Researchers at USC are developing a groundbreaking approach that combines artificial intelligence with brain-machine interfaces to read, understand, and potentially restore human memory. Associate Professor Dong Song has published a framework in Advanced Science that could help people with Alzheimer's disease, traumatic brain injury, and post-traumatic stress disorder by using AI to decode the neural patterns that create memories, then using electrical stimulation to reinforce or modify those patterns. What Is Episodic Memory and Why Does It Matter? Episodic memory is your brain's autobiographical record, the mental movie reel of where you were, what happened, and when. It's what allows you to remember waking up yesterday, the smell of coffee, a conversation with a friend, or where you parked your car. Scientists have known for decades that a brain structure called the hippocampus is critical to forming these memories, but the exact mechanisms have remained a mystery. The challenge is that lived experience throws an enormous number of variables at the brain simultaneously: sights, sounds, smells, locations, people, and time, all intertwined. Traditional lab experiments sidestep this complexity by stripping memory down to something manageable, like showing a person a word list and asking what they recall minutes later. But that approach doesn't reflect how memory actually works in real life. "You are largely what you remember. Episodic memory is really essential for your identity. What you did yesterday, when you met someone last time, all the important things to maintain a normal life," said Dong Song. Dong Song, Associate Professor, Department of Neurological Surgery at the Keck School of Medicine of USC and Alfred E. Mann Department of Biomedical Engineering at the Viterbi School of Engineering How Can AI Help Scientists Understand Memory Formation? Song's solution is to use AI as a translator between two worlds: lived experience and the brain signals that encode it. Imagine someone going about a typical day wearing a brain sensor while a camera and microphone record what they see and hear. Those recordings are fed into an AI system along with brain data. The system searches for patterns, moments when bursts of neural activity reliably coincide with something in the outside world, such as a familiar face, a place, or a sound. Over time, the AI builds a dictionary linking brain signals to real-world events. The ultimate goal is to move beyond correlation and test causation. Song proposes using a brain-machine interface (BMI), a device that can both read brain activity and send small electrical pulses back into the brain. First, AI would identify the specific pattern of brain activity linked to a particular memory. Then the BMI would recreate that pattern in the brain using brief electrical pulses, even though the person wouldn't actually experience a thing. In effect, the device would plant an artificial memory. "You manipulate that code through brief stimulation, you write in that code, and then you ask this person, 'Did you remember that?' The person says yes. Then this is conclusive. This is what I mean by causation, not just pure correlation," explained Song. Dong Song, Associate Professor, Department of Neurological Surgery at the Keck School of Medicine of USC and Alfred E. Mann Department of Biomedical Engineering at the Viterbi School of Engineering Which Conditions Could Benefit From Memory Engineering? The most immediate candidates for this technology are people living with specific neurological and psychiatric conditions. Song's team has already demonstrated pieces of this approach in rodents, nonhuman primates, and human epilepsy patients, where AI-generated stimulation patterns enhanced recall. - Alzheimer's Disease: Gradually destroys the brain's ability to convert short-term experiences into lasting memories. A patient might recall a conversation from 40 years ago but not one from this morning. If scientists can pinpoint the neural patterns responsible for that conversion, a BMI might step in to reinforce them, serving as a prosthetic for a failing memory system. - Traumatic Brain Injury: Can impair the formation of new memories. The technology could help restore the neural patterns necessary for memory encoding and consolidation. - Post-Traumatic Stress Disorder: The technology might work in reverse, selectively weakening traumatic memories. Song called that possibility theoretical but plausible. Song estimated that meaningful improvements for some patients could arrive within five to 10 years, though the timeline depends on advances in recording technology, more refined AI models, and experiments in real-world settings. Steps to Advance Memory Engineering Research Song's team is following a structured approach to bring this technology from concept to clinical reality: - Improve Recording Technology: Develop more precise sensors that can capture brain activity with greater spatial and temporal resolution, allowing AI systems to identify neural patterns with higher accuracy. - Refine AI Models: Create more sophisticated artificial intelligence systems that can detect subtle correlations between brain signals and real-world events, moving beyond simple pattern matching to causal relationships. - Conduct Real-World Experiments: Move beyond controlled lab settings to study how memory forms and can be modified during everyday life, capturing the full complexity of human experience. What Ethical Concerns Does Memory Engineering Raise? The power to read, write, and alter human memory raises profound questions about personal identity, the authenticity of one's recollections, and the potential for misuse. Song's paper argues that advances in memory engineering must be matched by equally rigorous ethical and regulatory guardrails. "This is a really sensitive topic. This potential needs to be realized very carefully, in a very rigorous ethical framework," noted Song. Dong Song, Associate Professor, Department of Neurological Surgery at the Keck School of Medicine of USC and Alfred E. Mann Department of Biomedical Engineering at the Viterbi School of Engineering The research is supported by DARPA's Restoring Active Memory program, the NIH/NIDA BRAIN Initiative's Theories, Models and Methods program, and DARPA's Investigating how Neurological Systems Process Information in REality program. As this technology advances, the scientific community will need to establish clear ethical guidelines to ensure it's used responsibly and for therapeutic benefit rather than exploitation or coercion.
