Computational Models Could Aid ADHD Diagnosis

Nadja Ging-Jehli

While a combination of computational models with neurocognitive tests are used increasingly more to understand the cognitive characteristics of ADHD patients, it is not understood whether computational models refine the understanding of cognitive concepts such as slow processing speed and inhibition failures.

A team, led by Nadja Ging-Jehli, a graduate student in psychology at Ohio State, synthesized and summarized findings from available studies using 3 commonly applied computational models-- diffusion decision models, absolute accumulator models, ex-Gaussian distribution models.

How ADHD is Currently Diagnosed

Currently, the majority of mental health conditions are diagnosed for and treated based on clinical interviews and questionaries, as well as cognitive testing, which can give clinicians a better understanding on why an ADHD patient is behaving in a certain way.

However, in ADHD cognitive tests do not identify the variety of symptoms and deficits, including selective attention, poor working memory, altered time perception, difficulties maintaining attention, and impulsive behavior commonly associated with the disorder.

The Value of Computational Models

By using computational tests, researchers believe they can better capture the complexity of the symptoms associated with the disease.

The investigators examined 50 studies from a broad range of ADHD tests. Overall, the research team identified 4 areas can improve the utility of neurocognitive testing for ADHD—the requirements for appropriate application of the computational models, the consideration of sample characteristics and neurophysiological measures, the integration of findings from cognitive psychology into the literature of cognitive testing to reconcile mixed evidence, and future directions for the study of ADHD endophenotypes.

The results of the study show sample characteristics and integrating findings from computational models and neurophysiological measures could yield evidence for ADHD endophenotype-specific cognitive characteristics.

Examining the cognitive characteristics of ADHD endophenotypes can exist beyond the scope of existing research, mainly because some tests lack the sensitivity needed to detect clinical characteristics, analysis methods do not allow the study of subtle cognitive differences, and the precategorization of participants restricts the study of symptom severity on a continuous spectrum.

Future Uses

Children with ADHD often take longer to make decisions while they perform tasks. While tests that rely on average response times explain this, computational methods can provide additional data on the patient that could be leveraged for different treatment.

Ultimately, the research team believes computational testing can offer better characterizing ADHD and any accompanying mental health diagnoses including anxiety and depression, improving treatment outcomes, and potentially forecast which adolescent patients will "lose" the ADHD diagnosis as adults.

The researchers also found a broader range of externally evident symptoms, along with subtle characteristics are difficult to detect with common testing methods.

However, by understanding the many biologically based differences in children with ADHD shows a single task-based test is not sufficient to make a diagnosis.

"We can use models to simulate the decision process and see how decision-making happens over time - and do a better job of figuring out why children with ADHD take longer to make decisions," Ging-Jehli said in a statement.

The researchers suggested various new methods leveraging the technology to better diagnose ADHD patients. By combining cognitive testing with other measures, such as eye-tracking and EEGs, clinicians could have enough data to make a diagnosis more reliable and shape treatment decisions.

The study, “Improving neurocognitive testing using computational psychiatry—A systematic review for ADHD,” was published online by the American Psychological Association.