Power spectral distribution analysis for detection of freezing of gait in patients with Parkinson's disease

Currently it is estimated that in Europe there are between 120.000 and 200.000 new cases of Parkinson's Disease each year and around 2% of the population over 65 is affected. Freezing of gait (FOG) is a common and disabling symptom of Parkinson's disease. During FOG hastening of gait occurs and the person loses the control of gait. It is an episodic and non predictable phenomenon that affects most commonly the gait in the form of start hesitation, sudden immobility and hastening or festination. This often results in falls and consequent injuries reducing quality of life. A pilot study has been carried out with 4 patients to see if FOG can be reliably detected during daily activity using non invasive sensors such as accelerometers and gyroscopes. This can be done monitoring the power spectral distribution (PSD) and to do so, the short-time Fourier transform (STFT) was used. A multivariate linear discriminant analysis was applied to the extracted features of the PSD to classify each frame of time into normal gait or freezing gait. The extracted PSD features can be used to determine the actual walking state (FOG presence) for a subject but changes significantly from subject to subject.