Some hopeful news for those who can’t remember new people’s names: stimulating the brain with a magnetic pulse can bring back forgotten short-term memories, as long as we know that we’ll need that information later.

Scientists used to think that we had to keep consciously thinking about something, like a new name, to remember it short-term. From the point of view of a brain scan, this means that all the neurons, or brain cells, involved in remembering that name would fire continuously and light up on the scan. But in a study published today in the journal Science, researchers led by University of Notre Dame psychologist Nathan Rose discovered that people remember things short-term even when the neurons aren’t firing. This suggests that short-term memory works in several different ways. And crucially, transcranial magnetic stimulation (TMS), a non-invasive way of stimulating the brain, can even bring that memory back.

People have been wondering for a decade short-term memory might work when the neurons aren’t firing, says Edward Ester, a cognitive scientist at the University of California-San Diego who was not involved in the study. Until now, though, nobody could show this using direct experiments in humans. The research involving TMS is exciting, too, he adds, and could one day help people with long-term memory problems, too.

First, Rose’s team told participants to remember (so they could later recognize) a certain image, like a face. Researchers analyzed participants’ brain activity while they tried to do so. This way, they could see the exact pattern of neurons firing when the participant was thinking about that face, versus thinking about other images. Next, they told participants to try to remember and recognize a different image. This distracted them, and those “face” neurons stopped firing. This is supposed to mean that they forgot the image entirely. But when Rose applied TMS to the brain area where the “face” neurons had been firing before, those very same neurons began firing again in the same pattern, as if they remembered the image after all.

There were changes in behavior, too. In another experiment, the researchers showed the participants a lot of different images and told them they would be tested on whether they recognized two different ones, like a face and a word. Researchers then applied TMS to an area of the brain used to remember an image that participants weren’t going to be tested on. If the TMS really brought the memory back, it would be harder for the participant to figure out that this was the wrong image. There would be more false positives when they had to choose whether a new image matched the one they were supposed to remember. That’s exactly what happened.

The most interesting part is that the researchers used whether two times, and the false positives only happened after the first time. “When the TMS was applied after the second when the item was no longer relevant, we did not see that same increase in false alarms,” says Rose. That’s because after the first test, you know that you might still need to remember all this other information for the second test. But after the second time, you know you won’t be tested anymore and there’s no real point in continuing to remember.

This sort of neuroimaging is hard to do, so the study is limited by a small sample size of around six to 20 participants, according to Ester, the UCSD cognitive scientist. But the combination of imaging techniques they used was impressive, he says, with the findings potentially having a lot of applications. Now that we know it’s possible to reactivate memories, we may be able to try continuously reactivating a memory to improve performance, or manipulate traumatic memories to help people with phobias or PTSD. “That’s obviously quite a few steps away, but this is nice work that does point to a whole bunch of these readily available techniques that could be used,” he says.