That doesn't seem right to me. (current staying the same) Shading does reduce the current and it also reduces the voltage a little. I have some numbers in my notes somewhere from when I was playing with a towel on parts of my panel.
Imp is just what you get when you divide input watts by the Vmp. I see it as a result not a cause. (Maybe that is backwards though) I was wondering where the input watts (VA) came from in the first place and how it is affected by shading one panel.
I don't really like the current bottleneck explanation, because Imp happens after the array, not amongst the array where a bottleneck might exist at one panel. I am thinking it is two different sets of current. After a while it will come clear as these things do, but there is no reason for you guys to get all tangled up.