Neuroscience is like:

• Trying to learn computer graphics by playing lots of video games (behaviour studies);
• Trying to explore the file system by looking at a computer with night-vision goggles (MRI);
• Trying to learn network architecture from a packet stream without knowing what a computer is (electrophysiology);
• Trying to learn CPU architecture by soaking a laptop in acid, slicing it with a band saw, and then studying the pieces under a microscope (connectomics);
• Trying to construct cohomology theories by using nothing but the category of sets (explanations of high-level cognition by studying raw neural circuits);
• etc.

And, despite all that, all of these things have led us, over the course of just a century, from zero to brain simulations.