There is a particular kind of memory that researchers don’t have a clean name for yet, but astronauts show it constantly. They can tell you, fifteen years after a mission, about the flight surgeon who remembered their kid’s birthday. They cannot tell you what the mission commander said about their performance in the post-flight debrief. One memory is crisp. The other has dissolved.

I know this pattern from the research literature. I also know it because I lived inside it for twenty years without recognising what it was doing to me. I could recall, in perfect detail, every kindness my ex-wife ever showed me — the coffee left on my desk during a deadline, the flight she rebooked when my conference ran over. I could not recall a single compliment she paid me. Not one. The asymmetry was not a quirk. It was a fault line, and eventually my marriage fell along it.

In interviews with astronauts and crew members, this same asymmetry appears consistently. It is not a quirk of the astronaut personality. It is a cognitive pattern that shows up across populations, across professions, across isolation studies. And it has measurable consequences for how people hold up under stress.

The asymmetry is real, and it has a name

The phenomenon psychologists call the Fading Affect Bias describes the tendency for the emotional intensity of unpleasant memories to decay faster than that of pleasant ones. It is usually framed as adaptive. A mechanism that lets us carry our lives forward without being crushed by every old humiliation.

But the bias is not uniform across everyone. And it is not uniform across types of positive memory either. Some people show a robust fading of negative affect. Others show the opposite — the emotional charge of criticism stays vivid for decades while the compliments fade like cheap ink.

The people who remember every small kindness but cannot recall a single compliment about themselves fall into a specific cognitive profile. Their memory system is not broken. It is selectively tuned. Acts of kindness directed outward — what someone did for someone else, what a colleague did for them in a practical sense — get encoded and retained. Compliments about the self, statements of praise, affirmations of worth, do not survive the encoding process intact.

Why kindness lands and compliments don’t

The distinction between remembering kindness and remembering compliments is not trivial. It reflects two different encoding pathways.

A kindness is an event. Someone did something. It has a time, a place, a concrete action. It can be segmented into the memory structure cleanly — research on event segmentation describes this as a discrete unit that can be retrieved later. The flight surgeon remembered the birthday. You can picture the card, the office, the expression. The event has edges.

A compliment is different. A compliment is a claim about you. To retain it, you have to accept it as accurate information about yourself. If your self-model says you are mediocre, a compliment is an anomaly — and anomalies get filtered. They don’t fit the schema. They get encoded weakly, if at all, or they get reinterpreted on arrival. People often dismiss compliments by telling themselves the person is just being polite or that similar praise is given to everyone.

The compliment is edited before it can be archived. The kindness is archived before it can be edited.

What the data shows, and what it misses

Research on cognitive biases has examined whether they prospectively predict anxiety and depression — and the key word is prospectively. The biases come first. The symptoms come after. This is not simply depressed people remembering sad things because they are depressed. The memory pattern itself is a risk factor.

Knowing about this does not protect you from it. If anything, the intellectual framework can make you more efficient at dismissing your own experience. You can explain your symptoms to yourself faster than you can feel them. I did this for years — reading the literature on cognitive bias by day, driving home at night unable to remember a single nice thing anyone had said about my work. Understanding the mechanism does not deactivate it.

The isolation chamber version

This pattern plays out in isolation studies with notable consistency. Crews in Mars analogue facilities, winter-over teams in Antarctica, parabolic flight candidates in week-long confinement protocols. After the mission, when participants are asked in structured debriefs to describe what had sustained them during the hardest weeks, the answers are almost always about acts done by others.

A crewmate who quietly took over a task. A ground controller who changed the music playlist when she noticed someone was flagging. Someone who saved the last of the good coffee for a colleague who had a bad night. These were remembered with almost photographic specificity.

When the same participants are asked what their crewmates had said about them — how they had been praised, what strengths had been named — they struggle. Research suggests that some participants in isolation studies fail to recall positive feedback even when records show they received it, or cite the actual study with proper attribution if it exists.

Research suggests this may be a developmental adaptation — people who grew up in environments where warmth was scarce learned to record it carefully. The archivists are often people who had, somewhere in their history, needed to know exactly where affection came from and how to find it again.

The cost of the asymmetry

An adaptive pattern in childhood is not necessarily adaptive in adulthood. A person who archives kindness and discards praise is carrying an incomplete dataset about themselves. They know, in fine detail, that others have been generous. They do not know, in any detail, that they are respected, competent, or loved in a general sense.

This produces a specific kind of exhaustion. The person appears to everyone else as functional, even thriving. People often describe them as doing fine or appearing functional. Internally they are running on a model that tells them they are getting away with something, that the kindness they receive is not quite earned, that the next assessment could go either way.

In long-duration crews, this produces a specific fault line. The person with asymmetric memory is often the most conscientious crew member. They notice other people’s needs. They remember what was done for them. They are disproportionately valuable in the social functioning of the group. And they are disproportionately likely to deteriorate in the third quarter of a mission, when the sustaining social feedback they need is present but uncollectable.

The mechanism underneath

There is a neuroenergetic dimension to this that is increasingly persuasive. Research on cognitive biases in chronic illness and on stress-related encoding suggests that high cognitive load and chronic physiological stress reduce the fidelity of contextual encoding. The brain, under sustained demand, gets worse at laying down rich, distinctive memory traces.

When encoding fidelity drops, what survives tends to be what the existing schema confirms. Negative self-assessments, already well-rehearsed, get reinforced. Compliments, which require the schema to update, get filtered out as noise. The asymmetry deepens under stress, which is precisely when the person needs the countervailing data most.

This is why astronauts in the third quarter of long missions often become harder on themselves, not easier, even as their performance metrics remain strong. The environment supplies no shortage of positive feedback. The encoding system has narrowed to the point where the positive feedback cannot enter.

What I missed in my own life

I will say the rest of it plainly because it is relevant. I experienced a divorce in part because I had prioritised my research career over my marriage — years of choosing the work, the next study, the next mission protocol, over being genuinely present. I remembered every kindness my partner showed me. I could not remember a single compliment. And I could not, it turned out, offer the sustained attention that a marriage actually requires when my encoding system was tuned entirely toward the work.

The asymmetry was part of my problem. It took years of therapy to notice that I was carrying a memory system that was selectively deaf to one class of information, and that I had built an entire life around the kind of focus that made the deafness worse. I had read the literature on cognitive bias. I had not applied it to myself.

What can actually be done

The clinical literature is cautious about claiming that cognitive biases can be easily retrained. But there are some things that seem to help, both in astronaut training contexts and in therapeutic settings.

External storage is one. If you cannot retain compliments internally, retain them externally. A file. A notebook. Specific statements, with attribution, with date. When the internal model tells you that no one has ever said anything good about you, the external record disagrees. It does not feel emotionally real at first. Over time, rereading the record can weaken the schema enough that new praise gets through.

Segmentation is another. Because event-based memory is stronger than evaluative memory, compliments that are tied to a specific action tend to stick better than abstract praise. Abstract praise like telling someone they were excellent tends to fade, whereas specific feedback tied to concrete actions (such as explaining exactly what someone did well in a particular moment) creates an event structure that memory can attach to more easily.

Slow relationships are the third. We’ve written about the people who maintain one close friendship for decades, and one of the underappreciated functions of such a relationship is that it becomes an external memory for your own worth. A person who has known you for thirty years remembers the times you were competent, generous, or brave, even when you do not. You can borrow their memory until yours works better.

The connection to trust

In my recent piece on the hardest part of being trusted, I wrote about the astronauts who stop being checked on because they are assumed to be fine. The asymmetric memory pattern compounds this. Someone who cannot retain compliments is particularly vulnerable to the silence of competent performance. They are not receiving internal feedback, and the external environment has stopped providing it because they look like they don’t need it.

The quietest people in high-functioning crews are often the ones running lowest on archived positive data. They look self-sufficient. They are not. They are running on a depleted record and no one is topping it up because no one knows the record is depleted.

This is where memory asymmetry and trust converge into something genuinely dangerous. Trust, in its healthiest form, is built on a shared record — I remember what you have done for me, and I remember what you have said I am worth to you. But asymmetric memory corrupts that record. You trust that someone has been kind, because you remember the kindness. You do not trust that someone values you, because the evidence of that valuation has been silently discarded by your own encoding system. The result is a person who can trust in the goodness of others but not in their own place within that goodness. They believe in generosity as a principle. They cannot believe it is directed at them in particular.

And so the most trusted people — the ones everyone assumes are fine, the ones who remember every small kindness others have shown — become the least able to trust the one piece of information that would actually sustain them: that they are valued not for what they do, but for who they are. The record of that fact exists in other people’s memories. It has never survived in their own.

What I want readers to take from this

If you recognise yourself in this description — if you can describe in detail the kindnesses you have received and cannot, under any conditions, remember a single specific compliment — you are not broken and you are not uniquely flawed. You are running a memory system that was probably adaptive at some point in your life and is no longer serving you.

The literature on stress and cognition suggests that these patterns are more modifiable than they feel. Not by insight alone. Insight is insufficient. By deliberate practice of external recording, by asking for specific rather than general feedback, and by the slow patient building of relationships with people who will reflect your worth back to you until you start to believe them.

The people around you have probably said more generous things about you than you remember. That is not a failure of their affection. It is a feature of your encoding. And encoding, unlike history, can be changed.

Photo by Katya Wolf on Pexels