EEG classification with BSA spike encoding algorithm and evolving probabilistic spiking neural network

Authors:
Nuttapod Nuntalid;Kshitij Dhoble;Nikola Kasabov
Affiliations:
Knowledge Engineering and Discovery Research Institute, Auckland University of Technology, Auckland, New Zealand;Knowledge Engineering and Discovery Research Institute, Auckland University of Technology, Auckland, New Zealand;Knowledge Engineering and Discovery Research Institute, Auckland University of Technology, Auckland, New Zealand
Venue:
ICONIP'11 Proceedings of the 18th international conference on Neural Information Processing - Volume Part I
Year:
2011

Citing 11
Cited 3

Real-time computing without stable states: a new framework for neural computation based on perturbations

Neural Computation
Evolving Connectionist Systems: The Knowledge Engineering Approach

Evolving Connectionist Systems: The Knowledge Engineering Approach
Person Authentication Using Brainwaves (EEG) and Maximum A Posteriori Model Adaptation

IEEE Transactions on Pattern Analysis and Machine Intelligence
EEG Based Biometric Framework for Automatic Identity Verification

Journal of VLSI Signal Processing Systems
Neural networks letter: To spike or not to spike: A probabilistic spiking neuron model

Neural Networks
A new supervised learning algorithm for multiple spiking neural networks with application in epilepsy and seizure detection

Neural Networks
Real-time epileptic seizure detection on intra-cranial rat data using reservoir computing

ICONIP'08 Proceedings of the 15th international conference on Advances in neuro-information processing - Volume Part I
Towards spatio-temporal pattern recognition using evolving spiking neural networks

ICONIP'10 Proceedings of the 17th international conference on Neural information processing: theory and algorithms - Volume Part I
Multilevel Darwinist Brain (MDB): Artificial Evolution in a Cognitive Architecture for Real Robots

IEEE Transactions on Autonomous Mental Development
Collective stability of networks of winner-take-all circuits

Neural Computation
Probabilistic Computational Neurogenetic Modeling: From Cognitive Systems to Alzheimer's Disease

IEEE Transactions on Autonomous Mental Development

NeuCube evospike architecture for spatio-temporal modelling and pattern recognition of brain signals

ANNPR'12 Proceedings of the 5th INNS IAPR TC 3 GIRPR conference on Artificial Neural Networks in Pattern Recognition
2013 Special Issue: Dynamic evolving spiking neural networks for on-line spatio- and spectro-temporal pattern recognition

Neural Networks
NeuCube: A spiking neural network architecture for mapping, learning and understanding of spatio-temporal brain data

Neural Networks

Quantified Score

Hi-index	0.00

Visualization

Abstract

This study investigates the feasibility of Bens Spike Algorithm (BSA) to encode continuous EEG spatio-temporal data into input spike streams for a classification in a spiking neural network classifier. A novel evolving probabilistic spiking neural network reservoir (epSNNr) architecture is used for the purpose of learning and classifying the EEG signals after the BSA transformation. Experiments are conducted with EEG data measuring a cognitive state of a single individual under 4 different stimuli. A comparison is drawn between using traditional machine learning algorithms and using BSA plus epSNNr, when different probabilistic models of neurons are utilised. The comparison demonstrates that: (1) The BSA is a suitable transformation for EEG data into spike trains; (2) The performance of the epSNNr improves when a probabilistic model of a neuron is used, compared to the use of a deterministic LIF model of a neuron; (3) The classification accuracy of the EEG data in an epSNNr depends on the type of the probabilistic neuronal model used. The results suggest that an epSNNr can be optimised in terms of neuronal models used and parameters that would better match the noise and the dynamics of EEG data. Potential applications of the proposed method for BCI and medical studies are briefly discussed.