Stanford study finds aging occurs in bursts rather than gradual decline
Researchers at Stanford University have found that aging may unfold through sudden transitions rather than a steady, gradual decline. The findings come from an experiment that tracked nearly every movement of small fish throughout their lives and showed that early behavioral patterns can predict how long an individual will live.
The study, published on March 12 in the journal Science, followed 81 African turquoise killifish continuously from early adulthood until natural death. Scientists analyzed billions of video images to build what they describe as a complete behavioral map, or “behaviome,” of aging in a vertebrate species.
The research team, led by postdoctoral scientists Claire Bedbrook and Ravi Nath along with senior authors Anne Brunet and Karl Deisseroth, created an automated monitoring system. Each fish lived alone in an aquarium equipped with cameras that recorded every moment of its life.
Using computer vision and machine learning tools, the system identified roughly 100 distinct “behavioral syllables.” These represent fundamental units of activity and rest that together form the animals’ full behavioral repertoire.
The killifish, which naturally live between four and eight months, did not show a smooth decline over time. Instead, they experienced two to six rapid transitions between relatively stable behavioral stages. The pattern resembles phase transitions, with long periods of stability interrupted by sudden reorganizations in behavior.
These results mirror findings from molecular studies of aging in mammals, including humans, where waves of biomolecular changes have been observed in midlife and later adulthood.
The study also revealed that signs of longevity appear early. Between about 70 and 100 days of age, fish that eventually lived longer already behaved differently from those with shorter lifespans.
Long lived fish tended to swim more vigorously, reach higher speeds and concentrate most of their sleep during the night. Fish that died sooner showed more daytime naps and irregular activity patterns even in relatively early adulthood.
Bedbrook said the findings suggest that behavioral changes early in life may provide clues about long term health and lifespan.
The researchers trained a machine learning model described as a “behavioral clock.” Using only a few days of behavioral data from midlife, the model could reliably estimate how much time a fish had left to live.
Molecular analysis supported the behavioral findings. Fish with accelerated aging patterns showed coordinated changes in liver gene expression related to protein synthesis and cellular maintenance, indicating a biological basis for the observed behavioral differences.
The research raises the possibility that continuous monitoring of human behavior could one day reveal early signals of aging trajectories. With wearable devices already tracking movement patterns and sleep quality, scientists believe similar principles might eventually be applied to human health.
Bedbrook said the team now plans to investigate whether aging trajectories can be altered through interventions such as dietary changes, genetic modifications or adjustments to sleep patterns, with the goal of intervening before significant decline begins.
-
23:40
-
23:20
-
23:00
-
22:40
-
22:20
-
22:00
-
21:40
-
21:20
-
21:00
-
20:40
-
20:20
-
20:00
-
19:40
-
19:20
-
19:00
-
16:20
-
16:15
-
16:00
-
15:58
-
15:50
-
15:45
-
15:30
-
15:20
-
15:15
-
15:00
-
14:53
-
14:50
-
14:45
-
14:30
-
14:22
-
14:20
-
14:15
-
14:07
-
14:00
-
13:50
-
13:45
-
13:43
-
13:30
-
13:20
-
13:15
-
13:00
-
12:50
-
12:45
-
12:30
-
12:20
-
12:00
-
11:50
-
11:20
-
10:50
-
10:20
-
09:50
-
09:20
-
08:50
-
08:20
-
07:50
-
07:20
-
07:00
