The FDA approved an EEG based marker as a diagnostic for ADHD but how well does it really perform? The evidence thus far is mixed and inconsistent offering both caution and opportunity.

ADHD is characterized by a set of behaviors that include impulsivity, and the inability to sustain attention and concentration. Although its reported prevalence varies, one review in 2008 suggested that 3-7% of school-age children exhibit traits of ADHD.  Diagnosis of ADHD typically relies on the assessment of a child’s behavior through clinical interviews, self-report questionnaires and observations. Like any assessment that relies on answers to multiple questions and subjective observation to throw up a single score or ‘diagnosis’, there are a vast number of combinations of answers (and therefore behavioral profiles) that can yield the same ‘diagnosis’.  A critical question is whether the broad set of traits or behaviors that have been categorized under the umbrella of ADHD have a physiological correlate.

See related post The Difficulty of Diagnosing Depression

The Theta/Beta ratio: An FDA approved marker for ADHD

 The link between ADHD and attention has prompted investigation with EEG to try and come up with a suitable metric for diagnostic purposes. Significantly, the Food and Drug Administration (FDA) in 2013, approved a particular EEG-marker – the theta/beta ratio – for the diagnosis of ADHD as part of the Neuropsychiatric EEG-Based Assessment Aid for ADHD or “NEBA” system.  This is the first instance of an FDA approved EEG based diagnostic for a mental health condition.

The NEBA system involves recording from a single electrode in the Cz position with the idea that it can be used to reduce diagnostic error and refine ADHD treatment plans. The recommendation from the FDA is that it is not used as a standalone diagnostic tool but instead in conjunction with clinical assessments and patient self-report measures. The precise details of the theta/beta ratio used within the system are not publically available because they are considered to be commercially sensitive.

See related post Neurofeedback and ADHD

Inconsistent results across studies

With FDA approval, you might imagine that there is considerable evidence that unilaterally supports the legitimacy of this metric for use in ADHD diagnosis. But a review of the data suggests that the research is in fact quite inconsistent.

The theta/beta power ratio (TBR) for ADHD was first suggested by Joel Lubar from the University of Tennessee as an increase in theta power (4-7 Hz) relative to a decrease in beta power (13-30 Hz), primarily across fronto-central regions. In a confirmatory study with 482 participants aged 6-30 years old, Vincent Monastra and colleagues in the US showed that when measuring the TBR at the Cz electrode you could determine whether someone had ADHD or not with a sensitivity of 86% and a specificity of 98%. A subsequent review of nine studies (n=1498) in 2006 showed a similar level of sensitivity and specificity (94%), again supporting the idea that the theta/beta ratio is a common characteristic of ADHD.

But subsequent studies and analyses have not been so conclusive. For example, a meta analysis conducted in 2012 by researchers from Utrecht University in the Netherlands together with Duke University and Stanford University from the US reviewed 9 studies comparing 1253 children and adolescents with ADHD and 517 without ADHD. They specifically looked at the theta/beta ratio in the context of diagnosing ADHD and concluded that there was not sufficient evidence to warrant it’s use as a diagnostic tool across the whole domain of ADHD. However they did find that there are particular subgroups of children with ADHD who do display this marker and suggested that for these groups it could be used to predict the effectiveness of different therapeutic interventions.

From Arns et al 2013 (above meta analysis)

 

In the same year, a review of the clinical utility of EEG for diagnosing ADHD was published by Sandra Loo and Scott Makeig from the University of California Los Angeles. They concluded that further studies were needed to validate the use of the theta/beta power ratio before it could be used for clinical diagnosis because of inconsistencies in the results across studies.

A further review in 2015 by Jacqueline Saad and colleagues from the University of Sydney also concluded that further validation studies were still needed to clarify the application of the theta/beta ratio in ADHD, and specifically to support the utilization of the NEBA systems as a diagnostic tool. They highlighted the fact that standard EEG wavebands are based on an adult EEG and their boundaries shift through development. The changes over time in the brain of a child with ADHD may simply represent these developmental trajectories rather than be specifically related to the ADHD. They also suggest a modification to the TBR which doesn’t use generic frequency bands but instead uses bands which are tailored to the individual’s EEG. In the past this has been done by anchoring to the peak alpha frequency but difficulties with this approach has led to alternative suggestions. One of these is a method proposed by Wolfgang Klimesch called transition frequency which makes is easier to identify the boundary between theta and alpha at an individual level – a potential confound in previous studies.

Moving towards more specific markers of behavioral traits

Together the data highlights the difficulties that arise when trying to use a single univariate metric for the diagnosis of a complex spectrum of behaviors such as ADHD. For now the latest recommendations offered by a recent report in 2016 from the American Association of Neurology states that the technique should not be used as a stand-alone diagnostic tool, does not replace a standard clinical evaluation and that its high level of false positives has the potential to cause significant harm to patients due to misdiagnosis.

The good news, however, is that as a direction this holds tremendous promise.  Eventually, one should be able to define behavioral traits and mental health conditions based on clear physiological metrics rather than subjective assessments.  However this will have to take into account the vast individual variability both in the EEG and in the spectrum of traits associated with ADHD and the condition under which they arise.