Elastic Registration of Single Subject Task Based fMRI Signals

Global Diffeomorphic Phase Alignment of Time-Series from Resting-State fMRI Data

We present a novel method for global diffeomorphic phase alignment of time-series data from resting-state functional magnetic resonance imaging (rsfMRI) signals. Additionally, we propose a multidimensional, continuous, invariant functional representation of brain time-series data and solve a general global cost function that brings both the temporal rotations and phase reparameterizations in alignment. We define a family of cost functions for spatiotemporal warping and compare time-series warps across them. This method achieves direct alignment of time-series, allows population analysis by aligning time-series activity across subjects and shows improved global correlation maps, as well as z-scores from independent component analysis (ICA), while showing new information exploited by phase alignment that was not previously recoverable.

Deep Learning of Warping Functions for Shape Analysis

Rate-invariant or reparameterization-invariant matching between functions and shapes of curves, respectively, is an important problem in computer vision and medical imaging. Often, the computational cost of matching using approaches such as dynamic time warping or dynamic programming is prohibitive for large datasets. Here, we propose a deep neural-network-based approach for learning the warping functions from training data consisting of a large number of optimal matches, and use it to predict optimal diffeomorphic warping functions. Results show prediction performance on a synthetic dataset of bump functions and two-dimensional curves from the ETH-80 dataset as well as a significant reduction in computational cost.