How we know
In 2007, neuroscientist David Eagleman ran the cleanest test of the slow-motion claim ever published. The paper, Stetson, Fiesta and Eagleman in PLoS ONE, recruited 20 volunteers at a Dallas free-fall attraction called SCAD. Riders are released backwards from a 31-metre platform into a net below. The fall lasts roughly 2.5 seconds. People report it as the most frightening thing they have ever done.
Each volunteer wore a perceptual chronometer on the wrist. The device flashes a number, then its inverse, alternating faster than normal vision can resolve. Read the actual number, you have beaten your own perceptual frame rate. The team calibrated each participant's frame rate on the ground first. Then they fell.
If time genuinely slowed during fear, the falling participants should have read the numbers. They did not. Their accuracy in free-fall matched their accuracy on the ground. The clock in their heads ticked at the same speed all the way down.
What did change was their memory of the event. Asked afterwards to estimate the duration of their own fall, participants reported it as 36 percent longer than the actual time. The fall felt longer in retrospect but ran at normal speed live.
What it means
The mechanism is memory density, not real-time perception. High-arousal events flood the amygdala, which modulates the hippocampus to consolidate memory at a higher resolution than usual. You record more details per second of a frightening event. When you replay it from memory, the additional frames take longer to play back, so it feels stretched. This is the Cahill and McGaugh 1998 amygdala-consolidation finding, applied to subjective time.
Two practical implications. First, the experience of slow motion in a car skid is a memory artefact, not a perceptual superpower. You did not see more frames during the event. You stored more frames of the event. Second, if you want a vivid memory of any moment, novelty and emotion both increase memory density. Boring weeks blur. Frightening seconds stretch.
The watch did not lie
The chronometer made the difference. Every previous study of slow-motion-under-fear relied on self-report after the event, which cannot separate live perception from memory of the event. Eagleman's wrist chronometer ran during the fall, so it could measure perception alone. The test failed cleanly. Time did not slow.