Philip Seymour Hoffman, the great actor, was found dead a few days ago with a needle in his arm. Last year, Cory Montieth, the actor, died in similar circumstances. Why did they die? It was hardly the first time they’d taken heroin.
Starting in the 1970s, Shepard Siegel, a psychology professor at McMaster University, did a series of rat experiments that showed that drug tolerance and craving involved a large amount of Pavlovian conditioning. Repeated exposure (e.g., injection) of Drug X in Situation Y (e.g., your bedroom at 11 p.m.) will cause learning of an association between X and Y. This association has two effects. First, when exposed to Y, you will crave X. Second, when you take Drug X in Situation Y, the effect of the drug is diminished. You become “tolerant” to it.
You have probably experienced both effects — they occur with caffeine, chocolate and alcohol, for example — but are unaware of Siegel’s explanation: Situation-drug associations. After you learn the association, Siegel said, the situation generates a response in your body that opposes (reduces) the drug effect. If the drug makes you less sensitive to pain, the conditioned response is more pain sensitivity. If the drug makes you awake, the conditioned response makes you sleepy. It’s easy to realize that drug craving involves associative learning: You notice that you crave coffee or whatever in familiar situations but not unfamiliar ones. You crave Drug X at college but not at home. It’s much harder to grasp that the tolerance involves associative learning.
Drug users, such as Hoffman and Montieth, as they become “tolerant”, take larger and larger doses, not realizing that their “tolerance” depends on a learned association. The situation becomes more and more dangerous because if you take the drug without eliciting the conditioned compensatory response, you may die. Without the conditioned response, a drug amount you survived yesterday may kill you today.
Siegel explored the possibility that this actually happened — that drug addicts died of “drug overdose” because they took the drug under unusual circumstances that didn’t evoke the compensatory response. He found plenty of support for this idea. One sort of support was interviews with “overdose” survivors. They often described unfamiliar circumstances at the time of the injection — e.g., there’s usually music, but this time there wasn’t. Another sort of support was a rat experiment with heroin. After developing tolerance to heroin, rats injected with a really large dose in a strange situation were much more likely to die than rats injected with the same dose in the same situation where they became tolerant.
I am pretty sure Hoffman and Montieth didn’t know about this research. There is a connection with the Shangri-La Diet. In both cases, Pavlovian learning has a big effect on something we care a lot about (life/death in the drug case, body weight in the diet case) and this connection is highly non-obvious.