His boss, the famed neuroscientist Joseph LeDoux, couldn’t have been more discouraging. “I told Karim he was wasting his time,” LeDoux says. “I didn’t think the experiment would work.” To LeDoux, the reason was obvious: Even if Nader blocked protein synthesis during recall, the original circuitry would still be intact, so the memory should be too. If Nader could induce amnesia, it would be temporary. Once the block was removed, the recall would return as strong as ever. And so LeDoux and Nader made a bet: If Nader failed to permanently erase a set of fear memories in four lab animals, he had to buy LeDoux a bottle of tequila. If it worked, drinks were on LeDoux. “I honestly assumed I’d be spending a bunch of money on alcohol,” Nader says. “Everyone else knew a lot more about the neuroscience of memory. And they all told me it would never work.”

He taught several dozen rats to associate a loud noise with a mild but painful electric shock. It terrified them—whenever the sound played, the rats froze in fear, anticipating the shock. After reinforcing this memory for several weeks, Nader hit the rats with the noise once again, but this time he then injected their brains with a chemical that inhibited protein synthesis. Then he played the sound again. “I couldn’t believe what happened,” Nader says. “The fear memory was gone. The rats had forgotten everything.” The absence of fear persisted even after the injection wore off.

The secret was the timing: If new proteins couldn’t be created during the act of remembering, then the original memory ceased to exist. The erasure was also exceedingly specific. The rats could still learn new associations, and they remained scared of other sounds associated with a shock but that hadn’t been played during the protein block. They forgot only what they’d been forced to remember while under the influence of the protein inhibitor.