NeurosciencebioRxivPreprint
Prediction of fMRI activity using vector autoregressive models: a comparison of sparse and low-rank approaches
Xinle Tian, Alex J. Gibberd, Sandipan Roy, Matthew A. Nunes
This paper compares sparse and low-rank approaches for predicting fMRI activity using VAR models, finding that low-rank pre-smoothing reduces prediction error and improves individual-level parameter estimation.
DOI 10.64898/2026.06.11.731556