Using meta-analytic techniques, the authors surveyed the literature to quantify neuropsychological test sensitivities and determine deficit profiles in a number of dementing disorders and neuropsychiatric conditions. The review makes use of investigations that compare the performances of pathological and normal groups on commonly used neuropsychological tests and mood inventories. Compelling arguments are made for why "the effect size" associated with a given psychological test is the yardstick of interest rather than the more traditional probabilistic test of significance, since the former enables better appreciation for the degree of separation in test score distributions between pathological and normal groups, that is, the sensitivity of the test for detecting the diagnostic condition. A "heuristic benchmark criterion" of less than 5% overlap between the pathological and normal distribution scores is proposed to identify core features of a syndrome. Disorders reviewed are dementia of the Alzheimer type, frontotemporal dementia, primary progressive aphasia, progressive supranuclear palsy, Parkinson disease, Huntington disease, multiple sclerosis, major depressive disorder, obsessive-compulsive disorder, and mild traumatic brain injury. Neuropsychological test domains included attention/concentration, memory acquisition, delayed recall, cognitive flexibility and abstraction, language, visuospatial and visuomotor skills, and manual dexterity.
Neuropsychological Differential Diagnosis, vol 3. Arch Neurol. 2000;57(9):1375–1379. doi: