This credibility of this phenomenon is not based on prisoner testimony. It is based on observation of nearly 100% of prisoners, other people (an example is Terry Anderson, Iranian hostage), and test subjects, and has been observed and documented since the 1800's.
Well, it really is, because the symptoms are self-reported -- unless someone actually kills himself.
There isn't an objective way to verify "feelings of anxiety and nervousness, headaches and the like . . . nightmares, heart palpitations and fear of impending nervous breakdowns . . . hallucinations and . . . suicidal ideation." Now, I haven't spent time in a prison, but I suspect most prisoners wouldn't hesitate to lie if they might stand to gain something.
Long-term solitary confinement may not be the ideal, but what's the alternative?