Section A is a systematic review of emotional processing in people with Anorexia Nervosa (AN). By drawing together all self-report research, it outlines the specific emotional processing difficulties experienced by this population. It concludes that theoretical assertions that poor emotional processing maintains AN are tentatively empirically supported, but data is limited and largely cross-sectional. Section B Background. Effective treatments for AN in adults remain elusive. Recent research suggests poor socio-emotional and neurocognitive functioning may act as underpinning maintaining mechanisms for AN. These difficulties are therefore indicated as treatment targets; however, their roles as maintaining factors and the benefits of including socio-emotional and neurocognitive difficulties as foci for therapy lack empirical evidence. Methods. A randomised control trial design was employed to compare the Maudsley AN Treatment for Adults (MANTRA) based on a novel maintenance model of AN seeking to target key socio-emotional factors (maladaptive emotion beliefs and over-control of emotion, alongside poor social cognition) and neurocognitive factors (poor set-shifting and extreme detail-focus), against a control treatment mimicking treatment as usual (specialist supportive clinical management; SSCM). It was hypothesised that: (1) baseline socio-emotional and neurocognitive difficulties would predict outcomes, highlighting their role as maintenance factors; and (2) MANTRA would significantly improve these factors over SSCM leading to better clinical outcomes, therefore confirming them ‘causal’ maintenance factors. Results. Emotional avoidance (behavioural/cognitive avoidance, low acceptance) and some other elements of control (submissive behaviour) predicted clinical outcomes. Social cognitive (emotion recognition, emotional Theory of Mind) and neurocognitive functioning (set-shifting, detail-focus) had limited predictive ability. There were socio-emotional (social comparison, cognitive avoidance) and neurocognitive (set-shifting) improvements following MANTRA over SSCM, but no clinical advantage. Conclusions. Emotional avoidance and submissive behaviours may represent maintenance factors for AN. Social cognitive and neurocognitive variables appear less important. No ‘causal’ maintenance factors could be concluded from the data. Implications for adapting MANTRA are discussed. Section C is a critical appraisal of this research reflecting on key learning points and the limitations of the study, as well as the impact of the study on clinical work and in planning future research.