Loss of schizophrenia risk gene XPO7 disrupts neuronal excitability and network regularity via altered Na(+) channel dynamics in human neurons.

Cognitive behavioral therapy (CBT) is a first‐line treatment for eating disorders characterized by recurrent binge eating, with strong empirical support across diagnoses (Cuijpers et al. ; Grilo ). The enhanced transdiagnostic form of CBT (CBT‐E) is structured and modular, typically beginning with psychoeducation and regular eating, followed by targeted strategies addressing key maintaining mechanisms such as dietary restraint, systematic problem solving, exposure to feared foods, body image disturbance, mood intolerance, and, where indicated, supplementary modules targeting perfectionism, interpersonal difficulties, and core low self‐esteem (Fairburn ). Although CBT‐based treatments reliably reduce symptoms at the group level, substantial variability in individual response is common. Some individuals experience marked improvement, others show limited change, and a subset deteriorate (Cuijpers et al. ; Messer et al. ). This heterogeneity highlights the need for approaches that move beyond average treatment effects and instead identify which individuals are most likely to benefit from specific therapeutic components or delivery formats. One promising approach for improving treatment selection is the personalized advantage index (PAI), which estimates how much better (or worse) an individual is expected to respond to one treatment compared with an alternative (DeRubeis et al. ). The PAI is derived by fitting multivariable prediction models that estimate individual‐level outcomes under each treatment condition based on baseline characteristics, and then comparing these predicted outcomes to determine which treatment is expected to yield the superior result for a given person (DeRubeis et al. ; Meinke et al. ). PAI models have been generated using both traditional multivariable linear regression approaches and more complex machine learning algorithms capable of modeling non‐linear relationships and higher‐order interactions (Meinke et al. ). By translating multivariable prediction models into a single, clinically interpretable metric, the PAI offers a structured, data‐driven method for guiding treatment selection (Cohen and DeRubeis ). In principle, this approach may reduce reliance on trial‐and‐error decision making and improve the efficiency of care allocation. The utility of the PAI approach has been evaluated across several mental health conditions, with promising results. Studies comparing treatment options for depression, post‐traumatic stress disorder (PTSD), and somatic symptom disorder have found significantly better outcomes among those who received their PAI‐indicated treatment compared to those who did not, particularly when the predicted difference in treatment response was large (Bauer‐Staeb et al. ; Delgadillo and Gonzalez Salas Duhne ; DeRubeis et al. ; Huibers et al. ; Lopez‐Gomez et al. ; Wade et al. ). Notably, most PAI investigations have compared treatment options that differ in therapeutic orientation (e.g., CBT versus IPT) or treatment modality (e.g., psychotherapy versus pharmacotherapy) using self‐reported or interview‐based sociodemographic and clinical data as input. For example, in the first application of the PAI, DeRubeis et al. () developed a generalized linear model to match individuals with depression to either antidepressant medication or CBT using five identified moderators of treatment response (marital status, employment status, life events, comorbid personality disorder, and prior medication trials). Among participants for whom a clinically significant treatment advantage was predicted (60% of the sample), those who received their predicted optimal treatment had significantly lower symptom severity at post‐intervention compared to those who received a non‐optimal assignment (DeRubeis et al. ).