Breaking habits is hard. We all know this, whether we’ve failed our latest diet (again), or felt the pull to refresh our Instagram feed instead of making progress on a work project that is past due. This is largely because we are constantly barraged by stimuli engineered to make us crave and consume, stimuli that hijack the reward-based learning system in our brains designed initially for survival.

Put simply, reward-based learning involves a trigger (for example, the feeling of hunger), followed by a behavior (eating food), and a reward (feeling sated). We want to do more of the things that feel good and less of the things that feel bad — or stressful. These three components (trigger, behavior, and reward) show up every time we smoke a cigarette or eat a cupcake. This is especially true at work. Each time we try to soothe ourselves from a taxing assignment we reinforce the reward, to the point where unhealthy distractions can become habits.

So why can’t we just control ourselves and decide to replace bad habits with good ones? The doctrine of self-control has been promulgated for decades, despite the fact that researchers at Yale and elsewhere have shown that the brain networks associated with self-control (e.g. the prefrontal cortex) are the first to go “offline” when faced with triggers such as stress. Still, in medical school, I was taught to pass self-control rhetoric on to my patients. “Need to lose weight? Quit eating junk food. Trying to quit smoking? Stop cold turkey or use a nicotine replacement.”

When I started actually practicing medicine, however, I quickly learned that it doesn’t work this way in real life.

Self-control theories have missed something critical: reward-based learning is based on rewards, not behaviors. How rewarding a behavior is drives how likely we are to repeat that behavior in the future, and this is why self-control as an approach to breaking habits often fails.