E-PHATE models the interactions between brain activation and exogenous information about participants using multi-view manifold learning. (IMAGE)
Caption
In this schematic, the first view of E-PHATE takes as inputs a vector of brain activation for each participant and computes a PHATE-based affinity matrix. The second view takes a vector of environment scores for each of those participants and builds an environment-based affinity matrix. These two views are combined into the E-PHATE diffusion matrix, which now captures both brain and environmental relations among participants and can be embedded into lower dimensions for visualization. Here, participants’ coordinates in E-PHATE dimensions visually reflect individual differences along externalizing problem scores.
Credit
Biological Psychiatry: Cognitive Neuroscience and Neuroimaging
Usage Restrictions
Please include attribution.
License
Original content