Time Series Generative Learning with Application to Brain Imaging Analysis

TL;DR

This paper introduces a novel deep learning framework for generating realistic brain imaging time series data, aiding in understanding brain aging and improving disease detection through data augmentation.

Contribution

It proposes a nonparametric time series generator based on an f-divergence minimax formulation, with theoretical convergence guarantees and extension to panel data scenarios.

Findings

Generated brain MRI sequences resemble real data.

Data augmentation improves Alzheimer's disease detection accuracy.

Method effectively models sequential brain imaging data.

Abstract

This paper focuses on the analysis of sequential image data, particularly brain imaging data such as MRI, fMRI, CT, with the motivation of understanding the brain aging process and neurodegenerative diseases. To achieve this goal, we investigate image generation in a time series context. Specifically, we formulate a min-max problem derived from the $f$-divergence between neighboring pairs to learn a time series generator in a nonparametric manner. The generator enables us to generate future images by transforming prior lag-k observations and a random vector from a reference distribution. With a deep neural network learned generator, we prove that the joint distribution of the generated sequence converges to the latent truth under a Markov and a conditional invariance condition. Furthermore, we extend our generation mechanism to a panel data scenario to accommodate multiple samples. The effectiveness of our mechanism is evaluated by generating real brain MRI sequences from the Alzheimer's Disease Neuroimaging Initiative. These generated image sequences can be used as data augmentation to enhance the performance of further downstream tasks, such as Alzheimer's disease detection.