Project Summary Anorexia nervosa (AN) is a severe psychiatric disorder. However, we lack neurobiological models and interventions to explain and treat the core characteristics of food restriction, feeling fat, and body size overestimation. While research has made progress in understanding brain function involved in AN pathophysiology, translating those results into biological therapies has been challenging. Studies have suggested that metabolic factors contribute to developing and maintaining AN pathophysiology. Specifically, brain glucose utilization and metabolism may be altered in AN, interfere with brain energy homeostasis, and contribute to illness development and maintenance. A small study indicated that ketosis might be therapeutic for AN core behaviors such as eating and shape concerns. In this application, we will study individuals weight recovered from AN, establish biological targets as diet-related metabolic markers for AN (R61 phase), and replicate and link those targets to AN-specific behaviors (R33 phase). The weight- recovered AN group will also be compared with a healthy control sample. The R61 Phase Specific Aim of the project is to establish target engagement for a therapeutic ketogenic diet (TKD) in AN after weight recovery and establish safety and tolerability. We hypothesize that TKD will be associated with reduced brain glucose metabolism using [18F]fluorodeoxyglucose ([18F]FDG) and positron emission tomography (PET) (target engagement). Higher blood ketosis levels will be associated with a more significant reduction of the brain [18F]FDG glucose metabolism rate (dose dependency). We hypothesize that TKD will be well tolerated, that participants will remain within the normal weight range (tolerability), and that study participants will be able to adhere to TKD as indicated by regular blood ketosis measurements (treatment fidelity). There will be an initial indication that TKD and associated biological measures correlate with behavioral measures derived from eating disorder-specific assessments. The go/no-go criterion for the transition from the R61 to the R33 phase is determined by a significant change in [18F]FDG metabolism rate between before and after TKD in the frontal cortex (target engagement). The R33 Phase Specific Aim is to replicate target engagement in a larger weight-recovered AN cohort and associate brain response with AN- specific behaviors (functional outcome). Similarly to the R61 Specific Aims, we hypothesize that TKD will decrease brain [18F]FDG metabolism in the larger cohort (target engagement replication). The level of ketosis and magnitude of [18F]FDG uptake decrease will be associated with decreased eating restraint, eating and weight concerns, and clinical impairment based on clinical assessments (functional outcome).