Researchers discovered how the brain decides what to remember

A song from childhood can hit you out of nowhere. A smell can carry you back decades. Then again, yesterday’s errands fade overnight. Scientists have long known that short-term memories become lasting ones only when the brain builds new proteins through gene activity. What had stayed unclear was how some memories endure for weeks, months or even a lifetime.

New research in mice now shows that memory lasts because of a carefully timed handoff across brain regions. The process moves from the hippocampus, to the thalamus, and then to the cortex. Each stop uses a different genetic program to extend how long a memory lives. The work was published in Nature and led by Priya Rajasethupathy at the Skoler Horbach Family Laboratory of Neural Dynamics and Cognition.

“This is a key revelation because it explains how we adjust the durability of memories,” Rajasethupathy said. “What we choose to remember is a continuously evolving process rather than a one-time flipping of a switch.”

To watch memory form and fade, the team built a virtual reality task for mice. The animals ran on a foam ball and moved through digital hallways filled with colors, sounds and smells. Some scenes ended with a sip of water. Another came with a puff of air to the nose.

Two scenes brought rewards. One appeared often. The other showed up less. A third predicted the air puff. Over seven days, the mice learned to lick in the reward zones and hold back for the air.

Right after training, both reward scenes were remembered. Three weeks later, only the one that appeared most often stayed in mind. Repetition had decided which experience would last.

This design gave the team a clear split. They could compare brains that kept a memory with those that lost it.

The scientists then tested which parts of the brain mattered most over time. They focused on the hippocampus, the thalamus and the anterior cingulate cortex.

Using light to briefly quiet regions during the task, they saw a pattern. Silencing the hippocampus blocked early recall. Quieting the cortex hurt long-term recall but spared short-term memory.

The surprise came from the thalamus. Turning off the path from the thalamus to the cortex during learning did not hurt early memory. Weeks later, the memory was gone. When the team boosted that same route, weak memories became strong.

The message was clear. The thalamus acts like a gatekeeper. It decides what reaches the cortex and earns a long life.

Next, the researchers inspected gene activity in single cells from the thalamus and cortex. They analyzed more than 320,000 cells across days and weeks.

They did not find one magic gene. They saw waves.

In the thalamus, genes tied to flexible wiring turned on in mice who remembered. These genes help neurons grow and strengthen links. The burst was short-lived.

In the cortex, the story changed. Genes that reshape how DNA is packed switched on. These changes can persist. They keep memory programs running long after an event ends.

So memory starts with changes in circuits and ends with chemical marks on DNA packaging.

To place cells along a timeline, the team used a tool called pseudotime analysis. It sorts cells by how similar their gene patterns are, which traces a path through time.

In the thalamus, memory-related cells split. Some drifted toward short-term states. Others moved into early and late phases linked to persistence. Only strong memories took this route.

In the cortex, cells marched into a stable state and stayed there. The path did not reset.

“These results suggest that long-term memory is not maintained by a single molecular on or off switch,” Rajasethupathy said. “It is driven by a cascade of gene-regulating programs that unfold over time.”

The team then searched for the genes that run the show.

In the thalamus, two transcription factors stood out. CAMTA1 supported memory at mid-range. TCF4 sustained it later.

In the cortex, one enzyme ruled long storage. ASH1L alters histones that control access to DNA.

Using CRISPR, the researchers removed each gene. Learning still worked. Remembering did not.

Taking out CAMTA1 caused forgetting within days. Removing TCF4 erased memory weeks later. Deleting ASH1L looked harmless at first but collapsed memory long after.

None of the three were needed to learn. All were needed to keep.

The team also tracked brain signals. Without CAMTA1 or TCF4, communication between thalamus and cortex weakened and grew erratic. Signals lost rhythm. Memory followed.

Further tests showed what each gene handles. CAMTA1 runs genes that reshape synapses. TCF4 steers cell structure and stickiness. ASH1L keeps helpful genes open for business.

The researchers charted a ladder. CREB1 in the hippocampus supports early memory. CAMTA1 and TCF4 in the thalamus extend it. ASH1L in the cortex preserves it.

For years, models focused on two regions. The hippocampus made short memories. The cortex stored long ones. The idea was tidy but thin.

Rajasethupathy said older views imagined “transistor-like memory molecules that act as on or off switches.” Reality appears richer.

In 2023, her group reported that the thalamus links short memory to long storage. This new work shows how. The thalamus does more than relay. It judges.

“Unless you promote memories onto these timers,” she said, “you are primed to forget.”

She added that ASH1L belongs to a family used across biology. Immune cells use it to remember infections. Developing cells use it to keep their identity. The brain may be borrowing ancient tools to remember life.

The lab now wants to learn how the timers start and how long they run. What marks an event as important? How does the thalamus weigh value?

“We’re interested in understanding the life of a memory beyond its initial formation in the hippocampus,” Rajasethupathy said. “We think the thalamus and its parallel streams of communication with cortex are central.”

This work opens paths for treating memory loss. If doctors can target the gene programs that preserve memory, they may one day strengthen weak memories or slow fading ones.

The findings also hint at new ways to route information around damaged circuits after injury or in diseases like Alzheimer’s.

Understanding the thalamus as a decision hub could guide therapies that help the brain choose better what to keep. Over time, this knowledge may improve learning strategies, rehabilitation methods and care for aging minds.

Research findings are available online in the journal Nature.

Like these kind of feel good stories? Get The Brighter Side of News' newsletter.