A new automatic method for the detection of limb movements and the analysis of their periodicity. (April 2016)

• Record Type: Journal Article
• Authors: Alvarez-Estevez, Diego
• Is Part Of: Biomedical signal processing and control. Volume 26(2016)
• Page Start: 117
• Page End: 125
• Publication Date: 2016-04
• Subjects: Signal processing -- Periodicals; Biomedical engineering -- Periodicals; Signal Processing, Computer-Assisted -- Periodicals; Image Processing, Computer-Assisted -- Periodicals; Biomedical Engineering -- Periodicals; 610.28
• DOI: 10.1016/j.bspc.2016.01.008
• Languages: English
• ISSNs: 1746-8094

Abstract

Highlights: A new algorithm for the automatic detection of (periodic) limb movements. The algorithm offers robust detection capabilities in the presence of artefacts. Extensive validation including both ambulatory and clinical recordings. Good performance results are achieved against human experts scorings. The detector beats the performance of other state-of-the-art automatic approaches. Abstract: We present a new algorithm for the automatic detection of periodic and non-periodic limb movements in polysomnographic (PSG) sleep recordings. A set of 70 PSG recordings obtained in the course of common practice were randomly selected for the validation of the proposed approach. The dataset includes 35 recordings that were acquired in ambulatory conditions and 35 that were carried out under the supervision of clinicians at our sleep centre. The algorithm includes robust mechanisms to handle the presence of artefacts, and has the ability to adjust its detection thresholds to dynamically adapt to changing signal conditions. The validation results in our dataset, which also include the comparison with another two automatic methods available in the literature, support the validity of our approach, and its utility as a valuable tool to help the clinician in the scoring task.