Moran's I using lisa function in ncf package

I'm using the lisa function in the ncf package to calculate spatial autocorrelation among elevation values.
I have a lidar (raster) layer, and have created a set of 100 random points (in ArcMap), so each point has a latitude, longitude, and elevation. I loaded these point data into R Studio.
I've successfully run the lisa function on the database with the 100 points, but I don't completely understand how the calculation works.
As I understand Local Moran's I, each sample point is compared to many other points (at varying distances) in the data, so each of the 100 points would be paired with 100 (or some number) of other points in the elevation raster.
In ArcMap, the clustering function compares each point to other values in the elevation raster.
In R, the lisa function seems to use just one dataset ( the 100 pts in this case), so there is no access to the rest of the raster data. The function seems to work, but I don't understand what data each of the 100 pts is compared to.