Resilience Is Brain Network Reorganization Following a Stressor

Fred Shaffer
47 minutes ago
11 min read

Watanabe and colleagues' (2026) provocative study, "Neural Signatures of Human Psychological Resilience Driven by Acute Stress," challenges our understanding of resilience. It shifts our framing from "how well we handled a stressor" to "how effectively our brain reorganized itself during the hour following a stressor."

The takeaway

The popular view treats resilience as something you display in the moment — the person who stays calm under pressure, keeps their head during a crisis, or bounces back quickly the moment the threat passes. On that view, the biologically resilient individual is the one whose stress response is smaller, shorter, or better controlled while the stressor is actually happening. Assessment, training, and clinical intuition have largely followed that assumption: catch someone at peak stress and you have your window.

Watanabe et al. (2026) challenge that framing directly. Resilience is not the absence of stress, nor a snap-back reflex.

In Watanabe et al.'s (2026) multimodal imaging study, the most diagnostically useful neural markers of resilience did not emerge during the stressor itself. They emerged roughly 1 hour afterward.

At that point, more resilient participants showed greater default mode network and posterior hippocampal activity, whereas less resilient participants showed stronger salience network connectivity and higher high-beta and gamma EEG power.

For biofeedback and neurofeedback practitioners, that timing entirely reframes the clinical question.

The moment that tells you the most about a client's regulatory capacity is not the first spike of distress. It is the recovery window, when the brain actively reorganizes its networks.

What's the science?

Acute stress is a short-lived challenge that sets off a coordinated mobilization across the brain and body. Two of its best-known drivers are the sympathetic nervous system, which triggers rapid physiological arousal, and the hypothalamic-pituitary-adrenal (HPA) axis, a hormone cascade that culminates in cortisol release. Beyond these classic stress pathways, acute stress also reshapes how large-scale brain networks communicate with one another.

The salience network detects what matters most in the environment, the default mode network supports internally directed thought and autobiographical processing, and prefrontal regions modulate attention and emotional responses. Under stress, the balance among these networks shifts in ways that differ markedly between individuals (de Kloet et al., 2005; Hermans et al., 2011).

The broader neuroscience literature had already established that stress biology unfolds across distinct temporal phases. Corticosteroids can exert rapid effects through non-genomic mechanisms, acting within minutes without requiring gene transcription, and slower effects through genomic mechanisms that alter cellular signaling over the course of an hour or more.

In human neuroimaging work, those slower cortisol effects were linked to normalized responding to aversive stimuli and altered medial prefrontal coupling, patterns more consistent with recovery than with initial alarm.

More recent work confirms that only the slow cortisol window improved the voluntary downregulation of negative emotion (Henckens et al., 2010; Pan et al., 2023). That context makes the one-hour window identified by Watanabe and colleagues biologically coherent rather than coincidental.

How did they do it?

To trace recovery across time, the authors enrolled approximately 100 participants in a structured laboratory stress protocol and measured multiple biological signals before the stressor, immediately after, and at intervals up to 1.5 hours later.

They combined self-report measures of perceived stress with two complementary brain imaging methods: functional magnetic resonance imaging (fMRI), which estimates neural activity by tracking blood oxygen level-dependent (BOLD) changes, and electroencephalography, or EEG, which captures scalp-recorded electrical activity with millisecond precision.

They also monitored heart rate, respiration, pupil diameter, and cortisol across the same time course. Acute stress was induced with a cold pressor test, a standardized laboratory task in which participants submerged a hand in ice-cold water for 2 minutes.

What did they find?

The immediate post-stress period was not the most revealing window. Self-reported stress levels did not map cleanly onto concurrent biological signals, and brain and body measures showed only modest differentiation between groups right after the stressor.

The clearest separation between more resilient and less resilient participants emerged approximately one hour later. At that point, the two groups showed opposing patterns across major networks: more resilient participants exhibited greater default mode network and posterior hippocampal activity, while less resilient participants showed stronger salience network connectivity and elevated high-beta and gamma EEG power.

Machine-learning classification confirmed that the one-hour time point provided the most information for distinguishing resilience profiles in the full multimodal dataset.

The EEG findings reinforced the imaging picture. Changes were most pronounced over prefrontal regions, the anterior cortical territory that supports planning, working memory, and top-down emotion regulation.

That convergence across modalities suggests that resilience is not simply a matter of having a milder initial reaction. It depends on active regulatory processes that take hold during recovery, and those processes leave measurable traces in both network connectivity and cortical oscillations.

What's the context?

Hermans et al. (2011) demonstrated that stress-related noradrenergic activity, meaning arousal signaling driven by noradrenaline, prompts a large-scale reconfiguration of neural network interactions.

In a prospective longitudinal design, Zhang et al. (2022) showed that acute stress-induced changes in salience network coupling predicted later increases in perceived stress and clinician-rated trauma symptoms, whereas hormonal and subjective measures did not carry the same predictive weight.

The timing finding most directly impacts clinical and performance practice. De Kloet et al. (2005) argued that corticosteroids shape adaptive brain responses by acting through receptor systems that influence both network dynamics and downstream neural processing.

Henckens et al. (2010) showed that hydrocortisone rapidly dampened amygdala reactivity but took longer to normalize responses to negative emotional material through altered medial prefrontal coupling.

Pan et al. (2023) similarly found that only the slow cortisol window was associated with improved downregulation of negative emotion. Taken together, these studies argue that the hour following a stressor is a neurobiologically active phase in which the brain navigates the transition from vigilance to regulation, and the quality of that transition distinguishes resilient from less resilient individuals.

Across this body of work, a consistent picture emerges: acute stress does not simply turn arousal up and then down again. Hermans et al. (2011) established that noradrenergic signaling during stress triggers a large-scale reconfiguration of neural network interactions, meaning the brain reorganizes its functional architecture rather than merely amplifying a single alarm signal. Zhang et al. (2022) extended that insight prospectively, showing that stress-induced changes in salience network coupling predicted later trauma symptoms and perceived stress more reliably than hormonal or subjective measures did. Watanabe et al. (2026) add the critical temporal dimension: the most informative neural markers of resilience do not peak during or immediately after the stressor but emerge roughly one hour later, when more resilient individuals show a shift toward default mode and hippocampal activity while less resilient individuals remain locked in salience network dominance and high-frequency oscillatory patterns.

The cortisol literature explains why that hour matters biologically. De Kloet et al. (2005) argued that corticosteroids shape adaptive brain responses through receptor systems that influence both network dynamics and downstream neural processing across different time scales. Henckens et al. (2010) demonstrated this concretely: hydrocortisone rapidly dampened amygdala reactivity, but the normalization of responses to negative material through altered medial prefrontal coupling required substantially more time.

Pan et al. (2023) confirmed that only this slower cortisol window was associated with improved downregulation of negative emotion.

Together, these findings converge on a single conclusion: the hour following a stressor is not a period of passive recovery. It is a neurobiologically active transition phase, and the quality of that transition — how effectively the brain moves from vigilance toward regulation — is what distinguishes resilient from less resilient individuals.

What's the impact on our understanding of resilience?

These results extend a substantial literature showing that acute stress reorganizes brain network architecture rather than simply amplifying a uniform arousal signal. If the most discriminating neural markers of resilience appear during recovery rather than during stress, then resilience may be less about the character of your stress response and more about the quality of your regulatory rebound.

The question shifts from "how did you handle the stressor?" to "how effectively did your brain reorganize itself in the hour that followed?" That is a substantially different target, both conceptually and clinically, and it aligns with the cortisol timing literature, which shows that genomic hormone effects on emotion regulation unfold over a delayed window rather than at the moment of peak arousal.

In practice, a clinician working with a medical student after a high-stakes simulation, or with a manager after a tense negotiation, may gain more clinically useful information from a structured HRV recovery check 45 to 75 minutes after the event than from an immediate post-event snapshot alone (Goessl et al., 2017; Schumann et al., 2021; Thayer et al., 2012).

For neurofeedback practitioners, the EEG findings point toward a clinically actionable target: supporting clients in moving out of sustained high-arousal oscillatory states and toward more flexible prefrontal regulation and network balance. Direct evidence for neurofeedback in everyday acute stress recovery remains limited, but the PTSD literature offers a useful translational bridge.

A practical application is to evaluate whether neurofeedback improves recovery trajectories over the 45 to 75 minutes following a stressor, not only whether clients feel calmer during the training session itself (Askovic et al., 2023; Nicholson et al., 2020).

In clinical practice, this reframing changes how sessions might be structured. A client who leaves a demanding oral presentation saying "I'm fine" may still carry measurable physiological activation an hour later. A first responder who looks composed after a high-stress call may show impaired autonomic recovery or persistently elevated fast-wave EEG activity well into the subsequent hour.

The implication is operational: assessment and training may be substantially more informative when they explicitly capture the delayed recovery window, not just the stress peak.

Five Takeaways

Resilience appears to be a delayed recovery process rather than an instant rebound, with the most informative biological differences emerging about 1 hour after the stressor, not immediately.
The period approximately 1 hour after acute stress may be the highest-yield window for identifying adaptive versus maladaptive recovery patterns, and clinical assessments timed to that window may outperform immediate post-stressor snapshots.
Large-scale network balance among the salience network, default mode network, hippocampus, and prefrontal cortex appears central to resilience, and disruption of this balance during recovery is associated with prospective risk of symptoms.
Immediate self-report is an imprecise index of biologically meaningful recovery dynamics; subjective stress ratings often fail to capture the network-level reorganization that distinguishes resilient from less resilient individuals.
Biofeedback and neurofeedback may offer the most clinical leverage when they track and train the recovery window rather than focusing exclusively on the moment of peak stress.

Glossary

biofeedback: a training method that provides real-time physiological information so clients can learn to regulate bodily processes more deliberately. In stress applications, common signals include respiration, heart rate, and HRV. blood oxygen level-dependent (BOLD) signal: the small changes in MRI signal intensity that occur when neural activity increases blood flow to a region, delivering oxygenated blood faster than local neurons consume it. Because oxygenated and deoxygenated hemoglobin have different magnetic properties, fMRI can detect changes in their ratio and use this as an indirect proxy for neural activity.

cold pressor test: a standardized laboratory stress induction task in which a participant submerges a hand in ice-cold water for a brief, timed period to elicit an acute physiological stress response.

cortisol: a glucocorticoid hormone released via the HPA axis in response to stress. Cortisol exerts both rapid, non-genomic effects and slower, genomic effects on brain function; the latter are particularly relevant during the recovery phase.

default mode network: a large-scale brain network associated with internally directed thought, autobiographical memory, and self-referential cognition. In Watanabe et al., greater activity in this network during delayed recovery was a marker of greater resilience.

functional connectivity: a measure of the degree to which activity in different brain regions or networks fluctuates together over time. It is widely used to characterize how stress reorganizes the brain's large-scale network architecture.

functional magnetic resonance imaging (fMRI): a brain imaging method that estimates neural activity indirectly by detecting blood oxygen level-dependent changes in regional cerebral blood flow.

gamma oscillations: very fast EEG rhythms, typically above 30 Hz, often interpreted as markers of intense or localized neural processing. In Watanabe et al., higher gamma power during the delayed recovery period was more characteristic of less resilient participants.

genomic effects: slower hormonal effects that depend on changes in gene transcription following receptor binding, typically unfolding over 30 to 90 minutes. These effects help explain why some stress-related brain changes emerge well after the initial hormonal response.

heart rate variability (HRV): the normal variation in timing between successive heartbeats. Greater and more contextually flexible HRV is broadly associated with stronger autonomic and emotional regulatory capacity, though interpretation always depends on the measurement context.

hippocampus: a medial temporal lobe structure central to memory encoding and the contextual processing of experience. Watanabe et al. found that greater posterior hippocampal activity during the delayed recovery window distinguished more resilient participants.

high-beta oscillations: EEG rhythms in the approximate range of 20 to 30 Hz, often associated with heightened alertness, effortful cognitive processing, and arousal. In Watanabe et al. (2026), elevated high-beta power during delayed recovery was characteristic of less resilient individuals.

neurofeedback: a training method that uses real-time brain activity signals, typically EEG, to help individuals learn to modify their own neural oscillatory patterns. Clinical evidence is most developed in PTSD, though protocol heterogeneity and variable study quality limit firm conclusions.

non-genomic effects: rapid hormonal effects that do not require gene transcription, operating through membrane-associated receptors and second-messenger pathways. They can appear within minutes of hormone release and differ functionally from the slower genomic recovery effects.

prefrontal cortex: the anterior region of the cerebral cortex that contributes to executive functions, including planning, attentional control, working memory, and top-down regulation of emotion. Watanabe et al. identified EEG changes centered on this region during the delayed recovery window.

salience network: a large-scale brain network, anchored in the anterior insula and anterior cingulate cortex, that detects behaviorally relevant stimuli and coordinates the reallocation of attention and neural resources. Stronger salience network connectivity after stress has been associated with lower resilience and with prospective trauma symptom risk.

References

Askovic, M., Soh, N., Elhindi, J., & Harris, A. W. F. (2023). Neurofeedback for post-traumatic stress disorder: Systematic review and meta-analysis of clinical and neurophysiological outcomes. European Journal of Psychotraumatology, 14(2), 2257435. https://doi.org/10.1080/20008066.2023.2257435

de Kloet, E. R., Joëls, M., & Holsboer, F. (2005). Stress and the brain: From adaptation to disease. Nature Reviews Neuroscience, 6(6), 463–475. https://doi.org/10.1038/nrn1683

Goessl, V. C., Curtiss, J. E., & Hofmann, S. G. (2017). The effect of heart rate variability biofeedback training on stress and anxiety: A meta-analysis. Psychological Medicine, 47(15), 2578–2586. https://doi.org/10.1017/S0033291717001003

Henckens, M. J. A. G., van Wingen, G. A., Joëls, M., & Fernández, G. (2010). Time-dependent effects of corticosteroids on human amygdala processing. The Journal of Neuroscience, 30(38), 12725–12732. https://doi.org/10.1523/JNEUROSCI.3112-10.2010

Hermans, E. J., van Marle, H. J. F., Ossewaarde, L., Henckens, M. J. A. G., Qin, S., van Kesteren, M. T. R., Schoots, V. C., Cousijn, H., Rijpkema, M., Oostenveld, R., & Fernández, G. (2011). Stress-related noradrenergic activity prompts large-scale neural network reconfiguration. Science, 334(6059), 1151–1153. https://doi.org/10.1126/science.1209603

Nicholson, A. A., Ros, T., Densmore, M., Frewen, P. A., Neufeld, R. W. J., Théberge, J., Jetly, R., & Lanius, R. A. (2020). A randomized, controlled trial of alpha-rhythm EEG neurofeedback in posttraumatic stress disorder: A preliminary investigation showing evidence of decreased PTSD symptoms and restored default mode and salience network connectivity using fMRI. NeuroImage: Clinical, 28, 102490. https://doi.org/10.1016/j.nicl.2020.102490

Pan, D., Jentsch, V. L., Langer, K., Hagedorn, B., Höffken, O., Wolf, O. T., & Merz, C. J. (2023). What a difference timing makes: Cortisol effects on neural underpinnings of emotion regulation. Neurobiology of Stress, 25, 100544. https://doi.org/10.1016/j.ynstr.2023.100544

Schumann, A., de la Cruz, F., Köhler, S., Brotte, L., & Bär, K.-J. (2021). The influence of heart rate variability biofeedback on cardiac regulation and functional brain connectivity. Frontiers in Neuroscience, 15, 691988. https://doi.org/10.3389/fnins.2021.691988

Thayer, J. F., Åhs, F., Fredrikson, M., Sollers, J. J., III, & Wager, T. D. (2012). A meta-analysis of heart rate variability and neuroimaging studies: Implications for heart rate variability as a marker of stress and health. Neuroscience & Biobehavioral Reviews, 36(2), 747–756. https://doi.org/10.1016/j.neubiorev.2011.11.009

Watanabe, N., Yoshida, S., Keerativittayayut, R., & Takeda, M. (2026). Neural signatures of human psychological resilience driven by acute stress. Proceedings of the National Academy of Sciences of the United States of America, 123(13), e2524075123. https://doi.org/10.1073/pnas.2524075123

Zhang, W., Kaldewaij, R., Hashemi, M. M., Koch, S. B. J., Smit, A., van Ast, V. A., Beckmann, C. F., Klumpers, F., & Roelofs, K. (2022). Acute-stress-induced change in salience network coupling prospectively predicts post-trauma symptom development. Translational Psychiatry, 12, 63. https://doi.org/10.1038/s41398-022-01798-0

About the Author

Fred Shaffer earned his PhD in Psychology from Oklahoma State University. He earned BCIA certifications in Biofeedback and HRV Biofeedback. Fred is an Allen Fellow and Professor of Psychology at Truman State University, where has has taught for 50 years. He is a Biological Psychologist who consults and lectures in heart rate variability biofeedback, Physiological Psychology, and Psychopharmacology. Fred helped to edit Evidence-Based Practice in Biofeedback and Neurofeedback (3rd and 4th eds.) and helped to maintain BCIA's certification programs.