Lehrer repeated this warning about the limitations of fMRI in later stories. And yet, both in How We Decide and Imagine, fMRI is Lehrer’s deus ex machina. No supermarket decision or sneaker logo or song lyric is conceived without “lighting up” a telltale region of the brain. Here comes the anterior cingulate cortex, and there goes the superior temporal sulcus, and now the amygdala has its say. It reads like a symphony—magical, authoritative, deeply true.

This contradiction was pointed out back in March in a critique of Imagine published at the literary site the Millions by Tim Requarth and Meehan Crist. “It’s baffling that in Imagine Lehrer makes statements so similar to ones he thoroughly discredits” elsewhere. Then they offer an analogy to explain what’s wrong with drawing vast conclusions from pretty fMRI pictures. “Brain regions, like houses, have many functions,” they write, and just because there are people at someone’s house doesn’t mean you know what they’re doing. “While you can conclude that a party means there will be people,” they write, “you cannot conclude that people means a party.”

Rebecca Goldin, a mathematician-­writer who often criticizes “neurobabble,” points out that this is exactly what’s so enticing about this brand-new science: its mystery. Imagine that fMRI is a primitive telescope, and those clumps of neurons are like all the beautiful stars you can finally see up close, but “may in fact be in different galaxies.” You still can’t discern precisely how they’re interacting. Journalist David Dobbs recently asked a table full of neuroscientists: “Of what we need to know to fully understand the brain, what percentage do we know now?” They all gave figures in the single digits. Imagine makes it look like we’re halfway there.

If Lehrer was misusing science, why didn’t more scientists speak up? When I reached out to them, a couple did complain to me, but many responded with shrugs. They didn’t expect anything better. Mark Beeman, who questioned that “needle in the haystack” quote, was fairly typical: Lehrer’s simplifications were “nothing that hasn’t happened to me in many other newspaper stories.”

Even scientists who’ve learned to write for a broad audience can be fatalistic about the endeavor. Kahneman had a surprise best seller in 2011, Thinking, Fast and Slow. His writing is dense and subtle, as complicated as pop science gets. But as he once told Dan Ariely, his former acolyte, “There’s no way to write a science book well. If you write it for a general audience and you are successful, your academic colleagues will hate you, and if you write it for academics, nobody would want to read it.”

For a long time, Lehrer avoided the dilemma by assuming it didn’t apply to him, writing not for the scientists (who shrugged off his oversimplifications) or for the editors (who fixed his most obvious errors) but for a large and hungry audience of readers. We only wanted one thing from Jonah Lehrer: a story. He told it so well that we forgave him almost ­everything.

Related:

• Archive: “Features”
• Articles by Boris Kachka
• Table of Contents: Nov 5, 2012 issue of New York | Subscribe!