Every day on the dot of noon, Jane* would eat her 150-calorie lunch: nonfat yogurt and a handful of berries. To eat earlier, she felt, would be “gluttonous.” To eat later would disrupt the dinner ritual. Jane's eating initially became more restrictive in adolescence, when she worried about the changes her body was undergoing in the natural course of puberty. When she first settled on her lunchtime foods and routine—using a child-size spoon to “make the yogurt last” and sipping water between each bite—she felt accomplished. Jane enjoyed her friends' compliments about her “incredible willpower.” In behavioral science terms, her actions were goal-directed, motivated by achieving a particular outcome. In relatively short order, she got the result she really wanted: weight loss.

Years later Jane, now in her 30s and a newspaper reporter, continued to eat the same lunch in the same way. Huddled over her desk in the newsroom, she tried to avoid unwanted attention and feared anything that might interfere with the routine. She no longer felt proud of her behavior. Her friends stopped complimenting her “self-control” years ago, when her weight plummeted perilously low. So low that she has had to be hospitalized on more than one occasion.

The longed-for weight loss did not make her feel better about herself or her appearance. Jane's curly hair, once shiny and thick, dulled and thinned; her skin and eyes lost their brightness. There were other costs as well—to her relationships, to her career. Instead of dreaming about a great romance, Jane would dream of the cupcakes she could not let herself have at her niece's birthday party. Instead of thinking about the best lead for her next story, she obsessed over calories and exercise.

Jane's ritualized and restrictive approach to food, her obsession with calories and her painfully low body weight are common symptoms of anorexia nervosa, a dangerous eating disorder that affects roughly one in 200 American women. Anorexia has a high relapse rate and ranks among the deadliest of all psychiatric conditions. Individuals with the disorder, about 10 percent of whom are men, enter a state of starvation that can cause numerous medical complications, including heart ailments, anemia, bone loss, infertility, and more. A young woman with this illness faces six times the average risk of death for someone her age, according to a 2011 meta-analysis of 36 studies, and mortality rises by 5 percent for every decade of illness.

Anorexia nervosa is often misunderstood by a public that tends to glorify thinness and view rule-ridden eating as an act of enviable self-control. This is nothing new. In the Middle Ages, a handful of religious figures, including Saint Catherine of Siena, were admired for engaging in extreme self-starvation—a condition termed “holy anorexia.” Today we see self-starvation in the name of a culturally sanctioned pursuit of thinness. But there is nothing glorious about this disease, nor does it provide any actual measure of true control. Rigid, behavioral routines gradually close in on the afflicted individual until life becomes entirely about numbers on a food label, or a scale, or a clothing tag.


A new line of research suggests that the core of Jane's condition—her low weight—is not simply a matter of self-control. Rather her routines now occur almost automatically without regard for the outcome. Jane weighs herself each morning before she showers and again before she leaves for work. At each meal, she reads and rereads food labels for their nutritional breakdown. She cuts food into tiny pieces without thinking. In behavioral science terms, her mind has been overtaken by habit.

Habits can be incredibly useful. They allow the mind to multitask and in so doing enable efficiency. Behaviors get linked together into a routine, and once the chain of action is initiated, the rest follows with little mental effort. Yet sometimes habits take hold when they are not useful. We and others in the field are learning that this may occur with anorexia nervosa.

The habit-based model of the disorder offers a compelling explanation for why patients such as Jane struggle for years through chapters of outpatient and inpatient treatment without finding lasting recovery. If her illness is even partially explained by hijacked habit learning, it suggests that habit-busting techniques could be part of the solution. Habit-reversal therapy, for example, is well supported for conditions such as trichotillomania (hair-pulling disorder) and tic disorders. This type of treatment helps people become more aware of the cues that set their habits in motion and develop competing responses. For example, those with an urge to pull hair might be instructed to occupy their hands by imagining they are squeezing a lemon. We have adapted this approach for patients with anorexia in an intervention called REACH (regulating emotions and changing habits).

Jane worked with us in the REACH framework. The habit hypothesis made sense to her and helped her to feel better about why she had been stuck in routines that she knew were not healthy. We shared with her results from a brain-imaging study, published last year in Nature Neuroscience, that one of us (Steinglass) co-authored. It showed that when people with anorexia nervosa make decisions about what to eat, they use a different part of the brain—the dorsal striatum—than those without eating disorders. Studies in both animals and humans have shown that this deep-brain structure is involved with many aspects of behavior, one of which is habitual behavior.

In individual psychotherapy sessions, we helped Jane identify a number of habits that served the eating disorder better than they served her. At home and work she kept track of these routines and paid attention to their earliest cues. For example, Jane noticed that her mealtime rituals began with washing her hands. In therapy, she identified another action to try when faced with this cue. She began to bypass the sink, altering her route to the dining table. This small change made a difference in the subsequent chain of behaviors. Jane practiced moving her water glass out of arm's reach at the start of a meal; with improved awareness, it became easier to resist taking sips between each bite. Behaviorists refer to this as stimulus control: altering the environment to encourage an alternative behavior. In other instances, Jane developed competing responses—simple, motor-based counteractions—that made it harder to act out of habit. For example, she practiced picking up her utensils with her nondominant hand to help her take bigger and less “perfect” bites.

As new behaviors helped her break old habits, Jane tackled other routines of illness. For years she had kept a written record of what she ate at every meal. Jane decided to switch the location of her food journal, putting it out of her line of sight after meals. Instead of reaching for the journal, she turned to friends and family after eating—by phone or e-mail or in person if possible—which also provided an element of distraction. Nevertheless, this change provoked anxiety. To manage it, her therapist taught her a muscle-relaxation exercise—tensing one muscle at a time and then letting it go.

Most important, Jane learned that reversing or replacing old habits brought good outcomes. This was an essential element because behaviors that are associated with reinforcement grow stronger over time. During meals, Jane felt more present, and she found, to her pleasure, that she could participate more fully in conversation during and after eating. As she spent less time logging calories in her journal, she could focus instead on reading for work and leisure. Breaking these routines felt frightening at first, but loosening the grip of old preoccupations also brought an unanticipated element of relief. Jane's weight slowly improved, and although this change felt scary, she described feeling more motivated and able to maintain her new behaviors because they led to clearly positive rewards.

Encouraged by success with our initial patients, we have begun a small, randomized controlled trial to compare our habit-breaking approach with routine treatment for anorexia nervosa. By linking treatment directly with mechanisms of illness—in this case, the neural circuitry of habit—we hope to better understand this puzzling disorder, improve treatment and free more patients like Jane from the prison of habit.

*Jane is a pseudonym. Some details of her story have been altered to further protect her identity.