Title: Imputation for longitudinal neuroimaging studies
Abstract:
The infant brain undergoes a remarkable period of neural development that is crucial for the development of cognitive and behavioral capacities. Longitudinal magnetic resonance imaging (MRI) is able to both characterize the developmental trajectories as well as predict later outcome such as cognitive development or the emergence of disorders like ASD and ADHD. However, missing data at different time points is an unavoidable occurrence in longitudinal studies owing to participant attrition and scan failure. Compared to dropping incomplete data, data imputation is considered a better solution to address such missing data in order to preserve all available samples. Here I will present recent research of my lab in deep learning based data imputation in developmental neuroimaging studies, from simple approaches like auto encoders to more advanced ones like perceptive generative adversarial networks (GAN). Our results show that effective data imputation is possible, and that the significantly improved data size leads to enhanced outcome in regression and classification tasks.
Posted by: Hong Xu