Michael Kahana, a psychologist at the University of Pennsylvania, has studied memory for more than 30 years: how it works and what happens when it doesn't.
He's not just interested in memory loss from traumatic brain injury, which affects more than 5 million people in the U.S., or the roughly 7 million Americans with Alzheimer's disease, but his research also focuses on memory disorders that affect everyone, regardless of cognitive ability.
“We all have bad memories from time to time,” Kahana told The Washington Post. “They fluctuate throughout the day and even from moment to moment. That's just the way our brains are wired. Once I realized that, it got me asking, how can I keep my brain in good working order all the time?”
Kahana's work on memory culminated in a groundbreaking study published last January, when he and his team administered a computerized intervention to 47 people with epilepsy, delivering electrical pulses directly to their brains at the moment memory loss was likely to occur, via electrodes implanted directly into the patients' brains as part of their epilepsy treatment.
These electrodes (there are 100 to 200 per person) can recognize brain signals when a patient is trying to remember something and send precisely timed electrical messages to the lateral temporal cortex, the part of the brain used to store and process memories.
The results exceeded Kahana's expectations — brain stimulation improved memory by 28% — and while he's cautiously optimistic, he can't help but be enthusiastic.
“I think we're on the cusp of a new era in human neuroscience and neurotherapeutics,” he said.
Kahana isn't the only one exploring the potential of brain-computer interfaces: Scientists across the country are developing brain-computer interfaces (BCIs) that could be used to treat everything from memory loss to speech disorders to paralysis.
Last year, a Stanford University School of Medicine study found that patients with brain implants were so amazed at how much their memory improved after 90 days of treatment that some refused to turn the devices off.
And in August, Neuralink, a neurotechnology startup owned by Elon Musk, announced plans to implant a second subject with a BCI designed to give paralyzed patients the ability to use digital devices using only their thoughts.
Noland Arbaugh, a 30-year-old Arizona man who was paralyzed from the neck down in a diving accident eight years ago, received the first Neuralink implant in January. During a livestream on X in March, Arbaugh demonstrated how he could control a computer cursor with his thoughts, play games and text. In May, the device suddenly began to slip out of Arbaugh's skull, but the company announced the problem had been resolved.
Musk said: Predicted Within a few years, hundreds of people will own Neuralinks, and within a decade millions will own them.
In August, researchers at the Swiss Federal Institute of Technology in Lausanne unveiled a brain that can convert thoughts into text with 91% accuracy, and it's even smaller than Neuralink's chip.
Progress is coming in leaps and bounds, and the FDA will be hosting a workshop on clinical outcomes evaluation of BCIs later this month.
“If the preliminary results are replicated, meaningful assistive technology for people with severe illnesses and disabilities may be years, not decades, away,” said Anna Wechsler, a professor at the Perelman School of Medicine who studies the ethical, legal and social issues surrounding emerging technologies.
When we think of computers helping people with ALS (formerly known as Lou Gehrig's disease) speak, the first image that comes to mind is of the famous theoretical physicist Stephen Hawking, who spoke to a computer powered by an Intel microprocessor, but whose voice sounded metallic, like a robot in a science fiction movie.
But for 45-year-old Casey Harrell, who lost the ability to speak due to ALS, a brain-computer interface called BrainGate2 has given him back his voice — his natural voice.
This allows Harrell to communicate with his 5-year-old daughter.
“She was barely able to communicate with me for about two years…” Harrell He told Scientific American In my August 2024 post, I said, “I can help her mother raise her. I can build a deeper relationship with her and tell her what I think. I can just tell her how much I love her.”
David Brandman, a neurosurgeon at the University of California, Davis, who helped develop the brain chip, said the BCI interprets brain signals and replicates them in the voice assistant software.
“The system is about 97 percent accurate and can pronounce words from a dictionary of 125,000 words,” Brandman told The Washington Post. “Using artificial intelligence, we've also recreated the sound of his voice so that the computer can read text aloud in the way that he would have sounded before he was diagnosed with ALS.”
When it comes to memory, the challenge is a bit murkier: our memory waxes and wanes, so the problem isn't always consistent. According to Dr. Brent Roeder, the goal isn't necessarily to improve memory in general, but “to improve your memory for specific important or urgent information, like 'Did I take my medicine this morning?'”
Roeder, a research associate in the Translational Neuroscience department at Wake Forest University School of Medicine, is studying ways to replicate the individual codes within hippocampal activity for specific memory information.
He and his research colleagues accomplished this by using a “memory prosthesis” in which electrodes inserted into the brain interact with the hippocampus and create neural recordings as patients perform specific memory tasks. “Once these unique memory codes were created, we determined whether we could stimulate them during memory tasks to improve patients' memory,” says Roeder.
In other words, they encoded memories for future reference, creating a Post-it to remind your brain what you've forgotten.
The treatment, researchers found, helped patients remember very specific information — not just overall memory improvements (though there were indeed 11-54 percent improvements), but specifically improvements in memory problems that interfere with everyday life, like where they put their car keys or whether they turned off the stove.
One advantage of this kind of approach is that it's not limited to any particular condition, says Roeder: “Once it's ready for clinical use, we hope it could be used to treat any condition that impairs memory function, from traumatic brain injury to dementia and Alzheimer's disease.”
While the research is intriguing, questions remain about how the technology will be used, or, in Wechsler's words, “blurring the line between therapeutic and augmentative BCIs.”
“Most people probably wouldn't be interested in an implantable BCI that allowed them to type as fast as they could type with their fingers or type with their voice,” Wexler said, “but if it really did provide a significant, measurable improvement, which is something that remains to be demonstrated, it would be interesting.”
That seems to be exactly what Musk is hoping for: In a video posted to X on July 10, he claimed Neuralink's long-term goal is to “give people superpowers” and provide capabilities “well beyond those of a normal human.”
But scientists like Lauder don't share those ambitions. “The focus of our research has always been to restore function that's been lost through disease or injury,” he told the Post. “We see it as a superpower to give someone back something they've lost.”
Just getting the technology to be widely deployed will be no easy feat. It involves brain surgery, after all. Tom Oxley, CEO of brain-interface startup Syncron, said: He spoke at a TED Talk in 2022.“The brain doesn't really like being stuck with needles.”
Kahana agrees that this is a hurdle. “You can't modulate the brain from a distance with a light gun,” he says. “So for this to work, you have to get inside the brain.” But he adds that it's becoming increasingly safe to do so. “A lot has changed in the last few years. The imaging has improved, the electrodes have gotten smaller. When the time comes, I wouldn't hesitate to undergo this procedure myself.”
He co-founded Nia Therapeutics, a company funded by the Defense Advanced Research Projects Agency to help commercialize brain implants, as part of an effort to help veterans with brain injuries. But it's also personal for him.
“My son can't speak, he can't articulate words. He uses a device to communicate, but as you can imagine, it's very inconvenient. He searches through menus to find the right words. He knows what he wants to say, but how do you translate those brain patterns into spoken language?,” Kahana explained. “You and I do it so easily, we take it for granted. But if someone could develop the technology to decode those brain signals, well… that would be really amazing.”
