TY - GEN

T1 - Prediction of Anxiety Levels Based on Spatial-Frequency Patterns of EEG Activity During Perception of Another Person's Face

AU - Lozhnikov, Victor

N1 - Conference code: 26

PY - 2025/8/8

Y1 - 2025/8/8

N2 - This study proposes a method for predicting anxiety levels by analyzing the spatial-frequency characteristics of electroencephalogram (EEG) data. The experiment investigated spectral power density (PSD) shifts triggered by exposure to unfamiliar facial stimuli versus a neutral baseline (blank screen), with data collected from 61 healthy Russian and Chinese students. To mitigate electrode configuration dependencies, electroencephalogram data were normalized via interpolation onto a uniform grid. A regression model (Ridge, α=0.012) revealed a relationship between spatial-frequency patterns and State-Trait Anxiety Inventory (STAI) scores (MAE = 0.16, R2 = 0.3). Features associated with the beta frequency range (17-32 Hz) in the parietal and right temporal regions contributed most significantly to anxiety prediction. Although the model's statistical reliability is insufficient to draw definitive conclusions about anxiety levels, it identifies electroencephalogram biomarkers (beta-band oscillations) and cortical regions linked to heightened anxiety. These findings offer insights into neurophysiological mechanisms underlying anxiety and potential pathways for anxiolytic interventions.

AB - This study proposes a method for predicting anxiety levels by analyzing the spatial-frequency characteristics of electroencephalogram (EEG) data. The experiment investigated spectral power density (PSD) shifts triggered by exposure to unfamiliar facial stimuli versus a neutral baseline (blank screen), with data collected from 61 healthy Russian and Chinese students. To mitigate electrode configuration dependencies, electroencephalogram data were normalized via interpolation onto a uniform grid. A regression model (Ridge, α=0.012) revealed a relationship between spatial-frequency patterns and State-Trait Anxiety Inventory (STAI) scores (MAE = 0.16, R2 = 0.3). Features associated with the beta frequency range (17-32 Hz) in the parietal and right temporal regions contributed most significantly to anxiety prediction. Although the model's statistical reliability is insufficient to draw definitive conclusions about anxiety levels, it identifies electroencephalogram biomarkers (beta-band oscillations) and cortical regions linked to heightened anxiety. These findings offer insights into neurophysiological mechanisms underlying anxiety and potential pathways for anxiolytic interventions.

KW - anxiety disorder

KW - eeg

KW - eeg signal processing

KW - machine learning

KW - regression

KW - spectral analysis

UR - https://www.scopus.com/pages/publications/105014203073

UR - https://www.mendeley.com/catalogue/78c8584d-9842-3277-b2c7-49afa13586c7/

U2 - 10.1109/EDM65517.2025.11096655

DO - 10.1109/EDM65517.2025.11096655

M3 - Conference contribution

SN - 9781665477376

T3 - International Conference of Young Specialists on Micro/Nanotechnologies and Electron Devices, EDM

SP - 1850

EP - 1853

BT - International Conference of Young Specialists on Micro/Nanotechnologies and Electron Devices, EDM

PB - IEEE Computer Society

T2 - 2025 IEEE 26th International Conference of Young Professionals in Electron Devices and Materials (EDM)

Y2 - 27 June 2025 through 1 July 2025

ER -