A5044706165- Neural dynamics and brain function
- EEG and Brain-Computer Interfaces
- Visual perception and processing mechanisms
- Advanced Algorithms and Applications
- Functional Brain Connectivity Studies
- Advanced Sensor and Control Systems
- Advanced Memory and Neural Computing
- Advanced MIMO Systems Optimization
- Optical Wireless Communication Technologies
- Millimeter-Wave Propagation and Modeling
- Energy Harvesting in Wireless Networks
- Wireless Power Transfer Systems
- Advanced Wireless Communication Techniques
University of Southern California
2024
Institute for Basic Science
2019-2023
Samsung (South Korea)
2022
Sungkyunkwan University
2019-2020
To handle the scarcity and heterogeneity of electroencephalography (EEG) data for Brain-Computer Interface (BCI) tasks, to harness power large publicly available sets, we propose Neuro-GPT, a foundation model consisting an EEG encoder GPT model. The is pre-trained on large-scale set using self-supervised task that learns how reconstruct masked segments. We then fine-tune Motor Imagery Classification validate its performance in low-data regime (9 subjects). Our experiments demonstrate...
When we interact with our environment, there is often a significant amount of variations in behavioral responses to incoming sensory input even when inputs are identical. Variations sensory-motor behavior can be caused by several factors, including changes cognitive status and intrinsic neural the brain. The correct identification sources important for understanding underlying mechanisms practical applications (e.g., development precise brain-computer interface). However, studies on humans...
Abstract Humans integrate multiple sources of information for action-taking, using the reliability each source to allocate weight data. This reliability-weighted integration is a crucial property Bayesian inference. In this study, participants were asked perform smooth pursuit eye movement task in which we independently manipulated target motion and direction-of-motion cue. Through an analysis initiation multivariate electroencephalography activity, found neural behavioral evidence...
Abstract Visually-guided smooth pursuit eye movements are composed of initial open-loop and later steady-state periods. Feedforward sensory information dominates the motor behavior during pursuit, a more complex feedback loop regulates pursuit. To understand neural representations motion direction pursuits, we recorded electroencephalography (EEG) responses from human observers while they tracked random dot kinematograms as targets. We estimated population tuning curves multivariate EEG...
This paper presents a new physical random access channel (PRACH) preamble design suitable for the next-generation 6G cellular communication systems operating in sub-Terahertz (THz) bands. In these extremely high frequency channels, subcarrier spacing (SCS) of orthogonal frequency-division multiplexing (OFDM) based waveform should be significantly increased at least to MHz-level mitigate severe effect phase noise, thus resulting OFDM symbol length several hundreds nano-seconds. Under this...