Neurofeedback is controversial because its core claim, that people can learn to change their own brainwave patterns and improve mental health symptoms, has not been convincingly proven in the kind of rigorous trials scientists use to separate real treatments from placebo effects. Multiple sham-controlled studies have found that fake neurofeedback produces the same improvements as the real thing, and the field has struggled to design experiments that settle the question. At the same time, neurofeedback sessions are expensive, rarely covered by insurance, and delivered by practitioners with widely varying qualifications.
The Placebo Problem
The single biggest issue in neurofeedback research is that when scientists compare real neurofeedback to fake versions (where a person sits through the same setup but receives random or someone else’s brainwave signals), both groups often improve at the same rate. This pattern has appeared repeatedly in ADHD research, which is the condition neurofeedback is most commonly marketed for.
In one trial, Lansbergen and colleagues found “significant improvements of ADHD symptoms over time, but changes were similar for both groups.” Arnold et al. reported that both real and sham groups showed significant improvement in ADHD symptoms, with no effect specific to neurofeedback. Van Dongen-Boomsma and colleagues found the same: total ADHD symptoms improved over time, but there was no significant treatment effect. Schönenberg et al. reported that “self-reported ADHD symptoms decreased substantially for all treatment groups between pretreatment and the end of 6 month follow-up, independent of treatment condition.”
Critics point to these results and argue that the benefits people experience from neurofeedback come from sitting in a calm environment, receiving personal attention from a practitioner, and believing the treatment works. Supporters counter that the sham designs themselves are flawed, arguing that even fake feedback can trigger some degree of brain learning, making it an unfair comparison. This back-and-forth has played out across dozens of published papers without clear resolution.
Why Better Studies Are Hard to Run
In drug trials, researchers can give one group a sugar pill and neither the patient nor the doctor knows who got what. Neurofeedback doesn’t lend itself to that kind of blinding. The practitioner typically adjusts reward thresholds manually during each session based on what the patient’s brain is doing in real time. Automating that process for a blinded study may not work as effectively, which means the study design itself could undermine the treatment’s performance.
There are also ethical concerns. If a researcher believes neurofeedback works, withholding it from a control group for the 30 to 40 sessions a typical course requires raises questions about depriving participants of a potentially helpful treatment. These practical barriers mean the field has far fewer high-quality controlled trials than you’d expect for a treatment that has been around since the 1960s.
How It Compares to Standard Treatment for ADHD
A meta-analysis of head-to-head trials comparing neurofeedback directly to stimulant medication for ADHD found that medication was significantly more effective at reducing core ADHD symptoms, including inattention and impulsivity. One notable finding, though, was that dropout rates were significantly lower in the neurofeedback groups, suggesting people tolerate it better or find it more acceptable than medication.
At follow-up time points, medication still tended to outperform neurofeedback, though the results were less consistent depending on who was rating the symptoms (parents, teachers, or clinicians). This matters because parent ratings can be influenced by expectations. A parent who invested thousands of dollars in neurofeedback may perceive more improvement than a blinded teacher observing the same child in a classroom.
Unclear Biological Mechanisms
Neurofeedback is based on the idea of operant conditioning applied to brainwaves: your brain produces a desired pattern, you get a reward signal (a beep, a brighter screen), and over time your brain learns to produce that pattern more reliably. Research in neural plasticity supports the general principle that the brain can be trained through feedback loops, and animal studies have shown that neurons can learn to change their firing patterns when rewarded.
The controversy is whether this laboratory-level mechanism actually translates into meaningful, lasting clinical change in humans. Producing a slightly different brainwave pattern during a training session is not the same as rewiring the brain in a way that reduces anxiety or improves attention weeks later. Supporters argue the mechanism is well-established; skeptics say the leap from basic neuroscience to clinical claims is too large given the current evidence.
Side Effects Are Real but Underreported
Neurofeedback is often marketed as a “natural” or side-effect-free alternative to medication. Research tells a different story. A randomized, sham-controlled, double-blind study found that most participants reported experiencing some side effects during neurofeedback training. These effects were transient, but they were real, and they varied depending on the specific brainwave protocol being used. One protocol targeting sensory-motor rhythms appeared particularly prone to causing side effects. The fact that side effects differ by protocol actually supports the idea that neurofeedback does something biologically, but it also undermines the “completely safe” marketing language many clinics use.
Regulation and Practitioner Quality
Neurofeedback devices are classified by the FDA as Class II biofeedback devices, a category that covers instruments providing visual or auditory signals tied to physiological activity like brainwaves. When battery-powered and indicated only for relaxation training and muscle reeducation, they’re exempt from the premarket notification process. This means the FDA has not evaluated or approved neurofeedback devices for treating ADHD, PTSD, depression, or any other specific psychiatric condition, even though many clinics advertise it for exactly those uses.
On the practitioner side, the Biofeedback Certification International Alliance (BCIA) offers board certification in neurofeedback, with different tiers for licensed clinicians, non-clinical performance practitioners, and technicians working under supervision. The technician pathway does not require a specific educational background. No U.S. state requires a neurofeedback-specific license to practice, so what you’re legally allowed to offer depends entirely on your state’s general practice act. This means the person delivering your neurofeedback sessions might be a licensed psychologist with years of training or a technician with a short certification course.
The Cost Factor
A single neurofeedback session typically costs between $100 and $200, with most practitioners charging $120 to $150. A standard course of treatment involves 30 to 40 sessions, putting the total cost somewhere between $3,000 and $8,000. Most health insurance plans do not cover neurofeedback directly, though some PPO plans offer partial out-of-network reimbursement, and flexible spending or health savings accounts may qualify it as an eligible expense.
This cost structure is central to the controversy. Families spending thousands of dollars on a treatment without strong placebo-controlled evidence are taking a significant financial risk. Practitioners argue that the clinical improvements they see in their offices are real regardless of what controlled trials show. Critics argue that anecdotal clinical experience is exactly the kind of evidence most vulnerable to placebo effects and confirmation bias, especially when the practitioner’s income depends on the treatment working.
Where the Evidence Is Stronger
Not all neurofeedback applications are equally controversial. A systematic review and meta-analysis of neurofeedback for PTSD found that EEG-based training was more effective for PTSD symptoms than for anxiety or depression, suggesting it may work better for some conditions than others. The evidence base for PTSD is still smaller than for ADHD, but early results have been promising enough to sustain ongoing research interest.
The core tension in the neurofeedback debate is that the treatment has a plausible biological basis, decades of clinical use, and many patients who report real improvement, but it has consistently failed to outperform sham treatment in controlled trials. Whether that reflects a genuinely ineffective treatment, a flawed way of testing it, or something in between remains an open question that the field has not yet answered.