Posturography in differential diagnosis of normal pressure hydrocephalus and brain atrophy

Differentiation between normal pressure hydrocephalus (NPH) and brain atrophy is difficult in clinical practice. The purpose of this paper was to apply two advanced statistical, pattern recognition methods: discriminant analysis (DA) and k-nearest neighbour (K-NN) for the classification of NPH and atrophy patients to approach computer aided differential diagnosis. The classification is based on a few measures of the center of foot pressure (COP) movements (radius, area, and length). The posturography method gives a measure of current postural stability by a quantitative evaluation of postural sways. Measurements have been performed in the standing upright position in two conditions: with eyes open (EO) and closed (EC). The study comprises 18 patients (mean age 64 ±13 years) diagnosed as normal pressure hydrocephalus and qualifying for shunt implantation. The patients were evaluated by static posturography twice: before and after surgery. The NPH patients were compared with 36 atrophy patients (mean age 64±13 years) and 47 healthy persons (mean age 60 ±7 years). There were two basic dissimilarities in the NPH patients before surgery in comparison with the other groups: very large sways and their independence from vision. Over 90% of the NPH cases both before and after surgery were correctly classified. There also were over 90% of correctly classified patients if we compared the before surgery NPH and atrophy patients. Further posturographic measurements and data collection are needed to verify these results.