In a soon-to-be published meta-analytic study*, researchers report that neurofeedback benefits people with treatment refractory (resistant) epilepsy. The authors’ review of current research revealed that almost 50 million people worldwide have epilepsy, or 0.8% of the general population, and that approximately 33% of patients with this unfortunate medical condition do not benefit from traditional medical treatments (often pharmacotherapies). The researchers hypothesized that neurofeedback (or “EEG biofeedback”) might be an effective alternative treatment given the existing positive published data, albeit from mostly case designs or uncontrolled studies. A meta-analytic study design was used to combine these published studies into a single, more powerful analysis.
Meta-Analytic Study Design
Published studies from 1970 to 2005 that investigated neurofeedback for epilepsy/seizure using SMR (sensorimotor rhythm) neurofeedback or slow cortical potential (SCP) neurofeedback were candidates for inclusion. The SMR protocol is more popular in the United States, while SCP training is more commonly used in Europe. The researchers focused on individual pre/post treatment seizure rates (seizure frequency) due to a limited number of studies with adequate comparison groups (i.e., assignment of participants with dangerous medical conditions to placebo groups is ethically problematic, though more recent novel neurofeedback studies have worked around this limitation; for example, see here). Few studies reported this information, and unfortunately, this reduced the more than 60 available studies to 10 with a final total of 87 participants.
The group seizure rate was significantly reduced by SMR or SCP neurofeedback. For example, 79% of participants who received SMR neurotherapy decreased seizure frequency. The authors specifically cited one study were the median seizure rate was reduced by 61%, or 13 less seizures per month in 50% of participants. The overall effect size, depending on the reported model, hovered around .20. Qualitatively, this is best described as a “small” effect; however, from a statistical standpoint, this modest effect size comes as no surprise as the relatively small number of participants severely limited the chance of larger effects.
The researchers concluded that,
“Based on these consistent findings, the practical value of neurofeedback should be recognized. Medication, while commonly helpful, generally provides effective control of seizure for only two-thirds of patients. In addition to side effects, long-term use of many anti-seizure drugs has significant health risks [Bohan, Mansuri, & Wilson, 2007; Petty, O’brien, & Wark, 2007]. Neurofeedback offers an attractive alternative to neurosurgery or implantation of vagal nerve stimulators. Antiepileptic drugs are also teratogenic: prenatal exposure to anticonvulsant medication puts children at risk for developing autism” or fetal anticonvulsant syndrome which manifests as major physical birth defects and infant mortality [Pennell, 2002]. Neurofeedback offers women of child-bearing age a possibility of controlling epilepsy without risking the health and well-being of her newborn.”
This meta-analytic study provides preliminary evidence that neurofeedback may be