Navigation

Recent Publications

More...

Home

How does non-invasively measured neural activity reflect the learning process and ability?

ouyangg's picture

ouyangg - Posted on 21 November 2017

Project Description: 

Understanding and optimizing the process of learning constitute key subjects in educational research. Knowing the neural characteristics of fundamental learning processes and their relationships with higher order cognitive performance may bring us insights on learning optimization. We aim to examine the fast temporal neural dynamics of learning and the link between the neural characteristics of learning with general cognitive abilities, to seek potential neural feedback for learning optimization. We will employ high temporal-resolution brain electroencephalogram (EEG) recordings to investigate the neural underpinnings of fast time-scale learning processes during learning tasks. To systematically capture learning-related neural activities, we will utilize high performance computation power to extract neural dynamic information including component amplitude, latency, morphology, scalp distribution, spectrum, and time frequency across a wide range of parameters.

Researcher name: 
Guang Ouyang
Researcher position: 
Assistant Professor
Researcher department: 
Faculty of Education
Researcher email: 
ouyangg@hku.hk
Research Project Details
Project Duration: 
09/2017 to 09/2020
Project Significance: 
We will analyze high density brain EEG signals from 100 participants during basic learning tasks and extract the multi-dimensional dynamic information of neural activities at both the scalp sensor and cortical neural source levels. Upon the parameterization of the multi-dimensional neural activities, we will explore the relationship between the neural parameters and learning outcome to build the link between neural dynamics and learning performance. Furthermore, we will evaluate the reliability and validity of single trial-based neural traits during basic learning processes in predicting cognitive abilities (assessed by psychometric tasks) based on structural equation modeling. Linking neural correlates of learning with general cognitive abilities will frame the study into the context of education. The research paradigm will also prepare the foundation for further investigating the relationships between the neural underpinnings of other learning domains and intellectual abilities.
Remarks: 
HPC will supply the computation power to: 1. Conduct independent component analysis based on second-order blind identification (SOBI) on the high-resolution EEG data under different parameters settings (different filtering, different temporal and spatial sampling rate). 2. Estimate dynamic parameters from reinforcement learning data based on probabilistic programming.