SNN Cognitive Load Analysis Using LWDLP EEG Encoding
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for real-time cognitive load analysis using EEG signals face challenges due to high power consumption, complex computations, and noise susceptibility, making them unsuitable for wearable devices with limited resources.
Innovation Solution
A Spiking Neural Network (SNN) based method that employs a Light-Weight-Lossless-Decoder less-Peak-based (LWDLP) encoding technique to convert EEG signals into spike trains, which are then processed by an SNN architecture for cognitive load classification, reducing computational complexity and power consumption.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If deep learning based architectures are used for cognitive load assessment, then classification accuracy is improved, but power consumption and computational complexity increase
Solution Approach 1:
The patent transforms continuous EEG signals into discrete spike train representations, fundamentally changing the data parameter format. This transformation enables the use of Spiking Neural Networks that process events asynchronously, dramatically reducing computational operations and power consumption while preserving classification accuracy for cognitive load assessment
Solution Approach 2:
The patent replaces conventional deep learning mechanical computation systems with Spiking Neural Networks that mimic biological neuron behavior. The SNN architecture uses event-driven spike processing instead of continuous matrix multiplications, reducing computational overhead and energy consumption while maintaining classification performance
2Reliability
If conventional EEG signal processing with artifact removal is performed, then signal quality is improved, but computational complexity and processing time increase
Solution Approach 1:
The patent extracts only the essential spike events from continuous EEG signals using threshold-based detection, discarding redundant continuous signal data. This extraction approach eliminates the need for complex artifact removal algorithms while retaining the critical information needed for cognitive load classification
Solution Approach 2:
The Spiking Neural Network architecture inherently handles noisy EEG signals through its event-driven spike processing mechanism. The network's temporal integration and asynchronous processing naturally filter out artifacts without requiring separate preprocessing steps, making the system self-sufficient in handling signal quality
3Measurement precision
If high-performance models with large network parameters are used, then assessment accuracy is improved, but computational requirements and power consumption increase
Solution Approach 1:
The patent changes the fundamental parameter representation from continuous floating-point values in conventional neural networks to discrete spike timing events. This parameter transformation enables efficient event-driven computation in SNNs, reducing the computational burden of processing large networks while maintaining assessment accuracy through precise temporal coding of EEG features
Data Source
AI summary
State of art techniques, need a decoder following the encoder to encode EEG signals, whose morphology is undefined. Embodiments herein disclose a method and system for a Spiking Neural Network (SNN) based low power cognitive load analysis using electroencephalogram (EEG) signal. The method receives a raw EEG signal from multichannel EEG set up, wherein each of the raw EEG signal is re-referenced and encoded into a spike train using a Light-Weight-Lossless-Decoder less-Peak-based (LWDLP) encoding. Further, the spike trains are processed by the SNN architecture using backpropagation based supervised approach, wherein the spatial information and the temporal information are learnt by the SNN in form of neuronal activity and synaptic weights. Post learning the SNN architecture applies an activation function on the neuronal activity for classifying a cognitive load level experienced by a subject from among a plurality of predefined cognitive load levels using a SNN classifier.