, is untested outside this single programme of research.

In practice, when I see patients with suspected MCI who report difficulty remembering appointments or medication schedules, I rely on validated neuropsychological assessments and clinical history rather than experimental EEG markers. What this thesis does well is point toward a direction that future research should pursue. For now, the most defensible clinical approach remains structured cognitive assessment, attention to real-world functional decline, and careful longitudinal monitoring rather than waiting for a biomarker that has not yet been validated.

Clinical Perspective

This thesis sits very early in the research arc for EEG-based PM biomarkers in MCI. It fills a genuine gap by being reportedly the first body of work to directly examine the neurophysiology of prospective memory in older adults with MCI using EEG. The finding that healthy older adults may compensate neurally for PM demands is consistent with broader cognitive reserve literature, and the observation that MCI-related PM impairment may involve attentional and metacognitive deficits aligns with existing models of executive dysfunction in early neurodegeneration. However, the cross-sectional design, unverified sample sizes, and absence of peer review for most component studies mean that none of these findings can be considered established. The evidence does not currently support any patient-facing clinical recommendation regarding EEG-based PM assessment.

From a practical standpoint, no pharmacological or safety considerations arise directly from this work since it is purely observational and diagnostic in orientation. Clinicians should be aware that patients may encounter media coverage framing EEG or machine learning as near-term diagnostic tools for early dementia, which substantially overstates where this science currently stands. One concrete step clinicians can implement now is to include explicit prospective memory assessment in cognitive evaluations for patients reporting real-world planning failures, using validated instruments such as the Memory for Intentions Screening Test (MIST), rather than relying solely on retrospective memory measures that may miss early functional decline.

Study at a Glance

What Kind of Evidence Is This

This is a doctoral thesis comprising multiple cross-sectional, observational EEG studies and an exploratory computational modelling component. As a thesis, it has not undergone independent peer review equivalent to journal publication, placing it below peer-reviewed observational studies in the evidence hierarchy. The most important inference constraint is that no causal or predictive claims can be drawn: the cross-sectional design cannot establish whether observed neural differences precede, accompany, or follow PM decline, and the absence of verified sample sizes makes it impossible to assess statistical power or the reliability of group differences.

How This Fits With the Broader Literature

The thesis aligns with and extends prior behavioural research showing that PM is sensitive to early cognitive decline, including work by Blanco-Campal and colleagues (2009) demonstrating PM impairment in amnestic MCI and meta-analytic findings from van den Berg and colleagues (2012) documenting age-related PM effects. The novelty lies in the EEG neurophysiology: prior PM research in MCI has relied predominantly on behavioural paradigms without concurrent neural measurement. The application of spiking neural networks to EEG data draws on the NeuCube framework developed at Auckland University of Technology, which has been applied to other neurological conditions but not previously to prospective memory.

However, without published effect sizes or replication, this work remains hypothesis-generating rather than confirmatory. It opens a promising research direction but does not yet constitute the kind of converging evidence needed to shift clinical practice.

Common Misreadings

The most likely overinterpretation is that EEG can now detect or diagnose early MCI through prospective memory tasks. This thesis does not establish diagnostic sensitivity, specificity, or predictive validity for any EEG marker. Equally problematic would be concluding that the spiking neural network modelling represents a validated clinical tool; it is an exploratory analytical method applied to a single dataset without external validation. Finally, the finding that healthy older adults were “not impaired” on PM tasks should not be read as evidence that PM is robust in ageing generally, since the thesis does not report the specific conditions under which this held or whether ceiling effects in the task design contributed to this result.

Bottom Line

This thesis makes a credible first attempt to characterize the neurophysiology of prospective memory in MCI using EEG, identifying plausible neural markers and proposing a novel computational framework. However, it is early-stage, cross-sectional, largely unverified by peer review, and lacking reported sample sizes. It generates hypotheses worth pursuing but establishes nothing that should change clinical assessment or patient counselling at this time.

References

1. Chou, Y. H. (2020). Investigating prospective memory in ageing and mild cognitive impairment using EEG and spiking neural networks [Doctoral thesis, Nottingham Trent University]. Nottingham Trent University Repository.
2. Blanco-Campal, A., Coen, R. F., Lawlor, B. A., Walsh, J. B., & Burke, T. E. (2009). Detection of prospective memory deficits in mild cognitive impairment of suspected Alzheimer’s disease etiology using a novel event-based prospective memory task. Journal of the International Neuropsychological Society, 15(1), 154-159.
3. van den Berg, E., Kant, N., & Postma, A. (2012). Remember to buy milk
   
   
   

   Physician-Led, Whole-Person Care

   A doctor who takes the time to truly understand you.

   Personal care that starts with listening and is guided by experience and ingenuity.

   Health, Longevity, Wellness

   One-on-One Cannabis Guidance

   Metabolic Balance

   Leave a Message
   Metabolic Care
   Medical Consulting
   Cannabis Care