Are You Reading the Room, or an Old Wound?

In her book How Emotions Are Made, neuroscientist Lisa Feldman Barrett tells a story of going on a date with a guy from her lab that she wasn’t really into. As the date progressed, however, she found herself increasingly more attracted to him. Her face flushed, her heart rate increased, and she started feeling butterflies in her stomach.

After agreeing to a second date, Barrett went home, got into bed, and had the flu for a week.

In other words, those bodily sensations were an incipient illness, not signs of sexual attraction. But when her brain had to interpret those sensations, it did so through the context of a dinner date. So it predicted, “You’ve got the hots for this dude.”

She wasn’t lying to herself. She was simply doing what all of us do, all the time: constructing reality from incomplete data.

This is the predictive brain at work.

The Predictive Brain

You don’t experience the world raw, without filters. (You’d be bombarded with so much data, you’d be completely unable to function.)

Instead, you experience a model of the world that your brain constructs on the fly, assembled in about half a second from sensory data, memory, and your brain’s best guess about what’s happening, what it means, and what you need to do about it.

Most of the time, the model is close enough for you to function successfully. You bring your fork to your mouth instead of your eye, you don’t walk in front of traffic, and you laugh at the punch line of the joke.

And that model does get things wrong — a lot. Most of the time, that’s not a problem. Your brain compares incoming sensory data with its prediction and updates its model.

Prediction, comparison, prediction error, updated prediction.

That loop is your brain at its best — learning in real time.

But sometimes prediction errors don’t get corrected. Say you interpret a harmless glance as accusatory, or the innocent absence of an instantaneous response to a text message (“my phone died”) as anger.

Or your stomach drops before a team all-hands, and your brain codes that as Something's wrong with the vibe in here — what are they not telling me? rather than I just drank too much coffee.

When that happens, you don’t feel wrong. Your brain doesn’t update. To the contrary; you feel certain. Certain that you’re under attack, and there’s a threat that needs to be dealt with.

You feel that certainty as evidence of objective truth, when it’s actually nothing more than evidence of your brain’s confidence in its own prediction.

Instead of learning, you’re stuck in a self-sealing feedback loop.

What’s The Origin of Distorted Predictions?

Borrowing from Michael Singer’s The Untethered Soul, I call it a “thorn”: an old wound, encoded under conditions of real threat or loss, that got locked into your prediction algorithm as “the most important thing to consider here.”

What Andy Clark refers to in his book The Experience Machine, in the charmingly technical terminology of neuroscience, as “a heavily weighted prior.”

A heavily weighted prior — a thorn — isn’t a memory that you can visit; it’s a pair of glasses you see through without even realizing that you’re wearing them.

And because you can't see the glasses, you assume what you're seeing is the world. As meditation teacher and author Richard Dixey puts it in his book Three Minutes a Day, “We react to the display we ourselves have created as if it is unreservedly real, actually the case.”

Your brain mistakes its own construction for ground truth. It refuses to update because it’s unable to find evidence that disconfirms its prediction.

That’s what I refer to as a “glitch”: not a broken brain, but a brain unable to replace outdated data with new learning.

When Predictions Harm

Barrett’s flu da