Multiscale Lempel–Ziv complexity for EEG measures

Classification of gait rhythm signals between patients with neuro-degenerative diseases and normal subjects: Experiments with statistical features and different classification models

“…In current research, it also behaved very well when representing the variability of the gait time series. As for the Lempel-Ziv complexity feature, it seemed that it did not play a very important role as expected, even though it has had considerable success in other applications such as machine health monitoring [17] and biomedical signal interpretation [41].…”

Combined nonlinear metrics to evaluate spontaneous EEG recordings from chronic spinal cord injury in a rat model: a pilot study

“…Kolmogorov complexity is conceptually different from entropy measures, as it measures the amount of information by description length (size) (Teixeira et al 2011). In recent years, Kolmogorov complexity has been widely used in biomedical applications to estimate the temporal complexity of neurophysiological time series (Gao and Hu 2011;Gao et al 2013;Ibáñez-Molina et al 2015;Xiong et al 2013). A time series with complex temporal patterns tends to have a large value of Kolmogorov complexity.…”

Inspecting temporal scales with non-linear signal features: A way to extract more information from brain activity?

“…The article by Ibáñez-Molina et al (2015) in this issue of Clinical Neurophysiology represents one of the latest steps in this quest. Ibáñez-Molina et al (2015) propose an elegant modification of a widespread non-linear signal feature -the Lempel-Ziv Complexity (LZC, Lempel and Ziv, 1976) -to consider multiple temporal scales related to specific frequencies. The resulting multiscale LZC, hereafter mLZC, is able to characterise the EEG signals better by capturing the complexity of both fast and slow rhythms (Ibáñez-Molina et al, 2015).…”

“…However, the same binarisation process that lies at the heart of those advantages of the traditional LZC is also responsible for the limitation discussed by Ibáñez-Molina et al (2015). Ideally, the binary sequence should reflect the non-linear behaviour of the original signal.…”