Springe direkt zu Inhalt

Boyan Rong

image_6483441

Neural Dynamics of Visual Cognition

PhD Candidate

Address
Habelschwerdter Allee 45
Room JK 25/223
14195 Berlin

I got my bachelor’s at Xidian University in China, majoring in electrical engineering. After that, I finished my Master’s at Columbia University in the USA with a focus on data analysis and machine learning. Then, my interest in neuroscience led me to join an 1-year internship at Peking University in China, where I was involved in MEG data collection, analysis and research using DNN models to study neural dynamics underlying global precedence in visual perception. I joined the lab in November 2021. During free time, I enjoy traveling around, watching movies, cooking and eating. I also like watching football as a fan of FC Barcelona.

General research interests

My main interest lies in how cognitive training changes visual representation in the human brain and how it can be related to deep neural networks, e.g. transfer learning. Currently, I am using both MEG and fMRI to study this topic combined with encoding models and RSA. I am collaborating with Prof. Dr. med. Emrah Düzel and Panagiotis Iliopoulos at Otto-von-Guericke-Universität Magdeburg in this main project, which is funded by DFG until 2024. 

Research projects

·      

Project 1:

In this project, we use a longitudinal design by training participants for 3 months on mnemonic discrimination. With pre- and post- training scanning with MEG and 7T-fMRI, we can investigate meso-scale underpinnings of neural resources underlying pattern separation and pattern completion in terms of both visual processing and memory. In particular, we will apply machine learning methods and representational similarity analysis to quantify the spatial (fMRI) and temporal (MEG) profile of object discriminability.

·     

Project 2:

Based on a huge dataset of human EEG responses to around 17000 naturalistic images collected by my colleague Alessandro Gifford, we are aimed to build an end-to-end encoding model based deep neural networks. We apply different pretrained convolutional models and also try to combine them with recurrent neural networks that process temporal dynamics.

Student supervision & opportunities 

Except for people mentioned above, I am also currently working with two research assistants at Otto-von-Guericke-Universität Magdeburg, Divya Sasidharan and Reshma Rethy, who are helping us to develop the website for our training sessions.

Free to email me if you’re interested in the projects mentioned above.