> Michael Persinger of Laurentian University has done some promising research
> by applying computer-controlled selenoid magnets to the heads of test
> subjects. He alters brain activity in the temporal lobe, and has managed to
> produce a variety of altered states, including hallucinations.
This is known as transcranial magnetic simulation. The head is surrounded by an array of coils (solenoids) which can be pulsed with very powerful current spikes, and induce local currents in the conductive (electrolyt) brain tissue, which interferes with normal neuroelectrical activity. The same technique works with muscles: a spike in a coil applied to muscle will cause it to twitch, obviously a minor variation on Galvani's classical frog experiments.
The utility of this technique is very limited, as it is primarily disruptive (though influence of complex timed firing patterns is as of yet uncharted). Much better spatial resolution can be achieved in toying around with pulse sequences in fMRI.
> Since the temporal lobe is arguably where the brain "creates" reality, I
> wonder how far way we are from achieving a "full-immersion" virtual
> environment by way of precise electromagnetic fields to the brain. How small
> would the 'trodes have to be--assuming it's possible to create an interactive
> hallucination without using intrusive surgery?
I don't think it's realistic to talk about true I/O without considering at least 10^5..10^7 electrode arrays, either wired very selectively or requiring lots of assistance (complex treatment to increase neuroplasticity and/or implantation very early in life).