[Geezer]

You could always try oxygen in a hyperbaric chamber.

[Lmnre]

I knew a physics teacher who tried to do the experiment with CO2, but the reflex response to this was so extreme that he immediately breathed it out again with such force that he was unable even to try speaking. I don't know if his account is true, though I see no reason to disbelieve it. Perhaps it could be done with a mixture of CO2 and air without going to the same extreme.

I worked in lab where we kept dry ice in a horizontal freezer. One day, I bent into the freezer to reach a piece at the bottom. I inhaled, but it did not result in a reflex response, although it smelled slightly metallic. I assumed the gas in the freezer was mostly CO₂ that sublimated from the dry ice.

[David Cooper]

How about trying it with a liquid? Many years ago a mouse was shown on Tomorrow's World swiming about in a container filled with a liquid which carried oxygen much more efficiently than water - the mouse could breathe by inhaling this liquid instead of air. Now, if you put a person in such a liquid, they could breathe it in safely (I hope, unless the mouse was slowly being poisoned) and they could try speaking too. I wonder if that mouse squeaked deeply while it swam...

[RD]

The youtube link now leads to: "This video is no longer available due to a copyright claim by NBC Universal."

something similar ...

[CliffordK]

Excellent film.
I liked the tin foil boat.

I'm surprised the host allowed the people to take more than a couple of breaths.

[Savo]

Sulfer Hexaflouride.

[Rob]

It can be done using Nitrous Oxide, Diflouroethane, Xenon, Sulfur Hexaflouride, or any inert gas that is heavier than air.  Many different gases will work to produce this effect, but some can kill you or cause other negative reactions within the body.  If you are going to attempt this, I would recommend using Xenon which is a safe, inert, noble gas, but still runs risks due to the gas density displacing oxygen in the lungs.  Sulfur hexafluoride is slightly easier to obtain from universities with strong electrical engineering or research departments, but you still run the same risk.