Input to Output ghost elements

Dear friends, it is the strangest thing I have seen and the weirdest question to ask. Consider the following model:

The input is a matrix [t_in; x_in], where t_in is a vector of time stamps, and x_in = [0 1 0 1 1 1 1 0 0 1 0 0 1 1 1 1 0 1 0 0 0 0];

Reading from input.mat directly into output.mat you might expect getting just the same stuff, but hey, lets plot the input and the output together:

Wait a second. The output array has an extra '1' around the 17th element and then the whole thing is just shifted one element to the right.

So, the question is 'why', where did that '1' ghost come from?