Meta-Analysis Assesses Neurofeedback for Executive Function in Pediatric ADHD

A new systematic review and meta-analysis, published in a Nature Portfolio journal, examines whether neurofeedback training improves executive function in children with attention-deficit/hyperactivity disorder (ADHD). The study consolidates the available controlled-trial evidence on a training approach that clinicians in pediatric neurology and child psychiatry have debated for more than a decade.

Neurofeedback in the ADHD context almost always means EEG neurofeedback: children view a real-time display of their own scalp-recorded brain activity — most commonly over central and frontal sites, targeting sensorimotor rhythm (SMR) and theta/beta ratio protocols — and learn to modulate those rhythms across repeated sessions. The technique holds obvious appeal for families seeking alternatives or adjuncts to stimulant medication. Yet its efficacy for the cognitive deficits that underlie daily impairment, rather than for behavior rating scales alone, has remained contested.

Executive function is precisely where that question matters most. Children with ADHD characteristically show weaknesses in working memory, inhibitory control, and cognitive flexibility — domains measured in trials with performance-based instruments such as continuous performance tests, Stroop-type tasks, and n-back paradigms, alongside parent and teacher rating inventories. A meta-analysis that pools these outcomes across studies can address a question individual trials, often underpowered and heterogeneous in protocol, cannot: does training-related change in EEG rhythms translate into measurable cognitive gains?

The authors conducted the review according to standard systematic-review methodology: a structured literature search, predefined eligibility criteria for randomized and controlled trials of neurofeedback training in children with ADHD, risk-of-bias assessment of included studies, and quantitative pooling of executive-function outcomes with between-study heterogeneity analysis. By synthesizing results across the eligible trials, the analysis aims to determine the consistency and magnitude of any training effect on executive-function measures and to evaluate how study quality and design characteristics shape the estimates.

For reading-room and lab practice, the clinical stakes are concrete. EEG neurofeedback is already offered in many outpatient neurophysiology and biofeedback settings, frequently as a paid service, while its evidence base for cognitive outcomes has lagged behind its commercial availability. A rigorous pooled estimate helps clinicians distinguish what the intervention has actually demonstrated — on which outcome domains, in which populations, and under which recording and training protocols — from the broader claims that circulate in consumer marketing. It also matters for guideline discussions: standard-of-care recommendations for ADHD rely on behavioral therapy and pharmacotherapy, and neurofeedback's position in that hierarchy depends on exactly this kind of quantitative synthesis.

The review also contributes to a methodological conversation within the field. Earlier meta-analyses of neurofeedback in ADHD have diverged in their conclusions depending on whether they isolated blinded-evaluator outcomes, separated standardized protocols from individualized ones, and distinguished active-control comparisons from no-treatment or waitlist controls. How the present analysis handled those design variables — and how large the pooled sample of children proved to be — will determine how much weight its executive-function estimates can carry.

The findings arrive at a time when interest in neurophysiology-based interventions for neurodevelopmental disorders is growing, driven in part by wider availability of consumer EEG hardware and remote-training platforms. Whether that interest will be matched by evidence strong enough to move neurofeedback from investigational adjunct to recommended treatment for the executive-function deficits of pediatric ADHD is the question this meta-analysis was designed, in part, to answer — and one that subsequent adequately powered, actively controlled trials will need to settle.