Automatic Sleep Spindle Detection in Patients with Sleep Disorders

In this paper, we present a new automatic method for sleep spindle detection. It consist of a generalisation of the Schimicek's method [12] that takes more types of artefacts into account and uses variable thresholds regarding the statistical properties of the signal. Validity of our process is examined on the basis of visual spindle scoring performed by an expert. Results obtained are compared to those obtained by Schimicek's method. For a specificity of 90%, we obtain a sensitivity of 76.9% while Schimicek's method has a sensitivity of 70.4%. Moreover an increase of the area under the ROC curve is observed and confirms that the detection process is improved.