Stress Changes in Decision-Making Mark Suicide Risk in Young Adults
In young adults at elevated suicide risk, stress appears to alter decision-making in ways that are clinically meaningful and measurable. A longitudinal preprint from the University of Pittsburgh followed 233 participants ages 18 to 30, including 55 with recent suicide attempts, 59 with suicidal ideation, 71 psychiatric controls, and 48 healthy controls, and tracked behavior at baseline, 3, 6, and 12 months. The central signal was not just whether participants made risky choices, but how those choices shifted when stress was introduced. Over time, those with a history of suicide attempts showed greater reductions in risky choices under stress than healthy controls, while diminished expected value sensitivity under stress predicted later suicidal behavior.
The design paired behavioral testing with biological sampling, giving the work a stronger mechanistic frame than a simple symptom survey. Investigators measured hair cortisol concentration and circulating inflammatory markers, including CRP, IL-6, IL-10, IL-1ra, and TNF-alpha, alongside the Stressful Cups Task and clinical assessments. Hair cortisol did not show significant associations with the decision-making outcomes, which narrows attention toward immune signaling rather than chronic cortisol exposure as the more relevant biological correlate in this cohort. The implication is straightforward: stress biology may be shaping how a person weighs risk and reward in the moment, not simply how stressed they feel in the abstract.
Behavioral Shifts Under Stress Were Not Random
The most striking pattern was the stress-related drop in risky choices among participants with prior suicide attempts. Rather than moving toward a more impulsive profile under stress, this group became less willing to choose risky options compared with healthy controls. In a clinical sense, that may reflect a form of maladaptive overcorrection, where stress disrupts the calibration between danger, reward, and anticipated consequence. Suicide risk is often discussed through the lens of impulsivity, but this pattern points to something more specific: altered decision architecture when the environment becomes psychologically loaded.
One sentence captures the clinical tension. Stress did not simply increase risk-taking; it changed how value was computed.
Expected value sensitivity was the more consequential signal. Reduced sensitivity to EV under stress relative to neutral conditions predicted subsequent suicidal behavior, suggesting that when the brain no longer tracks the weighted payoff of choices as efficiently under strain, later self-harm risk may rise. That finding matters because EV sensitivity is closer to the mechanics of choice than to a broad psychiatric label. A young adult who cannot reliably integrate probability, reward, and punishment during stress may be more vulnerable when faced with rapid, emotionally charged decisions.
Inflammation Matched the Risk Profile More Closely Than Cortisol
Inflammation showed several associations with choice behavior, and those associations were directionally informative. Higher CRP corresponded with fewer risky choices under both neutral and stress conditions. Higher IL-6 was associated with fewer risky choices under neutral conditions. Higher IL-10 tracked with greater reductions in risky choices under stress relative to neutral conditions over time. No significant links emerged for hair cortisol concentration.
These markers do not provide a complete causal map, but they fit a biologically plausible pathway. Systemic inflammatory signaling can influence frontostriatal circuits involved in valuation, inhibitory control, and threat appraisal. In that framework, elevated inflammation may bias decision-making away from exploration and toward caution in some contexts, while stress may interact with immune activity to narrow the range of available responses. The pattern is especially relevant because IL-10, an anti-inflammatory cytokine, behaved differently from the pro-inflammatory markers, hinting that immune balance rather than inflammation alone may shape stress responses.
A Behavioral Biomarker With Bedside Relevance
The study suggests that suicidal behavior risk in young adults may be better understood through stress-sensitive cognition than through baseline behavior alone. A patient can appear relatively stable in neutral settings and still lose decision-making precision when pressure mounts. That gap is what makes the work clinically provocative. It points toward assessments that deliberately probe cognition under stress, rather than relying only on static questionnaires or interviews completed in calm conditions.
For clinicians, the practical consequence is a more nuanced risk conversation. Stress exposure may reveal vulnerabilities that are not obvious during routine encounters. If future studies confirm these associations, decision tasks like the Stressful Cups Task could help identify patients whose choice behavior collapses under pressure, particularly when paired with inflammatory biomarkers. The combination could eventually inform more targeted monitoring or intervention strategies for young adults with recent suicidal thoughts or attempts.
The cohort also included psychiatric controls and healthy controls, which helps separate suicide-related effects from general psychiatric burden. Even so, the data remain preprint findings, and the work should be treated as early evidence rather than a practice-changing signal. The sample was modest, the study was observational, and the biological markers were measured alongside behavior rather than manipulated experimentally. None of those limits erase the result, but they do restrain how far it can be pushed.
A final point stands out. Hair cortisol did not explain the behavioral pattern, so the stress-suicide link here is not simply a story about chronic HPA-axis activation. The better fit is a model in which acute stress-related decision disruption, immune signaling, and suicidal behavior intersect. Follow-up work will need to establish whether these markers hold across broader populations, whether they predict imminent risk as well as longer-term outcomes, and whether they can be used in clinical settings without adding noise to already complex suicide assessments.