Localized semi-nonnegative matrix factorization (LocaNMF) of widefield calcium imaging data

• DOI: 10.1371/journal.pcbi.1007791
• Publication Date: 2020-04-13T17:35:26Z
• Authors: Shreya Saxena; Ian Kinsella; Simon Musall; Sharon H. Kim; Jozsef Meszaros; David N. Thibodeaux; Carla Kim; John Cunningham; Elizabeth M. C. H...; Anne Churchland; Liam Paninski
• Keywords: Non-negative Matrix Factorization; Calcium imaging

Widefield calcium imaging enables recording of large-scale neural activity across the mouse dorsal cortex. In order to examine relationship these signals resulting behavior, it is critical demix recordings into meaningful spatial and temporal components that can be mapped onto well-defined brain regions. However, no current tools satisfactorily extract different regions in individual mice a data-driven manner, while taking account mouse-specific preparation-specific differences. Here, we introduce Localized semi-Nonnegative Matrix Factorization (LocaNMF), method efficiently decomposes widefield video data allows us directly compare multiple by outputting localized functional are significantly more interpretable than traditional decomposition techniques. Moreover, provides natural subspace correlation maps dynamics behaviors, mice, experimental conditions, identification task- movement-related