As the world watches images of destroyed buildings, displaced families, frightened children, and exhausted clinicians, lives are lost and uncertainty becomes the condition of everyday life. Beyond the immediate humanitarian crisis, armed conflict may leave a and less visible imprint on brain health, through disruptions with consequences that extend beyond the end of hostilities or the crossing of a border. Whether these experiences contribute to lasting cognitive vulnerability, and through which pathways, remains an open scientific question.
Throughout this perspective, “armed conflict” refers to organized armed violence, an editorial convention rather than a legal classification. Material deprivation may precede, accompany, or follow conflict and must be measured separately as this distinction is critical because people living through the same conflict may experience very different variations of violence, displacement, hunger, injury, and loss of access to care.
We propose that specific disruptions help explain differences in cognitive trajectories within conflict-affected populations. Table 1 organizes this proposition into four distinct pathways: (1) interruption of vascular care, (2) interrupted education, (3) displacement and social disconnection, and (4) traumatic stress or injury. Each links a measurable exposure to a proposed intermediate process, a defined brain health outcome, and an observation that would weaken the explanation. These pathways provide the broader framework for this perspective; the supporting evidence varies, and the complete causal sequences remain to be established.1–4 Within this framework, we examine the traumatic stress and injury pathway in greater depth, focusing primarily on persistent traumatic stress and post-traumatic stress disorder (PTSD) in relation to cognition and diagnosed dementia. These studies include veterans, civilians, and disaster responders, and help frame questions for conflict-affected populations without establishing a causal effect of armed conflict.1–3 Traumatic brain injury (TBI) remains a distinct exposure within the same pathway and should be assessed separately. We use the term “brain health risk environment” to describe the concentration of these exposures alongside the loss of protective resources, without assuming that their effects necessarily compound one another. Similar pathways may arise after interpersonal violence, natural disasters, or severe material adversity. Armed conflict warrants focused study because threats may recur across whole communities while the institutions needed for protection and recovery are damaged. This perspective therefore proposes comparative, prospective research and care grounded in current symptoms and functional human needs. It neither estimates a conflict-attributable dementia burden nor supports automatic cognitive screening based on conflict exposure alone.
The Pathways Linking Armed Conflict and Brain Health
Armed conflict can alter the conditions that sustain brain health long before cognitive impairment becomes apparent. A child may lose access to education; an older adult may lose access to regular treatment; a displaced family may lose the relationships and resources on which daily life depends. The scientific question is how these disruptions influence lifespan development, cognitive aging, brain health, and which consequences persist when safety and support are restored. The 2024 Lancet Commission offers a starting point because many of its 14 potentially modifiable dementia risk factors may be potentially impacted by armed conflict including education, vascular health, depression, sensory impairment, and social connection.4 These conditions provide context for interpreting trauma-related associations, while each remains a separate pathway to investigate either individually or in combination.
Vascular care interruption is the first pathway, providing a concrete way to examine how the loss of an established protective resource may influence brain health. For example, when access to medication and primary care is disrupted, people with hypertension may lose a source of protection that existed before the conflict. There is some evidence of adverse cardiovascular consequences across conflict settings.5 Studies in Gaza linked geographically located violence to physical impairment and chronic disease, with suggestive contributions from income loss and reduced access to healthcare; separate research associated reported traumatic events with persistently very high blood pressure trajectories.6,7 These findings connect conflict exposure with vascular vulnerability, but they do not show that verified medication interruptions caused blood pressure deterioration or subsequent cognitive decline.
The link to cognition comes from different sources. In a longitudinal Chinese community cohort, antihypertensive use and adherence were associated with slower decline in global cognition and memory.8 This observational finding cannot establish the effect of restoring treatment during conflict, and adherence across survey waves differs from documented days without medication. Taken together, the studies justify a focused question: does conflict-related interruption of vascular care precede poorer blood pressure control and, in turn, a less favorable cognitive trajectory? The complete sequence remains to be tested.
Interrupted education defines the second pathway, which operates on a different timescale. Conflict can disrupt the opportunities through which children learn, yet poorer school achievement does not necessarily indicate impaired cognitive development. Conflict experience during the Second Intifada has been linked to later primary-school achievement, with the analysis implicating non-cognitive skills rather than measured cognitive skills as the channel.9 This distinction matters for the proposed pathway: school attendance, distress, educational attainment, and cognitive performance should be measured separately. The immediate question concerns learning and development; any subsequent effect on adult cognitive reserve or dementia requires longer follow-up.
Displacement and social disconnection form the third pathway. Displacement may separate people from social networks and familiar sources of support, while food insecurity and material deprivation may accompany this loss of stability. Each exposure requires a corresponding measure and outcome rather than being absorbed into a general category of conflict exposure. The direction of change also matters: food insecurity may lead to undernutrition rather than obesity, and restrictions on movement or alcohol availability need not alter risk uniformly. The value of the Lancet framework is therefore to organize specific questions, while the extent to which risks accumulate or interact remains an empirical issue.
Traumatic stress and injury comprise the fourth pathway and the main focus of this perspective. Violence and repeated loss of safety may be followed by persistent PTSD symptoms, whereas blast exposure or other physical injury may cause TBI or hearing loss. These exposures can coexist but require separate histories and assessments. The stress-related hypothesis concerns whether cumulative symptom burden, potentially involving sleep or vascular changes, precedes changes in learning or working memory. Longitudinal PTSD studies motivate this question without establishing the complete causal sequence.10,11 The sections that follow examine the traumatic stress component in greater depth within the broader four-pathway framework.
These pathways depend on the resources available before and after exposure. Mostert et al. propose broadening dementia risk models to include poverty, wealth shocks, income inequality, and HIV, with attention to sex, gender, and populations underrepresented in existing evidence.12 Their proposal directs attention to structural conditions that may shape both exposure and access to protection. In conflict research, this means distinguishing prior deprivation from conflict-related losses, household wealth changes from area-level inequality, and, where relevant, HIV infection from interruption of its treatment. These considerations complement the Commission framework; they do not provide a validated conflict-specific model or a dementia burden attributable to armed conflict.
Who Is at Risk?
Within this four-pathway framework, the evidence on traumatic stress comes largely from observational studies of PTSD. Large veteran cohorts provided some of the earliest findings linking PTSD with later diagnosed dementia. In a landmark U.S. Veterans Affairs study of more than 180,000 veterans aged 55 and older without dementia at baseline, PTSD was associated with increased incident dementia over follow-up13 Subsequent work found associations involving prisoner-of-war status and PTSD, with the highest dementia risk among veterans exposed to both.14
In a large civilian healthcare system, PTSD was associated with increased dementia risk in both women and men, with higher estimates among those with comorbid depression.15 A Korean matched cohort also associated broader stress-related disorders with subsequent dementia, indicating that the relevant exposure landscape extends beyond combat and PTSD alone.16
Thus, an overall association between PTSD and dementia is supported but causal relationships have not been established. A 2020 meta-analysis reported a pooled hazard ratio of 1.61 for all-cause dementia (95% CI 1.43–1.81), with high heterogeneity (I2 = 85.8%) associated with PTSD.1 A recent 2026 review included 18 studies overall and reported an adjusted pooled hazard ratio of 1.56 (95% CI 1.27–1.91; I2 = 95.5%).2 The very high heterogeneity of these estimates limits generalization across populations, although population type was not a significant moderator in the later meta-analysis.
The PTSD literature highlights symptom-related vulnerability, while studies that directly measure armed-conflict exposure help situate these vulnerabilities within a broader framework. This evidence remains limited, but newer studies move beyond a single category of conflict exposure. The pooled war/Holocaust-related estimate reported by Severs et al. derives from three heterogeneous studies and does not isolate violence from injury, deprivation, PTSD, or selective survival.3,14 In the Vietnam Health and Aging Study, severe childhood hunger and wartime environmental hardship were associated with poorer late-life cognitive performance, while combat duties were associated with higher scores.17,18 Among 548 older Vietnamese Americans, associations were found that differed by exposure and cognitive domain: displacement was associated with poorer executive function and family endangerment with poorer visuospatial performance, whereas some early-life adversity associations were in the opposite direction.19 Both analyses were cross-sectional and cannot establish cognitive decline or incident dementia, and recall, selection, and survival remain critical explanations. Their heterogeneity makes the decomposition of exposure scientifically necessary.
Within armed-conflict settings, clinically relevant exposure extends beyond combat. It includes bombardment, detention, bereavement, sexual and interpersonal violence, forced displacement, family separation, repeated loss of safety, and work under sustained threat. Healthcare workers, humanitarian responders, civilians, and military personnel may differ substantially in exposure pattern, baseline vulnerability, and access to care. These distinctions should be measured explicitly rather than collapsed into a single category of “war trauma.”
Children may be especially vulnerable to cumulative and developmentally timed adversity. Systematic review data in refugee children show psychological distress shaped by premigration trauma and postmigration stressors.20 A review of 155 publications on armed conflict and child health and development, documented diverse exposures and outcomes while identifying major gaps in developmental, longitudinal, and protective-factor research.21 Prolonged conflict may disrupt schooling, emotional regulation, and physical health.22 These findings justify protecting development and access to care.
PTSD also rarely occurs in isolation. Depression, sleep disturbance, substance use, hypertension, diabetes, social isolation, and TBI may coexist with trauma and contribute to cognitive vulnerability through partly overlapping pathways.4,15,23 Clinical priority should therefore reflect current cognitive or functional change together with age, symptom persistence, comorbidity, and barriers to care. This profile offers a practical basis for clinical attention, although it is not a validated dementia prediction score. The brain may remember what the body survives, but epidemiology must still distinguish memory, mechanism, and marker. Table 2 summarizes what the available evidence supports and where inference remains limited.
PTSD as an Epidemiologic Risk Marker
The closer examination of the trauma pathway requires distinguishing PTSD as a risk marker from a demonstrated causal mechanism. PTSD is not one of the 14 potentially modifiable dementia risk factors identified by the 2024 Lancet Commission.4 Observational studies nevertheless associate it with higher rates of diagnosed dementia.1,2 We therefore consider persistent traumatic stress a candidate process to investigate alongside vascular, sensory, developmental, and social risks. Whether PTSD symptoms contribute causally to cognitive decline, or mainly identify other vulnerabilities, remains unresolved.
A risk marker identifies an observed association; it does not establish that changing the marker will change the outcome. Testing a pathway requires an exposure that precedes the proposed mediator and outcome, while addressing plausible alternative explanations. Care remains justified by present symptoms and established health needs, even when its effect on future dementia is unknown.
Several limitations constrain causal inference. Most studies are retrospective or based on administrative diagnoses; exposure severity, treatment history, socioeconomic conditions, sleep, education, and premorbid cognition are incompletely measured. Greater healthcare contact may increase dementia detection, whereas disrupted access may reduce it. Residual confounding, diagnostic misclassification, selection into clinical care, competing mortality, and survival bias may all influence estimates.1–3 Depression, sleep disturbance, and vascular illness may precede or follow PTSD; treating them uniformly as confounders can obscure mediation or other temporal relationships.
Reverse causation also requires explicit consideration. Early or prodromal neurodegenerative changes may influence psychiatric presentation, coping, or recurrence of remote traumatic memories. Delayed-onset or recurrent PTSD phenomena has been described in some people with dementia and earlier trauma.30 Although lagged analyses reduce this concern, the authors cannot fully eliminate long prodromal periods or shared vulnerability. This does not imply that dementia causes remote historical trauma; it requires separating exposure timing, PTSD onset or recurrence, and cognitive change.
Within this trauma-focused evidence, repeated cognitive performance offers a more immediate outcome than dementia. A systematic review of PTSD in older adults that included 18 studies, only three were longitudinal, and these reported poorer global cognition and memory with mixed executive-function findings.24 A multilevel meta-analysis of 53 studies found a small association between PTSD and cognitive impairment after adjustment for publication bias.31 In a study of 12,270 trauma-exposed women, higher PTSD symptom burden was associated with faster cognitive decline; a subsequent analysis of the same cohort found differing baseline profiles for remitted and unresolved PTSD, while both groups showed faster learning/working-memory decline than trauma-exposed women without PTSD.10,32 In a separate sample of 11,961 older adults, including 179 with lifetime PTSD, associations with declines in working-memory were found.11
Biomarker evidence is also unsettled. Self-reported posttraumatic stress symptoms were associated with higher amyloid burden in a cross-sectional analysis of 4,228 cognitively unimpaired older adults; exposure was measured with the Impact of Event Scale rather than a diagnostic interview.33 In 87 Vietnam-era veterans without dementia, subjective cognitive decline was found more strongly related to amyloid and tau PET burden among those without PTSD than among those with PTSD.34 These studies ask different questions and neither establishes a temporal PTSD-to-Alzheimer’s disease pathway. Cognitive complaints, objective performance, incident dementia, and biomarker-defined pathology must remain distinct outcomes.
Biologically plausible pathways provide a coherent, non-exclusive framework for research. Chronic traumatic stress may be associated with dysregulation of stress-response systems, inflammatory signaling, vascular burden, sleep disruption, and reduced engagement with healthcare.23 These pathways are not specific to armed conflict, but armed conflict may bring them into contact with TBI, malnutrition, interrupted treatment, and social disruption. This convergence is a plausible reason to study conflict-affected populations, while remaining hypothesis-generating. Study-level moderator findings concerning hypertension and diabetes support attention to these comorbidities but do not establish individual effect modification or a validated prediction rule.2
We also do not know whether treating PTSD prevents dementia. Trauma-focused treatment should be offered for current suffering and impairment. Managing vascular and metabolic illness, harmful substance use, and sensory barriers has independent clinical value, without implying equal dementia-prevention evidence for every intervention or a proven neuroprotective effect of PTSD treatment.4,23
Comparative trauma and recovery
The uncertainty surrounding PTSD and cognition makes comparison with other forms of adversity essential. Examining trauma beyond armed conflict helps test whether the associations discussed here reflect shared pathways or features of particular settings. Natural disasters, prolonged economic insecurity, societal instability, and forced migration unrelated to armed conflict can also disrupt safety, education, and healthcare. The Armenian earthquake literature is an essential comparator. In the 23-year follow-up, 725 survivors were assessed: earthquake-attributed PTSD was reported in 11.6% at follow-up versus 48.7% at the earlier assessment. Later burden was associated with exposure intensity and subsequent adversities, while housing assistance and social support were associated with lower PTSD burden.25 This documents both persistence in a subgroup and substantial improvement, emphasizing conditions for recovery beyond the initial event.
A smaller prospective study of 78 adults compared earthquake exposure and political violence and found no significant differences in PTSD severity, profile, or course between the severely exposed groups.26 Neither study assessed incident dementia and therefore informs trauma trajectories rather than equivalence of dementia risk. The comparison tempers claims of biological uniqueness, while leaving open an operational distinction: armed conflict can be prolonged, recurrent, intentionally violent, population-wide, and accompanied by sustained institutional damage. Whether that combination produces additional dementia risk beyond the component adversities remains unknown.
Trauma also does not produce a uniform trajectory. Resilience describes relatively maintained functioning after adversity, whereas recovery involves improvement after an initial disruption.35 Posttraumatic growth refers to positive psychological change arising through the struggle with trauma, including changes in meaning, relationships, or priorities; it is not simply return to baseline.36 Perceived growth may coexist with distress and should not be equated with prospectively measured change.37 These concepts broaden the framework, but the cited evidence does not show that resilience or growth protects against dementia. Survivors should neither be portrayed as inevitably damaged nor expected to demonstrate growth.
A child whose schooling is interrupted may lose opportunities for learning, while a displaced older adult separated from community and medication may face worsening vascular control or functional difficulties. These examples underscore the need for immediate protection while raising distinct scientific questions. School attendance and developmental performance can be assessed now; repeated cognition and clinically adjudicated dementia require different follow-up periods. Neither short-term distress nor a low cognitive score is a substitute for incident dementia.
From Conflict Exposure to Testable Questions
Turning this evidence into prospective research requires defining the exposure experienced by each person, measuring its proposed consequences over time, and testing the explanation against plausible alternatives. As the trauma pathway is the primary focus here, we propose measuring the timing and intensity of threat or violence, cumulative PTSD symptom burden and duration, sleep disturbance, and repeated learning or working-memory performance. We hypothesize that greater symptom burden over time may precede faster decline in these domains than in similarly trauma-exposed people with low lifetime symptom burden and comparable baseline cognition. Longitudinal studies motivate this question, but do not establish the causal sequence or its applicability to conflict-affected populations.10,11 Prior symptom burden should remain distinct from current remission, and verified TBI requires a separate exposure history and analysis.
The World Trade Center literature illustrates this principle. Studies followed cognitive outcomes in first responders while measuring on-site work duration and PTSD symptoms; later work examined incident dementia before age 65 using exposure categories that incorporated dust and debris, response activities, work duration, and protective equipment.38,39 This approach turns a large-scale event into questions about specific exposures and subsequent outcomes. Its observational findings in a selected responder population cannot be transferred directly to armed conflict or used to isolate the effects of PTSD, injury, and pollution.
Applying this approach to conflict requires locating each exposure in a person’s history: where it occurred, how intense and prolonged it was, and at what age. Residential mobility, incomplete reporting, and effects extending beyond the immediate conflict area can complicate reconstruction of these histories. Luetke and colleagues show that different spatial and temporal aggregations of conflict events can change the observed association with health.40 Follow-up should therefore be anchored to the exposure under investigation, such as a verified medication interruption or displacement episode. Pre-exposure information, wherever available, is needed to distinguish prior differences from subsequent change.
Testing the pathways set out in Table 1 requires recognizing their different levels of supporting evidence. Vascular studies address separate links in the proposed sequence, educational research concerns achievement, displacement evidence includes cross-sectional cognitive performance, and PTSD research includes longitudinal cognitive trajectories.7–11,19 Testing these explanations requires reliable repeated measures and comparisons between people with sufficiently different component exposures but comparable baseline vulnerability. The follow-up interval and the cognitive or developmental difference considered meaningful should be specified before analysis. Only then can a sufficiently precise absence of that difference inform the hypothesis; a nonsignificant result alone is insufficient. A null association among survivors would be less informative if mortality or diagnostic access differed substantially between exposure groups.
Trauma Care as Brain Health Infrastructure
The clinical implications of the trauma pathway extend beyond treatment of PTSD symptoms, because care heavily depends on the services and resources available after exposure. In armed-conflict settings, damaged services, displaced clinicians, lost records, and interrupted medication supplies can undermine continuity of care. These consequences can be investigated during active conflict. A longitudinal medical-record study documented reduced healthcare use and altered prescribing among older adults with and without pre-existing dementia after conflict onset.41 Falling rates of recorded diagnoses or visits may reflect reduced access and ascertainment rather than improved health.
A proportionate response should therefore be layered and feasible. The first priorities are safety, basic needs, social protection, continuity of primary care, culturally and linguistically appropriate trauma-informed support, and treatment of severe psychiatric symptoms. At the same time, primary and community services can address sleep, depression, substance use, hypertension, diabetes, and TBI whenever resources permit.4,23,42,43 These priorities are justified by immediate clinical and public-health needs, irrespective of whether PTSD or armed conflict has a causal effect on dementia.
Cognitive assessment should begin with assessment of change from the person’s usual functioning, rather than with conflict exposure or a risk-factor count. A planned assessment is warranted when the person, an informant, or a clinician reports a new or worsening cognitive difficulty, with a recorded example wherever possible: repeated errors with available medication, getting lost on a familiar route, or difficulty retaining information needed for usual tasks. Persistent or progressive complaints warrant evaluation even without an available informant. Functional deterioration or a safety concern increases urgency. Acute confusion, fluctuating attention or arousal, and new focal neurological signs require prompt medical evaluation for acute illness, delirium, or neurological injury. These are proposed operational triggers for assessment, not validated conflict-specific diagnostic thresholds. No fixed symptom duration or cognitive-test cut-off can be inferred from the evidence reviewed here.44
Age, persistent PTSD, TBI, vascular-metabolic illness, and interrupted care can help prioritize people who already have concerning changes; these features alone do not justify an automatic dementia-screening programme. Assessment should investigate acute illness, medication effects, depression, sleep loss, and sensory barriers, and arrange reassessment after remediable contributors are addressed when feasible. Progressive deterioration or safety concerns should not be deferred while waiting for stress to resolve. A single low score during acute distress neither establishes dementia nor demonstrates neurodegeneration. Diagnostic clarification should combine history, examination, contextualized testing, and follow-up or referral. This approach follows general assessment principles, while its implementation and net benefit in conflict settings require evaluation.23,44
Intervention research can also proceed during ongoing conflict. A recent scoping review identified mental health and psychosocial interventions delivered in these settings, with heterogeneous designs and outcomes.45 Digital approaches to consultation, training, self-management, and data services were described in fragile Middle East and North African states, alongside important infrastructure and privacy constraints.46 Where locally feasible, remote follow-up or secure medication records could help evaluate continuity, symptom burden, and functioning. These uses require accessible non-digital alternatives and safeguards for sensitive data. Neither review establishes dementia prevention or validates remote cognitive testing in conflict settings.
The same principles should apply outside armed conflict. Populations affected by armed conflict may warrant operational priority when needs are high and protective systems have collapsed, but the criteria should remain based on symptoms, functional change, comorbidity, and feasibility rather than on the label of armed conflict alone. Box 1 summarizes a feasible response grounded in these principles.
A Long View
The four-pathway framework identifies a broader research agenda, within which this perspective has examined traumatic stress and PTSD in greater depth. This emphasis reflects the evidence reviewed, rather than a claim that trauma is the dominant cause of cognitive vulnerability after conflict. Our central proposition remains pathway-specific: disruption of protective resources and persistence of measurable risks may help explain adverse cognitive trajectories in some conflict-affected populations. For the trauma pathway, prospective comparisons should determine whether cumulative symptom burden precedes domain-specific cognitive change and which intermediate processes account for it. An additional effect of armed conflict beyond these components is a separate, unresolved question.
Comparative longitudinal studies should measure both conflict intensity and the component exposure of interest. Within affected populations, comparisons can distinguish people with preserved versus interrupted care or sustained versus transient displacement. Across settings, disaster-exposed or economically disadvantaged populations with similar component exposures can test whether an explanation generalizes beyond conflict. Repeated measures should establish temporal ordering, distinguish baseline performance from subsequent decline, and separate cognitive change from incident dementia. Models must address prior disadvantage, TBI, education, competing mortality, differential diagnosis, and loss to follow-up. A prespecified causal model should distinguish confounders from mediators; routine adjustment for every consequence of conflict can remove the pathway being investigated. The mortality estimand, survivor selection, and cross-cultural comparability of the measures should be specified before interpreting a null or adverse association.
The humanitarian crisis of armed conflict is immediate. Research on its consequences for brain health should not wait for a dementia endpoint decades later. During active conflict, locally led studies can assess medication continuity, blood pressure, symptoms, social connection, cognition when feasible, and daily functioning. Longer follow-up can determine whether these changes predict persistent impairment or clinically adjudicated dementia. Protecting education, primary care, and mental health remains justified by present needs while these longer-term questions are examined.4,23,41,45,46
Armed conflicts reshape nations, institutions, families, and lives. They may also shape trajectories of brain health through disrupted care, interrupted education, fractured social connections, and traumatic stress or injury. Uncertainty about these pathways should sharpen the questions we ask, not delay care for those living through conflict.
DISCLOSURES
The author reports no conflicts of interest.
FUNDING
This study received no external funding.