The location of memories in the brain may not be fixed, but may physically change over time. A new neuroscience study conducted at Northwestern University has revealed that memory is not merely a mental process but also a structure that physically moves and changes.

Memories Physically Move in the Brain

Experiments conducted on mice indicate that neural patterns formed over time, particularly in the hippocampus, the region responsible for spatial memory, do not remain fixed.

The research team created a controlled experimental field by placing mice on treadmills in a virtual environment. Environmental factors were kept constant throughout the experiment; a maze shaped by the mice’s movements was displayed on the screen, a specific odor was used, and white noise was used in the background. Under these conditions, neural activity in the hippocampus was monitored in real time using high-resolution imaging techniques.

The observations revealed the existence of a phenomenon called representational shift. This phenomenon indicates that even when the same experience is repeated, the brain does not produce the same neural pattern; instead, it reorganizes in a slightly different way.

Such shifts were previously thought to be due to environmental changes. However, the study found that these changes persisted even when environmental conditions were constant. The findings suggest that this movement represents an internal restructuring process in the brain.

Daniel Dombeck and his team believe this movement holds an important clue about how the brain stores and organizes information. It’s been reported that familiar but distinct experiences—for example, visiting a favorite restaurant one has visited repeatedly—are recategorized each time. This suggests that memories can physically shift location in the brain over time.

This discovery could have groundbreaking implications, particularly in understanding how memory disorders arise in humans and redefining treatment approaches. Proving that memory is not fixed but dynamic could pave the way for more realistic modeling of biological memory systems in the field of artificial intelligence.