Analyzing the Creative Potential of Subjects Using EEG-Induced Capsule Graph Neural Network

The paper introduces an innovative approach to evaluate the creative potential of individuals based on their analogical reasoning abilities, utilizing an Electroencephalography based data acquisition system. The brain signals recorded during analogical problem-solving tasks undergo pre-processing an...

Full description

Saved in:
Bibliographic Details
Published in:Proceedings of ... International Joint Conference on Neural Networks pp. 1 - 8
Main Authors: Ghosh, Sayantani, Konar, Amit, Nagar, Atulya K
Format: Conference Proceeding
Language:English
Published: IEEE 30.06.2024
Subjects:
brain connectivity
Brain modeling
Electroencephalogram (EEG)
Feature extraction
graph neural network
Graph neural networks
Heuristic algorithms
Orbits
Scientific creativity
Statistical analysis
Technological innovation
ISSN:2161-4407
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:The paper introduces an innovative approach to evaluate the creative potential of individuals based on their analogical reasoning abilities, utilizing an Electroencephalography based data acquisition system. The brain signals recorded during analogical problem-solving tasks undergo pre-processing and transformation into brain connectivity networks using the Pearson's correlation coefficient technique. Through centrality-based feature analysis, the brain's hub lobes involved in the cognitive task are identified, showcasing the active participation of the bilateral orbito-frontal cortex and the right supra-marginal gyrus for successful problem solvers. The extracted centrality features are then classified into two categories, creative (CRT) and non-creative (NCRT), employing a novel Capsule Graph Neural Network (CapsGNN). The uniqueness of the proposed classifier model lies in the utilization of a new activation function, Hyperbolic Tangent Exponential (TanhExp), designed to expedite the model's convergence rate. Furthermore, an attention module has been introduced to accentuate crucial information within the primary capsule layer, thereby enhancing the model's discriminative capabilities during the classification task. Additionally, a novel Sigsoftmax function-based dynamic routing algorithm has been incorporated into the classifier model to improve the coupling strength between the primary and higher capsule layers. Performance analysis and comparisons with existing algorithms underscore the superiority of the proposed classifier. Consequently, this innovative technique proves to be a valuable tool for identifying and recruiting creative individuals for various research-intensive sectors.
ISSN:2161-4407
DOI:10.1109/IJCNN60899.2024.10650325