METHODS: We did a concurrent, triple-blind, randomised, controlled trial using authorised deception in adults with
ADHD from one centre (University of Tübingen) in Tübingen, Germany. Participants were eligible if they fulfilled the DSM-IV-TR criteria for
ADHD, were aged between 18 years and 60 years, and had no or stable use of medication for at least 2 months with no intention to change. We excluded participants who had comorbid
schizophrenia or
schizoaffective disorder,
bipolar disorder,
borderline personality disorder,
epilepsy, or
traumatic brain injury;
substance abuse or dependence; or current or planned other psychological treatment. Those eligible were randomly assigned to three groups: a
neurofeedback group which received 30 verum θ-to-β
neurofeedback sessions over 15 weeks, a
sham neurofeedback group which received 15
sham followed by 15 verum θ-to-β
neurofeedback sessions over 15 weeks, or a meta-cognitive group therapy group which received 12 sessions over 12 weeks. Participants were assigned equally to one of the three interventions through a computerised minimisation randomisation procedure stratified by sex, age, and baseline symptom severity of
ADHD. Participants were masked as to whether they were receiving
neurofeedback or
sham neurofeedback, but those receiving meta-cognitive therapy were aware of their treatment. Clinical assessors (ie, those assessing outcomes) and research staff who did the
neurofeedback training were masked to participants' randomisation status only for
neurofeedback and
sham neurofeedback. The primary outcome was symptom score on the Conners' adult
ADHD rating scale, assessed before treatment, at midtreatment (after 8 weeks),
after treatment (after 16 weeks), and 6 months later. All individuals with at least one observation after randomisation were included in the analyses. This trial is registered with ClinicalTrials.gov, number NCT01883765.
FINDINGS: Between Feb 1, 2013, and Dec 1, 2015, 761 people were assessed for eligibility. 656 (86%) were excluded and 118 (15%) were eligible for participation in this study. Eligible participants were randomly assigned to
neurofeedback (38 [32%]),
sham neurofeedback (39 [33%]), or meta-cognitive therapy (41 [35%]). 37 (97%) individuals for
neurofeedback, 38 (97%) for
sham neurofeedback, and 38 (93%) for meta-cognitive therapy were included in analyses. Self-reported
ADHD symptoms decreased substantially for all treatment groups (B=-2·58 [95% CI -3·48 to -1·68]; p<0·0001) between pretreatment and the end of 6 month follow-up, independent of treatment condition (
neurofeedback vs
sham neurofeedback B=-0·89 [95% CI -2·14 to 0·37], p=0·168;
neurofeedback vs meta-cognitive therapy -0·30 [-1·55 to 0·95], p=0·639). No treatment-related or trial-related serious adverse events were reported.
INTERPRETATION: Our findings suggest that
neurofeedback training is not superior to a
sham condition or group psychotherapy. All three treatments were equivalently effective in reducing
ADHD symptoms. This first randomised,
sham-controlled trial did not show any specific effects of
neurofeedback on
ADHD symptoms in adults.
FUNDING: German Research Foundation.