Healing From Grief & Loss | Dr. Mary-Frances O'Connor

The episode begins by separating two words that ordinary language merges. Grief, in O'Connor's usage, is the state a person could report if asked to rate it from one to ten in a given moment. Grieving is the way that state changes over months and years. The distinction is not semantic housekeeping. It blocks a common inference: that a wave of grief arriving late is proof the earlier grieving failed.

grieving is the way that grief changes over time

Her analogy is a market index. Daily moves can be violent while the yearly trend still has a direction. Applied to bereavement, the relevant question is not whether the wave ever returns but what trajectory the waves sit inside. That is how she can hold two claims at once: grief never disappears because it is a human emotion, and grieving is still a process with an outcome. A person who cries on the anniversary of a death twenty-five years later has not necessarily failed to grieve; the wave and the learning are different time scales.

The explanatory frame is attachment, and O'Connor treats attachment as prior to grief rather than as one coping style among many. Her claim is that loss cannot be modelled without first modelling the bond, because the intensity of the response is proportional to what was formed. This is why the episode keeps returning to the bond rather than to the funeral, the diagnosis, or the stages.

Attachment, as she describes it, writes a contract: I will always be there for you, you will always be there for me, and time or distance will not change that. The contract is not a conscious promise. It is closer to a default setting that organises attention, and it explains why absence normally increases pursuit rather than acceptance.

In an ordinary absence, the correct response is to think about the missing person more, to put energy into finding them, or to make enough noise that they find you. Death is the case the algorithm cannot represent. The search has no terminal state, because the person cannot be found and cannot answer. O'Connor calls the resulting configuration the gone but also everlasting theory: the cortical record says the death happened, while the attachment system continues to represent the person as merely elsewhere.

I think of this as the gone but also everlasting theory

A wave of grief, on this account, is what conflict between the two streams feels like when it becomes conscious. The person is gone and the person is everlasting, and both propositions cannot be true at the same time. The model predicts that reminders which force the contradiction will be the most painful, and it explains why the pain is not proportional to how much the person is missed in the abstract but to how specifically the absence is encountered.

Anticipated death complicates the picture without dissolving it. O'Connor reports that closure conversations in hospice care are helpful later, because they leave something to reflect on: saying I love you, thank you, I forgive you, please forgive me. But such conversations do not necessarily change the attachment biology. In her studies, a participant will describe a long palliative course, a terminal diagnosis, and then answer that the death was sudden, because the belief in the person's continued existence does not update on evidence.

the belief that they will always be there does not respond to logical thinking

This is also why a bereaved person may still reach for the phone to text someone who has been dead for months. The behaviour is not a failure of memory. It is what a representation looks like when it was built to survive distance: they do not have to be in my time and space for them to exist. The clinical task is therefore not to prove the death again, but to transform the relationship so that it can continue without the body.

The neuroimaging work gives the model a measurable centre. Participants rated how much they were yearning for the deceased, then viewed photographs of that person and of a stranger. The contrast that mattered was not person versus stranger in general but the participant's own person, the one being yearned for. The more yearning a participant reported, the more activity appeared in the nucleus accumbens, a reward-learning region in the ventral striatum. All bereaved participants also showed memory, emotion, emotion-regulation, and autonomic areas, which is unsurprising when the stimulus is a photograph dense with history. What varied with yearning was the accumbens response.

The interpretation is motivational rather than hedonic. Yearning is the hallmark of grieving, and yearning is a form of wanting. Dopamine is described here not as pleasure but as the effort a person would invest to obtain a wanted thing. The question the accumbens is answering is not how good the loved one feels but how much effort would be spent to see them again, one more time. That reframing turns grief from a passive emotion into an action tendency with a target that cannot be reached.

how much effort would you put in to get this thing you want

O'Connor resists the addiction label that the accumbens finding invites. Her counterexample is thirst. Lost in the desert without a water bottle, a hiker can think of nothing but water, yet no one would call that addiction. Water is a homeostatic need, and the obsession is the signal that the need has become urgent. Attachment figures, she argues, belong to the same class of requirement: we need them as we need food and water, and we fail to thrive without them. The accumbens signal is the cue to reach out, not evidence of a hijacked reward system.

We need our attachment figures like we need food and water.

The addiction contrast is still instructive, because it marks where the analogy stops. Drugs of abuse override the same circuits, but they do so by changing receptor number or affinity, narrowing what is rewarding until only the drug can fill the receptors. A loved one does not narrow the world in that way. The deficit after death is not a new stressor piled onto an already full plate; it is the removal of a capacity that was built into the person. That is why the practical consequence is not detoxification but replacement or redistribution: the attachment need has to be met somewhere, whether through another relationship, a continuation of the internal bond, or another form of connection.

Bowlby's division of grief responses into protest and despair is the episode's behavioural spine. Protest is the toddler missing in the grocery store: the body is primed to search, and the feeling is the go signal. Despair is the letter addressed to the spouse who will never open it: the recognition that the search cannot succeed, accompanied by withdrawal and lethargy. The information is the same in both cases. What differs is whether the system treats the absence as a problem to be solved or as a fact to be absorbed.

Despair is usually treated as the failure state, but O'Connor gives it a function. It stops the search, and the search is metabolically expensive: blood pressure, cortisol, and sustained action all cost energy. Withdrawal conserves resources when the caregiver or mate is gone and the time of return is unknown. The argument is evolutionary rather than moral. A system that kept searching indefinitely would exhaust itself, so despair imposes a pause during which another way of living can be learned.

despair has the function of stopping us from searching