Rxivist logo

SPARSE INFOMAX BASED ON HOYER PROJECTION AND ITS APPLICATION TO SIMULATED STRUCTURAL MRI AND SNP DATA

By Kuaikuai Duan, Rogers F. Silva, Jiayu Chen, Dongdong Lin, V. D. Calhoun, Jingyu Liu

Posted 07 Mar 2019
bioRxiv DOI: 10.1101/570317

Independent component analysis has been widely applied to brain imaging and genetic data analyses for its ability to identify interpretable latent sources. Nevertheless, leveraging source sparsity in a more granular way may further improve its ability to optimize the solution for certain data types. For this purpose, we propose a sparse information maximization (infomax) algorithm based on nonlinear Hoyer projection, leveraging both sparsity and statistical independence of latent sources. The proposed algorithm iteratively updates the unmixing matrix by infomax (for independence) and the sources by Hoyer projection (for sparsity), finally taking the sparse sources as the input data for the next iteration. Consequently, sparseness propagates effectively through infomax iterations, producing sources with more desirable properties. Simulation results on both brain imaging and genetic data demonstrate that the proposed algorithm yields improved pattern recovery, particularly under low signal-to-noise ratio conditions, as well as improved sparseness compared to traditional infomax.

Download data

  • Downloaded 224 times
  • Download rankings, all-time:
    • Site-wide: 101,416
    • In bioengineering: 1,949
  • Year to date:
    • Site-wide: 130,038
  • Since beginning of last month:
    • Site-wide: 130,011

Altmetric data


Downloads over time

Distribution of downloads per paper, site-wide


PanLingua

Sign up for the Rxivist weekly newsletter! (Click here for more details.)


News