No, we don't see causation at the neurological level. The "correlation ≠ causation" meme doesn't just undermine overspeculative econometric conclusions: it undermines overspeculative biological conclusions too. The article may show correlations at the econometric level (I'm not convinced). It reports correlations at the medical level. But correlation still ≠ causation, and so it's not clear that lead causes crime "neurologically". Indeed it's not clear that any single factor trumps the human element in committing crime. At some level, crime is always a human choice. That's why we call it "crime." The overall fallacy of the article, therefore, is that statistical analysis—with its weak ability to establish causation—can shed more light on crime than human factors can, when crime is clearly a highly complex, volitional, and "human" behavior that resists analysis of any kind.