Cross-Session EEG Normalization via ​ SSVEP-Based Linear Calibration

Description:
  • Uses steady-state visually evoked potentials (SSVEPs) as endogenous reference signals for cross-session EEG calibration.
  • Creates a subject-specific linear mapping that aligns raw EEG recordings to a common neural baseline.
  • Reduces the impact of session dependent recording differences to support consistent longitudinal EEG analysis.

Abstract

This technology uses steady-state visually evoked potentials (SSVEPs) as calibration references for cross-session EEG normalization. Fixed-frequency visual stimuli (e.g., 10Hz, 15Hz) evoke stable responses used to create a linear mapping—via ordinary least squares or ridge regression—that calibrates raw EEG data to a common reference. Aligning to this baseline reduces inter-session variability, improves event-related potential (ERP) fidelity, and achieves up to 80% similarity in neural profiles, supporting reliable cognitive state decoding for clinical diagnostics and brain-computer interface (BCI) systems.​

Benefit

  • Improves EEG consistency without invasive methods or strict electrode placement​
  • Enables mental state decoding with ~80% reproducibility across sessions​
  • Reduces noise using endogenous calibration signals​
  • Supports scalable BCI deployment with minimal recalibration​

Other

  • Experimentally validated​
  • Available for exclusive and non-exclusive license​

Patent Information:

  • Title: Cross-Session Alignment of Neural Recordings Using Sensory Tasks
  • App Type: Utility
  • Country: United States
  • Serial No.: 18/946,641
  • Patent No.:  
  • File Date: 11/13/2024
  • Issued Date:  
  • Expire Date:  
  • Patent Status: Patent Pending