As is the case with many physiological functions being autonomic (circulatory, digestive, metabolism and most of respiratory), a majority of cognitive functions are also "automatic" and are performed outside the control of the conscious mind. In many cases this background processing is tasked with integrating various sensory inputs into a comprehensive sense of reality. Providing unnatural or unexpected inputs can confuse these processes. As pilots, a common experience with this phenomenon is the spatial disorientation experienced during unusual attitude training. The disorientation is the product of conflicting sensory inputs from the inner ear (balance), propriosensic (seat of the pants) and vision (instruments).  Vision usually dominates, and would in this case, except that the steam gauges and needles require additional processing  that degrades the "believability" of the visual information so the incorrect signals dominate.

Fast forward to the modern age of high fidelity glass. The synthetic images displayed require less processing and are therefore more believable to the autonomic cognitive systems. It is easy and natural to for your brain to "lock" onto the image as if it were reality and ignore conflicting information. This makes for safer IFR flight as it requires less brain processing bandwidth, leaving more for other flight tasks (communicating, navigating, monitoring, etc.)

The problem occurs  when the Glass is wrong. Even when the image on the screen does NOT represent reality then there is still an overwhelming visual input. Imagine that you were flying on instruments in VFR conditions. Your scan was on the gauges but your peripheral vision allowed you to see outside.  Now imagine what would happen if the view outside started to tumble while the gauges and your butt are telling you that you are still in level flight. Since the outside view is more compelling, the natural reaction would be to ignore the needles and attempt to follow the compelling visual picture. Now imagine the same scenario except that you are in IFR conditions and "compelling" visual image is from your high fidelity glass cockpit.

I can tell you from personal experience (development flight tests) that when the glass starts to roll, you really want to roll with it  and even when you have identified the that there is a problem and are using the standby gauges the "wrong" picture in your peripheral vision continues to be a distraction. My strong advice is that if the glass is in doubt, turn it off (power, dimming or cover it) and go to your standby gauges.

I was motivated to write this after reading the following account:
http://www.ntsb.gov/ntsb/brief.asp?ev_id=20050131X00119&key=1

There is no way to determine if a glass failure induced this accident and I would not want to speculate as to the cause. I was reminded by this account of the disorientation I experienced when the power supply in a new AHRS I was testing went TU and induced a 30 degree per second roll on the display. Reverting the PFD to Moving Map immediately "fixed" my disorientation as I was then able to use the standby attitude indicator without the conflicting rolling image on the PFD.  It was NOT enough to just "know" the glass was wrong and ignore it as my autonomic brain wanted to lock onto the more compelling image.

If you have incorrect information displayed, cover it or remove it from view. Ignoring it doesn't work.

Regards
Brent Regan