EEG Visual Attention Decoding via Temporal Modulation

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Current methods for decoding EEG signals to determine visual attention are limited by their reliance on specific stimuli, such as flashing lights, and require prior classification of signals, making them unsuitable for complex visual environments like text, images, and menus.

Innovation Solution

A hybrid decoding method that combines stimulus reconstruction from EEG signals with excitation methods using temporally characterized stimuli, allowing for increased sensitivity and robust real-time analysis by calculating statistical dependence between reconstructed modulation signals and generated stimuli, adaptable to various graphic objects without prior signal classification.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If prior classification methods using descriptors and discrimination algorithms are used, then EEG signals can be classified into stimulus classes, but the method becomes dependent on descriptor relevance and discrimination method choice, limiting robustness and applicability

Engineering Contradiction:
Improverobustness of decodingVSAvoidcomplexity of signal classification system
Core Design Contradiction:
ReliabilityVSDevice complexity

Solution Approach 1:

The patent extracts the temporal modulation characteristics from EEG signals directly, separating the essential stimulus identification task from complex descriptor-based classification systems. By focusing only on temporal modulation patterns rather than multiple descriptor types and discrimination algorithms, the method achieves robust decoding without relying on irrelevant descriptors or complex discrimination methods.

Inventive Principle:
Principle #2Taking out (Extraction)

Solution Approach 2:

The temporal modulation analysis method is universally applicable to various stimulus types (flashing lights, rotating objects, moving objects) without requiring stimulus-specific descriptors or discrimination methods. This single approach serves multiple functions: stimulus identification, attention detection, and cross-stimulus generalization, eliminating the need for separate classification systems for different stimulus categories.

Inventive Principle:
Principle #6Universality (Multi-functionality)

2Adaptability or versatility

If flashing light stimuli are used for EEG analysis, then stimulus identification can be achieved, but the scope of application is severely limited to simple visual stimuli

Engineering Contradiction:
Improvescope of visual stimulus applicationVSAvoidlimitation to specific stimulus types
Core Design Contradiction:
Adaptability or versatilityVSReliability

Solution Approach 1:

The temporal modulation analysis method is universally applicable to various stimulus types including flashing lights, rotating objects, moving objects, and other dynamically changing visual stimuli. By extracting temporal modulation characteristics that are common across all these stimulus types, the patent achieves broad applicability without being limited to single stimulus category.

Inventive Principle:
Principle #6Universality (Multi-functionality)

Solution Approach 2:

The method changes the analysis parameter from static spatial patterns to temporal modulation characteristics. This parameter transformation allows the same analysis framework to handle diverse stimulus types that differ in their spatial and temporal properties, as long as they exhibit temporal modulation during observation.

Inventive Principle:
Principle #35Parameter changes

3Measurement precision

If multiple descriptors and discrimination methods are used for EEG classification, then stimulus identification can be performed, but the system becomes dependent on descriptor relevance and discrimination method choice

Engineering Contradiction:
Improveaccuracy of stimulus identificationVSAvoidnumber of descriptors and discrimination methods
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

The patent extracts only the essential temporal modulation characteristics from EEG signals, eliminating the need for multiple descriptors (spatial frequency, temporal frequency, amplitude modulation) and complex discrimination methods. This extraction approach maintains measurement precision by focusing on the most relevant temporal patterns while significantly reducing system complexity.

Inventive Principle:
Principle #2Taking out (Extraction)

Solution Approach 2:

The method segments the EEG signal analysis into distinct temporal modulation components (different temporal frequencies), allowing each component to be analyzed independently. This segmentation simplifies the overall system by breaking down complex stimulus identification into manageable temporal frequency analyses, reducing the need for multiple descriptors and discrimination methods.

Inventive Principle:
Principle #1Segmentation

Data Source

PatentEP3678532B1Decoding the visual attention of an individual from electroencephalographic signals
Publication Date: 2024.10.09 NEXTMIND SAS
  • EP3678532B1 patent drawingFigure 1A
  • EP3678532B1 patent drawingFigure 1B
  • EP3678532B1 patent drawingFigure 2A

AI summary

A method for determining the focusing of the visual attention of an individual from electroencephalographic signals. At least one visual stimulus to be displayed is generated (411) from at least one graphic object, a visual stimulus being an animated graphic object obtained by applying, to a graphic object, a time sequence of elementary transformations temporally parameterised by a corresponding modulation signal. A modulation signal is reconstructed (414) from a plurality of electroencephalographic signals produced by the individual focusing his visual attention on one of the visual stimuli. A visual stimulus corresponding to the modulation signal is identified (415), for which the degree of statistical dependence with the reconstructed modulation signal is greater than a first threshold.