Cookies on this website

We use cookies to ensure that we give you the best experience on our website. If you click 'Accept all cookies' we'll assume that you are happy to receive all cookies and you won't see this message again. If you click 'Reject all non-essential cookies' only necessary cookies providing core functionality such as security, network management, and accessibility will be enabled. Click 'Find out more' for information on how to change your cookie settings.

Undersampling of functional MRI (fMRI) data leads to increased temporal resolution, as it allows shorter acquisition time per frame. High quality reconstruction of fMRI data from undersampled measurements requires proper modeling of the fMRI data. Recent publications suggest that the fMRI signal is a superposition of periodic and aperiodic signals. In this paper we develop an fMRI reconstruction approach based on this modeling. The fMRI data is assumed to be composed of two components: a component that holds a sum of periodic signals which is sparse in the temporal Fourier domain and an component that holds the remaining imaging information (consisting of the background and aperiodic signals) which has low rank. Data reconstruction is done by solving a constrained problem that enforces a fixed, moderate rank on one of the components, and a limited number of temporal frequencies on the other. Our approach is coined PEAR - PEriodic and ApeRiodic signal separation for fast fMRI. Experimental results are based on fMRI reconstruction using realistic timecourses. Evaluation was performed both quantitatively and visually versus ground truth. Results demonstrate PEAR's improvement in estimating the realistic timecourses versus state-of-the-art approaches at acceleration ratio of R=16.6.

Original publication




Journal article


Conf Proc IEEE Eng Med Biol Soc

Publication Date





505 - 508


Algorithms, Magnetic Resonance Imaging