Clues emerged in a feedback experiment among AN patients and controls.
It is a challenging question: Why do many patients with anorexia nervosa (AN) continue to restrict their calories after their weight is restored and many of their psychiatric symptoms have improved? One possible reason, according to Karin Foerde, PhD and Joanna E. Steinglass, MD of Columbia University, is that AN patients may experience reduced learning from feedback or reinforcement. In their recent study, the same pattern was not seen among healthy controls (Int J Eat Disord. 2017; 50:415).
The researchers wanted to explore why reward processing among AN patients differs from that of normal control patients, and why learning from feedback is not effective. To do so, Drs. Foerde and Steinglass studied patients between 16 and 45 years of age who had DSM-5 diagnoses of AN-restrictive or binge-purge subtype AN (36 patients; 35 females and 1 male). All the subjects were inpatients at the New York State Psychiatric Institute, and were compared with 26 healthy controls (24 females, 2 males). The two groups did not differ in age, educational level, gender, ethnicity, or general cognitive functioning.
After 2 sessions of an association task, individuals with AN showed poorer learning from feedback in comparison with the healthy controls; this pattern was reported both before and after weight was restored. The reduction in feedback learning was associated with eating disorder psychopathology and suggested by the association with some eating disorder symptom severity scales and illness duration—but not with body mass index, or BMI. However, AN patients could generalize from what they had learned, just as the healthy controls did. Their psychological symptoms did get better with weight restoration treatment, but feedback learning did not. Both groups were well matched on global cognition, including IQ, working memory, and attention.
Although the authors’ study involved a behavioral task and no direct studies of the brain, their results were consistent with those of studies that have evaluated various forms of striatal-based learning among AN patients. (J Cognitive Neuroscience (2003;Myers et al. 2013). This suggested the possibility of frontal striatal abnormalities in AN patients. In a small study conducted in 2003, Lawrence et al. used a learning task sensitive to striatal function and a memory test sensitive to medial temporal lobe damage in a small group of AN patients tested while underweight. Foerde and Steinglass found this same pattern in their study after weight was regained.
Maladaptative behaviors may arise during development
If feedback learning is compromised in AN, how could this be related to the maladaptive behavior in AN? One answer, according to the authors, might be maladaptive behaviors established during a developmental “window” of learning in adolescence. Maladaptive eating behaviors and accelerated learning are most common during adolescence. Or, AN may affect neurocognitive processes so that developmental learning worsens the longer a person is ill, making it ever harder to counteract maladaptive behavior.
Decreased learning from feedback may also be important to the patient’s response to treatment. According to the authors, problems with feedback learning seen at baseline, without regard to how and when the behavior is acquired, and that are associated with failure to respond to treatment, could be a marker for a more chronic course of illness.