The summer day on Okinawa Island of Japan was so warm that those us who were there to teach had given up on going to the beach in the afternoon and instead had decided to try some early morning tennis. On this particular early morning, a distinguished scientist and friend had joined our group, and there he was holding the ball and warming up for his serve. As he tossed the ball up, he bent his body sideways and then back and twisted upwards and finally had his racquet make contact with the ball, delivering a pretty good serve.
I stood there marveling at how he had learned to serve this way. He said: “Well, I learned on my own, and despite lots of coaches who have tried to break it down and rebuild it, I haven’t been able to change it much.”
Memories, like that of how to hit a tennis serve, can become so persistent that the brain seems unable to change them. This, at the surface, may not appear that important, as the cost is looking a little silly and not being able to do something as efficiently as possible. But what if you are traveling with family on a peaceful day and stop at a gas station, and suddenly the smell of petroleum brings back memories of combat, paralyzing you with fear? What if you are watching a movie where the hero is climbing the face of a rock and when she reaches the peak, she stands and looks down, and you find your knees shaking? Does the brain have a mechanism in place to rebuild or even erase unwanted habits and fear-inducing memories?
Until about 15 years ago, it was generally assumed that when the brain learns something new, the newly acquired memory is initially in a labile state and can be readily changed, but after a short period of time (hours), it becomes ‘consolidated’, meaning that it becomes resistant to change. For example, when rats were given a single pairing of a tone with a food-shock, this made it so that the next time they heard that tone, they got scared and stopped moving. If a drug was given to them that disrupted the molecular pathways that are involved in consolidation (protein synthesis inhibitors), the next day when they heard the tone they were not scared of it. However, the drug had to be given soon after the animal’s first experience of the tone-shock pairing. If it was given even a few hours after the first experience, it did not have much of an effect; the animal still feared the tone. And so it seemed that once an emotional or fear-inducing memory was acquired, there was little that could be done to change it.
The basis for this idea was a century of work that had described how memories form. Neurons communicate with each other via their synapses, tiny junctions where one neuron sends and receives messages from another neuron. Eric Kandel, a Columbia University neuroscientist, had shown that short-term memories, things that last for a few minutes, are due to transient changes to the synapse to make it more efficient, but these changes were sustained only for a short period of time. To make memories last, the changes at the synapse had be sustained indefinitely, and this required manufacture of new proteins. If the initial experience was strong enough, with passage of time these new proteins were made by the neuron and the memory was maintained, apparently becoming permanent.
But in the year 2000, Karim Nader, an Egyptian born neuroscientist who was raised in Canada, made a discovery that completely overturned this idea. He was working in Joseph LeDoux’s laboratory in New York University where he took rats and gave them a single pairing of a tone with a foot-shock, and indeed, the next day he found that when they heard the tone, the rats froze in their tracks (rats express fear by ‘freezing’). However, right after they heard this tone, he injected into their amygdala (a region of the brain involved in storing fearful memories) a drug that inhibits protein synthesis. Amazingly, he found that a day later, when they heard the tone their fear was reduced by half (measured by the time spent ‘freezing’). Interestingly, if the drug was given without the reactivation of the memory (that is, on day 2 don’t play the tone), it had no effect. And if the animal heard the tone but 6 hours later was given the drug, it still feared the tone the next day. So the key idea was that the fear-inducing memory could be weakened if the drug was given right after the memory was reactivated, but it could not be weakened if the drug was given alone, or if the memory was reactivated but without the drug.
Unfortunately, protein synthesis inhibitors cannot safely be given to humans, and so until recently, it was unclear whether this new understanding could be applied to fear-inducing memories in people. In 2009, Merel Kindt and colleagues in Amsterdam asked a few undergraduate students to look at a picture of a spider and then a few seconds later played a loud sound, followed by a mild shock to the hand. When the students heard the loud sound, they had a startle reflex, producing an eye blink. They also showed them a picture of another spider, followed by another loud sound, but no shock to the hand. So the students learned to fear the picture of the 1st spider, but not the second. The amount of fear was measured by how they reacted to the loud sound. Indeed, the students feared the 1st spider more than the second. The students returned on Day 2, and Kindt showed them the picture of the 1st spider, but did not shock them. Right after this, they gave them a drug called propranolol, which is often used to prevent stage fright, and works to inhibit actions of norepinephrine. When the students returned on the next day, they did not show fear of the spider. Importantly, if they gave the drug but did not show them the picture of the spider, the fear-inducing memory remained.
So it seems possible that in humans, certain fear-inducing memories can be weakened by a combination of reactivation of that memory and consumption of certain drugs like propranolol. Later work from the Kindt group showed that the key step is that during recall of the memory, there must be a prediction error. That is, during recall, the brain appears to predict that a bad thing is going to happen (a shock), and if it does not happen, and the drug is present, then the memory is weakened. Both the prediction error and the presence of the drug seem to be required, as one without the other is much less effective.
These approaches are now being studied for treatment of PTSD. In a recent study, propranolol was given to people who were involved in a serious car accident. Those people were less likely to develop PTSD symptoms in the following 3 months compared to people who were given placebo.
Notice, however, that all the successes have been on weakening newly formed memories. What about the old fear-inducing memories? The news there is less clear. Older memories may be less likely to be affected when they are reactivated. Which brings me to one of my favorite quotes from Margaret Thatcher, who was quoting her father when she said:
thoughts for they become words.
Watch your words for they become actions.
Watch your actions for they become habits.
Watch your habits for they become your character.
And watch your character for it becomes your destiny.
Kindt M, Soeter M, Vervliet B (2009) Beyond extinction: erasing human fear responses and preventing the return of fear. Nature Neurosci 12:256-258.
Nader K, Schafe GE, Le Doux JE (2000) Fear memories require proten synthesis in the amygdala for reconsolidation after retrieval. Nature 406:722-726.
Sevenster D, Beckers T, Kindt M (2013) Prediction error governs pharmacologically induced amnesia for learned fear. Science 339:830-833.