Early Screening Tools for Dyslexia Using Neurophysiological and Behavioural Indicators

Abstract

Dyslexia is a common neurodevelopmental disorder that has problems in reading, spelling and phonological processing which may cause academic problems in the long term unless early diagnoses are made. The conventional methods of diagnosis mainly use behavioral assessment methods, which can take long to detect any problems in learning until major problems are observed. The present paper suggests a new EEG-based early detection model, which will combine the neurophysiological signal examination with the state-of-the-art machine learning algorithms in order to be able to correctly and promptly identify dyslexia. The offered system is based on the organized pipeline that comprises of EEG signal acquisition, preprocessing, feature extraction, dimensionality reduction, and classification. EEG signals are analyzed to extract time-domain, frequency-domain, and connectivity-based features to reflect the neural activity related to language and cognitive processing. Principal Component Analysis (PCA) is used to down-sample feature dimensions whereas Spiking Neural Networks (SNNs) are used to classify data since they are effective in processing time-related information. Visualization and decision-support mechanisms are also used to support the model to make it more interpretable. Through experimental analysis, the proposed approach shows a better performance over traditional machine learning models, where the accuracy, precision, recall and F1-score are high. Also, the framework facilitates neurofeedback-based intervention where the process of continuous monitoring and cognitive enhancement can be conducted. These findings reveal the usefulness of a combination of neurophysiological cues with intelligent models of computation in the early detection of dyslexia. The study helps to develop scalable, objective and real-time screening systems, and it can be applied in educational and clinical settings. Future research is on multimodal fusion and application to the real world through wearable EEG devices.