Leaving a toxic relationship is supposed to feel like freedom. And in many ways it does, but for a significant number of survivors, freedom arrives accompanied by an unwanted passenger: a nervous system that cannot stop scanning for danger. Doors that open too fast, a raised voice in the other room, a partner’s unexplained silence, stimuli that would hardly mean anything to an average person but can trigger a chain reaction of anxiety in the mind of someone who has experienced repeated relational abuse. This is hypervigilance, and it does not fade away just because the threat has been eliminated.
What the new neuroscience can make evident is that this is not weakness, hypersensitivity, or failure to move on. It is a quantifiable, biological re-shuffling of the brain threat-detection system – one that starts at the amygdala, proceeds through the hypothalamic-pituitary-adrenal (HPA) axis, and reinvents the very patterns in which attention is distributed on a moment-to-moment basis. Understanding how toxic relationships produce this neurobiological signature and what it means for individuals seeking to build new lives after them is the aim of this essay.
Read More: Psychologist’s Insights on Hypervigilance: Causes, Symptoms, and Solutions
The Brain That Learned to Expect Harm
The amygdala is the centre of threat perception; it is an almond-shaped organ located deep within the temporal lobe whose main function is to detect potential danger and prepare the body to respond. Under ordinary circumstances, the amygdala acts as a calibrated alarm: sensitive enough to detect genuine threat, but held in check by prefrontal regulatory circuits that evaluate context and inhibit unnecessary responses. This calibration is lost in patients who are survivors of chronic relational trauma.
Harb et al. (2024) provided compelling prospective evidence of this disruption. The childhood maltreatment in a study of adult trauma survivors who were scanned two weeks following the traumatic event and followed up 6 months later revealed that the amygdala had significantly high resting-state functional connectivity with the precuneus, identified as a region of self-referential thought, mental imagery, as well as autobiographical memory. Critically, this heightened connectivity statistically mediated the relationship between childhood abuse and anxiety symptoms at the six-month follow-up. It was specific to anxiety rather than PTSD or depression.
The conclusion is that the early relational damage does not merely create the psychological scars; it causes the long-lasting upregulation of the neural circuit connecting the threat detection and the self-related processing. To adults who survived toxic relationships, the implication is pronounced. An already sensitised brain is likely to encode relationship damage more profoundly. And to preserve the state of hypervigilance of threat well beyond the termination of the relationship.
Read More: Hypervigilance as a Post-Traumatic Adaptation
Early Development of Threat Bias
The developmental foundations of this sensitivity are traced in the comprehensive review by Fu and Pérez-Edgar (2018), who document how threat-related attention bias emerges as early as five to seven months of age, supported by the amygdala-orbitofrontal cortex network. Their framework proposes that early-appearing attention bias serves as an “experience-expectant” scaffold for socioemotional development, one shaped by temperament, parental characteristics, and the broader relational environment. When that environment is chronically threatening or unpredictable, the system does not simply adapt. It recalibrates around danger as a baseline expectation. What begins as adaptive vigilance can, over time and repeated exposure, become entrenched as a default attentional style.
Attention That Cannot Settle: The Variability Finding
Popular accounts of trauma often describe hypervigilance as a fixed, sustained state of heightened alertness, a kind of permanent lookout. The empirical picture is considerably more interesting, and arguably more distressing. Research using the dot-probe paradigm, in which participants respond to probes appearing in the location of either threatening or neutral stimuli, has revealed that what characterises post-traumatic attention is not a stable bias toward threat but an unstable oscillation between vigilance and avoidance.
Naim et al. (2015) examined this phenomenon, termed Attention Bias Variability (ABV), across seven independent datasets spanning combat veterans, motor vehicle accident survivors, soldiers post-deployment, and clinical anxiety samples. Using a refined moving-average measurement technique, they found that elevated ABV was specific to PTSD and could not be explained by anxiety alone. PTSD groups showed significantly higher ABV than social anxiety disorder, high trait anxiety, and healthy combat-exposed soldiers (all p < .001). ABV correlated with PTSD symptom severity across multiple validated measures, was trauma-type independent (combat and accident survivors showed equivalent levels), and crucially was specific to threatening stimuli rather than emotional stimuli generally, ruling out a simple explanation in terms of impaired general attention.
Naim et al. (2015) interpreted ABV as reflecting a conflict between two competing post-traumatic tendencies. Hypervigilance pulls attention toward threat, while avoidance pulls it away—an attentional “tug-of-war” that creates instability. Swick and Ashley (2017) added a complementary perspective, demonstrating in a sample of combat veterans that ABV was strongly associated with the ex-Gaussian tau parameter of reaction time distributions, an index of excessively slow, attentional-lapse responses, and with self-reported cognitive failures more broadly.
Their interpretation: the instability reflects not threat-specific attentional fluctuation but a more pervasive deficit in executive control and sustained attention. Both stories resonate with anyone who has gone through the experience of living in a post-toxic relationship. The survivor who flinches at some tone of voice and then mentally discounts it as an overreaction, but then flinches again several seconds later, is living the ABV discovery in the present.
Read More: How Emotional Distractions Impact Sustained Attention and Memory
The Body Keeps the Score: HPA Axis Dysregulation
Hypervigilance is not just a cognitive and attentional event. It gets embedded into the architecture of the stress-response in the body. The main biological process through which psychological threat comes to be converted into physiological preparedness involves the HPA axis, or hormonal cascade, involving the hypothalamus, pituitary gland, and adrenal cortex. Its main effector hormone, cortisol, not only mobilises the body to act, but on a chronic basis, also sensitises the very neural pathways, such as the amygdala, which produce the threat response in the first place.
Inslicht et al. (2006) have evaluated neuroendocrine activity in female survivors of intimate partner violence (IPV), both those who have developed lifetime PTSD and those who have been exposed to IPV and never developed PTSD. Cortisol levels of women with lifetime PTSD are much higher throughout the day, despite adjusting for factors related to age, severity of depression, and features of the abuse itself. The result indicates that cortisol elevation is a biomarker of IPV-related PTSD – a physiologic response that continues even after the acute stage of trauma and probably indicates permanent alterations in the functioning of the HPA axis.
Read More: Trauma Bonds Explained: Why Abuse, Fear, and Pain Create Emotional Loyalty
Cortisol Patterns in Chronic Trauma
Notably, Alhalah and Falatah (2020) observed a different trend when hair cortisol. A retrospective biomarker indicating cumulative activity of HPA over months was used as the dependent variable in a sample of Saudi women with IPV. Higher levels of IPV predicted lower hair cortisol levels, which is in line with the well-reported effect of HPA blunting or hypocortisolism in states of chronic and intense stress.
These two findings are not necessarily contradictory: combined, they point to a direction that early or acute PTSD could be characterised by cortisol elevation, whereas long-term chronic exposure could bring the system to an underactive state. In both, the normal level of HPA regulation is disturbed, and the amygdala, the one directly sensitive to cortisol, is not working within its calibrated range.
This creates a self-reinforcing biological loop: a sensitised amygdala produces more threat responses. This triggers more cortisol release, which in turn decreases the threshold of the amygdala to alarm. After a toxic relationship, this loop can unconsciously shape perceptions and reactions before rational thinking intervenes.
Read More: 9 Signs of a Toxic Relationship
When Hypervigilance Enters the Next Relationship
The consequences of post-traumatic hypervigilance do not remain confined to the individual. They radiate outward into relationships, including, and perhaps especially, the ones survivors build after leaving harm. Conducting an extensive literature review of PTSD and relationship functioning, Campbell and Renshaw (2018) found emotional numbing and hyperarousal as the most consistent clusters of symptoms related to relationship distress, rather than, perhaps intuitively, the re-experiencing symptoms such as flashbacks. But instead, the more diffuse and pervasive changes in affective functioning at baseline.
This result coincides with the larger body of relationship studies indicating that inadequately positive functioning in the form of warmth, responsiveness, and emotional availability is more likely to predict the decline of a relationship than the actual existence of conflict (Gottman and Levenson, 2000, as cited in Campbell and Renshaw, 2018).
Interpreting Safety Through a Threat Lens
An emotionally withdrawn, readily startled, misinterpretation of neutral cues as threats, and unable to maintain intimacy, survivor is not willfully difficult; they are showing the relational manifestation of a re-organised protective nervous system. Campbell and Renshaw (2018) also document the role of partner cognitions as a moderating factor: when partners understand that these behaviours reflect a disorder rather than personal failing or deliberate withdrawal, the negative impact on relationship satisfaction is significantly attenuated.
The interpersonal aspect of hypervigilance is further complicated by the results of a study on the role of appraisals in psychologically abusive relationships conducted by Matheson et al. (2023): people who had more personal experience of abuse were found to be more sensitive to detecting abuse signs, especially in ambiguous contexts, and engaged in more avoidant coping behaviour, irrespective of the situational appraisal. This implies that there is a direct behavioural equivalent to the hypervigilant attentional system at the neurobiological level. A tendency to respond to relational uncertainty with withdrawal and avoidance, even when no threat exists in a new relationship.
What Recovery Actually Looks Like
Neuroscience of hypervigilance carries an important implication for how we should understand recovery. Post-relationship hypervigilance reflects genuine neurobiological reorganisation of sensitised amygdala circuits, disrupted HPA regulation, and unstable attentional processing. Then recovery cannot simply be a matter of rational reassurance or willpower.
The system changed, and changing it back requires targeted intervention at the level of the system itself. Naim et al. (2015) reported that two randomised controlled trials using attention control training designed to stabilise rather than simply redirect threat-monitoring resulted in significant reductions in PTSD symptoms in combat veterans, and that these reductions were mediated by decreases in ABV. This finding points toward attentional stability, rather than the elimination of threat sensitivity, as a meaningful therapeutic target. The goal is not to make the nervous system naive, but to restore its ability to evaluate rather than simply react.
Fu and Pérez-Edgar’s (2018) developmental framework adds further nuance. Because threat-detection systems are shaped through accumulated experience across development, they can also be reshaped through new experience. Safe, predictable relational environments, whether therapeutic or interpersonal, provide the repeated disconfirmation of threat expectancy that allows the system to recalibrate.
Campbell and Renshaw (2018) similarly point toward couple-based treatments that incorporate psychoeducation, communication training, and explicit attention to emotional numbing as promising approaches, noting that interventions targeting the relational context of PTSD may reach dimensions of the disorder that individual therapy does not.
Conclusion
Hypervigilance following a toxic relationship is not irrational or permanent. But, really, it has its basis in observable alterations in the threat architecture of the brain, the hormonal control of the body, and the distribution of attention on a moment-by-moment basis. The amygdala sensitisation reported by Harb et al. (2024), the attentional instability reported by Naim et al. (2015) and Swick and Ashley (2017), and the HPA dysregulation reported by Inslicht et al (2006). Together, they constitute a consistent neurobiological picture of the brain learning that closeness is harmful. Alhalah and Falatah (2020)
The Path Toward Neural Recalibration
Recovery, the evidence suggests, is not a matter of forgetting or deciding to trust again. It is the slow work of giving a reorganised nervous system enough evidence of safety to gradually stand down. That process is made harder by the relational avoidance and emotional numbing that hypervigilance itself produces, a cruel irony in which the very adaptation that protected the survivor now interferes with the conditions needed for healing. Understanding this mechanism, clearly and without judgment, may be the first step toward interrupting it.
References +
Alhalah, E., & Falatah, R. (2020). Intimate partner violence and hair cortisol concentration: A biomarker for HPA axis function. Psychoneuroendocrinology, 117, 104686. https://doi.org/10.1016/j.psyneuen.2020.104686
Campbell, S. B., & Renshaw, K. D. (2018). Posttraumatic stress disorder and relationship functioning: A comprehensive review and organisational framework. Clinical Psychology Review, 65, 152–162. https://doi.org/10.1016/j.cpr.2018.08.003
Fu, X., & Pérez-Edgar, K. (2018). Threat-related attention bias in socioemotional development: A critical review and methodological considerations. Developmental Review, 51, 31–57. https://doi.org/10.1016/j.dr.2018.11.002
Harb, M., Lebois, L. A. M., van Rooij, S. J. H., Ely, T. D., Jovanovic, T., Ressler, K. J., & Stevens, J. S. (2024). Childhood maltreatment and amygdala-mediated anxiety and posttraumatic stress following adult trauma. Biological Psychiatry Global Open Science, 4(1), 100268. https://doi.org/10.1016/j.bpsgos.2023.100268
Inslicht, S. S., Marmar, C. R., Neylan, T. C., Metzler, T. J., Hart, S. L., Otte, C., & Baum, A. (2006). Increased cortisol in women with intimate partner violence-related posttraumatic stress disorder. Psychoneuroendocrinology, 31(7), 825–838. https://doi.org/10.1016/j.psyneuen.2006.03.007
Matheson, A., Decker, M. R., & Smyth, J. M. (2023). Assessment of appraisals of dating relationship conflicts and perceptions of appropriate coping strategies with psychologically abusive interactions. Frontiers in Psychology, 14, 1074787. https://doi.org/10.3389/fpsyg.2023.1074787
Naim, R., Abend, R., Wald, I., Eldar, S., Levi, O., Fruchter, E., Ginat, K., Halpern, P., Sipos, M. L., Adler, A. B., Bliese, P. D., Quartana, P. J., Pine, D. S., & Bar-Haim, Y. (2015). Threat-related attention bias variability and posttraumatic stress. American Journal of Psychiatry, 172(12), 1242–1250. https://doi.org/10.1176/appi.ajp.2015.14121579
Swick, D., & Ashley, V. (2017). Enhanced attentional bias variability in post-traumatic stress disorder. Scientific Reports, 7, 16559. https://doi.org/10.1038/s41598-017-16891-y
