Attention Research Update

January 2003

"Helping parents, professionals and educators stay informed about new research on ADHD"

David Rabiner, Ph.D.  Senior Research Scientist, Duke University


Neurofeedback - also known as EEG Biofeedback - is an approach for treating ADHD that has been studied and practiced for a number of years.  In neurofeedback treatment, individuals are provided with real-time feedback about their brainwave activity and taught to use that feedback information to alter brainwave actvity.  As a treatment for ADHD, neurofeedback is based on findings that measurements of brain activity in many individuals with ADHD indicate reduced activity in the prefrontal region and frontal lobes (i.e. cortical slowing).  (For a review of studies that have documented a pattern of cortical slowing in individuals with ADHD, go to - see the article titled "An objective procedure for diagnosing ADHD?" and )

Neurofeedback treatment is designed to train individuals to increase the production of brainwave patterns that reduce or eliminate this cortical slowing, and thus reduce or eliminate many associated ADHD symptoms.  Specifically, in the most widely used neurofeedback training protocol, individuals are trained to increase the frequency of electrophysiological activity within either the 12-15 HZ or 16-20 HZ range, while attempting to decrease slower cortical activity in the 4-8 Hz range.  Such alterations in are intended to eliminate the EEG differences that are frequently found between individuals with and without ADHD, and it is hypothesized that when this occurs, improvement in attention and reductions in hyperactive/impulsive behavior will result.

Within the medical and scientific communities, a diverse range of viewpoints exists on the utility of neurofeedback treatment for ADHD.  At one extreme, there are prominent researchers who argue that, although neurofeedback treatment is consistent with current theories about the biological underpinnings of ADHD, there is a lack of scientific data documenting the efficacy of this approach.  In fact, in a recently published survey of 100 ADHD experts, relatively little support for neurofeedback as an appropriate treatment option was expressed.  (For a review of this survey, go to and see the article titled New eavluation and treatment guidelines from expert panel.)

On the other hand, some neurofeedback researchers and practitioners have argued that published studies clearly establish the effectiveness of this treatment.  In fact, there have been numerous cases study reports and two controlled investigations of neurofeedback treatment for ADHD in which positive results have been reported.  Unfortunately, even for the controlled studies, the judgment of many is that limitations in the research design prevent any firm conclusions about the efficacy of neurofeedback treatment from being made.  Thus, although promising results have been reported, there is a compelling need for additional controlled studies of neurofeedback treatment for ADHD so that the potential benefits of this approach can be more clearly determined.

A study that was recently published in the journal Applied Psychophysiology and Biofeedback represents an important step in this direction (Monastra et al. 2002. The Effects of Stimulant Therapy, EEG Biofeedback and Parenting Style on the primary symptoms of ADHD. 27, 249.)   Participants were 100 children (83 boys and 17 girls) with an average age of 10 (range 6-19). Each child was diagnosed with ADHD and treated at the Family Psychology Clinic, a private outpatient psychological clinic in upstate New York.  The ADHD diagnosis was established using a structured interview and standardized parent and teacher behavior rating scales.

In addition to meeting all DSM-IV behavioral criteria for ADHD, partcipants were required to show impaired performance on a computerized test of sustained attention, and, to have QEEG scans that fell in an abnormal range relative to same age peers (i.e. to show the pattern of cortical slowing described above that has been found to characterize many individuals with ADHD). These last two requirements were included so that treatment changes for "objective neurophysiologocal measures" in addition to changes in parent and teacher ratings could be examined.

After each child/teen was diagnosed, his or her parents were informed of a comprehensive treatment approach that included stimulant medication, parent counseling, school consultation to establish and monitor a program of academic support, and neurofeedback. Approximately half of the participants (n=51) opted to include neurofeedback as part of their child's treatment.

Because this study was conducted in a clinic setting where parents paid for treatment (as opposed to an academic research setting where treatment is often provided at no charge), it was not possible to randomly assign children to receive neurofeedback as part of their treatment package.  One limitation associated with the lack of random assignment is the possibility that children in the two groups (i.e. those whose treatment included neurofeedback and those whose did not), or their families, may have differed in a systematic way prior to the beginning of treatment.  If this were true, than any group differences found at the end of treatment may have reflected differences that existed prior to when treatment began, rather than resulting from the inclusion vs. exclusion of neurofeedbac.

Fortunately, however, this does not seem to have been true. Information presented indicates that the two groups (i.e. children whose treatment included neurofeedback and those who did not) did not differ in age, gender composition, IQ, or socioeconomic status.  In addition, the representation of the different ADHD subtypes (i.e. inattentive and combined) was virtually identical.  Finally, pre-treatment parent and teacher ratings of ADHD symptoms, scores on the TOVA (a computerized test of sustained attention that is often used in ADHD evaluations and to monitor the effects of medication treatment), and results of a QEEG scan were also equivalent.


Participants received treatment over the course of 12 months.  The different treatment components are described below.

Stimulant medication: All participants received treatment with Ritalin throughout the year.  The average daily dose was 25 mg (10 mg in the morning, 10 mg at midday, and 5 mg in the late afternoon) for children in both groups.  All children were treated with Ritalin and received the lowest dose at which normal results on a computerized test of sustained attention were obtained.  Once this dose was determined, it reamined constant over the course of the year.

Parent Counseling: Parents participated in a ten-session parenting class, followed by individual consultation on an "as needed" basis.  The parenting class was designed to increase parents' understanding of ADHD and help them increase the use of systematic reinforcement strategies and positive parental attention. Information on nutrition, problem solving with teens, and the educational rights of children with ADHD was also presented.   The average number of clinical contact hours during the 12 months (parenting classes and subsequent individual consultation) totaled 25 for parents in the neurofeedback group and 27 for the other parents.

School Consultation: At the conclusion of the diagnostic evaluation, parents were informed about procedures to obtain special educational services for their children under the appropriate federal regulations.  In accordance with applicable laws, school districts evaluated each child and developed, revised, and implemented an individualized educational program (IEP) or a plan of academic support/accommodation ("504 Plan") for each with the assistance of the treating clinician.  For each group, the mean number of on-site consultations during the treatment year was 3.

Neurofeedback: For children whose parents elected to include neurofeedback in their child's treatment (n=51), "attention training" sessions lasting 30 to 40 minutes were conducted on a weekly basis.  Periodic QEEG scans were used to determine training effectiveness and training continued until the child exhibited a degree of cortical slowing on the QEEG scan that fell within the age appropriate normal range patient no longer exhibited abnormal cortical slowing.  The average number of sessions required to reach this criterion was 43, and highlights the duration of time that may be required for successful neurofeedback treatment.

As is evident from the above discussion, the overall treatment regimens for the two groups of children appear to have been virtually identical, except for the inclusion of neurofeedback treatment in one of the groups. Children in both groups received the same medication treatment, parents of these children received an identical parent-counseling program, and school consultation services appear to have been the same as well. Because the groups did not differ in any systematic way before treatment began, the researchers could evaluate whether including neurofeedback training made any appreciable difference in the children's outcomes.


A comprehensive set of treatment outcome measures was collected on each child one year after treatment had begun.  These measures included: 1) parent and teacher ratings of inattentive and hyperactive/impulsive behavior using the Attention Deficit Disorder Evaluation Scale (ADDES), a widely used standardized behavior rating scale; 2) the children's scores on the TOVA; and 3) the children's Attention Index score, based on a QEEG scan.  Each measure was collected twice: once when participants were still on medication and a second time after they had been off medication for an entire week.  This same set of measures had been collected prior to the beginning of treatment, so that change over the course of treatment could be determined.

The authors predicted that children in both groups would show behavioral improvements and "normalized" TOVA (i.e. the TOVA is the computerized test of sustained attention that was used in this study) results when medication treatment was still in place.  After medication had been discontinued for one week, however, these gains were only expected to persist in children whose treatment included neurofeedback.  This prediction was based on prior case study reports suggesting that neurofeedback treatment is associated with sustained improvements in ADHD symptoms while the benefits associated with medication typically do not extend beyond when medication is being actively administered.  The researchers also expected that only children who received neurofeedback would show normalized Attention Index scores on the QEEG scan.

Outcomes when children were still on medication

As predicted, TOVA scores at the first outcome assessment were well within the normal range for both groups.  Thus, there was evidence that medication was continuing to be helpful in normalizing children's performance on a laboratory measre of attention.

In contrast to expectations, however, parent and teacher ratings of ADHD symptoms remained in the clinical range for children who had not received neurofeedback.  For participants whose treatment included neurofeedback, however, average parent and teacher ratings of ADHD symptoms were all well within in the normal range.  These results are shown below.  Scores below 7 are considered to indicate significant difficulty and higher scores actually represent fewer problems.  The numbers reported represent the average score for each group.  As can be seen scores for children whose treatment included neurofeedback were consistently above 7 while scores for children who did not receive neurofeedback as part of their treatment had scores that were consistently below 7.

Parent inattention ratings with/without neurofeedback

Parent hyperactivity ratings with/without neurofeedback

Teacher inattention ratings with/without neurofeedback

Teacher hyperactivity ratings with/without neurofeedback

Outcomes after medication was discontinued

Similar results were obtained on the outcome measures taken after the children had been without medication for an entire week.  As before, children whose treatment had not included neurofeedback continued to show significant ADHD symptoms according to parent and teacher ratings.  In addition, the TOVA results for these participants fell in the clinical range on 3 of the 4 subscales.  As expected, QEEG scans taken on these children continued to show a pattern of cortical slowing that is characteristic of individuals with ADHD.

In contrast, parent and teacher ratings for children in the neurofeedback group all remained within the normal range.  In fact, ratings obtained after medication was discontinued were virtually identical to those provided when medication was still being administered.  In addition, TOVA results remained well within the normal range and average QEEG scan results no longer indicated significant cortical slowing.  Thus, on both behavior and physiological measures, these children continued to show marked improvement even after medication had been discontinued.

The impact of parenting

Although the primary focus of this study was to document the value of including neurofeedback in a comprehensive treatment program for ADHD, the authors were also interested in whether parenting style would moderate the results obtained.

To examine this issue, parents were interevied at the 12-month outcome period to learn whether they had been using the systematic behavior management principals that had been the focus of the parent-counseling program.  Parenting style was rated as "systematic" if parents reported the consistent use of time out, removal of privileges, and use of rewards to encourage appropriate behavior.  Parenting style was rated as "unsystematic" if the parents reported that they had not used these behavior management strategies "most of the time" but had instead resorted to physical punishment, giving in to the child's demands, or avoiding their child.

Among children whose treatment did not include neurofeedback, there was no evidence that use of a "systematic" parenting style impacted treatment outcome.  Thus, regardless of whether or not parents reported consistent use of behavior management strategies, neither parent nor teacher ratings of children's ADHD symptoms showed any difference.

For children whose treatment included neurofeedback, however, there was evidence that parenting style was a significant influence of treatment outcome.  Thus, parent ratings of children's behavior showed significant improvement among parents who reported the use of systematic parenting, but were less robust when this parenting style had not been adopted.  For teachers' ratings, in contrast, signficant behavioral improvement was evident - both before and after medication was discontinued - regardless of the parenting style parents reported using.


These results provide strong suggestive evidence that incorporating neurofeedback into a comprehensive treatment approach for ADHD can yield important benefits.  As discussed above, only the participants whose treatment included attention training via neurofeedback showed behavioral improvement upon follow up, and these benefits were evident even after medication was discontinued. These children were doing substantially better --according to both parents and teachers--than participants who had not received neurofeedback.

In addition to these behavioral changes, there was evidence of improvement on computerized tests of attention for neurofeedback children that persisted beyond the administration of medication.  This is important because such measures are not influenced by the baises that can affect parent and teacher behavior ratings.  Similarly, the pattern of cortical slowing that is found in many individuals with ADHD, and which is specifically targeted by neurofeedback, was no longer evident.  This suggests that the gains associated with neurofeedback training were unlikely to have occurred because of a placebo effect, but instead reflect meaningful changes in EEG activity that are known to be associated with ADHD symptoms.

This is an impressive set of findings, but as with any study, it is important to recognize its limitations.  First, it is surprising that no significant gains in parent and teacher ratings were obtained for the non-neurofeedback group, even when medication treatment was still in place.  Recently published results from the MTA study (for a review of this study, go to document substantial benefits from medication treatment alone, and in combination with behavioral interventions, over a 14-month period.

Based on these results, as well as results from other studies, improved symptom ratings from parents and teachers would have been anticipated. As the authors note, the failure to document similar improvement in this study may have occurred because no procedure was in place to adjust children's medication treatment as needed.   Carefully monitoring the ongoing effectiveness of medication treatment is critically important to optimize treatment benefits, and even when careful procedures are used to determine the best starting dose for a child, adjustments are often needed.  (For a recent study of this issue, go to ).

It is quite likely that the absence of ongoing monitoring in this study accounts for the failure to find improvement in children whose treatment did not include neurofeedback.  This does not call into question the benefits found in this study of including neurofeedback as part of a comprehensive treatment program.  It does, however, raise the question of whether these incremental benefits would have been detected if the benefits provided by the other intervention components occurred as expected.

It also should be noted that because random assignment to treatment conditions was not used in this study, one cannot rule out the possibility that parents who opted to include neurofeedback in their child's treatment were a more highly motivated group of parents, and this is why their children did better.  After all, this was a time-consuming and expensive addition.  The fact that there was no differential attendance in parent counseling sessions between the two groups mitigates these concerns, however.  In addition, the EEG changes revealed by the QEEG scan seem to make it less likely that enhanced parent motivation alone could explain the differential treatment results.  This is because these results were associated with documented changes in neurophysiological processes known to be associated with ADHD, and it seems unlikely that parents' motivation, or other extraneous factors, would have produced such changes.

A related issue is the absence of a "true" control group.  The strategy adopted by these researchers was to determine whether the addition of neurofeedback treatment to a comprehensive set of interventions was associated with incremental gains. As described above, this was found to be the case. In addition to the neurofeedback treatment these children received, however, there was also the considerable extra therapist attention associated with the delivery of neurofeedback treatment. It is possible that the general effects of additional attention, and not the specific effects of neurofeedback, resulted in the more postive outcomes for patients receiving neurofeedback.

Although providing the non-neurofeedback participants with equivalent therapist attention was not possible in a study in which participants are actual clients paying for services, it would be important to include such controls so that this possibility can be completely ruled out.  This, of course, would likely require a grant-funded study in which participants received treatment at no cost.  While this would be a useful control to include, it is also important to note that individual therapy approaches for ADHD children are widely regarded to be ineffective in alleviating core symptoms. Thus, it seems unlikely that a non-specific factor like therapist attention could explain the incremental gains in the neurofeedback treated group.

It is also important to emphasize that neurofeedback was delivered as part of a comprehensive treatment plan that included three other components. There is thus no basis for determining whether neurofeedback alone would have yielded positive results.  And, although gains were sustained for one week beyond the active use of medication, it is unclear whether these gains would persist without ongoing intervention.  These issues would be important to address in subsequent research.

Finally, in addition to the outcome measures collected in this study, it would have been helpful to obtain information about children's academic success at school.  Academic functioning is often significantly impaired in children with ADHD, and one cannot assume that improvements in core ADHD symptoms necessarily translate into meaningful gains in this critically important area. Documenting that neurofeedback treatment is associated with enhanced real world academic success would thus be tremendously important. The need for this applies not only to neurofeedback treatment, however, as the positive impact of any ADHD treatment on long-term academic achievement remains to be conclusively demonstrated.

In summary, results from this study offer evidence that neurofeedback can make an important contribution to the treatment of individuals with ADHD.  However, the limitations noted above prevent any firm conclusions about the effectiveness of neurofeedback treatment from being made at this time, and additional research in which necessary controls are incorporated will be required before such conclusions can be made. One hopes that these studies will soon be forthcoming so that questions concerning the efficacy of neurofeedback can be conclusively answered.

Information presented in Attention Research Update is for informational purposes only, and is not a substitute for professional medical advice.  Although newsletter sponsors offer products and services that I believe will be of interest to subscribers, sponsorship of Attention Research Update does not constitute a specific endorsement or guarantee of any company's product or services.