Le neurofeedback dans le trouble déficit de l’attention/hyperactivité de l’enfant est-il efficace ? Depuis les études rigoureuses jusqu’aux bonnes pratiques cliniques.

Résumé Le neurofeedback par la création d’une boucle psychophysiologique rétroactive présente des avantages, en comparaison des autres techniques de remédiation cognitive utilisées dans la prise en charge des enfants souffrant de trouble déficit de l’attention/hyperactivité (TDA/H). Pourtant, bien que cette technique soit utilisée depuis près de 20 ans dans le TDA/H, son niveau de preuve d’efficacité reste débattu. L’évolution du nombre de publications recensées par la base de données électroniques PubMed par les termes medical subject headings (Mesh) « Neurofeedback » et « Attention-Deficit Disorder with Hyperactivity » permet de proposer une analyse de l’évolution de la littérature sur le neurofeedback et le TDA/H suivant deux périodes : avant 2011 et après 2011. Les premières études avant 2011, sur neurofeedback et TDA/H, ont mis essentiellement l’accent sur la qualité du protocole de neurofeedback et de l’effet d’apprentissage au cours des séances, aux dépens de la construction méthodologique du protocole en lui-même (absence de randomisation et de groupe témoin). Cette attention semble avoir diminué dans les études plus récentes, qui se concentrent sur la construction de protocoles plus rigoureux (avec groupes témoins de meilleure qualité et évaluations en aveugle). Cependant, les conditions de bonne pratique du neurofeedback lui-même sont le pendant indispensable aux études de preuve d’efficacité bien conduites. Ainsi les futures études d’efficacité du neurofeedback dans le TDA/H devront allier la qualité méthodologique des études randomisées contrôlées en aveugle à la qualité de la conduite des séances de neurofeedback. Background Neurofeedback, based on the concept of creating a retroactive psychophysiological loop, has advantages compared with other cognitive therapeutic techniques in the treatment of children with attention-deficit/hyperactivity disorder (ADHD). However, although this technique has been used for almost 20 years in ADHD, the level of evidence for its efficacy remains debated. This debate has recently been brought into focus in the literature following a published meta-analysis. Aims This article aims to review and classify existing literature on the efficacy of neurofeedback in ADHD. Methods Publications were identified through a literature search of the electronic database PubMed using the Medical Subject Headings (Mesh) terms “Neurofeedback” and “Attention-Deficit/Hyperactivity Disorder”. All relevant papers published in English or French were reviewed by the authors. These were separated into 2 groups: published before and after 2011. Results Prior to 2011, studies are characterized by low sample size and are often non-randomized and uncontrolled. In addition these are most often performed by practitioners who participated in the development of the first neurofeedback equipment and who have electrophysiological expertise, essential for effective training during sessions. In 2009 Arns et al. published the first meta-analysis on the efficacy of neurofeedback in ADHD. They found a large effect size, 1.02 (0.84 to 1.21) and 0.94 (0.76 to 1.12) respectively for the inattention and impulsivity dimensions, and a moderate effect size of 0.71 (0.54 to 0.87) on the hyperactivity dimension. Improved inattention dimension was proportional to the number of sessions and maintained in randomized trials, which was not the case for the hyperactivity dimension. After 2011, studies are characterized by larger samples, and methodology including randomized, controlled trials and blinded assessments. In 2013, Sonuga-Barke et al. published the second meta-analysis on the efficacy of neurofeedback in ADHD ( Am J Psych ). They found a smaller effect size of 0.59 (0.31 to 0.87) compared to the first meta-analysis of 2009. The effectiveness of neurofeedback was not confirmed in studies with blinded assessment. The effect size was 0.29 (−0.02 to 0.61) ( P = 0.07; NS). However the effectiveness of the training during neurofeedback sessions in some of the included studies has been called into question. Conclusion Methodological issues are likely to have a large impact on results obtained in studies of neurofeedback. Thus, it is critical that future trials implement adequately randomized, controlled, blinded designs that do not compromise the quality of neurofeedback session itself. [ABSTRACT FROM AUTHOR]