Reconstructing controllable faces from brain activity with hierarchical multiview representations

Ziqi Ren, Jie Li, Xuetong Xue, Xin Li, Fan Yang, Zhicheng Jiao, Xinbo Gao

Abstract: Highlights•A style-oriented decoding framework is proposed to reconstruct perceived faces from brain activity.•Facial information captured through a multi-regression model can be disentangled by feature hierarchy.•Reconstructions of different faces can be distinguished more accurately than the state-of-the-art.•Feature editing can be performed in fMRI domain for the first time to control human face attributes.