Figure 5

Assessment of Changes in Cell Density and Marker Expression following Injury Using tSNE. The machine-learning algorithm T-distributed Stochastic Neighbor Embedding (tSNE), which uses a nonlinear dimensionality reduction, was applied for visualization of the data. The islands created are determined by the algorithm contain events that share similar properties. The x- and y-axis and unites are determined by the algorithm (A) Presentation of total cell density changes comparing sham and injured with both contra-and ipsi-lateral hemispheres at 24 h after injury. The CCI ipsilateral shows a clear change with the formation of two distinct new islands. (B) Heat-maps of the mean values of the different markers used in the study are presented to visualize the changes the microglial population is undergoing at 24 h after injury. For comparison, the ipsilateral of sham and CCI are presented.