Objective diagnosis of attention-deficit/hyperactivity disorder by using load cell movement analysis under a smart chair in a simulated classroom: influence of sex and age.

backgroundAttention-deficit/hyperactivity disorder (ADHD) is a common neurodevelopmental disorder in children, typically characterized by persistent patterns of inattention or hyperactivity-impulsivity. Its diagnosis relies on criteria outlined in the Diagnostic and Statistical Manual of Mental Disorders, Fifth Edition, and is primarily based on subjective observations and information provided by parents and teachers. Despite the availability of assessment tools such as the Swanson, Nolan, and Pelham questionnaire, diagnosing ADHD in children remains challenging. Such scales predominantly offer subjective insights into the disorder. Therefore, in this study, we developed an objective method that employs load cells for the objective diagnosis of ADHD.

methodsA simulated classroom environment was constructed to replicate a real-world setting. The setup included a desk, chair, and large screen. Load cells, which deform under applied force, were integrated into the four legs of the chair to capture movement data. This study involved 30 children with ADHD (14 boys and 16 girls; mean age: 8 years and 1 month ± 1 year and 10 months) and 30 age- and sex-matched children without ADHD (mean age: 8 years and 3 months ± 1 year and 10 months). Participants were instructed to sit on the chair and watch an age-appropriate educational video on mathematics. Movement data, captured through the load cells, were analyzed to calculate the average trajectory length (ATL) as a measure of activity. For participants with ADHD, SNAP-IV questionnaires were completed by parents and teachers.

resultsThe ATL values for the ADHD and non-ADHD groups were 0.0378 ± 0.0191 and 0.0157 ± 0.0119 (p < 0.0001), respectively. In the ADHD group, boys exhibited a higher ATL (0.0443 ± 0.0100) than girls (0.0303 ± 0.0228; p = 0.0432). The SNAP-IV scores assigned by parents and teachers for participants with ADHD were 33.14 ± 13.75 and 30.95 ± 14.32, respectively. Decision tree classifiers incorporating sex as a variable demonstrated robust performance, achieving an accuracy of 90.67%, sensitivity of 92.33%, specificity of 89.00%, and area under the curve of 91.06%.

conclusionThe smart chair equipped with load cells is an interesting development in progress tool for the objective diagnosis of ADHD and can aid clinical physicians in making decisions regarding ADHD evaluation.