In the context of a video the terms mean the following:
3D Low Pass Filtering:
The three dimensions (3D) are x (horizontal), y (vertical), and time. Hopefully you know what a low pass filter is.
I did not find this term in the paper. Given the context, though, I would guess that they mean for a video frame at time $t$, the temporal (time) neighborhood would be all data (i.e. frames) at nearby times. Thus, if the length your spatio-temporal filter along the time dimension was $T$, then the time neighborhood would be all frames from time $t-T/2$ to $t+T/2$.
Filters that filter in both space and time. In this case you have two dimensions of space- x and y. Think of your data as being a three-dimensional matrix, where one dimension of the matrix is the x index, one dimension is the y index, and one dimension is the time index. The time index would be the frame numbers. The x and y indices would be the pixel positions. The spatio-temporal filter would simply be a three-dimensional filter that would operate on your data.