An assessment of EEG biofeedback for the remediation of attention-deficit/hyperactivity disorder

Abstract

Seven boys between the ages of 7 and 12 with Attention-Deficit Hyperactivity Disorder (ADHD), some also with Specific Learning Disabilities (LD) and/or elevated aggression scales on the Child Behaviour Checklist (CBCL), were involved in this study of the effectiveness of EEG Biofeedback for the remediation of ADHD. A multiple-baseline design with an embedded ABAB component was used to compare baseline, 12-15 Hz Sensorimotor Rhythm (SMR) EEG Biofeedback (“active”) and single-blind sham feedback (“placebo”) conditions. Seven dependent variables including the ADHD Rating Scale, Child Attention Profile, Children's Checking Task, Controlled Oral Word Fluency Test, paired associate learning and continuous performance tasks were combined as z-scores into a single composite repeated measure. Visual and multiple baseline analyses were inconclusive. The primary analysis was by randomisation test with statistical control for overall trend. Randomisation test results were combined across five subjects (two subjects having failed to complete the study) using techniques for combining independent probabilities; for the combined group the placebo vs. active contrast was significant (p < 0.02) with the active conditions displaying the more improved scores. An examination of effect sizes by subject and condition showed that the placebo vs. active contrast was significant (p < 0.05, based on the confidence intervals for effect estimates) with a large effect size if trend was ignored and the two “drop-outs” were excluded but non-significant with a small effect size if overall trend was controlled, and the two drop-outs included. Effect sizes calculated pre-post were medium on average, and Reliable Change Indices for CBCL parent rated attention were significant for five subjects (p < 0.001). There was no improvement in academic performance as assessed by pre-post Wide Range Achievement Test-Revised. There was no significant effect of condition on EEG Theta/SMR ratio, although there were some significant correlations with individual dependent variables. Four of six children continued to improve on the composite score from the end of the study to six-month follow-up. Results were discussed in light of non-specific treatment effects such as tacit behavioural training, Fried's respiration hypothesis, the cognitive energetic model of attention and confounds due to trend and subject mortality. It was concluded that EEG biofeedback appears to have beneficial effects on ADHD symptoms but these are due largely to non-specific treatment factors and confounds such as regression to the mean.