A continuous, life-long video of a vertebrate from puberty to death has uncovered that aging is not a slow, steady decline—but a series of abrupt shifts through stable stages.
🎧 ▶️ Press the play button below to listen in English.
The Problem with Watching Grass Grow
For centuries, we have described aging as a gradual process. Wrinkles appear slowly. Muscles weaken over the years. Memory fades bit by bit. But is that really what happens inside a living creature? Or is our snapshot view of aging—measuring someone at age 30, then again at 60, then at 90—missing something fundamental?
Imagine trying to understand a movie by looking at three random frames. You would see the beginning, middle, and end, but you would completely miss the plot twists, the sudden scene changes, and the quiet moments that define the story.
That has been the state of aging research until now.
The problem is simple but brutal: most vertebrates live too long. A mouse lives 2–3 years. A human lives nearly 80 years. Recording every single movement, every day, from adolescence to death, has been impossible. Scientists have had to settle for snapshots.
But a team at Stanford University found an extraordinary loophole.
The African Turquoise Killifish: Nature’s Time-Lapse Camera
The African turquoise killifish (Nothobranchius furzeri) lives its entire adult life in just 4 to 8 months. It matures from a juvenile to a breeding adult in three weeks. Within half a year, it will be old and dying.
This compressed lifespan makes the killifish a natural time-lapse camera for aging.
Claire Bedbrook, Ravi Nath, and their colleagues built a custom recording system: 12 individual tanks per camera, infrared lighting so the fish could be filmed in complete darkness, automated feeders, and cameras running 20 frames per second, 24 hours a day, from puberty until natural death.
They recorded over 30 billion frames of video.
Then came the real challenge: making sense of it.
Teaching Computers to Watch Fish
A human watching 24 hours of fish video would go mad. Watching 250 days? Impossible.
So the team trained deep neural networks—the same kind of artificial intelligence that powers facial recognition and self-driving cars—to track six specific points on each fish’s body: snout, midbody, endbody, tail, fan, and sidebody.
From these tracked points, the computer calculated 57 different “pose features” every single frame: velocity, acceleration, body curvature, direction changes, sleep states, and more.
Then they used a mathematical technique called a hidden Markov model (HMM) to break down the fish’s continuous movement into 100 repeating “behavioral syllables”—think of them as the letters in the alphabet of fish movement.
Some syllables were active: darting, flipping, glass-surfing. Others were passive: drifting, pausing, sleeping.
For the first time in any vertebrate, scientists could watch a complete behavioral dictionary unfold over an entire lifetime.
The First Surprise: Aging Is Not a Straight Line
When the researchers looked at how behavior changed with age, they expected to see a smooth, predictable decline. Older fish would simply slow down, sleep more, and move less.
That is not what they found.
Instead, individual fish followed dramatically different aging trajectories. Some aged “well.” Some aged poorly. And these trajectories diverged early in life—long before middle age.
At just 100 days old (roughly equivalent to a 40-year-old human), fish that would eventually live long lives were already different from those destined to die young:
- They were more active during the day.
- They had higher peak swimming speeds (sprinting ability).
- They showed tighter circadian timing—sleeping mostly at night and being active during the day.
Fish destined for short lives, by contrast, already slept more during the day and had lower peak velocities.
This was astonishing. Behavior at a relatively young age seemed to forecast the rest of an animal’s life.
Predicting the Future from a Few Days of Video
The team took this observation to its logical extreme.
They built machine learning models—behavioral clocks—that took a single day’s worth of behavior as input and predicted the fish’s age. The model was remarkably accurate, with a correlation of 0.94 between predicted and true age.
But the real test was forecasting remaining lifespan.
Using only behavior from fish at 70, 90, or 110 days old (well before middle age), the model could classify individuals as “short-lived” (less than 200 days total) or “long-lived” with better than 70% accuracy.
When they plotted the actual survival curves of fish predicted to be short-lived versus long-lived, the curves separated dramatically. The model was not guessing. It was reading the future in the way a fish moved, slept, and explored its tank.
The Molecular Echo: What the Liver Revealed
Behavior is the outward expression of what is happening inside the body.
At 150 days old, the researchers sacrificed a subset of fish and sequenced RNA from eight different organs: brain, gut, heart, kidney, liver, skin, spleen, and testis.
The liver stood out.
The liver transcriptomes of fish on short-lifespan trajectories were clearly different from those on long-lifespan trajectories. The short-lived fish showed increased expression of genes related to ribosome biogenesis—the cellular machinery that builds proteins.
Here is the fascinating part: the same genes that were elevated in short-lived fish were also elevated in old fish compared to young fish. In other words, short-lived fish looked molecularly older in their livers, even at the same chronological age as long-lived fish.
Immune-related genes also increased with age, but they did not differ between short-lived and long-lived fish. That suggests inflammation is a marker of aging but not necessarily a driver of lifespan differences. Ribosome biogenesis, however, may be a genuine driver.
The Second Surprise: Aging Happens in Sudden Jumps
Perhaps the most unexpected finding came when the researchers plotted behavioral syllables day by day across each fish’s entire life.
They expected a gentle slope.
Instead, they saw plateaus and cliffs.
A fish would maintain a stable set of behaviors for weeks. Then, abruptly—often within a single day—it would shift to a new, equally stable behavioral stage. These transitions happened at characteristic ages: youth, mid-life, late-life, and near death.
Between transitions, days looked remarkably similar to each other. Across transitions, days looked completely different.
The researchers used change-point detection algorithms to identify these shifts objectively. Most fish went through three or four abrupt transitions in their adult lives. Some went through five or six.
They built a hidden Markov model of “life stages” and found six distinct, stereotyped stages that fish progressed through in order. Young fish started in one stage. Then about half split into one mid-life stage and half into another. Later, they split again into two different old-age stages.
Aging, in this view, is not a continuous decline. It is a series of stable plateaus separated by rapid transitions—more like a staircase than a ramp.
This architecture of aging resembles the staged progression of embryonic development. Perhaps, the authors suggest, the body continues to reorganize itself in discrete steps throughout life, not just at the beginning.
Dietary Restriction Slows the Clock
One of the most robust lifespan-extending interventions across the animal kingdom is dietary restriction—eating less, or eating only during certain times.
The team put killifish on a restricted feeding schedule (three meals per day, all in the morning) compared to ad libitum feeding (seven meals spread throughout the day).
Dietary restriction did not change the shape of the aging trajectory. Fish still progressed through the same stages in the same order.
But they moved through those stages much more slowly.
At 110 days old, dietary-restricted fish were predicted by the behavioral clock to be only about 68 days old behaviorally. Their behavior aged at a significantly slower rate than ad libitum-fed fish.
This suggests that dietary restriction does not fundamentally rewrite the aging program. It simply turns down the speed dial.
Sex Matters: Females Age Differently
When the researchers tracked female killifish, they found a strikingly different pattern.
Females aged along a trajectory that resembled short-lived males, not the average male. They were more active when young, but then aged rapidly.
At ages over 100 days, the behavioral clock predicted that females were 50–200 days older than their true chronological age.
The model trained on male behavior still correctly classified 74% of females as short-lived based on their behavior at 70 days old. The behavioral signatures of short lifespan—disrupted circadian timing and increased daytime inactivity—were shared across sexes, even though the overall trajectories differed.
This is a crucial reminder that aging is not a one-size-fits-all process. Sex-specific biology matters profoundly.
What You Can Take Home: Lessons from a Tiny Fish
You are not a fish. But you share the same vertebrate biology. Here is what the killifish study teaches you—and what you can actually do about it.
1. Aging happens in sudden leaps, not a slow slide.
Stable periods are followed by abrupt transitions. Do not mistake feeling fine for staying fine forever.
What you can do: Monitor your energy, sleep quality, and physical capacity regularly. Small declines left unaddressed can become sudden cliffs.
2. Your midlife behavior forecasts your future.
Long-lived fish were more active during the day, slept mostly at night, and had higher peak speeds.
What you can do:
- Walk fast for 10–20 minutes daily. Peak velocity mattered more than average activity.
- Sleep 7–8 hours at night. Avoid daytime napping if it disrupts night sleep.
- Stay active during daylight hours. Morning and early afternoon movement matters most.
3. Your circadian clock is a health organ.
Disrupted timing—sleeping during the day, being inactive when you should be alert—was linked to shorter life.
What you can do:
- Get bright light within 30 minutes of waking.
- Dim lights and stop screens 1–2 hours before bed.
- Eat at consistent times each day. Consider time-restricted feeding (eating all meals within an 8–10 hour window).
4. More protein production is not always better.
Short-lived fish made more proteins but cleaned up less cellular debris. Balance matters.
What you can do: Prioritize quality protein (fish, eggs, legumes) over quantity. Support cellular cleanup with regular exercise, adequate sleep, and occasional fasting (12–16 hours overnight).
The bottom line
The fish had no choice in their environment. You do.
Protect your sleep. Move during the day. Eat on a schedule. Live like a long-lived fish.
Don’t Get Sick!
About Dr. Jesse Santiano, MD
Dr. Santiano is a retired internist and emergency physician with extensive clinical experience in metabolic health, cardiovascular prevention, and lifestyle medicine. He reviews all medical content on this site to ensure accuracy, clarity, and safe application for readers. This article is for educational purposes and is not a substitute for personal medical care.
💡 Support This Work
Creating well-researched articles, maintaining this website, and keeping the information free takes time and resources.
If you found this article helpful, please consider donating to support the mission of empowering people to live healthier, longer lives, without relying on medications.
🙏 Every contribution, big or small, truly makes a difference. Thank you for your support!
Follow me on Facebook, Gab, Twitter (formerly known as X), Instagram, and Telegram.
Related:
- The Counterclockwise Study: Proof That Mind Shapes The Body
- Sarcopenia: The Scourge of Aging
- Telomeres, Cancer, And Lifestyle: Unpacking The Anti-Aging Paradox
- Unlocking Music and Brain Health: Secrets to Aging Better
- Discover Hidden Brain Aging Signs You Must Address Now
- Telomeres And Daily Habits: A Simple Guide to Slower Aging
- Boost Survival Now: Protein’s Power In Aging CKD Kidneys
- Blood Sugar’s Hidden Aging Effects: From Wrinkles To Weakness
- The Effects of Kettlebell Exercises for Women 65 and OlderCoffee and Tea for Osteoporosis Prevention: A 10-Year Study
- How Exercise Speaks to the Brain and Prevents Alzheimer’s
References:
- Bedbrook, C. N., Nath, R. D., Zhang, L., Linderman, S. W., Brunet, A., & Deisseroth, K. (2025). Life-long behavioral screen reveals an architecture of vertebrate aging. bioRxiv. https://doi.org/10.1101/2025.11.21.688112. (Preprint)
- Bedbrook, C. N., Nath, R. D., Zhang, L., Linderman, S. W., Brunet, A., & Deisseroth, K. (2025). Life-long behavioral screen reveals an architecture of vertebrate aging. Science. https://doi.org/10.1126/science.aea9795. (Published Version)
Disclaimer:
This article is for educational purposes and is not a substitute for professional medical advice, diagnosis, or treatment. Always consult your physician before making health decisions based on the TyG Index or other biomarkers.
© 2018 – 2026 Asclepiades Medicine, LLC. All Rights Reserved
DrJesseSantiano.com does not provide medical advice, diagnosis, or treatment
As an Amazon Associate, I earn from qualifying purchases
Discover more from Don't Get Sick!
Subscribe to get the latest posts sent to your email.