Samantha Simons

Lempel-Ziv complexity of cortical activity during sleep and waking in rats

Understanding the dynamics of brain activity manifested in the EEG, local field potentials (LFP), and neuronal spiking is essential for explaining their underlying mechanisms and physiological significance. Much has been learned about sleep regulation using conventional EEG power spectrum, coherence, and period-amplitude analyses, which focus primarily on frequency and amplitude characteristics of the signals and on their spatio-temporal synchronicity. However, little is known about the effects of ongoing brain state or preceding sleep-wake history on the nonlinear dynamics of brain activity. Recent advances in developing novel mathematical approaches for investigating temporal structure of brain activity based on such measures, as Lempel-Ziv complexity (LZC) can provide insights that go beyond those obtained with conventional techniques of signal analysis. Here, we used extensive data sets obtained in spontaneously awake and sleeping adult male laboratory rats, as well as during and after sleep deprivation, to perform a detailed analysis of cortical LFP and neuronal activity with LZC approach. We found that activated brain states—waking and rapid eye movement (REM) sleep are characterized by higher LZC compared with non-rapid eye movement (NREM) sleep. Notably, LZC values derived from the LFP were especially low during early NREM sleep after sleep deprivation and toward the middle of individual NREM sleep episodes. We conclude that LZC is an important and yet largely unexplored measure with a high potential for investigating neurophysiological mechanisms of brain activity in health and disease.

Univariate and Multivariate Generalized Multiscale Entropy to Characterise EEG Signals in Alzheimer’s Disease

Alzheimer’s disease (AD) is a degenerative brain disorder leading to memory loss and changes in other cognitive abilities. The complexity of electroencephalogram (EEG) signals may help to characterise AD. To this end, we propose an extension of multiscale entropy based on variance (MSEσ2) to multichannel signals, termed multivariate MSEσ2 (mvMSEσ2), to take into account both the spatial and time domains of time series. Then, we investigate the mvMSEσ2 of EEGs at different frequency bands, including the broadband signals filtered between 1 and 40 Hz, θ, α, and β bands, and compare it with the previously-proposed multiscale entropy based on mean (MSEµ), multivariate MSEµ (mvMSEµ), and MSEσ2, to distinguish different kinds of dynamical properties of the spread and the mean in the signals. Results from 11 AD patients and 11 age-matched controls suggest that the presence of broadband activity of EEGs is required for a proper evaluation of complexity. MSEσ2 and mvMSEσ2 results, showing a loss of complexity in AD signals, led to smaller p-values in comparison with MSEµ and mvMSEµ ones, suggesting that the variance-based MSE and mvMSE can characterise changes in EEGs as a result of AD in a more detailed way. The p-values for the slope values of the mvMSE curves were smaller than for MSE at large scale factors, also showing the possible usefulness of multivariate techniques.

Classification of Alzheimer's disease from quadratic sample entropy of electroencephalogram

Currently accepted input parameter limitations in entropy-based, non-linear signal processing methods, for example, sample entropy (SampEn), may limit the information gathered from tested biological signals. The ability of quadratic sample entropy (QSE) to identify changes in electroencephalogram (EEG) signals of 11 patients with a diagnosis of Alzheimers disease (AD) and 11 age-matched, healthy controls is investigated. QSE measures signal regularity, where reduced QSE values indicate greater regularity. The presented method allows a greater range of QSE input parameters to produce reliable results than SampEn. QSE was lower in AD patients compared with controls with significant differences (p < 0.01) for different parameter combinations at electrodes P3, P4, O1 and O2. Subject- and epoch-based classifications were tested with leave-one-out linear discriminant analysis. The maximum diagnostic accuracy and area under the receiver operating characteristic curve were 77.27 and more than 80%, respectively, at many parameter and electrode combinations. Furthermore, QSE results across all r values were consistent, suggesting QSE is robust for a wider range of input parameters than SampEn. The best results were obtained with input parameters outside the acceptable range for SampEn, and can identify EEG changes between AD patients and controls. However, caution should be applied because of the small sample size.

Fuzzy Entropy Analysis of the Electroencephalogram in Patients with Alzheimer’s Disease: Is the Method Superior to Sample Entropy?

Alzheimer’s disease (AD) is the most prevalent form of dementia in the world, which is characterised by the loss of neurones and the build-up of plaques in the brain, causing progressive symptoms of memory loss and confusion. Although definite diagnosis is only possible by necropsy, differential diagnosis with other types of dementia is still needed. An electroencephalogram (EEG) is a cheap, portable, non-invasive method to record brain signals. Previous studies with non-linear signal processing methods have shown changes in the EEG due to AD, which is characterised reduced complexity and increased regularity. EEGs from 11 AD patients and 11 age-matched control subjects were analysed with Fuzzy Entropy (FuzzyEn), a non-linear method that was introduced as an improvement over the frequently used Approximate Entropy (ApEn) and Sample Entropy (SampEn) algorithms. AD patients had significantly lower FuzzyEn values than control subjects (p < 0.01) at electrodes T6, P3, P4, O1, and O2. Furthermore, when diagnostic accuracy was calculated using Receiver Operating Characteristic (ROC) curves, FuzzyEn outperformed both ApEn and SampEn, reaching a maximum accuracy of 86.36%. These results suggest that FuzzyEn could increase the insight into brain dysfunction in AD, providing potentially useful diagnostic information. However, results depend heavily on the input parameters that are used to compute FuzzyEn.

Distance-Based Lempel–Ziv Complexity for the Analysis of Electroencephalograms in Patients with Alzheimer’s Disease

The analysis of electroencephalograms (EEGs) of patients with Alzheimer’s disease (AD) could contribute to the diagnosis of this dementia. In this study, a new non-linear signal processing metric, distance-based Lempel–Ziv complexity (dLZC), is introduced to characterise changes between pairs of electrodes in EEGs in AD. When complexity in each signal arises from different sub-sequences, dLZC would be greater than when similar sub-sequences are present in each signal. EEGs from 11 AD patients and 11 age-matched control subjects were analysed. The dLZC values for AD patients were lower than for control subjects for most electrode pairs, with statistically significant differences (p < 0.01, Student’s t-test) in 17 electrode pairs in the distant left, local posterior left, and interhemispheric regions. Maximum diagnostic accuracies with leave-one-out cross-validation were 77.27% for subject-based classification and 78.25% for epoch-based classification. These findings suggest not only that EEGs from AD patients are less complex than those from controls, but also that the richness of the information contained in pairs of EEGs from patients is also lower than in age-matched controls. The analysis of EEGs in AD with dLZC may increase the insight into brain dysfunction, providing complementary information to that obtained with other complexity and synchrony methods.

Investigation of Alzheimer's Disease EEG Frequency Components with Lempel-Ziv Complexity

This pilot study applied Lempel-Ziv Complexity (LZC) to 22 resting EEG signals, collected using the 10-20 international system, from 11 patients with Alzheimer’s disease (AD) and 11 age-matched controls. This allowed for frequency band analysis as the EEG signals were first pre-filtered with a third order Hamming window in the ranges F to F+WHz with both F and W equal to 1-30Hz respectively. Control subjects were found to have a greater signal complexity than AD patients with statistically significant bands seen at various ranges in all 16 electrodes. The maximum statistical significance (Student’s t test, p<0.01) was increased over the findings with traditional signal filtering techniques allowing the whole range, with a maximum significance of p=3.50e− 6 at electrode T4 between 7-18Hz. Electrode F4 also showed significantly high statistically significant differences. The maximum accuracy, both controls and AD patients correctly identified, found with Receiver Operating Characteristic Curves was 95.45% (21 of 22 subjects correctly classified) at T4 (7-18Hz and 7-20Hz), Fp2 (8-32Hz) and F4 (6-21Hz), which is significantly more accurate than the most accurate methods previously applied to this dataset. The beta band (13-30Hz) was found to be most influential in separating the two test groups in this study with the best range suggested to be 5-26Hz, combining traditional theta, alpha and beta bands. These findings suggest pre-filtering has a significant effect on method outcomes and can be successfully tailored to improve the statistical effectiveness of LZC at distinguishing between these two EEG groups. However, more testing is required to investigate the effectiveness at distinguishing other signal dynamics.

Volume conduction effects on bivariate Lempel-Ziv Complexity of Alzheimer's disease electroencephalograms.

The spurious increase in coherence of electroencephalogram (EEG) signals between distant electrode points has long been understood to be due to volume conduction of the EEG signal. Reducing the volume conduction components of EEG recordings in pre-processing attenuates this. However, the effect of volume conduction on non-linear signal processing of EEG signals is yet to be fully described. This pilot study aimed to investigate the impact of volume conduction on results calculated with a distance based, bivariate form of Lempel-Ziv Complexity (dLZC) by analyzing EEG signals from Alzheimers disease (AD) patients and healthy age-matched controls with and without pre-processing with Current Source Density (CSD) transformation. Spurious statistically significant differences between AD patients and controls EEG signals seen without CSD pre-processing were not seen with CSD volume conduction mitigation. There was, however, overlap in the region of electrodes which were seen to hold this statistically significant information. These results show that, while previously published findings are still valid, volume conduction mitigation is required to ensure non-linear signal processing methods identify changes in signals only due to the purely local signal alone.

Permutation Entropy of the Electroencephalogram Background Activity in Alzheimerźs Disease

This pilot study applied Permutation Entropy (PE), a non-linear symbolic measure, and a novel modification (modPE), to investigate the regularity of electroencephalogram (EEG) signals from 11 Alzheimer’s disease (AD) patients and 11 age-matched controls given input parameters n (embedding vector), τ (coarse graining) and slide (difference between the start of two concurrent embedding vectors). PE discriminated better than modPE with controls showing reduced regularity over AD patients. Increasing τ identified the greatest differences between EEG signals. Longer embedding vectors were also more able to identify differences. The greatest difference between groups was at Fp1 with n,τ,slide = 3,10,1 (p=0.0112 Kruskal Wallis with Bonferroni). Subject and epoch based leave-one-out cross validation was carried out with thresholding from Receiver Operating Characteristic Curves. The greatest ability to correctly identify AD patients and controls were 81.82% (Fp2 n,τ,slide = 7,4,4, PE and modPE, F7 n,τ,slide = 3,10,1, PE and modPE) and 90.91% (Fp1 n,τ,slide = 3,10,1, PE and modPE), respectively. The maximum accuracy (both groups correctly identified) was 81.82% seen at many electrode and input combinations. All are with subject based analysis. This suggests that PE can identify changes in EEG signals in AD, given appropriate variables. However, modPE makes little improvement over PE.

Can Distance Measures Based on Lempel-Ziv Complexity Help in the Detection of Alzheimer’s Disease from Electroencephalograms?

This pilot study applied three distance-based bivariate Lempel-Ziv complexity (LZC) measures to investigate the changes in electroencephalogram (EEG) signals between 11 patients with Alzheimer’s disease (AD) and 11 age matched controls. These methods measure richness of complexity between pairs of signals. Complexity of control subjects’ EEGs was richer, i.e. signals were made from a greater number and greater range of subsequences, than those from AD patients in almost all cases in two non-normalized distance-based methods. Only some pairs including electrode T4 (2.1% of the total) occasionally showed the reverse result. Statistically significant differences were found with these two methods in 21 and 18 of 120 tested electrode pairs, respectively (Student’s t test, p<0.01). Receiver operating curves were used to calculate the sensitivity (number of correctly classified AD patients) and specificity (number of correctly classified controls). Accuracy is the combined correct classification of controls and AD patients. The maximum sensitivity found was 100%, specificity 90.9% and accuracy 86.4% at various electrode pairs with both non-normalized methods. The normalized method showed many electrode pairs with increased richness of complexity for AD patients than controls (67.5% of the total). It was found that this was due to the normalization procedure modifying the distribution of the original complexities from the electrode pairs. These findings suggest non-normalized distance-based bivariate LZC measures can be reliably applied to complex physiological signals such as human EEGs to further understand the effect of AD on the complexity of brain signals of patients. However, care must be taken when normalization procedures are applied.

Lempel-Ziv Complexity Analysis of Local Field Potentials in Different Vigilance States with Different Coarse-Graining Techniques

Analysis of electrophysiological signals recorded from the brain with Lempel-Ziv (LZ) complexity, a measure based on coarse-graining of the signal, can provide valuable insights into understanding brain activity. LZ complexity of local field potential signals recorded from the neocortex of 11 adult male Wistar-Kyoto rats in different vigilance states - waking, non-rapid-eye movement (NREM) and REM sleep - was estimated with different coarse-graining techniques (median, LZCm, and k-means, LZCkm). Furthermore, surrogate data were used to test the hypothesis that LZ complexity results reveal effects accounted for by temporal structure of the signal, rather than merely its frequency content. LZ complexity values were significantly lower in NREM sleep as compared to waking and REM sleep, for both real and surrogate signals. LZCkm and LZCm values were similar, although in NREM sleep the values deviated in some epochs, where signals also differed significantly in terms of temporal structure and spectral content. Thus, the interpretation of LZ complexity results should take into account the specific algorithm used to coarse-grain the signal. Moreover, the occurrence of high amplitude slow waves during NREM sleep determines LZ complexity to a large extent, but characteristics such as the temporal sequence of slow waves or cross-frequency interactions might also play a role.