The research team, led by Chenyi Chen and Yawei Cheng of National Yang Ming Chiao Tung University and collaborating institutions in Taiwan, recruited 40 female breast cancer survivors, 27 of whom had been treated with chemotherapy, along with 30 age-matched healthy controls. Chemotherapy regimens consisted primarily of the standard AC combination of doxorubicin and cyclophosphamide or the FEC protocol built around 5-fluorouracil, epirubicin, and cyclophosphamide, with some patients receiving additional agents such as docetaxel. All participants were right-handed Taiwanese women aged 35 to 75 with normal or corrected-to-normal vision, and anyone with a history of neurological or psychiatric disorders was excluded. Diagnoses were clinically confirmed by physicians, and the protocol was approved by the hospital’s institutional review board.

The centerpiece of the study was a face-word emotional Stroop task performed inside an fMRI scanner. On each trial, participants saw a photograph of a face displaying either a happy or fearful expression, with an emotional word superimposed in bold red font. Sometimes the face and word matched; sometimes they clashed. Participants had to identify the facial expression while ignoring the word, a deceptively simple instruction that forces the brain to manage two competing emotional signals at once. The task is a well-validated probe of emotional attention, the preferential allocation of mental resources to emotionally salient stimuli, and of emotional conflict regulation, the ability to resolve interference between incongruent affective cues. No prior study, the authors note, had used task-based fMRI with an affective paradigm specifically designed to examine these processes in breast cancer survivors.

The behavioral and questionnaire results already hinted at something strange. On standardized measures, the non-chemotherapy survivors reported the highest affective symptom load, scoring significantly higher on combined Beck Anxiety Inventory and Beck Depression Inventory-II indices than both the chemotherapy-treated group and the healthy controls. Yet on the Stroop task, the chemotherapy-treated survivors performed worse than controls overall, and the deficit was driven almost entirely by fearful faces: their accuracy on those trials dropped to about 87 percent, compared with nearly 96 percent in controls, while accuracy on happy faces did not differ between groups. In other words, the group reporting the least emotional distress was also the group least accurate at detecting fear, a dissociation that would be invisible to any screening program relying on self-report alone.

The brain imaging sharpened the picture. In the dorsomedial prefrontal cortex, a region that integrates cognitive control with emotion regulation, chemotherapy-treated survivors showed significantly reduced activation compared with both other groups, a result that survived false-discovery-rate correction. This region is a known target of chemotherapy-related neurotoxicity, and its hypoactivation aligns with earlier work linking the drugs to prefrontal alterations. Meanwhile, the non-chemotherapy survivors displayed the opposite configuration: heightened incongruency-related interference in the parahippocampal gyrus, which supports emotionally salient memory and contextual associations, and in the insula, a hub of interoceptive awareness and salience detection that is implicated in anxiety and mood disorders. Their salience networks appeared to be working overtime, amplifying the emotional weight of conflicting cues.

Perhaps the most intriguing finding concerned the dorsal anterior cingulate cortex, a node of the brain’s conflict-monitoring machinery. In healthy controls, greater dACC activation correlated with lower affective symptom scores, a normative inverse coupling consistent with the region’s regulatory role: the better the monitoring system works, the less distress accumulates. That coupling was absent in both survivor groups, suggesting that diagnosis or treatment decouples dACC function from emotional outcomes. In the chemotherapy group, the largest fearful-versus-happy dACC difference coincided with their worst behavioral accuracy, leading the authors to speculate, cautiously, about inefficient overinvestment of cognitive resources rather than effective regulation. They stress that no electrophysiological data were collected, so this remains a hypothesis for future combined fMRI-EEG work rather than a conclusion.

Putting these threads together, the researchers propose what they call an emotional attention disruption framework. Chemotherapy, they suggest, may alter the salience network nodes that detect, prioritize, and respond to emotionally significant stimuli, including the insula, parahippocampal cortex, anterior cingulate, and dorsomedial prefrontal cortex. The behavioral consequence would be reduced interference from negative distractors; the subjective consequence would be lower scores on standard distress questionnaires. The authors are careful to label emotional blunting an interpretive hypothesis rather than a directly measured variable. They did not collect trial-by-trial valence or arousal ratings, nor instruments for alexithymia, anhedonia, or emotional avoidance, and alternative explanations such as reduced task engagement, altered semantic-visual conflict resolution, or general cognitive inefficiency cannot be excluded. Low distress in an individual survivor may still reflect authentic adaptation.

The clinical implications, if the hypothesis holds, are substantial. Current distress screening in survivorship relies almost exclusively on self-report instruments such as the Beck inventories or the Distress Thermometer. A team using those tools alone would prioritize non-chemotherapy survivors for psychosocial support and deprioritize the chemotherapy-exposed, precisely the triage the neural data call into question. The authors frame their recommendations explicitly as hypotheses for intervention trials, not evidence-based guidelines. For chemotherapy-exposed survivors, approaches that re-engage emotional attention, such as emotion-focused therapy, exposure-based methods, or affect-labeling training, might restore adaptive salience detection. For non-chemotherapy survivors showing hyper-reactive salience processing, mindfulness-based stress reduction, cognitive-behavioral therapy, or attention-bias modification might better dampen reactivity. A single standardized protocol for all survivors, they argue, is likely suboptimal.

The study’s limitations are candidly acknowledged and considerable. The cross-sectional design precludes causal inference, and the non-chemotherapy subgroup contained only 13 participants, small enough that effect-size estimates are unstable and confidence intervals wide; the authors deliberately refrained from formal mediation analysis for that reason. Several key interactions did not survive false-discovery-rate correction and are presented as descriptive. Survivors were recruited from a single hospital clinic, raising selection-bias concerns, and the emotional Stroop paradigm itself rests on the contested assumption that standardized facial configurations map onto discrete emotion categories, an assumption challenged by cross-cultural research. Near-ceiling accuracy across groups made the behavioral findings conservative, and regression to the mean cannot be ruled out without pretreatment baselines. Fatigue, sleep, hormonal change, and inflammation, all plausible modulators, were not examined.

Even with those caveats, the study marks a conceptual shift in how survivorship care might be conceived. It positions divergent emotional outcomes along a single continuum of emotional attention: hyper-reactivity, as seen in the non-chemotherapy survivors, increases vulnerability to anxiety and depression, while hypo-reactivity, as seen after chemotherapy, lowers subjective distress at the cost of reduced neural and behavioral sensitivity to affective information. Both deviate from the balanced reactivity of healthy brains. The authors suggest that behavioral proxies of emotional attention, such as emotional Stroop performance, could eventually serve as feasible adjuncts to distress screening, helping stratify survivors for mechanism-matched support without requiring scanner time. Whether either intervention strategy actually improves outcomes in the corresponding phenotype is untested and awaits randomized trials. But the core message is already clear: in cancer survivorship, silence on a questionnaire is not always the sound of health, and the brain’s quietest survivors may be the ones most in need of a closer look.

Subject of Research: Affective neural phenotypes and emotional attention in chemotherapy-treated breast cancer survivors

Article Title: Distinct affective neural phenotypes in breast cancer survivors: implications for personalized supportive care after chemotherapy

Article References: Chen, C., Chiang, H.-J., Chen, Y.-C., Fan, Y.-T., Liao, T.-T., Chen, C.-Y., & Cheng, Y. (2026). Distinct affective neural phenotypes in breast cancer survivors: implications for personalized supportive care after chemotherapy. Supportive Care in Cancer, 34(10), Article 1048. https://doi.org/10.1007/s00520-026-11271-z

Image Credits: AI Generated

DOI: 10.1007/s00520-026-11271-z

Keywords: breast cancer survivorship, chemotherapy, emotional blunting, fMRI, emotional Stroop task, distress screening, salience network, dorsomedial prefrontal cortex, insula, anterior cingulate cortex, supportive care, emotion regulation