Memory is the difference between a week of illness and not noticing at all. It isn't one thing, but several cell populations, each providing a different form of protection.
What stays behind
| Cell | Where it lives | What it does |
|---|---|---|
| Long-lived plasma cells | Bone marrow | Make antibody constantly, for decades, with no need for a further signal. This is why you have antibody in your blood years later. |
| Memory B cells | Blood, spleen, lymph nodes | Silent until needed, then become plasma cells within days or refine further. |
| Central memory T cells | Lymph nodes | Long-lived, and able to multiply a lot on meeting the target again. |
| Effector memory T cells | Blood and tissues | Patrol and act fast, without so much need to multiply. |
| Tissue-resident memory T cells | Skin, gut, lung and so on | Stay in place where the infection came in, and act within hours. |
Why the second time is so much better
- More cells. Maybe a hundred or more times as many matching cells as before.
- Lower threshold. Memory cells wake more easily and don't need as much from dendritic cells.
- Better weapons. Antibodies have already been refined and class-switched.
- Already in position. Some cells are in the tissue, and antibody is already in the blood.
How long it lasts
It varies enormously. One long study of 45 people, followed for up to 26 years, estimated how quickly antibody declined.
| Antigen | Estimated antibody half-life |
|---|---|
| Measles, mumps | Over 200 years, so effectively lifelong |
| Chickenpox (varicella-zoster) | About 50 years |
| Diphtheria | About 19 years |
| Tetanus | About 11 years |
Those are estimates from a small group, but the pattern is clear. Live viral infections give stronger, longer memory than a protein vaccine does. Protection and blood antibody aren't always the same, either: a falling level may still be enough, or memory cells can make up the difference.
Why you get things again
When an illness comes back, it's usually not because you forgot.
| Reason | Example |
|---|---|
| The target changes | Flu mutates every year. Rhinoviruses, the main cause of colds, come in over 160 types. |
| The infection stays on the surface | RSV and common cold coronaviruses replicate in the lining of the airways, where blood antibody doesn't reach well, and mucosal IgA falls away within months. The virus can get going before the memory cells arrive. |
| The virus never left | Herpesviruses reactivate rather than reinfect. |
| The response was weak | Very young or very old people, or those on immune-suppressing drugs. |
| The first response stays dominant | Original antigenic sin, or imprinting: your response to a new flu strain may go back to the antibodies you made to the first strain you ever met, which may fit poorly. |
Memory can be wiped
Measles infects memory cells. A 2019 study in unvaccinated children found that measles erased between 11% and 73% of their existing antibody repertoire. Recovery only came through meeting those infections all over again. This is called immune amnesia, and it's a reason why measles raises the death rate from other infections for years. The MMR vaccine doesn't do this.
Can innate cells remember? Innate immunity was thought to have no memory, but some exposures, including the BCG vaccine, seem to leave innate cells more responsive for months, through lasting changes to how their DNA is packaged. This is called trained immunity. The cell-level finding is solid. Whether it protects people against unrelated infections in practice is still being argued over, and several large BCG trials have been disappointing.
Memory is cells waiting in several places. It's long-lasting for most things, and when you do get reinfected, it's usually because the virus changed, stayed on a surface, or was never gone.