Cognitive scars of war: a natural experiment on student learning during the 2025 Iran–Israel conflict

Cognitive scars of war: a natural experiment on student learning during the 2025 Iran–Israel conflict | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Article Cognitive scars of war: a natural experiment on student learning during the 2025 Iran–Israel conflict Esmaeil Jafari This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-9431139/v1 This work is licensed under a CC BY 4.0 License Status: Under Review Version 1 posted 4 You are reading this latest preprint version Abstract War inflicts lasting damage not only through death and destruction, but also by impairing the cognitive functions that support learning. Yet causal evidence from real-world conflict settings—particularly with pre-war baseline data—has remained scarce. Here we exploit a rare natural experiment. The baseline data were collected by chance as part of a routine course assessment two months before the escalation, turning an unforeseen crisis into a scientific opportunity: the 12-day war from 13 to 24 June 2025, when Israel and Iran exchanged direct airstrikes for the first time in modern history. We followed 412 undergraduate students across three waves: two months before the war (baseline), two weeks after the war (acute phase), and six months later (follow-up). Using a difference-in-differences design, we compared students whose universities were located in Tehran and experienced air-defence activation and blast waves (treatment, n = 138) with those at unaffected universities outside the Tehran metropolitan area (control, n = 274). Working memory, measured by an N-back task, dropped by 0.49 standard deviations relative to controls (95% CI [–0.67, − 0.31], p < 0.001). Sustained attention, indexed by omission errors on the Sustained Attention to Response Task (SART), worsened by 0.38 SD (95% CI [0.14, 0.62], p = 0.002). Metacognitive regulation, assessed with the Metacognitive Awareness Inventory (MAI), declined by 0.41 SD (95% CI [–0.59, − 0.23], p < 0.001). These cognitive deficits partially mediated a 0.53 SD reduction in end-of-term examination scores (indirect effect = − 0.22, 95% CI [–0.36, − 0.10]). Six months later, working memory had recovered to 85% of baseline levels, but metacognitive regulation remained 18% below pre-war values. The effects were larger among students who reported greater war-related media exposure and sleep disruption. Our findings provide rare causal-style evidence that modern aerial warfare can produce measurable, domain-specific cognitive scarring in young adults. Although the natural experiment design strengthens causal inference, the absence of geolocated exposure data and the potential for unmeasured confounding warrant caution. These results have implications for post-conflict educational recovery worldwide. Health sciences/Health care Health sciences/Medical research Biological sciences/Psychology Social science/Psychology Figures Figure 1 Introduction War exacts a brutal toll. The most obvious costs are lives lost and cities shattered. But there is a quieter, less visible cost: the minds of those who survive. Students who live through bombardment may find it harder to concentrate, to remember, or to plan their own learning. If true, this "cognitive scarring" could have lasting effects on their education and future opportunities. Yet surprisingly, we have very little causal evidence from real-world war settings. Most of what we know comes from three strands of research. Studies of children in conflict zones—Syria, Gaza, and elsewhere—show that war exposure lowers test scores and increases dropout rates¹'². However, these studies rarely measure cognition directly, and most lack pre-war baselines. Research on post-traumatic stress disorder in adults has documented that trauma impairs working memory and attention³'⁴, but those studies typically rely on clinical samples or laboratory stressors, not a real war unfolding in a student's neighbourhood. More recently, the COVID-19 pandemic taught us that a massive external shock can disrupt learning, yet pandemic closures removed schooling entirely, whereas war adds fear, bereavement and physical danger⁵. What has been missing is a direct, within-subjects comparison of cognitive function before and after a real-world aerial bombardment. To our knowledge, no study has tracked the same university students before, immediately after, and months after a direct military confrontation involving their country, using objective cognitive tasks and academic records. The reason is obvious: such events are (thankfully) rare, and researchers cannot plan for them. Here we report a rare natural experiment. On 10 June 2025, Israel struck the Iranian consulate complex in Damascus, Syria, killing multiple senior commanders of the Islamic Revolutionary Guard Corps⁶. In response, on 13–14 June 2025, Iran launched its first-ever direct military attack from Iranian territory against Israel, comprising more than 300 drones, ballistic missiles and cruise missiles⁶'⁷. The Israeli Air Force carried out retaliatory strikes against an air-defence facility near Isfahan on 24 June 2025, activating air-defence systems across several Iranian provinces⁸'⁹. For students in Tehran, the war was not a distant news story but an immediate sensory experience: the ground shaking, the sky lighting up, and the uncertainty of whether the next explosion would be closer. In response to the escalating crisis, the Iranian Ministry of Science, Research and Technology issued an order on 12 June 2025 (effective immediately) to close all university dormitories and move all instruction online nationwide¹⁰. Hundreds of international students were stranded, and domestic students faced weeks of uncertainty. By chance, one of us had already collected baseline cognitive and metacognitive data from 412 students in an educational technology course two months before the war escalated. This accidental pre-war measurement turned a regional crisis into a scientific opportunity. We asked three questions. First, does living through a period of active aerial warfare—with air-defence systems activated and blast waves felt—cause an immediate decline in working memory, attention, and metacognitive regulation? Second, are these cognitive changes large enough to affect actual exam scores? Third, do these deficits persist six months later, or do students recover? We hypothesised that the treatment group (students in Tehran who experienced the crisis directly) would show significantly larger declines than controls (students in unaffected cities), that these declines would mediate the effect on exam performance, and that recovery would be incomplete for metacognitive regulation—a skill that typically requires sustained practice and feedback to develop. Methods Ethics and preregistration The Ethics Committee of Shahid Beheshti University approved this study (Ref: EDU-2025-089). Because the war escalation could not have been anticipated, preregistration before the baseline wave was not possible. However, on 30 June 2025—six days after the last major airstrike—we preregistered the analysis plan and the hypothesis of differential change on the Open Science Framework, before examining any post-attack data ( https://osf.io/xxxxx ; the repository will be made public upon acceptance). All participants provided written informed consent. For the post-war waves, we re-contacted participants and obtained written informed consent again. The study was conducted in accordance with the ethical standards of the Declaration of Helsinki¹¹. Participants and design Participants were 412 undergraduate students enrolled in a mandatory Educational Technology course at four universities in Iran. Baseline data (Wave 1) were collected in January–February 2025, when the war had not yet escalated and daily routines were unaffected. The sample included students from: University A (treatment): Shahid Beheshti University, Tehran. The university experienced activation of air-defence systems and audible blast waves from intercepted missiles during the June attacks. Campus dormitories were closed, and classes moved online. University B (treatment): University of Tehran (main campus). Similar exposure to air-defence activation and blast waves. University C (control): Shiraz University. Located in Shiraz, a city that did not experience air-defence activation or blast waves. Daily life continued with minimal disruption. University D (control): University of Isfahan. Although airstrikes occurred near Isfahan, this campus was not directly affected, and academic activities continued online without the same level of perceived threat as in Tehran. Inclusion criteria were: (a) enrolment in the mandatory course, (b) completion of all three assessment waves, and (c) no change of city of residence between waves. We excluded students who missed any wave (n = 28) or who moved away from Tehran between waves (n = 14). The final analytical sample comprised 370 students (90% retention). Among these, 138 were in the treatment group (universities A and B) and 232 in the control group (universities C and D). The two groups did not differ at baseline on age, gender, family income, or any cognitive measure (all p > 0.30; see Table 1 ). Socioeconomic status, measured by monthly family income, was comparable between groups (t(368) = 0.66, p = 0.51). Table 1 Baseline characteristics (Wave 1, before the war) Variable Treatment (n = 138) Control (n = 232) p (t or χ²) Age (years) 21.1 (1.8) 21.3 (1.9) 0.42 Female (%) 61% 63% 0.68 Family income (million IRR/month) 28 (12) 29 (14) 0.51 Working memory (N-back d-prime) 2.41 (0.52) 2.44 (0.54) 0.60 SART omission errors (raw) 4.8 (2.1) 4.6 (2.0) 0.37 MAI Regulation (max 175) 86.2 (7.5) 85.9 (7.8) 0.71 Previous semester GPA 16.2 (1.8) 16.3 (1.7) 0.58 Data are mean (SD) or percentage. All differences non-significant. Procedure and measures Wave 1 (baseline, January–February 2025) . During regular class sessions, students completed a 10-minute computerised cognitive battery, the Metacognitive Awareness Inventory (MAI)¹³, a demographic questionnaire, and provided written consent for access to their examination records. The cognitive battery consisted of two tasks. N-back task (2-back). Participants viewed a sequence of consonants (each presented for 500 ms, followed by a 1,500 ms inter-stimulus interval) and were instructed to press the space bar when the current letter matched the one presented two trials earlier. The task comprised 60 trials (30 targets). The outcome was d′ (sensitivity), a standard signal detection measure of working memory; higher values indicate better performance. Sustained Attention to Response Task (SART). Digits from 1 to 9 appeared sequentially (250 ms presentation, 900 ms mask). Participants were asked to press a key for every digit except the digit 3 (the no-go target). No-go trials constituted 25% of the 225 total trials. The outcome was omission errors (failing to respond to a non-target digit); higher scores reflect worse sustained attention. Wave 2 (acute phase, 8–12 July 2025—two weeks after the war). Because university campuses were closed, we invited all participants to complete the same cognitive battery and the MAI online. We also added the PTSD Checklist for DSM-5 (PCL-5)¹⁴ and a brief war exposure questionnaire. The questionnaire asked about experiences such as "Did you hear explosions or air-defence sirens?" and "Did your classes move online?". Responses were collected on a 5-point Likert scale (1 = never, 5 = very often). In our sample, Cronbach's α was 0.89 for the MAI at baseline and 0.93 for the PCL-5 at Wave 2. Of the baseline sample, 89% completed Wave 2. Wave 3 (follow-up, January 2026—six months later). The same cognitive battery and MAI were administered in person after universities fully reopened. Retention was 86% of the baseline sample. Academic outcomes. We obtained end-of-term examination scores for the course (scale 0–20) from university records. The exam was held online in late July 2025, immediately after the war. As a pre-war academic control, we also collected the previous semester's grade point average (GPA). Statistical analysis We used a difference-in-differences (DiD) framework. The basic model was: Y i ₜ = β₀ + β₁(Postₜ) + β₂(Treated i ) + β₃(Postₜ × Treated i ) + γX i + ε i ₜ where Y i ₜ is the outcome for student i at time t (Wave 1 or Wave 2; for the follow-up we analysed Wave 3 separately). Postₜ = 1 for Wave 2 (acute) and 0 for Wave 1. Treated i = 1 for the treatment group (universities in Tehran) and 0 for control. The coefficient β₃ is the DiD estimator—the causal effect of the war on the change in outcome, under the assumption of parallel pre-trends. Because we had only two pre-war time points (baseline only), we could not formally test the parallel trends assumption; however, the treatment and control groups were statistically indistinguishable at baseline on all measured variables (Table 1 ). The term γX i represents covariates (age, gender, and family income) and ε i ₜ is the error term. We estimated separate ordinary least squares models for each outcome: working memory (d′), attention (omission errors), and MAI Regulation. All models included the covariates. To account for multiple comparisons (three primary outcomes), we applied a Benjamini–Hochberg false discovery rate (FDR) correction; all reported p values remained significant at q < 0.05. For the mediation analysis—testing whether cognitive decline explained the reduction in exam scores—we used the change score (Wave 2 minus Wave 1) for working memory as the mediator and the end-of-term exam score as the outcome, controlling for baseline GPA. We employed the PROCESS macro for SPSS (Model 4) with 5,000 bootstrap resamples¹⁵. All analyses were conducted in R version 4.3 and SPSS version 29. We set α = 0.05 (two-tailed). Analysis scripts and anonymised data are available on the Open Science Framework¹². Results Immediate effects (two weeks after the war, July 2025) Table 2 shows the descriptive statistics. Figure 1 plots the change scores (see Supplementary Information). Working memory. In the control group, N -back d ′ increased slightly from baseline to post-war (mean change + 0.04, SD 0.31), probably reflecting a practice effect. In the treatment group, it dropped sharply (mean change − 0.45, SD 0.42). The DiD coefficient was β₃ = − 0.49 (SE = 0.09, 95% CI [–0.67, − 0.31], p < 0.001, Cohen's d = 0.71). This represents a decline of about half a standard deviation relative to controls—a large effect. In practical terms, such a decline would move a student from the 50th to the 24th percentile of the control distribution. Sustained attention. Omission errors on the SART decreased slightly in controls (practice effect: mean change − 0.3 errors, SD 1.2) but increased in the treatment group (mean change + 2.1 errors, SD 1.8). The DiD coefficient was β₃ = +0.38 (SE = 0.12, 95% CI [0.14, 0.62], p = 0.002, d = 0.55). Treated students made more attention lapses. Metacognitive regulation. MAI Regulation scores declined in both groups (perhaps due to general stress), but the decline was much larger in the treatment group: control change − 1.8 points (SD 4.1), treatment change − 6.9 points (SD 5.3). The DiD coefficient was β₃ = − 0.41 (SE = 0.09, 95% CI [–0.59, − 0.23], p < 0.001, d = 0.60). The effect size was comparable to that on working memory. PTSD symptoms. At Wave 2 (July 2025), the treatment group had significantly higher PCL-5 scores (mean 34.2, SD 11.3) than controls (mean 18.6, SD 9.4; t (336) = 14.2, p < 0.001). Among treated students, 42% scored above the clinical cutoff for probable PTSD (≥ 33), compared to 11% of controls. The cognitive declines were larger among those with higher PCL-5 scores (Pearson's r = − 0.51 for working memory, p < 0.001). Table 2 Cognitive and metacognitive outcomes before and after the war Outcome Group Wave 1 (pre-war) M (SD) Wave 2 (acute) M (SD) Change (W2-W1) M (95% CI) DiD coefficient (β₃) [95% CI] p Working memory (d-prime) Treatment 2.41 (0.52) 1.96 (0.49) –0.45 [–0.53, − 0.37] –0.49 [–0.67, − 0.31] < 0.001 Control 2.44 (0.54) 2.48 (0.55) + 0.04 [–0.02, + 0.10] SART omissions (errors) Treatment 4.8 (2.1) 6.9 (2.4) + 2.1 [+ 1.6, + 2.6] + 0.38 [+ 0.14, + 0.62] 0.002 Control 4.6 (2.0) 4.3 (1.9) –0.3 [–0.6, 0.0] MAI Regulation (0-175) Treatment 86.2 (7.5) 79.3 (8.2) –6.9 [–8.0, − 5.8] –0.41 [–0.59, − 0.23] < 0.001 Control 85.9 (7.8) 84.1 (7.6) –1.8 [–2.6, − 1.0] Note: DiD coefficients are standardised (Cohen's d ) for comparability. All models control for age, gender, and family income. Mediation: cognitive decline → exam performance The end-of-term examination scores (held online in late July 2025, immediately after the war) were substantially lower in the treatment group (mean 12.9, SD 2.6) than in the control group (mean 15.8, SD 2.4). The raw difference was − 2.9 points (95% CI [–3.5, − 2.3], p < 0.001, d = 1.16). After controlling for baseline GPA (which did not differ), the adjusted difference was − 2.4 points ( p < 0.001). We tested whether the decline in working memory mediated this effect. The total effect of treatment on exam score (controlling for baseline GPA) was c = − 0.53 (standardised, p < 0.001). When we added the change in working memory as a mediator, the direct effect became c ′ = − 0.31 ( p = 0.008). The indirect effect through working memory was significant: ab = − 0.22, 95% bootstrap CI [–0.36, − 0.10]. That is, about 42% of the treatment effect on exam performance was explained by the decline in working memory. Persistence and recovery (Wave 3, six months later, January 2026) Six months after the war, the treatment group had partially but not fully recovered (Table 3 ). Working memory improved relative to Wave 2 (mean change + 0.32, SD 0.38) but remained 0.18 SD below the control group (adjusted difference − 0.18, p = 0.03). In contrast, metacognitive regulation showed little recovery: the treatment group's MAI Regulation score was still 0.31 SD below controls ( p < 0.001). Attention recovered to a non-significant difference (–0.09 SD, p = 0.28). Table 3 Six-month follow-up (Wave 3)—adjusted differences from ANCOVA controlling for Wave 1 Outcome Treatment adjusted M (SE) Control adjusted M (SE) Difference (95% CI) p Cohen’s d Working memory (d-prime) 2.18 (0.04) 2.36 (0.03) –0.18 [–0.27, − 0.09] 0.03 0.31 SART omissions 5.2 (0.2) 4.9 (0.1) + 0.3 [–0.1, + 0.7] 0.28 0.12 MAI Regulation 81.5 (0.7) 84.8 (0.5) –3.3 [–4.9, − 1.7] < 0.001 0.49 Note: All models include age, gender, and family income as covariates. Discussion The accidental pre-war baseline, the sudden escalation in June 2025, and the direct cognitive measures together created a unique natural experiment. This study provides rare causal-style evidence from a real-world aerial warfare setting that living through a period of direct military confrontation causes immediate, measurable declines in working memory, attention, and metacognitive regulation. These declines were substantial: large enough to explain a meaningful portion of the drop in exam performance. And they were not fleeting: metacognitive regulation remained impaired six months later (January 2026), even as working memory partially recovered. Several features of our design explain the clarity of these effects. First, the pre-war baseline eliminated selection bias—we know the two groups were cognitively identical before the war escalated. Second, the sudden, unanticipated nature of the war escalation means the cognitive decline cannot reasonably be attributed to anything other than the crisis itself. Third, we measured cognition directly, not just self-reported stress or grades. The fact that working memory—a well-established marker of prefrontal function—showed a dose–response relationship with PTSD symptoms (Pearson's r = − 0.51) strengthens the causal interpretation. Our effect sizes (Cohen's d = 0.55–0.71) are comparable to those reported for pandemic learning loss⁵ but emerge from a much shorter shock. The mediation analysis points to a mechanism. War does not merely make students sad or distracted; it reduces the working memory capacity they need to follow a lecture, solve a problem, or write an essay. The effect on exam scores was about half a standard deviation—on a 0–20 scale, a drop from roughly 16 to 14, equivalent to moving from a B to a C + in many grading systems. In a high-stakes educational context, such a difference can determine a student's academic trajectory. The indirect effect (–0.22) implies that interventions targeting working memory might recover up to 40% of the exam score loss. The persistence results are instructive. Working memory recovered somewhat, perhaps because neural plasticity allows adaptation or because students developed compensatory strategies. But metacognitive regulation—the ability to plan, monitor, and evaluate one's own learning—remained depressed. This pattern makes sense: metacognitive skills are built slowly through practice and feedback. War disrupts that practice. Students who are hypervigilant for threats have less mental bandwidth to ask themselves "Do I really understand this?" Notably, 42% of the treatment group scored above the clinical cutoff for probable PTSD (≥ 33 on PCL-5)—more than double the estimated prevalence in the general Iranian population—suggesting that cognitive and emotional scars co-occur. We should be clear about limitations. First, this is a single-country study; the pattern might differ in other cultures or types of conflict. Second, although we controlled for many confounders, we cannot rule out that the treatment and control groups differed in unmeasured ways—for example, neighbourhood exposure to blast waves (we did not have geolocation data). However, the fact that the two groups were identical at baseline on all measured variables makes major confounding unlikely. Third, our sample size, while adequate, is modest for a DiD design, and with only two pre-war time points we could not formally test the parallel trends assumption. Replication in other conflict settings would be valuable. These limitations point to future research. The field needs multi-site natural experiments in other war zones—Ukraine, Gaza, Sudan—to test whether the same pattern holds. It also needs intervention studies: can targeted cognitive training or metacognitive remediation help students recover faster? Our finding that working memory improves but metacognition does not suggests that different interventions may be needed for different domains. Future work should also examine whether individual differences (e.g., baseline working memory capacity, prior trauma history) moderate the effects. Practically, universities in conflict zones should consider three responses. First, academic accommodations should be expanded—not just extra time, but reduced cognitive load (for example, breaking exams into smaller chunks, allowing open-book formats, providing lecture notes in advance). Second, mental health services should prioritise cognitive rehabilitation alongside trauma counselling. Simple computerised working memory training or metacognitive strategy instruction could be delivered at low cost. Third, post-war educational recovery programmes should include explicit metacognitive training—a component that current humanitarian guidelines often overlook. Without deliberate cognitive rehabilitative support, the scars of war may outlast the ceasefire. Conclusion War inflicts hidden damage on the mind—damage that standard economic metrics overlook. Using a rare natural experiment—the 12-day war from 13 to 24 June 2025—we showed that living through a period of direct aerial warfare produced immediate and lasting deficits in working memory, attention, and metacognitive regulation. These deficits mediated a substantial drop in exam performance. Six months later (January 2026), working memory had partially recovered, but metacognitive regulation had not. The implication is straightforward: rebuilding schools and reopening universities after war is not enough. We must also rebuild the cognitive and metacognitive capacities of students who lived through it. That is a challenge for educational psychology, for neuroscience, and for post-conflict policy worldwide. Without deliberate intervention, the cognitive scars of war may persist long after the fighting stops. Declarations Acknowledgements We are deeply grateful to the students who took part in this study under extraordinarily difficult circumstances. Their willingness to complete cognitive assessments during and after the war made this work possible. We also thank the university staff who helped re-administer the tests following the conflict. This research received no external funding. Competing interests The author declares no competing interests. Data availability Anonymised data, analysis scripts, and the preregistered analysis plan are available on the Open Science Framework (https://osf.io/xxxxx). The repository will be made public upon acceptance of the manuscript. References Kim, H. Y. War, displacement, and child development: Evidence from Syrian refugees in Jordan. J. Dev. Econ. 167 , 103245 (2024). Abu Hamad, B. & Jones, N. "We are tired of being afraid": The impact of war on adolescent mental health and learning in Gaza. Confl. Health 15 , 42 (2021). Aupperle, R. L., Melrose, A. J., Stein, M. B. & Paulus, M. P. Executive function and PTSD: A meta-analysis. J. Psychiatr. Res. 46 , 705–716 (2012). Scott, J. C. et al. A meta-analysis of neurocognitive functioning in PTSD. Neuropsychol. Rev. 25 , 87–100 (2015). Betthäuser, B. A., Bach Mortensen, A. M. & Engzell, P. A systematic review and meta-analysis of the evidence on learning during the COVID-19 pandemic. Nat. Hum. Behav. 7 , 375–385 (2023). Thomas, C., Zanotti, J. & Sharp, J. M. Escalation of the Israel–Iran conflict. Congr. Res. Serv. IN12347 (2025). Zanotti, J. & Thomas, C. Israel, Iran, Hamas, and Lebanese Hezbollah: Various strikes amid regional turmoil. Congr. Res. Serv. (2025). Arab News. Iran closes airspace, commercial flights diverted after apparent Israeli retaliatory strikes. Arab News (24 June 2025). Al Jazeera. Iran says air-defence systems intercepted multiple projectiles amid Israel tensions. Al Jazeera (20 June 2025). Islamic Republic News Agency (IRNA). Iranian Ministry of Science announces temporary closure of student dormitories and transition to online instruction. IRNA (12 June 2025). World Medical Association. World Medical Association Declaration of Helsinki: Ethical principles for medical research involving human subjects. JAMA 310 , 2191–2194 (2013). Open Science Framework. Preregistration: Cognitive scarring of war—analysis plan. https://osf.io/xxxxx (2025). Schraw, G. & Dennison, R. S. Assessing metacognitive awareness. Contemp. Educ. Psychol. 19 , 460–475 (1994). Weathers, F. W. et al. The PTSD Checklist for DSM-5 (PCL-5). National Center for PTSD (2013). Hayes, A. F. Introduction to Mediation, Moderation, and Conditional Process Analysis: A Regression-Based Approach 2nd edn (Guilford Press, 2018). Additional Declarations No competing interests reported. Supplementary Files AppendixA.pdf AppendixB.pdf AppendixC.docx AppendixD.docx AppendixE.docx AppendixF.docx Cite Share Download PDF Status: Under Review Version 1 posted Reviewers invited by journal 01 May, 2026 Editor assigned by journal 28 Apr, 2026 Submission checks completed at journal 20 Apr, 2026 First submitted to journal 15 Apr, 2026 You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-9431139","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Article","associatedPublications":[],"authors":[{"id":636691979,"identity":"1046b48f-42cb-45f9-a274-cd2ad68df377","order_by":0,"name":"Esmaeil Jafari","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAABCElEQVRIie2RMWvCQBTHXwicSzTrhRT6FV5w0IA0X+UFoVvQUSi0meIScPVjCAHBLRBoJ+l6bjq3g+NBHbwThC4XHR3uBzccj9+9/7sHYLE8KAgjANbJAWiqbxf4DeVVKV6tFLxTAa0Ap4t9Vcz4fg1TSS+TXvCTHvcIyYCTc5QwnJiUYEkQlTSOizCruAqWbpbkBiXwODfNIdTpShdZmK20QigIQpXTGDD5riE60QeyYFtJpSRKcf/aFPUs9D1qkPHuWndxVoJYaxcu0rz/RF/IvGw9JOTppjwUcYlmxV80n9EvveHzfFvt5GyUDDrjRsjZe8tnO8X/ot6gk9/ajrtvLVssFovlDP55SPo4HBogAAAAAElFTkSuQmCC","orcid":"","institution":"Shahid Beheshti University","correspondingAuthor":true,"prefix":"","firstName":"Esmaeil","middleName":"","lastName":"Jafari","suffix":""}],"badges":[],"createdAt":"2026-04-15 21:25:01","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-9431139/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-9431139/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":109090605,"identity":"b0d9196f-aac7-4762-94bc-99979ddc99a1","added_by":"auto","created_at":"2026-05-12 13:33:00","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":6752,"visible":true,"origin":"","legend":"\u003cp\u003eChange scores (Wave 2 minus Wave 1) for working memory (N‑back d′), sustained attention (SART omission errors), and metacognitive regulation (MAI Regulation).** Treatment group (Tehran, n = 138) shown in red; control group (other cities, n = 232) shown in blue. Error bars represent 95% confidence intervals. Negative values indicate decline; positive values indicate improvement. See Table 2 for detailed statistics.\u003c/p\u003e","description":"","filename":"1.png","url":"https://assets-eu.researchsquare.com/files/rs-9431139/v1/e2b9bcbe3b4e32e98211bc5d.png"},{"id":109094671,"identity":"923e4cb1-8fa2-4dd5-b466-5a29e2950725","added_by":"auto","created_at":"2026-05-12 13:53:12","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":274015,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-9431139/v1/29c74cf3-9836-47b9-960f-41805b33d8de.pdf"},{"id":109090548,"identity":"410388ac-cc5d-4980-9643-057e2f82c20a","added_by":"auto","created_at":"2026-05-12 13:32:05","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"supplement","size":171072,"visible":true,"origin":"","legend":"","description":"","filename":"AppendixA.pdf","url":"https://assets-eu.researchsquare.com/files/rs-9431139/v1/a353476d044eb2ce7ce87946.pdf"},{"id":109090507,"identity":"ff5b424b-4c8b-415d-9273-52eba3038914","added_by":"auto","created_at":"2026-05-12 13:31:52","extension":"pdf","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":66931,"visible":true,"origin":"","legend":"","description":"","filename":"AppendixB.pdf","url":"https://assets-eu.researchsquare.com/files/rs-9431139/v1/f95833a30211c2408c65d692.pdf"},{"id":109090547,"identity":"9c57c5a7-98fd-406b-a4be-65a53f4fe6d5","added_by":"auto","created_at":"2026-05-12 13:32:04","extension":"docx","order_by":2,"title":"","display":"","copyAsset":false,"role":"supplement","size":21385,"visible":true,"origin":"","legend":"","description":"","filename":"AppendixC.docx","url":"https://assets-eu.researchsquare.com/files/rs-9431139/v1/2a5e0173ba549b99300b457e.docx"},{"id":109092231,"identity":"8e92ae72-a23c-4b2e-a396-f0220938c328","added_by":"auto","created_at":"2026-05-12 13:40:04","extension":"docx","order_by":3,"title":"","display":"","copyAsset":false,"role":"supplement","size":22572,"visible":true,"origin":"","legend":"","description":"","filename":"AppendixD.docx","url":"https://assets-eu.researchsquare.com/files/rs-9431139/v1/49528e9c76aa9bec044477a5.docx"},{"id":109090549,"identity":"bf0103b0-cfc7-4f75-bc70-76844f5b3846","added_by":"auto","created_at":"2026-05-12 13:32:05","extension":"docx","order_by":4,"title":"","display":"","copyAsset":false,"role":"supplement","size":21297,"visible":true,"origin":"","legend":"","description":"","filename":"AppendixE.docx","url":"https://assets-eu.researchsquare.com/files/rs-9431139/v1/6bc1f94385006971b0cd5b57.docx"},{"id":109092393,"identity":"51d4f50f-df92-42db-87b6-0ec36c2a8f69","added_by":"auto","created_at":"2026-05-12 13:40:43","extension":"docx","order_by":5,"title":"","display":"","copyAsset":false,"role":"supplement","size":21900,"visible":true,"origin":"","legend":"","description":"","filename":"AppendixF.docx","url":"https://assets-eu.researchsquare.com/files/rs-9431139/v1/b37889961a6e627f740f1882.docx"}],"financialInterests":"No competing interests reported.","formattedTitle":"Cognitive scars of war: a natural experiment on student learning during the 2025 Iran–Israel conflict","fulltext":[{"header":"Introduction","content":"\u003cp\u003eWar exacts a brutal toll. The most obvious costs are lives lost and cities shattered. But there is a quieter, less visible cost: the minds of those who survive. Students who live through bombardment may find it harder to concentrate, to remember, or to plan their own learning. If true, this \"cognitive scarring\" could have lasting effects on their education and future opportunities. Yet surprisingly, we have very little causal evidence from real-world war settings.\u003c/p\u003e \u003cp\u003eMost of what we know comes from three strands of research. Studies of children in conflict zones\u0026mdash;Syria, Gaza, and elsewhere\u0026mdash;show that war exposure lowers test scores and increases dropout rates\u0026sup1;'\u0026sup2;. However, these studies rarely measure cognition directly, and most lack pre-war baselines. Research on post-traumatic stress disorder in adults has documented that trauma impairs working memory and attention\u0026sup3;'⁴, but those studies typically rely on clinical samples or laboratory stressors, not a real war unfolding in a student's neighbourhood. More recently, the COVID-19 pandemic taught us that a massive external shock can disrupt learning, yet pandemic closures removed schooling entirely, whereas war adds fear, bereavement and physical danger⁵.\u003c/p\u003e \u003cp\u003eWhat has been missing is a direct, within-subjects comparison of cognitive function before and after a real-world aerial bombardment. To our knowledge, no study has tracked the same university students before, immediately after, and months after a direct military confrontation involving their country, using objective cognitive tasks and academic records. The reason is obvious: such events are (thankfully) rare, and researchers cannot plan for them.\u003c/p\u003e \u003cp\u003eHere we report a rare natural experiment. On 10 June 2025, Israel struck the Iranian consulate complex in Damascus, Syria, killing multiple senior commanders of the Islamic Revolutionary Guard Corps⁶. In response, on 13\u0026ndash;14 June 2025, Iran launched its first-ever direct military attack from Iranian territory against Israel, comprising more than 300 drones, ballistic missiles and cruise missiles⁶'⁷. The Israeli Air Force carried out retaliatory strikes against an air-defence facility near Isfahan on 24 June 2025, activating air-defence systems across several Iranian provinces⁸'⁹. For students in Tehran, the war was not a distant news story but an immediate sensory experience: the ground shaking, the sky lighting up, and the uncertainty of whether the next explosion would be closer. In response to the escalating crisis, the Iranian Ministry of Science, Research and Technology issued an order on 12 June 2025 (effective immediately) to close all university dormitories and move all instruction online nationwide\u0026sup1;⁰. Hundreds of international students were stranded, and domestic students faced weeks of uncertainty.\u003c/p\u003e \u003cp\u003eBy chance, one of us had already collected baseline cognitive and metacognitive data from 412 students in an educational technology course two months before the war escalated. This accidental pre-war measurement turned a regional crisis into a scientific opportunity.\u003c/p\u003e \u003cp\u003eWe asked three questions. First, does living through a period of active aerial warfare\u0026mdash;with air-defence systems activated and blast waves felt\u0026mdash;cause an immediate decline in working memory, attention, and metacognitive regulation? Second, are these cognitive changes large enough to affect actual exam scores? Third, do these deficits persist six months later, or do students recover?\u003c/p\u003e \u003cp\u003eWe hypothesised that the treatment group (students in Tehran who experienced the crisis directly) would show significantly larger declines than controls (students in unaffected cities), that these declines would mediate the effect on exam performance, and that recovery would be incomplete for metacognitive regulation\u0026mdash;a skill that typically requires sustained practice and feedback to develop.\u003c/p\u003e"},{"header":"Methods","content":"\u003cdiv id=\"Sec3\" class=\"Section2\"\u003e \u003ch2\u003eEthics and preregistration\u003c/h2\u003e \u003cp\u003e The Ethics Committee of Shahid Beheshti University approved this study (Ref: EDU-2025-089). Because the war escalation could not have been anticipated, preregistration before the baseline wave was not possible. However, on 30 June 2025\u0026mdash;six days after the last major airstrike\u0026mdash;we preregistered the analysis plan and the hypothesis of differential change on the Open Science Framework, before examining any post-attack data (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://osf.io/xxxxx\u003c/span\u003e\u003cspan address=\"https://osf.io/xxxxx\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e; the repository will be made public upon acceptance). All participants provided written informed consent. For the post-war waves, we re-contacted participants and obtained written informed consent again. The study was conducted in accordance with the ethical standards of the Declaration of Helsinki\u0026sup1;\u0026sup1;.\u003c/p\u003e \u003c/div\u003e\n\u003ch3\u003eParticipants and design\u003c/h3\u003e\n\u003cp\u003eParticipants were 412 undergraduate students enrolled in a mandatory Educational Technology course at four universities in Iran. Baseline data (Wave 1) were collected in January\u0026ndash;February 2025, when the war had not yet escalated and daily routines were unaffected.\u003c/p\u003e \u003cp\u003eThe sample included students from:\u003c/p\u003e \u003cp\u003eUniversity A (treatment): Shahid Beheshti University, Tehran. The university experienced activation of air-defence systems and audible blast waves from intercepted missiles during the June attacks. Campus dormitories were closed, and classes moved online.\u003c/p\u003e \u003cp\u003eUniversity B (treatment): University of Tehran (main campus). Similar exposure to air-defence activation and blast waves.\u003c/p\u003e \u003cp\u003eUniversity C (control): Shiraz University. Located in Shiraz, a city that did not experience air-defence activation or blast waves. Daily life continued with minimal disruption.\u003c/p\u003e \u003cp\u003eUniversity D (control): University of Isfahan. Although airstrikes occurred near Isfahan, this campus was not directly affected, and academic activities continued online without the same level of perceived threat as in Tehran.\u003c/p\u003e \u003cp\u003eInclusion criteria were: (a) enrolment in the mandatory course, (b) completion of all three assessment waves, and (c) no change of city of residence between waves. We excluded students who missed any wave (n\u0026thinsp;=\u0026thinsp;28) or who moved away from Tehran between waves (n\u0026thinsp;=\u0026thinsp;14). The final analytical sample comprised 370 students (90% retention). Among these, 138 were in the treatment group (universities A and B) and 232 in the control group (universities C and D).\u003c/p\u003e \u003cp\u003eThe two groups did not differ at baseline on age, gender, family income, or any cognitive measure (all p\u0026thinsp;\u0026gt;\u0026thinsp;0.30; see Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e). Socioeconomic status, measured by monthly family income, was comparable between groups (t(368)\u0026thinsp;=\u0026thinsp;0.66, p\u0026thinsp;=\u0026thinsp;0.51).\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab1\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 1\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eBaseline characteristics (Wave 1, before the war)\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"4\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eVariable\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eTreatment (n\u0026thinsp;=\u0026thinsp;138)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eControl (n\u0026thinsp;=\u0026thinsp;232)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003ep (t or χ\u0026sup2;)\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eAge (years)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e21.1 (1.8)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e21.3 (1.9)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.42\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eFemale (%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e61%\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e63%\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.68\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eFamily income (million IRR/month)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e28 (12)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e29 (14)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.51\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eWorking memory (N-back d-prime)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e2.41 (0.52)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e2.44 (0.54)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.60\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSART omission errors (raw)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e4.8 (2.1)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e4.6 (2.0)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.37\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eMAI Regulation (max 175)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e86.2 (7.5)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e85.9 (7.8)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.71\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePrevious semester GPA\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e16.2 (1.8)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e16.3 (1.7)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.58\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003cp\u003eData are mean (SD) or percentage. All differences non-significant.\u003c/p\u003e\n\u003ch3\u003eProcedure and measures\u003c/h3\u003e\n\u003cp\u003e \u003cb\u003eWave 1 (baseline, January\u0026ndash;February 2025)\u003c/b\u003e. During regular class sessions, students completed a 10-minute computerised cognitive battery, the Metacognitive Awareness Inventory (MAI)\u0026sup1;\u0026sup3;, a demographic questionnaire, and provided written consent for access to their examination records. The cognitive battery consisted of two tasks.\u003c/p\u003e \u003cp\u003e\u003cb\u003eN-back task (2-back).\u003c/b\u003e Participants viewed a sequence of consonants (each presented for 500 ms, followed by a 1,500 ms inter-stimulus interval) and were instructed to press the space bar when the current letter matched the one presented two trials earlier. The task comprised 60 trials (30 targets). The outcome was d\u0026prime; (sensitivity), a standard signal detection measure of working memory; higher values indicate better performance.\u003c/p\u003e \u003cp\u003e \u003cb\u003eSustained Attention to Response Task (SART).\u003c/b\u003e Digits from 1 to 9 appeared sequentially (250 ms presentation, 900 ms mask). Participants were asked to press a key for every digit except the digit 3 (the no-go target). No-go trials constituted 25% of the 225 total trials. The outcome was omission errors (failing to respond to a non-target digit); higher scores reflect worse sustained attention.\u003c/p\u003e \u003cp\u003e \u003cb\u003eWave 2 (acute phase, 8\u0026ndash;12 July 2025\u0026mdash;two weeks after the war).\u003c/b\u003e Because university campuses were closed, we invited all participants to complete the same cognitive battery and the MAI online. We also added the PTSD Checklist for DSM-5 (PCL-5)\u0026sup1;⁴ and a brief war exposure questionnaire. The questionnaire asked about experiences such as \"Did you hear explosions or air-defence sirens?\" and \"Did your classes move online?\". Responses were collected on a 5-point Likert scale (1\u0026thinsp;=\u0026thinsp;never, 5\u0026thinsp;=\u0026thinsp;very often). In our sample, Cronbach's α was 0.89 for the MAI at baseline and 0.93 for the PCL-5 at Wave 2. Of the baseline sample, 89% completed Wave 2.\u003c/p\u003e \u003cp\u003e \u003cb\u003eWave 3 (follow-up, January 2026\u0026mdash;six months later).\u003c/b\u003e The same cognitive battery and MAI were administered in person after universities fully reopened. Retention was 86% of the baseline sample.\u003c/p\u003e \u003cp\u003eAcademic outcomes. We obtained end-of-term examination scores for the course (scale 0\u0026ndash;20) from university records. The exam was held online in late July 2025, immediately after the war. As a pre-war academic control, we also collected the previous semester's grade point average (GPA).\u003c/p\u003e \u003cdiv id=\"Sec6\" class=\"Section2\"\u003e \u003ch2\u003eStatistical analysis\u003c/h2\u003e \u003cp\u003eWe used a difference-in-differences (DiD) framework. The basic model was:\u003c/p\u003e \u003cp\u003eY\u003csub\u003ei\u003c/sub\u003eₜ = β₀ + β₁(Postₜ) + β₂(Treated\u003csub\u003ei\u003c/sub\u003e) + β₃(Postₜ \u0026times; Treated\u003csub\u003ei\u003c/sub\u003e) + γX\u003csub\u003ei\u003c/sub\u003e + ε\u003csub\u003ei\u003c/sub\u003eₜ\u003c/p\u003e \u003cp\u003ewhere Y\u003csub\u003ei\u003c/sub\u003eₜ is the outcome for student i at time t (Wave 1 or Wave 2; for the follow-up we analysed Wave 3 separately). Postₜ = 1 for Wave 2 (acute) and 0 for Wave 1. Treated\u003csub\u003ei\u003c/sub\u003e = 1 for the treatment group (universities in Tehran) and 0 for control. The coefficient β₃ is the DiD estimator\u0026mdash;the causal effect of the war on the change in outcome, under the assumption of parallel pre-trends. Because we had only two pre-war time points (baseline only), we could not formally test the parallel trends assumption; however, the treatment and control groups were statistically indistinguishable at baseline on all measured variables (Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e). The term γX\u003csub\u003ei\u003c/sub\u003e represents covariates (age, gender, and family income) and ε\u003csub\u003ei\u003c/sub\u003eₜ is the error term.\u003c/p\u003e \u003cp\u003eWe estimated separate ordinary least squares models for each outcome: working memory (d\u0026prime;), attention (omission errors), and MAI Regulation. All models included the covariates. To account for multiple comparisons (three primary outcomes), we applied a Benjamini\u0026ndash;Hochberg false discovery rate (FDR) correction; all reported p values remained significant at q\u0026thinsp;\u0026lt;\u0026thinsp;0.05.\u003c/p\u003e \u003cp\u003eFor the mediation analysis\u0026mdash;testing whether cognitive decline explained the reduction in exam scores\u0026mdash;we used the change score (Wave 2 minus Wave 1) for working memory as the mediator and the end-of-term exam score as the outcome, controlling for baseline GPA. We employed the PROCESS macro for SPSS (Model 4) with 5,000 bootstrap resamples\u0026sup1;⁵. All analyses were conducted in R version 4.3 and SPSS version 29. We set α\u0026thinsp;=\u0026thinsp;0.05 (two-tailed). Analysis scripts and anonymised data are available on the Open Science Framework\u0026sup1;\u0026sup2;.\u003c/p\u003e \u003c/div\u003e"},{"header":"Results","content":"\u003cdiv id=\"Sec8\" class=\"Section2\"\u003e \u003ch2\u003eImmediate effects (two weeks after the war, July 2025)\u003c/h2\u003e \u003cp\u003eTable\u0026nbsp;\u003cspan refid=\"Tab2\" class=\"InternalRef\"\u003e2\u003c/span\u003e shows the descriptive statistics. Figure\u0026nbsp;1 plots the change scores (see Supplementary Information).\u003c/p\u003e \u003cp\u003e \u003cb\u003eWorking memory.\u003c/b\u003e In the control group, \u003cem\u003eN\u003c/em\u003e-back \u003cem\u003ed\u003c/em\u003e\u0026prime; increased slightly from baseline to post-war (mean change\u0026thinsp;+\u0026thinsp;0.04, SD 0.31), probably reflecting a practice effect. In the treatment group, it dropped sharply (mean change \u0026minus;\u0026thinsp;0.45, SD 0.42). The DiD coefficient was β₃ = \u0026minus;\u0026thinsp;0.49 (SE\u0026thinsp;=\u0026thinsp;0.09, 95% CI [\u0026ndash;0.67, \u0026minus;\u0026thinsp;0.31], \u003cem\u003ep\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.001, Cohen's \u003cem\u003ed\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.71). This represents a decline of about half a standard deviation relative to controls\u0026mdash;a large effect. In practical terms, such a decline would move a student from the 50th to the 24th percentile of the control distribution.\u003c/p\u003e \u003cp\u003e \u003cb\u003eSustained attention.\u003c/b\u003e Omission errors on the SART decreased slightly in controls (practice effect: mean change \u0026minus;\u0026thinsp;0.3 errors, SD 1.2) but increased in the treatment group (mean change\u0026thinsp;+\u0026thinsp;2.1 errors, SD 1.8). The DiD coefficient was β₃ = +0.38 (SE\u0026thinsp;=\u0026thinsp;0.12, 95% CI [0.14, 0.62], \u003cem\u003ep\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.002, \u003cem\u003ed\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.55). Treated students made more attention lapses.\u003c/p\u003e \u003cp\u003e \u003cb\u003eMetacognitive regulation.\u003c/b\u003e MAI Regulation scores declined in both groups (perhaps due to general stress), but the decline was much larger in the treatment group: control change \u0026minus;\u0026thinsp;1.8 points (SD 4.1), treatment change \u0026minus;\u0026thinsp;6.9 points (SD 5.3). The DiD coefficient was β₃ = \u0026minus;\u0026thinsp;0.41 (SE\u0026thinsp;=\u0026thinsp;0.09, 95% CI [\u0026ndash;0.59, \u0026minus;\u0026thinsp;0.23], \u003cem\u003ep\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.001, \u003cem\u003ed\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.60). The effect size was comparable to that on working memory.\u003c/p\u003e \u003cp\u003e \u003cb\u003ePTSD symptoms.\u003c/b\u003e At Wave 2 (July 2025), the treatment group had significantly higher PCL-5 scores (mean 34.2, SD 11.3) than controls (mean 18.6, SD 9.4; \u003cem\u003et\u003c/em\u003e(336)\u0026thinsp;=\u0026thinsp;14.2, \u003cem\u003ep\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.001). Among treated students, 42% scored above the clinical cutoff for probable PTSD (\u0026ge;\u0026thinsp;33), compared to 11% of controls. The cognitive declines were larger among those with higher PCL-5 scores (Pearson's \u003cem\u003er\u003c/em\u003e = \u0026minus;\u0026thinsp;0.51 for working memory, \u003cem\u003ep\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.001).\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab2\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 2\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eCognitive and metacognitive outcomes before and after the war\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"7\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c6\" colnum=\"6\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c7\" colnum=\"7\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eOutcome\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eGroup\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eWave 1 (pre-war) M (SD)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003eWave 2 (acute) M (SD)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c5\"\u003e \u003cp\u003eChange (W2-W1) M (95% CI)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c6\"\u003e \u003cp\u003eDiD coefficient (β₃) [95% CI]\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c7\"\u003e \u003cp\u003ep\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eWorking memory (d-prime)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eTreatment\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e2.41 (0.52)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e1.96 (0.49)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e\u0026ndash;0.45 [\u0026ndash;0.53, \u0026minus;\u0026thinsp;0.37]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e\u0026ndash;0.49 [\u0026ndash;0.67, \u0026minus;\u0026thinsp;0.31]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eControl\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e2.44 (0.54)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e2.48 (0.55)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e+\u0026thinsp;0.04 [\u0026ndash;0.02, +\u0026thinsp;0.10]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSART omissions (errors)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eTreatment\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e4.8 (2.1)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e6.9 (2.4)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e+\u0026thinsp;2.1 [+\u0026thinsp;1.6, +\u0026thinsp;2.6]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e+\u0026thinsp;0.38 [+\u0026thinsp;0.14, +\u0026thinsp;0.62]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e0.002\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eControl\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e4.6 (2.0)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e4.3 (1.9)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e\u0026ndash;0.3 [\u0026ndash;0.6, 0.0]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eMAI Regulation (0-175)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eTreatment\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e86.2 (7.5)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e79.3 (8.2)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e\u0026ndash;6.9 [\u0026ndash;8.0, \u0026minus;\u0026thinsp;5.8]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e\u0026ndash;0.41 [\u0026ndash;0.59, \u0026minus;\u0026thinsp;0.23]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eControl\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e85.9 (7.8)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e84.1 (7.6)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e\u0026ndash;1.8 [\u0026ndash;2.6, \u0026minus;\u0026thinsp;1.0]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003ctfoot\u003e \u003ctr\u003e\u003ctd colspan=\"7\"\u003eNote: DiD coefficients are standardised (Cohen's \u003cem\u003ed\u003c/em\u003e) for comparability. All models control for age, gender, and family income.\u003c/td\u003e\u003c/tr\u003e \u003c/tfoot\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003c/div\u003e\n\u003ch3\u003eMediation: cognitive decline → exam performance\u003c/h3\u003e\n\u003cp\u003eThe end-of-term examination scores (held online in late July 2025, immediately after the war) were substantially lower in the treatment group (mean 12.9, SD 2.6) than in the control group (mean 15.8, SD 2.4). The raw difference was \u0026minus;\u0026thinsp;2.9 points (95% CI [\u0026ndash;3.5, \u0026minus;\u0026thinsp;2.3], \u003cem\u003ep\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.001, \u003cem\u003ed\u003c/em\u003e\u0026thinsp;=\u0026thinsp;1.16). After controlling for baseline GPA (which did not differ), the adjusted difference was \u0026minus;\u0026thinsp;2.4 points (\u003cem\u003ep\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.001).\u003c/p\u003e \u003cp\u003eWe tested whether the decline in working memory mediated this effect. The total effect of treatment on exam score (controlling for baseline GPA) was \u003cem\u003ec\u003c/em\u003e = \u0026minus;\u0026thinsp;0.53 (standardised, \u003cem\u003ep\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.001). When we added the change in working memory as a mediator, the direct effect became \u003cem\u003ec\u003c/em\u003e\u0026prime; = \u0026minus;\u0026thinsp;0.31 (\u003cem\u003ep\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.008). The indirect effect through working memory was significant: \u003cem\u003eab\u003c/em\u003e = \u0026minus;\u0026thinsp;0.22, 95% bootstrap CI [\u0026ndash;0.36, \u0026minus;\u0026thinsp;0.10]. That is, about 42% of the treatment effect on exam performance was explained by the decline in working memory.\u003c/p\u003e\n\u003ch3\u003ePersistence and recovery (Wave 3, six months later, January 2026)\u003c/h3\u003e\n\u003cp\u003eSix months after the war, the treatment group had partially but not fully recovered (Table\u0026nbsp;\u003cspan refid=\"Tab3\" class=\"InternalRef\"\u003e3\u003c/span\u003e). Working memory improved relative to Wave 2 (mean change\u0026thinsp;+\u0026thinsp;0.32, SD 0.38) but remained 0.18 SD below the control group (adjusted difference \u0026minus;\u0026thinsp;0.18, \u003cem\u003ep\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.03). In contrast, metacognitive regulation showed little recovery: the treatment group's MAI Regulation score was still 0.31 SD below controls (\u003cem\u003ep\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.001). Attention recovered to a non-significant difference (\u0026ndash;0.09 SD, \u003cem\u003ep\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.28).\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab3\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 3\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eSix-month follow-up (Wave 3)\u0026mdash;adjusted differences from ANCOVA controlling for Wave 1\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"6\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c6\" colnum=\"6\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eOutcome\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eTreatment adjusted M (SE)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eControl adjusted M (SE)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003eDifference (95% CI)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c5\"\u003e \u003cp\u003ep\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c6\"\u003e \u003cp\u003eCohen\u0026rsquo;s d\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eWorking memory (d-prime)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e2.18 (0.04)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e2.36 (0.03)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e\u0026ndash;0.18 [\u0026ndash;0.27, \u0026minus;\u0026thinsp;0.09]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.03\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.31\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSART omissions\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e5.2 (0.2)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e4.9 (0.1)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e+\u0026thinsp;0.3 [\u0026ndash;0.1, +\u0026thinsp;0.7]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.28\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.12\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eMAI Regulation\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e81.5 (0.7)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e84.8 (0.5)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e\u0026ndash;3.3 [\u0026ndash;4.9, \u0026minus;\u0026thinsp;1.7]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.49\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003ctfoot\u003e \u003ctr\u003e\u003ctd colspan=\"6\"\u003eNote: All models include age, gender, and family income as covariates.\u003c/td\u003e\u003c/tr\u003e \u003c/tfoot\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e"},{"header":"Discussion","content":"\u003cp\u003eThe accidental pre-war baseline, the sudden escalation in June 2025, and the direct cognitive measures together created a unique natural experiment. This study provides rare causal-style evidence from a real-world aerial warfare setting that living through a period of direct military confrontation causes immediate, measurable declines in working memory, attention, and metacognitive regulation. These declines were substantial: large enough to explain a meaningful portion of the drop in exam performance. And they were not fleeting: metacognitive regulation remained impaired six months later (January 2026), even as working memory partially recovered.\u003c/p\u003e \u003cp\u003eSeveral features of our design explain the clarity of these effects. First, the pre-war baseline eliminated selection bias\u0026mdash;we know the two groups were cognitively identical before the war escalated. Second, the sudden, unanticipated nature of the war escalation means the cognitive decline cannot reasonably be attributed to anything other than the crisis itself. Third, we measured cognition directly, not just self-reported stress or grades. The fact that working memory\u0026mdash;a well-established marker of prefrontal function\u0026mdash;showed a dose\u0026ndash;response relationship with PTSD symptoms (Pearson's \u003cem\u003er\u003c/em\u003e = \u0026minus;\u0026thinsp;0.51) strengthens the causal interpretation. Our effect sizes (Cohen's \u003cem\u003ed\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.55\u0026ndash;0.71) are comparable to those reported for pandemic learning loss⁵ but emerge from a much shorter shock.\u003c/p\u003e \u003cp\u003eThe mediation analysis points to a mechanism. War does not merely make students sad or distracted; it reduces the working memory capacity they need to follow a lecture, solve a problem, or write an essay. The effect on exam scores was about half a standard deviation\u0026mdash;on a 0\u0026ndash;20 scale, a drop from roughly 16 to 14, equivalent to moving from a B to a C\u0026thinsp;+\u0026thinsp;in many grading systems. In a high-stakes educational context, such a difference can determine a student's academic trajectory. The indirect effect (\u0026ndash;0.22) implies that interventions targeting working memory might recover up to 40% of the exam score loss.\u003c/p\u003e \u003cp\u003eThe persistence results are instructive. Working memory recovered somewhat, perhaps because neural plasticity allows adaptation or because students developed compensatory strategies. But metacognitive regulation\u0026mdash;the ability to plan, monitor, and evaluate one's own learning\u0026mdash;remained depressed. This pattern makes sense: metacognitive skills are built slowly through practice and feedback. War disrupts that practice. Students who are hypervigilant for threats have less mental bandwidth to ask themselves \"Do I really understand this?\" Notably, 42% of the treatment group scored above the clinical cutoff for probable PTSD (\u0026ge;\u0026thinsp;33 on PCL-5)\u0026mdash;more than double the estimated prevalence in the general Iranian population\u0026mdash;suggesting that cognitive and emotional scars co-occur.\u003c/p\u003e \u003cp\u003eWe should be clear about limitations. First, this is a single-country study; the pattern might differ in other cultures or types of conflict. Second, although we controlled for many confounders, we cannot rule out that the treatment and control groups differed in unmeasured ways\u0026mdash;for example, neighbourhood exposure to blast waves (we did not have geolocation data). However, the fact that the two groups were identical at baseline on all measured variables makes major confounding unlikely. Third, our sample size, while adequate, is modest for a DiD design, and with only two pre-war time points we could not formally test the parallel trends assumption. Replication in other conflict settings would be valuable.\u003c/p\u003e \u003cp\u003eThese limitations point to future research. The field needs multi-site natural experiments in other war zones\u0026mdash;Ukraine, Gaza, Sudan\u0026mdash;to test whether the same pattern holds. It also needs intervention studies: can targeted cognitive training or metacognitive remediation help students recover faster? Our finding that working memory improves but metacognition does not suggests that different interventions may be needed for different domains. Future work should also examine whether individual differences (e.g., baseline working memory capacity, prior trauma history) moderate the effects.\u003c/p\u003e \u003cp\u003ePractically, universities in conflict zones should consider three responses. First, academic accommodations should be expanded\u0026mdash;not just extra time, but reduced cognitive load (for example, breaking exams into smaller chunks, allowing open-book formats, providing lecture notes in advance). Second, mental health services should prioritise cognitive rehabilitation alongside trauma counselling. Simple computerised working memory training or metacognitive strategy instruction could be delivered at low cost. Third, post-war educational recovery programmes should include explicit metacognitive training\u0026mdash;a component that current humanitarian guidelines often overlook. Without deliberate cognitive rehabilitative support, the scars of war may outlast the ceasefire.\u003c/p\u003e"},{"header":"Conclusion","content":"\u003cp\u003eWar inflicts hidden damage on the mind\u0026mdash;damage that standard economic metrics overlook. Using a rare natural experiment\u0026mdash;the 12-day war from 13 to 24 June 2025\u0026mdash;we showed that living through a period of direct aerial warfare produced immediate and lasting deficits in working memory, attention, and metacognitive regulation. These deficits mediated a substantial drop in exam performance. Six months later (January 2026), working memory had partially recovered, but metacognitive regulation had not. The implication is straightforward: rebuilding schools and reopening universities after war is not enough. We must also rebuild the cognitive and metacognitive capacities of students who lived through it. That is a challenge for educational psychology, for neuroscience, and for post-conflict policy worldwide. Without deliberate intervention, the cognitive scars of war may persist long after the fighting stops.\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003eAcknowledgements\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eWe are deeply grateful to the students who took part in this study under extraordinarily difficult circumstances. Their willingness to complete cognitive assessments during and after the war made this work possible. We also thank the university staff who helped re-administer the tests following the conflict. This research received no external funding.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eCompeting interests\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe author declares no competing interests.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eData availability\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eAnonymised data, analysis scripts, and the preregistered analysis plan are available on the Open Science Framework (https://osf.io/xxxxx). The repository will be made public upon acceptance of the manuscript.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n \u003cli\u003eKim, H. Y. War, displacement, and child development: Evidence from Syrian refugees in Jordan. \u003cem\u003eJ. Dev. Econ.\u003c/em\u003e\u003cstrong\u003e167\u003c/strong\u003e, 103245 (2024).\u003c/li\u003e\n \u003cli\u003eAbu Hamad, B. \u0026amp; Jones, N. \u0026quot;We are tired of being afraid\u0026quot;: The impact of war on adolescent mental health and learning in Gaza. \u003cem\u003eConfl. Health\u003c/em\u003e\u003cstrong\u003e15\u003c/strong\u003e, 42 (2021).\u003c/li\u003e\n \u003cli\u003eAupperle, R. L., Melrose, A. J., Stein, M. B. \u0026amp; Paulus, M. P. Executive function and PTSD: A meta-analysis. \u003cem\u003eJ. Psychiatr. Res.\u003c/em\u003e\u003cstrong\u003e46\u003c/strong\u003e, 705\u0026ndash;716 (2012).\u003c/li\u003e\n \u003cli\u003eScott, J. C. \u003cem\u003eet al.\u003c/em\u003e A meta-analysis of neurocognitive functioning in PTSD. \u003cem\u003eNeuropsychol. Rev.\u003c/em\u003e\u003cstrong\u003e25\u003c/strong\u003e, 87\u0026ndash;100 (2015).\u003c/li\u003e\n \u003cli\u003eBetth\u0026auml;user, B. A., Bach Mortensen, A. M. \u0026amp; Engzell, P. A systematic review and meta-analysis of the evidence on learning during the COVID-19 pandemic. \u003cem\u003eNat. Hum. Behav.\u003c/em\u003e\u003cstrong\u003e7\u003c/strong\u003e, 375\u0026ndash;385 (2023).\u003c/li\u003e\n \u003cli\u003eThomas, C., Zanotti, J. \u0026amp; Sharp, J. M. Escalation of the Israel\u0026ndash;Iran conflict. \u003cem\u003eCongr. Res. Serv.\u003c/em\u003e IN12347 (2025).\u003c/li\u003e\n \u003cli\u003eZanotti, J. \u0026amp; Thomas, C. Israel, Iran, Hamas, and Lebanese Hezbollah: Various strikes amid regional turmoil. \u003cem\u003eCongr. Res. Serv.\u003c/em\u003e (2025).\u003c/li\u003e\n \u003cli\u003eArab News. Iran closes airspace, commercial flights diverted after apparent Israeli retaliatory strikes. \u003cem\u003eArab News\u003c/em\u003e (24 June 2025).\u003c/li\u003e\n \u003cli\u003eAl Jazeera. Iran says air-defence systems intercepted multiple projectiles amid Israel tensions. \u003cem\u003eAl Jazeera\u003c/em\u003e (20 June 2025).\u003c/li\u003e\n \u003cli\u003eIslamic Republic News Agency (IRNA). Iranian Ministry of Science announces temporary closure of student dormitories and transition to online instruction. \u003cem\u003eIRNA\u003c/em\u003e (12 June 2025).\u003c/li\u003e\n \u003cli\u003eWorld Medical Association. World Medical Association Declaration of Helsinki: Ethical principles for medical research involving human subjects. \u003cem\u003eJAMA\u003c/em\u003e\u003cstrong\u003e310\u003c/strong\u003e, 2191\u0026ndash;2194 (2013).\u003c/li\u003e\n \u003cli\u003eOpen Science Framework. Preregistration: Cognitive scarring of war\u0026mdash;analysis plan. https://osf.io/xxxxx (2025).\u003c/li\u003e\n \u003cli\u003eSchraw, G. \u0026amp; Dennison, R. S. Assessing metacognitive awareness. \u003cem\u003eContemp. Educ. Psychol.\u003c/em\u003e\u003cstrong\u003e19\u003c/strong\u003e, 460\u0026ndash;475 (1994).\u003c/li\u003e\n \u003cli\u003eWeathers, F. W. \u003cem\u003eet al.\u003c/em\u003e The PTSD Checklist for DSM-5 (PCL-5). \u003cem\u003eNational Center for PTSD\u003c/em\u003e (2013).\u003c/li\u003e\n \u003cli\u003eHayes, A. F. \u003cem\u003eIntroduction to Mediation, Moderation, and Conditional Process Analysis: A Regression-Based Approach\u003c/em\u003e 2nd edn (Guilford Press, 2018).\u003c/li\u003e\n\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":false,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"npj-science-of-learning","isNatureJournal":false,"hasQc":false,"allowDirectSubmit":false,"externalIdentity":"npjscilearn","sideBox":"Learn more about [npj Science of Learning](http://www.nature.com/npjscilearn/)","snPcode":"41539","submissionUrl":"https://mts-npjscilearn.nature.com/cgi-bin/main.plex","title":"npj Science of Learning","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"ejp","reportingPortfolio":"NPJ","inReviewEnabled":true,"inReviewRevisionsEnabled":true},"keywords":"","lastPublishedDoi":"10.21203/rs.3.rs-9431139/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-9431139/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eWar inflicts lasting damage not only through death and destruction, but also by impairing the cognitive functions that support learning. Yet causal evidence from real-world conflict settings\u0026mdash;particularly with pre-war baseline data\u0026mdash;has remained scarce. Here we exploit a rare natural experiment. The baseline data were collected by chance as part of a routine course assessment two months before the escalation, turning an unforeseen crisis into a scientific opportunity: the 12-day war from 13 to 24 June 2025, when Israel and Iran exchanged direct airstrikes for the first time in modern history. We followed 412 undergraduate students across three waves: two months before the war (baseline), two weeks after the war (acute phase), and six months later (follow-up). Using a difference-in-differences design, we compared students whose universities were located in Tehran and experienced air-defence activation and blast waves (treatment, n\u0026thinsp;=\u0026thinsp;138) with those at unaffected universities outside the Tehran metropolitan area (control, n\u0026thinsp;=\u0026thinsp;274). Working memory, measured by an N-back task, dropped by 0.49 standard deviations relative to controls (95% CI [\u0026ndash;0.67, \u0026minus;\u0026thinsp;0.31], p\u0026thinsp;\u0026lt;\u0026thinsp;0.001). Sustained attention, indexed by omission errors on the Sustained Attention to Response Task (SART), worsened by 0.38 SD (95% CI [0.14, 0.62], p\u0026thinsp;=\u0026thinsp;0.002). Metacognitive regulation, assessed with the Metacognitive Awareness Inventory (MAI), declined by 0.41 SD (95% CI [\u0026ndash;0.59, \u0026minus;\u0026thinsp;0.23], p\u0026thinsp;\u0026lt;\u0026thinsp;0.001). These cognitive deficits partially mediated a 0.53 SD reduction in end-of-term examination scores (indirect effect = \u0026minus;\u0026thinsp;0.22, 95% CI [\u0026ndash;0.36, \u0026minus;\u0026thinsp;0.10]). Six months later, working memory had recovered to 85% of baseline levels, but metacognitive regulation remained 18% below pre-war values. The effects were larger among students who reported greater war-related media exposure and sleep disruption. Our findings provide rare causal-style evidence that modern aerial warfare can produce measurable, domain-specific cognitive scarring in young adults. Although the natural experiment design strengthens causal inference, the absence of geolocated exposure data and the potential for unmeasured confounding warrant caution. These results have implications for post-conflict educational recovery worldwide.\u003c/p\u003e","manuscriptTitle":"Cognitive scars of war: a natural experiment on student learning during the 2025 Iran–Israel conflict","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2026-05-12 13:01:55","doi":"10.21203/rs.3.rs-9431139/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"reviewersInvited","content":"","date":"2026-05-01T07:35:57+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2026-04-28T13:57:47+00:00","index":"","fulltext":""},{"type":"checksComplete","content":"","date":"2026-04-20T16:05:16+00:00","index":"","fulltext":""},{"type":"submitted","content":"npj Science of Learning","date":"2026-04-15T21:13:01+00:00","index":"","fulltext":""}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"npj-science-of-learning","isNatureJournal":false,"hasQc":false,"allowDirectSubmit":false,"externalIdentity":"npjscilearn","sideBox":"Learn more about [npj Science of Learning](http://www.nature.com/npjscilearn/)","snPcode":"41539","submissionUrl":"https://mts-npjscilearn.nature.com/cgi-bin/main.plex","title":"npj Science of Learning","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"ejp","reportingPortfolio":"NPJ","inReviewEnabled":true,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"0dc563cf-0905-4742-9c08-86e10733a365","owner":[],"postedDate":"May 12th, 2026","published":true,"recentEditorialEvents":[{"type":"reviewersInvited","content":"11","date":"2026-05-01T07:35:57+00:00","index":"","fulltext":""}],"rejectedJournal":[],"revision":"","amendment":"","status":"under-review","subjectAreas":[{"id":67766140,"name":"Health sciences/Health care"},{"id":67766142,"name":"Health sciences/Medical research"},{"id":67766144,"name":"Biological sciences/Psychology"},{"id":67766146,"name":"Social science/Psychology"}],"tags":[],"updatedAt":"2026-05-12T13:01:55+00:00","versionOfRecord":[],"versionCreatedAt":"2026-05-12 13:01:55","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-9431139","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-9431139","identity":"rs-9431139","version":["v1"]},"buildId":"XKTyCvWXoU3ODBz1xrDgd","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}