Most of the multi-tasks performed usually links to different parts of the brain. However, if you try to perform a multi-task that calls two regions of the brain in the same area, the communication becomes confusing and thus that effect.
We used fMRI to compare whether amplitude and directional interference during bimanual actions activate the same or specialized neural circuits. [...] In general, our data thus support the view that even though the encoding of amplitude and directional information converge at one point and activate the same neural substrate, additional independent mechanisms are involved in bimanual amplitude as compared to that in direction control. article
The most straightforward interpretation of the bottleneck in response selection is that the brain contains a single piece of machinery, akin to the central processing unit of a computer, that takes a single retrieval cue (the output of the process of perceiving a stimulus) and looks up in its memory the corresponding action to be performed. [...] If one imagines the brain working something like a digital computer or a clerk in a reference library, one is confronted with a paradox. The system appears able to take in two inputs and use them simultaneously to choose a single appropriate response from memory, but it cannot take two stimuli and use them simultaneously to choose two different responses. article