Attention Research Update

June 2002

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

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

Support for Attention Research Update is provided by Cogmed

Cogmed has developed a computerized training program to improve working memory, which is a frequent problem for children and adults with ADHD.  Research has shown that Cogmed's program can enhance working memory, and that improvements in working memory are associated with reductions in attention and learning problems.  You can learn more at  Clinicians interested in learning about the benefits of incorporating working memory training into their practice are invited to request an information package for professionals.

In this issue...

The effect of fatty acid supplementation on ADHD symptoms

Do children's expectations for medication
treatment effect treatment outcome?


Although stimulant medication is known to be an effective treatment for an estimated 70 to 90% of children with ADHD, there remains considerable interest in developing alternative treatments for several reasons.  First, not all children respond to stimulant medication and, even among children who are positive responders, there often remain residual difficulties that need to be addressed.  Second, some children experience intolerable side effects that preclude the ongoing use of stimulants.  Third, although stimulants are generally believed to be safe, even when used over long periods of time, conclusive data supporting their long-term safety is not yet available.  Finally, most studies documenting the beneficial effects of stimulant medication treatment are relatively short-term, and data showing that stimulant medication improves the long-term prognosis for children with ADHD is still scarce.

Dietary supplementation of long-chain fatty acids as an intervention for ADHD has generated considerable interest in recent years.  Certain highly unsaturated fatty acids (HUFAs) are known to play an important role in many aspects of physical health, and recently, attention has been focused on their potential role in a wide range of neuro-developmental and psychiatric conditions. For example, children with ADHD have been shown in several studies to have low blood levels of HUFAs. Because HUFAs are important in brain development and function, some researchers have suggested that increasing HUFA levels via dietary supplements could enhance brain functioning and reduce ADHD symptoms.

Study 1

This study was a randomized, double-blind, placebo-controlled trial of Docasahexaenoic acid (DHA, a type of long-chain fatty acid) supplementation in 63 ADHD children.  (Voight, R.G. et al. (2001). A randomized, double-blind, placebo controlled trial of docosahexaenoic acid supplementation in children with AD/HD. The Journal of Pediatrics, 139, 189-196.) Only children treated effectively with stimulant medication participated, and they continued on medication throughout the four-month study because of ethical concerns about stopping an effective treatment for this duration.  Comparisons between treatment and control participants thus determined whether DHA supplementation provided benefits beyond those already provided by medication.

After four months, participants completed a computerized test of attention, and parents completed both a Child Behavior Checklist and the Conners Rating Form, two widely used rating scales that assess parent perceptions of ADHD symptoms along with other behavioral and emotional difficulties.  The computerized attention test was administered when participants had been off medication for 24 hours.  Behavior ratings, however, were based on parents' observations during ongoing medication treatment.


Comparisons between children receiving DHA supplementation and those receiving placebo "treatment" revealed no differences on the computerized attention test or parent behavior ratings.  Based on this data, the authors conclude that the efficacy of DHA supplementation as an alternative or adjunctive treatment for ADHD is not supported.

The problem with this conclusion is that, because participants remained on medication during the trial, this was not a fair test of DHA supplementation for reducing ADHD symptoms.  Since all of the children were positive medication responders, their symptoms would already have been substantially reduced, thus making it difficult to demonstrate additional benefits.  In children not being treated with meds, however, DHA supplementation might still provide benefits that could not be detected here.

An analogy would be adding behavioral treatment for children already treated effectively with meds.  If no additional benefit was found, one could not conclude that behavioral treatment is not effective for ADHD.  A more reasonable explanation would be that, once effective medication treatment is implemented, behavioral treatment does not result in further gains.  In the absence of medication treatment, or with children who do not benefit from medication, however, it could still be associated with significant benefits.  (In actuality, there is some recent data from the MTA study to suggest that behavioral treatment does yield incremental gains in some areas for ADHD children.)

This concern would not apply to the computerized attention test results -- in which differences between treatment and control differences also were not detected -- because this was completed after meds had been withheld for 24 hours, enough time to clear from the children's systems.  These results thus provide a better indication that DHA supplementation may not be efficacious, although computerized attention tests are not generally considered to be the most valid indicator of treatment effectiveness.

The authors also note that supplementation of only a single fatty acid was provided in their study, and suggest that supplementation with a combination of fatty acids (specifically, omega 3 and omega 6 fatty acids) may provide a stronger test of this intervention.  In the next study reviewed, this is exactly what was done.

Study 2

The second study was conducted with 41 eight to twelve-year-old children with developmental dyslexia -- i.e. they had a specific learning disability in reading (Richardosn, A., & Puri, B.K. (2002). A radomized double-blind, placebo-controlled study of the effects of supplementation with highly unsaturated fatty acides on ADHD-related symptoms in children with specific learning disabilities. Progress in Neuro-Psychopharmacology & Biological Psychiatry, 26, 233-239). None had been formally diagnosed with ADHD, although all had above-average scores for high levels of ADHD based on parent responses to the Conners Rating Scale.

Children were randomly assigned to receive either HUFA supplementation, containing both omega 3 and omega 6 fatty acids, or an identical-looking placebo.  Treatment lasted 12 weeks and neither parents nor the children themselves had any knowledge of which they were receiving.  Twenty-nine participants completed the entire trial, with roughly equal numbers of treatment and control children withdrawing prematurely.


After12 weeks, parents completed the Conners Parents Rating Form a second time so that pre- and post-treatment ratings could be compared.  Prior to treatment, the treatment and placebo groups did not differ on any of the Conners subscales.  In contrast, post-treatment ratings of inattention and global ADHD symptoms were significantly lower in the treatment group than the placebo group.  Average scores for the treated children now fell towards the upper end of what would be considered "normal", while average scores for children in the placebo group remained elevated.

Within the placebo group, no significant reductions were found for any of the Conners scales relative to baseline levels.  Among treated children, however, statistically significant reductions were found for psychosomatic problems, cognitive problems, anxiety, attention problems, hyperactivity, and a global index measuring a broad range of behavior problems.  For several of these scales, the treatment effect sizes exceeded .50, indicating a reasonably robust effect.  Thus, there was clearly evidence of clinically significant benefits for the children receiving the supplements.


Under carefully controlled double-blind conditions, HUFA supplementation with a combination of fatty acids was significantly better than placebo in reducing a wide range of ADHD symptoms in children with reading disability and elevated ADHD symptoms.  This is an impressive result.

There are several limitations, however, that are important to note.  First, although all participants had elevated ADHD symptoms at baseline, none had been formally diagnosed with ADHD.  Whether similar effects would be found in clinically diagnosed ADHD children, particularly when reading disability was not also present, cannot be known with any certainty.

Second, because practical considerations did not permit blood sampling, it is not clear whether participants had fatty acid deficiencies to start with, or whether supplementation actually raised children's fatty acid levels.  No conclusions can thus be made about biochemical mechanisms to explain the improvements observed following HUFA supplementation.  However, the fact participants were not selected for low fatty acid status would be expected to weaken any positive effects of HUFA supplementation.  This makes the results even more impressive.

Finally, although significant results were obtained, the sample size was relatively small, reflecting the fact that this was a pilot study.  Replicating these results with a larger sample is thus required before more definitive claims about the benefits of HUFA supplementation for children with reading disabilities and elevated ADHD symptoms can be made.  This is a caution appropriately expressed by the authors.

Overall Conclusion

Study 1 has been widely cited as demonstrating that HUFA supplementation has no benefit for children with ADHD.  As indicated in the review, however, this conclusion seems inappropriate given that all participants also were treated with medication throughout the study.  In addition, as the authors acknowledge, limiting the supplement treatment to a single fatty acid (i.e. DHA) may also be problematic.  Results from Study 2 highlight that HUFA is potentially a helpful intervention for children with ADHD, and points to the need for a larger trial with a clinically diagnosed ADHD population.

At this point in our knowledge, it seems premature to conclude either that HUFA supplementation has no benefits for children with ADHD, or that such benefits are clearly established.  The authors of both papers are to be commended for doing the hard work required to move us closer to determining the benefits of this treatment.  Hopefully, research will soon be available that will permit a more definitive evaluation.


One concern frequently expressed about stimulant medication treatment for children with ADHD is this: could the medications encourage children to discount the importance of their own efforts when they experience success, and attribute positive outcomes to the medication?

It has been argued that the use of medication could inadvertently teach children to believe they need pills to succeed, and possibly that it contributes to the development of subsequent problems with substance use/abuse.

What are children's expectations about the effects of medication, and are their explanations for day-to-day successes and failures influenced by medication treatment?

These interesting questions were examined in a study published recently in the Journal of Consulting and Clinical Psychology (Pelham et al., (2002). Effects of methylphenidate and expectancy on children with ADHD: Behavior, academic performance, and attributions in a summer program and regular classroom settings. 70, 320-335.)

Participants were 136 boys, ranging in age from 7.6 to 12.7 years, who attended an intensive eight-week summer treatment program for ADHD that included both academic and recreational activities.  One aspect of the program was a Daily Report Card system in which individualized behavioral goals were developed for each child in four different domains: classroom behavior, classroom performance, adherence to rules, and behavior with peers.

At the end of each day, children were provided feedback about whether they had reached each goal for that day.  (The criteria required for the goals to be considered "met" were adjusted during the summer as the children progressed, thereby maintaining a consistent level of challenge.)

Two weeks into the summer program, participants began a trial period in which they received stimulant medication (i.e. methylphenidate).  Prior to beginning this medication trial, the children were told that on some days they would receive placebo pills in an effort to learn whether the medication really helped them.  They also were informed they would be told which type of pill -- real vs. placebo -- they received on a given day.  They were not told, however, that on half the days the information provided about which pill they had taken would be incorrect.

Thus, on some days children were given real medication and told it was real.  On other days, they were given real medication and told it was placebo.  Similarly, on some days they received a placebo and were told it was real. And, on other days, they received a placebo and were told it was a placebo.  The counselors and teachers who were working with the children, and providing daily ratings of their behavior, were not told whether the children had received active medication or placebo, nor what the children themselves had been told.


Several measures were collected each day to determine how well children expected to do in meeting their Daily Report Card goals, how well they actually did, and how they explained their daily successes and struggles.

First, after being told whether they had received real or "fake" medication, children were asked to predict whether they would succeed in meeting their goals that day.  Ratings were made on a 1 (very poorly) to 10 (very well) scale.   These ratings allowed the researchers to determine whether the children's medication-related expectations influenced their expectations for daily success.

Second, behavioral ratings were collected from the counselors and teachers, as were teacher ratings of academic performance.  Teacher and counselor ratings for the Daily Report Card also were examined.  These measures enabled the researchers to determine whether children's medication expectations (i.e. whether they were told they had received real medication) and actual medication condition (i.e. whether they actually received medication or placebo) influenced their behavioral and academic success during the day.

Finally, the researchers obtained children's attributions for their daily successes and failures.  After being told whether they had achieved their Daily Report Card goals, participants were asked why this had occurred.  For each goal, the children rated the degree to which their success or failure was attributable to: their ability, the difficulty of the goal, how hard they had tried, how much the medication helped, and how fairly teachers/counselors had treated them (or, in the peer domain, how nice the other children were.)

Each reason was rated on a 1 (really true) to 10 (not true at all) scale. These data enabled the researchers to examine how children's attributions of success or failure varied according to whether they had actually received medication and whether they believed they had received medication.


As anticipated, children anticipated more success meeting their goals when they were told they had received real medication than on days when they were told the medication was a placebo.  The differences in success predictions were not large, however, (average predictions of 8.4 vs. 8.0 on the 1-10 scale described above), and children clearly expected to be successful even when told they had received placebo medication.  Thus, even though they were somewhat more confident of succeeding when informed they had been given real medication, they were inclined to expect success regardless.

Results for the behavioral ratings were also interesting.  For counselors' ratings, children were reported to do significantly better on a wide range of behaviors (e.g. following rules, interrupting, etc.) when they had received real medication than on placebo days.  The same was true for teachers' behavior ratings and for the quality of the children's academic work.  In regards to the Daily Report Card, children were more than 4 times as likely to achieve their classroom behavior, classroom work, and getting-along-with-counselors goals on days when they had been given medication.  They also were almost 2.5 times as likely to meet their social goals with peers on the days they received real medication.

In contrast to these largely positive effects of medication, there were no significant effects for expectancy condition.  This means that telling children they have received medication when they were actually given a placebo does not result in an improvement in their performance.  And, telling children they were given placebo when they were actually given medication doesn't diminish the actual benefits of the medication.  In other words, what really matters in terms of children's success is whether or not they received active medication, regardless of what they were told.

The last issue examined was the attributions children made for their successes and failures in achieving their Daily Report Card goals.  Overall, there was little indication that boys regarded medication as an important reason for either their success or failure.  When boys succeeded in meeting their goals, they consistently rated the "pill" as the least important determinant for their success.  Instead, they rated their effort and ability level as far-more-important determinants of their success.

(Note: In addition to the results presented above, the researchers conducted a similar study using most of the same children, after they had completed the summer program and were back in their regular schools.  The purpose of this follow up study was to determine whether results obtained in the summer treatment program would generalize to children's regular classrooms.

This is exactly what was found.  As reported above, children behaved better, experienced greater classroom success, and were more likely to meet their daily goals on medication days than on placebo days, and this was true regardless of what they had been told.  In fact, behavioral differences between medication and placebo days were greater than those found during the summer program.  This probably reflects the fact that there was no carefully implemented behavioral program in the regular school setting, making it more difficult for children to succeed on placebo days.  Once again, there was no evidence that children regarded medication as an important reason for their success.  Instead, they again rated effort and ability level as being the most crucial factors for success in achieving their Daily Report Card goals.)


The results of this study have several interesting and important implications.  First, although children may expect more success when they believe they have received medication, boys with ADHD tend to expect success even when not on meds.  In fact, their expectations for success when not on meds may be overly optimistic.

On the one hand, it is encouraging that ADHD boys in this age group appear to expect positive outcomes and generally anticipate being able to accomplish their daily goals.  On the other hand, these optimistic expectations could create difficulties through boys failing to recognize what may be required in order for them to succeed.

(It would be interesting to know whether similarly positive expectations were found for girls with ADHD, as well as adolescents.  Because adolescents with ADHD have struggled for a longer time than the boys in this study, their expectations may not be so optimistic.)

Results from this study also highlight the significant benefits that many children with ADHD derive from medication treatment.  Recall that during medication days, boys were far more likely to achieve their daily goals than on placebo days, even though a very carefully implemented behavioral plan was in place throughout.  They also were doing better on virtually every dimension rated by counselors and teachers, including the quality of their academic work.  These findings highlight that, although behavioral intervention can be an effective treatment for children with ADHD, a multi-modal approach that includes carefully monitored medication treatment is likely to be more effective for many children.

Finally, despite concerns that ADHD children treated with medication will discount their own efforts as a reason for success, and instead attribute any success they experience to medication, this does not seem likely to occur.  As discussed above, boys consistently believed the most important reasons for their success were their effort level and ability, and they downplayed the role of medication.  Thus, although some children treated may come to believe that medication is the primary reason for their success -- and it is important to attend to this possibility -- there was no evidence that this is a frequent occurrence.  In fact, the authors raise the opposite concern, namely that boys with ADHD may minimize the role of medication treatment in their success.

The potential problem with this is that as children get older, they may decide to stop taking medication because they do not believe it is necessary for them to succeed.  In fact, follow-up data from other studies have shown that a large number of adolescents who have received medication treatment for ADHD during childhood discontinue taking medication, even though many could continue to benefit from it.  As a result, these teens may experience negative outcomes because they are no longer receiving a necessary treatment.

The authors suggest that providing children with a realistic understanding of the benefits of medication could be helpful in preventing this premature discontinuation.  They emphasize that parents, teachers, and health professionals should be careful not to inadvertently contribute to the emergence of "negative cognitive styles" by emphasizing to children the role of medication in their daily successes and failures.

Instead, for children who respond positively to medication (and remember, not every child with ADHD does), they suggest that adults should work to "develop realistic medication attributions in children that strike a balance between a belief that they are completely dependent on medication, on the one hand, and complete denial of medication effects, on the other hand."

This is a sound recommendation for parents and professionals to consider.

Thanks again to Cogmed for supporting this issue of Attention Research Update
(c) 2002 David Rabiner, Ph.D.

Information presented in Attention Research Update is for informational purposes only, and is not a substitute for professional medical advice.