VERA: Virtual Environments Recording Attention

Abstract

Children with Attention Deficit Hyperactivity Disorder (ADHD), present different symptoms binding for everyday life, e.g. difficulty to be focused, impulsiveness, difficulty to regulate motor functions, etc. The most commonly prescribed treatment is the medication that can present side effects. Another solution is behavioural treatment that does not seem to present better results than medication for a higher cost. A novel method with growing interest is the use of neurofeedback (NF) to teach the patient to self-regulate symptoms by herself, through the visualisation of the brain activity in an understandable form. Moreover, virtual reality (VR) is a supportive environment for NF in the context of ADHD. However, before proceeding the NF, it is important to determine the features of the physiological signals corresponding to the symptoms' appearance. We present here a novel framework based on the joint measurement of electroencephalogram (EEG) and sight direction by equipment that can be embedded in VR headset, the goals being to estimate attentional state. In parallel to the signal acquisition, attentional tasks are performed to label the physiological signals. Features have been extracted from the signals and machine learning (ML) models have been applied to retrieve the attentional state. Encouraging results have been provided from the pilot study with the ability to make the right classification in multiple scenarios. Moreover, a dataset with the labelled physiological signals is under development. It will help to have a better understanding of the mechanism behind ADHD symptoms.