How Grief Works: The Psychology and Neuroscience of Loss

What Is Grief?

Grief is the psychological, emotional, and physiological response to bereavement - the subjective experience of loss following the death of a significant person, encompassing sadness, yearning, anger, confusion, and often profound disorientation.

It is distinct from bereavement (the objective circumstance of loss) and mourning (the cultural and social expression of grief).

Contemporary grief science, developed through prospective longitudinal studies over the past three decades, has substantially revised older stage-based models, revealing that grief trajectories are highly variable, that resilience is the most common response, and that a minority of bereaved individuals develop a clinically significant condition called prolonged grief disorder.

In 1969, Elisabeth Kubler-Ross published On Death and Dying and introduced the world to the five stages of grief: denial, anger, bargaining, depression, acceptance.^[6] The book was intended to give voice to dying patients, to advocate for honest conversation about death in an era of medical paternalism.

It succeeded beyond any expectation. The five stages became one of the most widely known frameworks in psychology - mentioned in television shows, therapist offices, self-help books, and condolence cards.

They were also, as a model of bereavement, substantially wrong.

Not because Kubler-Ross was a bad observer - she was perceptive and humane, and the stages capture real emotional states that many bereaved people experience. But her stages were derived from dying patients coping with their own deaths, not bereaved individuals, and she never claimed they were a universal sequence.

Popular culture transformed her descriptive categories into a prescriptive journey, creating the implicit message that healthy grief should look a certain way, follow a certain order, and reach acceptance within a certain time.

The empirical science of grief, developed over the last three decades through prospective longitudinal studies of bereaved individuals, tells a different story - stranger, more hopeful in some ways, more challenging in others.

"Grief is not a disorder, a disease, or a sign of weakness. It is the price of commitment." - Colin Murray Parkes

Key Definitions

Bereavement - The objective situation of having lost someone through death. The state of having experienced the death of a significant person.

Grief - The psychological, emotional, and physiological response to bereavement: the subjective experience of loss, including sadness, yearning, anger, confusion, and sometimes profound disorientation.

Mourning - The social and cultural expression of grief: the rituals, behaviors, and observable processes through which grief is expressed in a cultural context. Mourning practices vary enormously across cultures.

Continuing bonds - The internal ongoing relationship that bereaved individuals maintain with the deceased, as described by Klass, Silverman, and Nickman (1996). The transformation of an external relationship to an internal one that provides psychological resources rather than requiring the person's physical presence.

Prolonged grief disorder (PGD) - A clinical condition (added to DSM-5-TR in 2022) in which grief symptoms remain severely impairing more than 12 months after loss (6 months for children), characterized by intense yearning, difficulty accepting the death, and significant functional impairment.

Formerly called complicated grief, pathological grief, or traumatic grief. Affects approximately 10-15% of bereaved individuals.

Dual process model - Stroebe and Schut's (1999) model proposing that adaptive grieving involves oscillation between loss orientation (confronting the grief) and restoration orientation (attending to the life adjustments loss requires).

Resilience trajectory - Bonanno's term for the most common grief trajectory: bereaved individuals who maintain relatively stable functioning and low distress throughout bereavement, without prolonged intense suffering. Found in approximately 40-65% of bereaved samples.

Grief work hypothesis - The historically dominant but empirically unsupported assumption that effective grieving requires active, sustained engagement with the pain of loss, and that failure to do so leads to delayed or complicated grief.

Disenfranchised grief - Grief for losses that are socially unsanctioned or unrecognized (loss of a pet, miscarriage, estranged relationship, public figure, divorce) - losses for which social support structures are not activated, potentially complicating adaptation.

What We Got Wrong: The Stage Model

The five stages of grief have been absorbed into culture so deeply that many bereaved people feel pressure to experience them in sequence, worry when they skip a stage, or feel inadequate when their grief doesn't follow the expected arc.

The scientific evidence does not support the stage model of bereavement.

George Bonanno, a clinical psychologist at Columbia University's Teachers College, has conducted the most rigorous large-scale prospective research on bereavement trajectories.

His method differs fundamentally from retrospective clinical observation: rather than asking people to recall their grief or observing clinical patients (who represent a selected, distressed subsample of the bereaved), he began with large community samples and assessed them repeatedly over time - before anticipated losses and following them.

The findings were surprising enough to reshape the field.

Grief Trajectories: What Actually Happens

Bonanno identified four to five distinct grief trajectories:^[1]

Resilience (~40-65%): The largest group. These individuals show relatively low distress before the loss (if anticipatable) and throughout bereavement. They grieve - they are sad, they feel the loss - but they maintain positive emotion, function effectively, and do not experience protracted suffering.

Recovery (~15-35%): Significant distress in the immediate aftermath of loss that gradually diminishes over 12-24 months, returning to pre-loss functioning.

Chronic grief (~10-15%): Consistently high distress that does not diminish over time. This is the population that benefits most from clinical intervention - the group for whom the stage model may be most relevant as a description.

Delayed grief (~5-10%): Low initial distress followed by later increase. This is the rarest trajectory and the one that has historically given rise to ideas about "unresolved grief" waiting to emerge. It exists but is uncommon.

Chronic depression (~5-8%): High distress that predates the loss and continues after, representing pre-existing depression complicated by bereavement rather than grief per se.

The most consequential finding: resilience is normal. Before Bonanno's research, clinicians widely assumed that low distress following significant bereavement reflected denial, suppression, or imminent delayed grief.

Bonanno found that resilient bereaved individuals were genuinely resilient - not suppressing emotion, capable of discussing their loss with full affect, maintaining psychological functioning, and not experiencing delayed breakdown.

This finding has direct clinical implications: attempting to induce grief in resilient bereaved individuals through mandatory counseling or pressure to "work through" the loss is not only unnecessary but potentially harmful.

The Neuroscience of Grief: What the Bereaved Brain Is Doing

Mary-Frances O'Connor at the University of Arizona has led the most rigorous neuroimaging research on grief, revealing that the bereaved brain is doing something more specific and more interesting than simply being sad.

The Reward System in Grief

O'Connor's fMRI studies (2008, 2012, 2021) found that when recently bereaved individuals viewed photographs of their deceased loved ones and reflected on the loss, the nucleus accumbens activated - a region central to reward processing, motivation, and craving.^[3]

This is not the expected finding for "sadness" or "depression." The nucleus accumbens activation suggests grief involves a craving component - the grief-stricken brain is, in part, driving toward reunion with the person who is gone.

This makes sense through the lens of attachment theory. John Bowlby proposed that the attachment behavioral system is a biologically based regulatory system evolved to maintain proximity to attachment figures - caregivers, partners, close bonds.

The system motivates proximity-seeking, generates distress when proximity is unavailable, and is soothed by reunion. When the attachment figure dies, the system continues firing: photographs, possessions, smells, sounds associated with the person trigger the yearning and proximity-seeking impulses of attachment, but reunion cannot occur.

Grief is, in part, an attachment system that cannot complete its circuit.

Complicated Grief and the Stuck Reward Circuit

O'Connor's follow-up studies found a striking difference between bereaved individuals who developed prolonged grief disorder (PGD) and those who did not. At 14 months post-loss, individuals with PGD showed stronger nucleus accumbens activation to photographs of the deceased than those without PGD.

This suggests a mechanism: in typical grief, the attachment-craving response gradually diminishes as the brain learns - through the repeated experience of not being reunited - that the person is no longer available. This is a process of extinction learning, updating the brain's predictions.

In PGD, this extinction learning is impaired: the reward-craving system remains hyperactive, continuing to generate intense yearning without the accompanying learning of absence.

This framework connects to the treatment implications: Complicated Grief Treatment includes structured exposure to grief-related stimuli (photographs, narratives of the loss) in a safe therapeutic context, potentially facilitating the extinction learning that PGD patients are failing to complete naturally.

Pain and Grief: The Overlap

Naomi Eisenberger's research on social pain has demonstrated that the neural substrates of social rejection and physical pain significantly overlap - particularly the dorsal anterior cingulate cortex (dACC), involved in the affective/distressing component of physical pain, and the anterior insula, which processes the felt quality of pain.^[7]

Grief activates both regions. This is not merely metaphorical: grief literally hurts in a neurologically meaningful sense, activating the same alarm systems that signal physical injury.

This has implications for how we talk about grief: the statement "my chest aches" from a bereaved person is physiologically meaningful, not simply a figure of speech. And it explains why attempts to "think your way out" of grief are limited - you cannot think your way out of a pain signal.

The Dual Process Model: Oscillation as Adaptation

Margaret Stroebe and Henk Schut, bereavement researchers at Utrecht University, developed the Dual Process Model (DPM) of grief in 1999 - now among the most empirically supported frameworks.^[2]

The DPM proposes that adaptive grief involves oscillation between two orientations, neither of which alone is sufficient:

Loss Orientation

Confronting the grief directly: allowing the pain to be felt, thinking about and yearning for the deceased, processing the relationship that has ended, working through feelings of guilt, anger, or regret, and building an ongoing internal relationship with the deceased in a new form.

This is what popular culture typically imagines as "doing grief work."

Restoration Orientation

Attending to the secondary consequences of loss: learning tasks the deceased performed, adjusting to a changed identity, managing practical demands, attending to new roles, and building new sources of meaning and connection.

Critically, this also involves temporarily stepping away from the grief - distracting oneself, experiencing positive emotion, not thinking about the loss.

The DPM's key insight is the oscillation: neither sustained immersion in loss orientation nor sustained escape into restoration orientation is adaptive. The bereaved person moves between them - engaging with the grief, then stepping back, then returning.

This captures what clinicians have long observed but stage models failed to represent: bereaved individuals do not grieve continuously. Even in early acute grief, there are moments of relief, humor, normal engagement. These are not failures to grieve properly; they are part of the adaptive process.

Studies testing the DPM find that oscillators - bereaved individuals who show dynamic movement between orientations - have better long-term outcomes than those who stay predominantly in one orientation. Persistent loss-orientation predicts complicated grief; persistent restoration-orientation predicts avoidance-related complications.

Continuing Bonds: Why "Letting Go" Is the Wrong Goal

For much of the twentieth century, the dominant therapeutic goal for grief was, explicitly or implicitly, "letting go." Freud's concept of "grief work" in "Mourning and Melancholia" (1917) proposed that the bereaved person must withdraw libidinal energy invested in the deceased - a painful process of recognizing the finality of death - and reinvest it in new objects.

Failure to do so was pathological.

This view shaped clinical practice for decades. Therapists encouraged bereaved clients to reduce their focus on the deceased, limit memorialization, and invest emotionally in present relationships.

In 1996, Dennis Klass, Phyllis Silverman, and Steven Nickman edited Continuing Bonds: New Understandings of Grief, collecting research from multiple cultures and clinical contexts that challenged the "letting go" model fundamentally.^[5]

The empirical observation was straightforward: most bereaved individuals who adapt well do not "let go." They maintain an ongoing internal relationship with the deceased - talking to them mentally, consulting an internalized sense of what they would have thought, feeling their presence, keeping meaningful objects, visiting graves, including them in family rituals.

Rather than withdrawing the bond, they transform it: from an external relationship requiring the person's physical presence to an internalized relationship carried within the bereaved person. The relationship changes form; it does not end.

Cross-cultural research was particularly important here. The "letting go" model reflected culturally specific Western assumptions. In many East Asian cultures, maintaining active ongoing bonds with ancestors - including ritual communication - is normative and expected.

In Mexico, Dia de Muertos practices maintain active ongoing engagement with deceased family members. In many African traditions, the deceased remain present members of the community.

These are not pathological failures to "move on"; they are culturally normative expressions of continuing bonds that appear to support adaptation.

The continuing bonds framework aligns with attachment theory: John Bowlby, late in his career, observed that what bereaved individuals achieve is not dissolution of the bond but creation of an internal representation of the attachment figure that functions as a "secure base" - a psychological resource rather than a physical absence.

Prolonged Grief Disorder: When Grief Becomes Entrenched

While most bereaved individuals adapt over time, approximately 10-15% develop Prolonged Grief Disorder (PGD) - grief that remains acutely impairing for 12 months or more.

Added to DSM-5-TR in 2022 and ICD-11, PGD is now recognized as a distinct clinical entity, separate from major depression (which is about global negative mood; PGD involves preserved positive affect in non-grief contexts and specific focus on the loss) and PTSD (though often co-occurring following traumatic bereavement).

Diagnostic Profile

The DSM-5-TR criteria require:

• Death of someone close (12+ months ago; 6+ months for children)
• Persistent intense yearning or longing for the deceased
• Intense grief reactions most days, severe enough to cause significant distress or functional impairment
• At least 3 of 8 additional symptoms (difficulty accepting the death, disbelief, emotional numbness, intense bitterness, feeling life is meaningless without the deceased, feeling part of self has died, inability to engage in activities or relationships)

Risk Factors

Treatment

Katherine Shear at Columbia University developed Complicated Grief Treatment (CGT), now called Grief-Focused Cognitive Behavioral Therapy, a 16-session structured psychotherapy specifically designed for PGD.^[4] A 2005 JAMA RCT found CGT produced 51% response rates versus 28% for interpersonal therapy - a significant advantage.

CGT components include:

• Revisiting: Structured narrative retelling of the circumstances of the death, similar to prolonged exposure therapy for PTSD - facilitating the extinction learning that PGD patients are failing to complete naturally
• Avoidance targeting: Identifying and gradually engaging with grief-related situations or activities the patient has been avoiding
• Situational restoration: Rebuilding life activities and goals around the new reality
• Promoting continuing bonds: Working toward a transformed, adaptive ongoing relationship with the deceased rather than demanding detachment

What Actually Helps: The Evidence on Grief Support

Research on grief interventions contains several counterintuitive findings.

Targeted, Not Universal Intervention

Meta-analyses of grief counseling interventions - particularly studies offering universal debriefing or counseling to all bereaved individuals regardless of clinical need - find neutral to slightly negative effects on average. This is initially surprising.

The explanation: when intervention is offered universally, the majority of participants (~60%) are following the resilience trajectory and do not need clinical intervention.

Treating them as though they should be more distressed, or as though their functioning represents unhealthy denial, may actually be iatrogenic - undermining natural resilience.

Universal grief intervention benefits the minority with genuine clinical need while doing little for or slightly harming the majority.

Targeted intervention for those showing high distress and functional impairment does show benefit. The implication: grief support should be offered, not mandated; sought, not prescribed.

Social Support Quality

Consistent evidence supports the importance of social support - but quality matters more than quantity. Bereaved individuals benefit from:

• Having people willing to listen without premature reassurance or minimization
• Social permission to mention the deceased, tell stories, maintain the memory
• Absence of pressure to grieve on a timeline or to "move on"
• Practical support addressing the secondary losses (tasks, roles, identity changes)

Disenfranchised grief - grief for losses that are socially unsanctioned (pet death, miscarriage, estranged relationships, non-romantic significant losses) - is associated with worse outcomes, likely because the absence of social recognition prevents the social support that facilitates adaptation.

Meaning-Making

Robert Neimeyer's research on meaning reconstruction in grief finds that finding or creating meaning from the loss is among the strongest predictors of adaptive outcomes.^[8]

This is not the toxic positivity of "everything happens for a reason" - it is the genuine reconstructive process of integrating the loss into an updated life narrative that can accommodate it.

Meaning-making can take many forms: religious or spiritual frameworks, finding purpose in advocacy or memorial activity, deriving wisdom from the relationship with the deceased, or simply coming to understand the loss as part of a coherent life story.

The Biology of Normal Grief

Grief is not purely psychological. It is a whole-body experience with physiological correlates.

Bereaved individuals show elevated cortisol, disrupted sleep architecture, impaired immune function, and increased inflammatory markers.

The highest-risk period for physical illness and mortality in bereaved spouses occurs in the weeks and months following loss - particularly for older widowers, who show dramatically elevated mortality rates compared to matched non-bereaved controls.

The mechanisms likely involve: grief-associated sleep disruption (sleep is a critical regulator of immune function and metabolic health), chronic stress-axis activation (elevated cortisol suppresses immune function), behavioral changes (reduced exercise, altered diet, alcohol use), and possibly direct neurological effects of grief on autonomic regulation.

The folk saying that people can "die of grief" has a physiological basis: bereavement-related mortality elevations are real, measurable, and largest in the period of most acute grief.

Grief Across Cultures and Time

Human grief is universal; its expression is culturally specific.

The continuing bonds pattern appears across virtually all cultures studied, but the forms it takes - ritual communication with ancestors, elaborate mourning ceremonies, anniversary observances, ongoing domestic presence of deceased household members - vary enormously.

The Western ideal of "moving on" and the diminishment of explicit bonds with the dead is itself a cultural artifact, not a biological imperative.

The duration of culturally prescribed mourning has varied dramatically across history. Victorian England prescribed mourning periods of years for widows; contemporary Western culture often implies a return to functioning within weeks.

Both extremes carry problems: excessive prescriptive mourning may entrench grief; excessive pressure to minimize it may disenfranchise legitimate need.

For related concepts, see what causes depression, attachment theory explained, why loneliness is deadly, and how to manage anxiety.

Sources & Further Reading

1. Bonanno, G. A. (2004). Loss, Trauma, and Human Resilience. American Psychologist, 59(1), 20-28. DOI: 10.1037/0003-066X.59.1.20
2. Stroebe, M., & Schut, H. (1999). The Dual Process Model of Coping with Bereavement. Death Studies, 23(3), 197-224. DOI: 10.1080/074811899201046
3. O'Connor, M. F., et al. (2008). Craving Love? Enduring Grief Activates Brain's Reward Center. NeuroImage, 42(2), 969-972. DOI: 10.1016/j.neuroimage.2008.04.256
4. Shear, K., et al. (2005). Treatment of Complicated Grief. JAMA, 293(21), 2601-2608. DOI: 10.1001/jama.293.21.2601
5. Klass, D., Silverman, P. R., & Nickman, S. (Eds.). (1996). Continuing Bonds: New Understandings of Grief. Taylor & Francis.
6. Kubler-Ross, E. (1969). On Death and Dying. Macmillan.
7. Eisenberger, N. I., Lieberman, M. D., & Williams, K. D. (2003). Does Rejection Hurt? An fMRI Study of Social Exclusion. Science, 302(5643), 290-292. DOI: 10.1126/science.1089134
8. Neimeyer, R. A. (2001). Meaning Reconstruction and the Experience of Loss. American Psychological Association.