The idea of disc dark matter is highly speculative, and requires a subdominant dark matter candidate with significant self-interactions (We know that most of the dark matter is in a roughly spherically symmetric halo).

Also, the correspondence between disc crossings and cratering looks tenuous at best...someone should check my work, but I plotted the tabulated impact pulses that the author associates with extinction events as a function of phase (assuming period = 30 Myr, t0 = now) at http://imgur.com/dfTkNEW . Running a Rayleigh statistic shows no evidence for non-uniformity (Rayleigh=0.16, N=13) and the mean phase is close to pi/2 too.

Okay, I have to ask, just how do we test his hypothesis?

I think the initial tests would be computer models. Take models of how the earth's core and mantle work (or the Oort cloud) and intersect that with models of how dark matter is spread through the galaxy. Are there inputs where the models behave as predicted, and are the inputs in any way realistic?

Wait 30 million years to cross the galactic disk again? /s

Maybe we could find a statistical difference between systems that are passing through the disk now and those that aren't.

Time to break out Betteridge's law again?

tome's law: In any discussion about an article whose title is a question, Betteridge's law is mentioned with probability 1.

Now that's a good law you got there, tome!

Yes--because it's the law.