The rat paper actually supports Mike's argument about He being worse than N2. The time to first EEG evidence of O2 toxicity was significantly shorter with the He experiments than the N2 ones.

That said, the issues in deep dives are more complex. One the one hand there are theoretical reasons why N2 should be protective against O2 seizures, narcosis works in a similar way to drugs - like benzodiazepines, which clearly do help with these fits. The rat study sort of showed this in the 'bimodal' response to N2 - implying some protection at inspired N2 at 2-6ATA.

On the other hand, N2 is much more dense than He so the higher the N2 the higher the work of breathing. High WOB produces more CO2 and makes CO2 elimination less effective. Narcosis also causes CO2 retention independantly of WOB. CO2 retention is almost certainly much more important than inert gas effects.

So whilst high He may be more likely (in and of itself) to be assosciated with CNS O2 toxicity, high N2 will cause more high CO2 which is a bigger problem.