Higher Prefrontal Responses to Threatening Words Linked to Later Suicide Attempts in Mood Disorders

Every year, more than 720,000 people die by suicide worldwide, and attempts are even more frequent than deaths. In 2021, suicide was the third leading cause of death among people aged 15 to 29 years, according to the World Health Organization. For psychiatrists treating major depressive disorder and bipolar disorder, two conditions strongly linked to suicidal thoughts and behaviors, the hardest question is not whether a patient is suffering but whether that suffering will translate into an attempt. Now, a small prospective study from Saitama Medical University in Japan offers a tantalizing clue: a noninvasive brain scan performed at a single baseline visit captured prefrontal responses to threatening words that differed between patients who later attempted suicide and those who did not.

The study, published in Frontiers in Neuroimaging, used near-infrared spectroscopy, or NIRS, a technique that shines near-infrared light at wavelengths of 695 and 830 nanometers through the scalp to track changes in oxygenated hemoglobin in the cortical surface beneath. Because NIRS requires only a seated patient and a probe array, it can be repeated with minimal burden, an advantage for people with severe depressive symptoms who may struggle to tolerate lengthy MRI sessions. The researchers fitted 52 channels in a 3 by 11 array over frontotemporal regions, aligned along the T3–Fpz–T4 line of the international 10–20 electroencephalography system, and estimated each channel’s anatomical location using Montreal Neurological Institute coordinates. Signals were computed with the modified Beer–Lambert law, sampled every 100 milliseconds, and cleaned of motion artifacts with a 5-second moving-average method.

While hemoglobin changes were recorded, participants performed an emotional Stroop task built from two-character Japanese kanji words categorized as happy, sad, threat, or neutral. Each word appeared in red, blue, yellow, or green for 2 seconds, and participants were explicitly told to attend to the word’s meaning while pressing a button corresponding to its color as quickly as possible. The interference between reading an aversive word and naming its color is the point: it forces the prefrontal cortex to exert cognitive control over emotional distraction. Each emotional condition consisted of a 30-second pre-task neutral block, a 60-second emotional-word block containing 30 words, and a 60-second post-task neutral block, with the happy, sad, and threat conditions administered in one of six predefined orders to reduce order and fatigue effects.

The analysis focused on the threat condition and, within it, on a specific window. The 60-second threat-word block was divided into 15-second periods designated a through d, and the researchers zeroed in on periods b and c, corresponding to 15 to 45 seconds after the block’s onset. Guided by their previous cross-sectional work in the same sample, they defined the main exploratory measure as the mean normalized oxygenated hemoglobin response across periods b plus c in the left dorsolateral prefrontal cortex, a region implicated in cognitive control during emotional interference. The corresponding measure in the left orbitofrontal cortex, involved in emotional evaluation and reward and punishment processing, was examined as an additional exploratory comparison. Because these regions and periods were selected based on findings from the same study sample, the authors stress that the analysis is exploratory rather than confirmatory.

Twenty-nine patients with major depressive disorder or bipolar disorder underwent baseline NIRS and clinical assessment. Diagnoses were established by psychiatrists using DSM-5 clinical interviews and the Mini-International Neuropsychiatric Interview, and exclusion criteria included personality disorder, neurological or organic brain disease, substance abuse or dependence, serious medical illness, left-handedness, and an estimated intelligence quotient below 80. Seven participants withdrew consent during follow-up, leaving 22 completers, five of whom made at least one actual suicide attempt during the 6-month period as determined by the Columbia Suicide Severity Rating Scale. Importantly, suicidal ideation, preparatory acts, aborted attempts, and interrupted attempts were not counted in this definition, sharpening the contrast between the groups.

The results were striking in direction. Patients who later attempted suicide showed significantly higher baseline oxygenated hemoglobin responses across periods b plus c in the left dorsolateral prefrontal cortex than patients who did not attempt suicide, with a Mann–Whitney U statistic of 7.00 and a Holm-adjusted p value of 0.022. The left orbitofrontal cortex showed a similar difference, with U equal to 5.00 and a Holm-adjusted p value of 0.020. Period-specific analyses in both regions and both periods pointed the same way. Behavioral performance, meanwhile, did not separate the groups: task accuracy and reaction times were statistically indistinguishable, which the authors note cannot rule out differences in cognitive effort or the efficiency of emotional processing even when outward performance looks identical.

What does a larger hemodynamic response mean? The authors caution that greater prefrontal activation does not necessarily indicate more effective functioning. Increased prefrontal oxygenation during demanding tasks has previously been interpreted as greater recruitment of cognitive-control resources, so the elevated response in the future-attempt group may reflect altered recruitment of prefrontal resources when processing threatening information. Whether that represents adaptive compensation, less efficient processing requiring greater cognitive effort, or heightened sensitivity to threat remains unclear. The dorsolateral prefrontal cortex regulates emotion through its influence on regions such as the amygdala and exerts control in conflict situations, and previous research has linked a history of suicide attempts to executive dysfunction associated with this region. Prior neuroimaging findings have been inconsistent, with some studies reporting reduced prefrontal activation and others, including work on adolescents processing angry faces and untreated depression patients making decisions, finding greater dorsolateral prefrontal activation in those with attempt histories.

A crucial caveat tempers the headline finding. At baseline, the future-attempt group was younger and carried higher scores on the 17-item Hamilton Depression Rating Scale, the Young Mania Rating Scale, and the Quick Inventory of Depressive Symptomatology, and four of its five members were in a depressive episode compared with seven of seventeen in the other group. Exploratory Spearman’s rank correlations showed that baseline QIDS scores remained significantly associated with the periods b plus c NIRS measures in both the left dorsolateral prefrontal cortex and left orbitofrontal cortex after Holm adjustment, with adjusted p values of 0.003 and 0.010. This suggests that baseline depressive symptom severity may have contributed to the observed group differences, meaning the prefrontal signal may partly reflect clinical severity rather than a suicide-attempt-specific neural alteration. The authors also emphasize that follow-up clinical scores, which were worse in the attempt group, may reflect status after baseline and may have been influenced by suicidal behavior itself rather than predicting it.

The limitations are substantial and the authors lay them out candidly. With only five attempts among 22 completers, multivariable modeling was impossible, and the effects of age, baseline symptom severity, diagnosis, and other clinical factors could not be reliably separated. The sample mixed major depressive disorder and bipolar disorder, conditions that may involve partially distinct neural mechanisms, and diagnosis-specific analyses were limited by the small numbers. Treatment during follow-up was not standardized, so medication, psychotherapy, or neuromodulation changes may have influenced both symptoms and suicidal behavior. Unmeasured confounders such as socioeconomic status, family history, and childhood adversity were not assessed. Seven of 29 participants dropped out, and although no significant baseline differences were detected between completers and non-completers, attrition-related selection bias cannot be excluded. Finally, NIRS measures only superficial cortical hemodynamics, cannot reach deep limbic structures such as the amygdala, and its signals can be contaminated by systemic physiological and extracerebral hemodynamic changes unrelated to neuronal activity.

Even with those constraints, the study is notable as one of the first prospective NIRS investigations to ask whether baseline prefrontal responses differ according to subsequent suicide attempts in mood disorders. Most neuroimaging work on suicidal behavior has been cross-sectional, capturing snapshots of people already ill or already attempters, whereas this design linked a baseline scan to outcomes months later. The authors frame their results as preliminary evidence that altered prefrontal hemodynamic responses during threat-related emotional interference may be associated with subsequent suicidal behavior, consistent with neurobiological models implicating emotional regulation, cognitive control, decision-making, and reward processing in the development of suicidal thoughts and actions. They explicitly reject any claim of a predictive biomarker or independent neural marker of suicide attempts. Replication in larger prospective studies with standardized treatment, separated diagnostic groups, and sufficient attempt numbers is required before the findings’ clinical relevance can be determined.

Subject of Research: Prefrontal hemodynamic responses to negative emotional stimuli and subsequent suicide attempts in patients with mood disorders, measured with near-infrared spectroscopy

Article Title: Prefrontal responses to negative emotional cognitive stimuli are associated with future suicide attempts in patients with mood disorders: a near-infrared spectroscopy study

Article References: Kamimura, H., Matsuoka, T., Okai, H., Shimizu, N., Harada, S., & Matsuo, K. (2026). Prefrontal responses to negative emotional cognitive stimuli are associated with future suicide attempts in patients with mood disorders: a near-infrared spectroscopy study. Frontiers in Neuroimaging, 5, Article 1955513. https://doi.org/10.3389/fnimg.2026.1955513

Image Credits: AI Generated

DOI: 10.3389/fnimg.2026.1955513

Keywords: near-infrared spectroscopy, suicide attempts, mood disorders, major depressive disorder, bipolar disorder, prefrontal cortex, dorsolateral prefrontal cortex, orbitofrontal cortex, emotional Stroop task, depression, neuroimaging, prospective study