Identification of peripheral neuropathy in type-2 diabetic subjects by static posturography and linear discriminant analysis

S. Fioretti, M. Scocco, L. Ladislao, G. Ghetti, R. A. Rabini

Research output: Contribution to journalArticlepeer-review


Background: An early diagnosis of peripheral neuropathy in diabetic patients is useful in order to slow down the progress of this complication. Nerve conduction tests are the gold standard for this diagnosis but they are challenging for the patients. This study examines whether it is possible to assess the presence of diabetic neuropathy at an early stage by static posturography tests. Methods: Static posturography tests were performed on 37 type-2 diabetic subjects (25 neuropathic patients and 12 non-neuropathic control subjects). Each subject was tested twice under two visual conditions: open and closed eyes. Both " global" (classic) and " structural" (model-based) posturographic parameters (PP) were derived from centre-of-pressure trajectories. A total of 65 PP were computed but only five were selected, normalized and fed to a linear classifier based on linear discriminant analysis. Results: This method correctly classified 86.5% of the patients. Five subjects were misclassified and only 2 false negatives out of 25 neuropathic subjects were erroneously diagnosed as control subjects. Conclusions: This paper shows that " global" and " structural" parameters derived by static posturography tests, and classic linear statistical approaches, can be used for the diagnosis of neuropathy provided PP are properly chosen and normalized.

Original languageEnglish
Pages (from-to)317-320
Number of pages4
JournalGait and Posture
Issue number3
Publication statusPublished - Jul 2010


  • Diabetic neuropathy
  • Static posturography
  • Statistical classification

ASJC Scopus subject areas

  • Biophysics
  • Orthopedics and Sports Medicine
  • Rehabilitation


Dive into the research topics of 'Identification of peripheral neuropathy in type-2 diabetic subjects by static posturography and linear discriminant analysis'. Together they form a unique fingerprint.

Cite this