Vol.:(0123456789) Discover Artificial Intelligence (2023) 3:33 | https://doi.org/10.1007/s44163-023-00082-4 1 3 Discover Artifcial Intelligence Research Early detection of dyslexia based on EEG with novel predictor extraction and selection Shankar Parmar 1  · Chirag Paunwala 2 Received: 9 August 2023 / Accepted: 11 October 2023 © The Author(s) 2023 OPEN Abstract Dyslexia is a learning disorder caused by difculties in the brain’s processing of letters and words. This study used EEG recordings to detect dyslexia at a young age. EEG recordings of 53 individuals, including 29 dyslexic and 24 normal individuals, were collected while they were engaged in two distinct mental activities known as the N-Back task and the Oddball task. Predictors were extracted using several methods and reduced using Principal Component Analysis (PCA). A relief-based strategy was applied to select predictors, and Support Vector Machine (SVM) classifer was used to achieve an average accuracy of 79.3% for dyslexia detection, which is better than the performance of its predecessors. The results indicate that EEG recordings and machine learning methods could be useful for identifying dyslexia in children. Keywords Dyslexia · EEG · Power moments · Principal component analysis (PCA) · Feature selection · Support vector machine (SVM) 1 Introduction Dyslexia afects one in fve people, and among those with learning problems, 80–90% have a reading disability as a result of dyslexia. Dyslexia is the most common of all neuro-cognitive problems [1]. Rudolf Berlin, an ophthalmologist in Stuttgart, Germany, invented the name "dyslexia" in 1887, combining the words "difcult to read" (from the English dys) and "to read" (from the Latin legere) [2]. The diagnosis of dyslexia is driven by the fact that it hampers a person’s ability to understand and process language, with mild to severe symptoms. Dyslexia is a global problem that afects people of all ages and walks of life. Even though there have been improvements in diagnosing, understanding, and teaching methods, too many dyslexics are still not recognized and taught. It is not possible to cure dyslexia because it is a condition that afects a person for their entire life and cannot be eradicated [1]. However, dyslexic people can achieve great success with early identifcation, screening, evidence-based reading intervention, and accommodations. Our goal is to identify dyslexic children at an early age, most likely when they frst enroll in school, in the hopes of protecting their professional future. The medical community can fnd that this detection technique, in addition to the conventional methods, is a benefcial tool for a higher level of accuracy. We have received funding under IEEE SPS- HAC to develop an early-detection dyslexia system for children aged 8–10 years. This system will help the medical community, the government in policy-making, and the children in the local region. * Chirag Paunwala, chirag.paunwala@scet.ac.in; Shankar Parmar, shankar.parmar@gtu.edu.in | 1 Electronics and Communication Department, Gujarat Technological University, Chandkheda, Ahmedabad 382424, Gujarat, India. 2 Electronics and Communication Department, Sarvajanik College of Engineering and Technology, Athwalines, Surat 395001, Gujarat, India.