Agreed, 747.
Tunnel vision, grey & black outs are normally caused by vertical acceleration. They are caused by lack of blood supply to the brain & can occur from 4 gn upwards, or 6 gn when supported by a "g-suite". Curiously, practise appears to improve tolerance. I'm told that tunnel vision & grey outs have happened at some "oval" circuits, notably Texas where drivers can experience 4 gn vertically, so that ties in with conventional wisdom. The fact that the Texas IRL race seems to have more than its share of late race incidents may well be significant.
I don't think much is known about tolerance to lateral acceleration, though tiredness must be a factor.
A few facts about accelerometers, if I may:
1. They do not measure acceleration. Most (that work down to DC) sense the displacement of a spring-retained mass.
2. An accelerometer fixed to the tub of an F1 race car is likely to experience +/-20gn engine-related "noise". This must be filtered electronically or mechanically (or, perhaps, both).
3. The "centre of gravity" of a race vehicle is difficult to define precisely. & it is probably even more difficult to attach an accelerometer at that location.
4. When a vehicle is manoeuvring close to its lateral limit, tyres often "stick/slip". This can cause lateral acceleration transients in excess of 1 gn.
For any or all of the above reasons, using quoted lateral gn figures to indicate the trajectory of a vehicle through a corner is usually misleading. I recall that somebody claimed to have recorded >6gn through Signes (Paul Ricard) "for several seconds". This would have implied a vehicle speed of 400 kph. Signes is a fast corner, but not that fast....