From my clinical experience it seems that while some panic attacks symptoms can be quite overt and easily noticed (like sweating, pallor, shaking) these may also be mis-attributed by observers and many less obvious symptoms may not even be noticed. I'm wondering if anyone is aware of published research that has attempted to measure how effectively observers can gauge levels of anxiety and panic in someone just by watching them perform a task?