Researchers from the universities of Málaga (UMA) and Alicante (UA) in Spain have developed an artificial intelligence tool to help with the early diagnosis of attention deficit hyperactivity disorder (ADHD), a condition that affects around 5 per cent of the population.
ADHD is a neurodevelopmental disorder that causes a massive deterioration in executive functioning – which is a group of mental abilities such as working memory, flexible thinking and self-control – and manifests itself in young children with symptoms such as attention deficit or hyperactivity and uncontrolled impulsivity.
However, these signs are often ‘the tip of the iceberg’ of other more complex symptoms, such as problems with decision-making, planning, organisation, retaining important information or difficulties in regulating emotions and motivation, the professors from the Faculty of Psychology and Speech Therapy at the University of Málaga Rocío Juárez and Rocío Lavigne, who carried out this work together with researchers Ignasi Navarro and Juan Ramón Rico from the University of Alicante, explained to EFE.
Early assessment of ADHD is crucial for the effective treatment of those affected, but it is a ‘long and complicated’ process that requires the intervention of professionals from different disciplines, such as neuropediatricians, child psychiatrists, psychologists or psycho-pedagogues, and the involvement of family members, teachers and other ‘observers’ close to the child.
According to the UMA professors, it is difficult to make a complete diagnosis of ADHD before the age of six: hence the idea of designing an instrument that can help specialists detect this condition as quickly as possible.
Researchers from UMA and UA have created a computer programme in which they have entered the parameters of 694 children aged six to 12 diagnosed with ADHD over the last decade in Spain.
When new patient data is entered into the software, it analyses the variables already incorporated, looks for common patterns and establishes a possible diagnosis.
‘Our machine learning model has skilfully predicted ADHD diagnoses in 90 per cent of cases and there is potential for further improvement with the expansion of our database,’ point out those responsible for the research in a scientific article published by the National Library of Medicine.
Rocío Lavigne said that the idea is to increase this sample to 1,500 or 2,000 subjects in Spain and even incorporate cases from abroad in order to extend the project to other European countries.
The tool is currently a pilot test that must be perfected ‘to make it even smarter and predict better’. In addition, it must be validated before it can be used by medical, psychological or educational professionals, and this is a process that may require a few more years of work.
Lusa