A Mistake in a Mouse Experiment Helped Reveal How Memories Might Be Manipulated

A surgical error in a mouse helped redirect neuroscientist Steve Ramirez’s career toward one of memory research’s most ambitious questions: whether memories can be physically identified and manipulated in the brain. In a 2013 study with MIT neuroscientist Xu Liu, Ramirez’s team used optogenetics, a technique that controls cell activity with light. Researchers tagged hippocampal cells active during the formation of a fearful memory, then reactivated those cells while the mice were in a safe setting. The first four mice showed no fear response. The fifth froze. When the researchers examined that animal’s brain, they found that the protein used to tag cells had been placed in a slightly different part of the hippocampus. The unexpected placement appeared to identify a brain area involved in the stored fear memory and demonstrated that the memory could be artificially reactivated. Ramirez, now an associate professor at Boston University’s Center for Memory and Brain, describes the work in his 2025 book, “How to Change a Memory: One Neuroscientist’s Quest to Alter the Past.” The book also draws on his personal experiences with loss, PTSD and addiction while examining the possible medical uses and ethical risks of memory research. A central concept in this field is the “engram.” Ramirez describes an engram as a theoretical term for the physical basis of a memory—the cellular building blocks that allow a memory to exist in the brain. Researchers do not yet have a precise map showing where an engram begins and ends or how its emotional, sensory and other components are arranged. Understanding that physical structure, Ramirez told Live Science, could eventually improve scientists’ ability to predict or prevent memory loss and other cognitive problems. He has also discussed the possibility of using such knowledge to develop treatments for disorders including PTSD or to help protect against dementia. For now, optogenetics is not widely used in human medicine, aside from potential applications involving the retina. Ramirez said future researchers are exploring less invasive ways to deliver optogenetic tools, but he also suggested that human therapies may not need to directly replicate experiments in rodents. Cognitive behavioral therapy and other noninvasive approaches could be considered before techniques such as deep brain stimulation or transcranial magnetic stimulation. Ramirez argues that any future use of memory manipulation should remain tied to restoring health and well-being, with safeguards designed to protect personal agency and prevent misuse. The technology remains a research tool in this context, not an established method for editing human memories.