Open Access

ARTICLE

Interactive Human Interface for ERP Component Extraction from Gifted Children

Kawther Benharrath¹, Amine Ben Slama^2,*, Balkine Khadoumi¹, Mounir Sayadi¹, Hervé Rix³, Olivier Meste³, Sophie Guetat⁵, Jérôme Lebrun³, Marie-Noële Magnie-Mauro^4,5

1 University of Tunis, ENSIT, SIME Laboratory, Taha Hussein Street, 1008, Tunis, Tunisia
2 University of Tunis El Manar, ISTMT, LR13ES07, LRBTM, Tunis, 1006, Tunisia
3 Laboratory of Computers, Signals and Systems I3S, UMR 7271 CNRS, Sophia-Antipolis, 06900, France
4 Functional Exploration Service of the Nervous System (EFSN), CHU Pasteur, Nice, 06000, France
5 Laboratory Bases, Corpus, Language (BCL), UMR 7320, UCA, Nice, 6357, France

* Corresponding Author: Amine Ben Slama. Email:

Intelligent Automation & Soft Computing 2022, 33(2), 1063-1080. https://doi.org/10.32604/iasc.2022.023446

Received 08 September 2021; Accepted 10 November 2021; Issue published 08 February 2022

Abstract

In the last century, scientists started to give importance to gifted children (GC) and to understand their behavior. Since then, research has pursued the various categories of these children and their early diagnosis in order to find the best control of their skills. Therefore, most researchers focus on recent advances in electroencephalogram (EEG) and cognitive events. The event-related brain potentials (ERPs) technique is generally used in the cognitive neuroscience process. However, it is still a challenge to extract these potentials from a few trials of electroencephalogram (EEG) data. The N400 ERP component is an important part of the studies of cerebral science and clinical neuropsychology. In this ongoing study, a new experimentation protocol and human tablet interactive equipment were assigned to analyze the brain activity. A combination of two techniques the Integral Shape Averaging (ISA) and Integral Shape Averaging applied on belated window (ISA-BW) was built to extract the semantic component from a single trial and to enhance the signal-to-noise ratio (S/N). The results obtained were compared with the most used method in the medical field Grand Average (GA). In addition, a statistical study was performed on a database for accurate characterization of children using feature reduction. The experimental results show the efficiency of the suggested approach which manifests the discriminant statistical feature extraction (J = 2.032) from ERP component dataset that can contribute to the recognition of GC. The proposed method is reinforced by a pilot device processed by an electrical engineer to improve the protocol simulation. The experimental procedure proves that the present approach is very interesting and helpful for improving the identification of such gifted children.

---

Keywords

---