Anavex Life Sciences Corp (AVXL)

[XenaLives]

I'm not going to pretend to understand all of this, but clearly EEG is being used heavily to study Alzheimer's and new techniques are being used. It is also obvious that Anavex isn't just testing a drug, but doing ground breaking research in Alzheimer's disease and how it develops:


https://anavex.com/wp-content/uploads/New-Exploratory-Alzheimers-Drug-Anavex-2-73-Changes-in.pdf

3.2. Non-linear EEG studies in AD

The basic concept of non-linear time series analysis of EEG rests on the assumption that the brain is a non-linear dynamical system thus its behavior follows the principles of chaos theory (71). EEG captures the large-scale spatio-temporal dynamics of electromagnetic fields in the brain which are thought to be generated by non-linear coupling interactions between different neuronal populations. According to Stam the trajectory of a non-linear dynamical system is determined by its initial state and the history of its evolution. It can be argued that such a system possesses memory since its current state is dependent on previous states (62). The first studies on non-linear dynamical properties of EEG have been published in the 80’s and focused on primate spontaneous EEG and human sleep EEG (72). In the following years, a revolution took place in the field of non-linear EEG analysis focusing on modeling non-linear neuronal dynamics and on the development of new methods with the aim to analyze noisy, non-stationary and high-dimensional EEG. Numerous methods have been developed including non-linear forecasting, non-linear cross prediction, and dimension density; many of the novel methods relied on embedding the EEG time series in a multidimensional state-space and observing different properties of resulting trajectories. For an exhaustive review see Stam (62).

It is well known that the pathological changes in AD affect not only specific brain regions but neural pathways between the major areas as well (73). Non-linear methods are commonly used in the research of AD comparing it to other psychiatric conditions, because of the widely accepted disconnection syndrome theory (59).

Correlation dimension (D2) is one of the most commonly used non-linear EEG parameter in AD research. D2is a measure of the assumed independent variables that are required to precisely define the complexity of cortical dynamics reflected in the EEG signal. An early study with 21 AD and 29 MCI patients has shown that AD patients have lower D2values almost on every EEG channel suggesting a globally reduced complexity of the electric brain activity in AD (74). Normal elderly have significantly increased D2values in eyes-open state compare to eyes-closed state, while the difference disappears in AD, suggesting the loss of brain reactivity to external stimuli in AD (75). Further studies pointed out decreased complexity even in eyes-closed state in 15 AD patients compared to normal controls (76). These studies aimed to characterize the EEG in a global frequency range; however, the separate frequency bands are thought to represent different brain dynamic systems (26). Numerous studies investigated the complexity of EEG in different frequency ranges or multiple time scales. They identified higher D2at higher frequency ranges in 20 AD (77) and 17 AD patients (78), while lower values have been demonstrated in the low frequencies. Yosimura and colleagues investigated mild AD patients with the so-called omega complexity method and found high complexity in a wide range of frequencies similarly to the findings of Czigler and colleagues involving 12 AD patients (78-79). Omega complexity estimates the number of the independent, uncorrelated brain sources. Thus, these studies suggest the functional disintegration of cortical networks, with lower coherence between sources and a higher number of independent generators. A decrease in the parameter that reflects flexibility of information processing, the so-called first positive Lyupanov exponent (L1),has been also reported in some studies (27), which may represent a difficulty to enter different states from the initial one during information processing in AD.

Some new methods, such as multiscale entropy (MSE) have been implemented in the recent years enabling to investigate the complexity of dynamic biological signals across a long-range of temporal scales. These studies in 11 AD patients (80) as well as 26 AD and 22 MCI patients (81) found that AD patients had less complexity at smaller temporal scales related to higher frequency bands and higher complexity at larger temporal scales regarding lower frequency bands (80-81). Higher complexity in large temporal scales was strongly associated to the extent of cognitive decline. Studies of global range complexity yielded diverging results. However, different frequency bands correspond to different brain functions (82). While higher-frequency bands are thought to support local neuronal communication in smaller neuron groups with short-range neural connectivity, slower oscillations likely arise from larger populations within wider-range networks (83). Although finding a physiological interpretation of complexity measures obtained with methods motivated by dynamical system theory is not straightforward, these results are in line with the notion that AD is characterized by a disruption of integration and segregation within distributed brain networks (61).

Complexity measurements have shown a strong correlation also with results of neuropsychological tests. Lower global dimensional complexity was associated with lower MMSE scores in 21 AD patients (74). The estimated severity of the disease also showed a strong relationship to non-linear measures in the study of Besthorn involving 50 AD patients (84). Region-specific changes were observed in the study of Ikawa et al., where a strong association was found between the reduced dynamical complexity (DC) in the mid-temporal, left frontal, central areas and cognitive status in 25 AD patients (85). In the same study a correlation has been found also between verbal memory performance and DC values at post-temporal, left-central and parietal regions. Interestingly, in the early phases of AD, increased predictability and reduced complexity are predominantly visible in frontal and temporal areas (76).

Using the relatively new measures of mutual information (MI) and synchronization likelihood (SL), it has been demonstrated that the interdependency between the distant electrodes, especially between the frontal and temporal regions, between the frontal and parietal and between the hemispheres are reduced (26). Significant decrease in the SL of beta and high alpha band was also indicated in the study of Stam with 24 AD patients, where SL of both frequency bands had a strong correlation to the MMSE scores (86), while the gamma band was not affected. Babiloni and his colleagues obtained similar results in MCI; namely they found reduced SL in the delta band in 109 AD and 88 MCI patients with fronto-parietal dominance (87). Phase lag index (PLI) measurements revealed frequency dependent results in 18 AD patients; in the theta band patients showed higher whole-brain PLI and in the alpha band lower whole-brain PLI compared to patients with subjective cognitive decline (88). Concluded minimum spanning trees (MSTs) analysis from PLI suggested that global efficiency loss was defined mostly by the parietal and occipital loss of network organization (89).

Overall, results suggest that earlier alterations in neural dynamics can be identified with spectral analysis methods which can identify more precisely the early stages of Alzheimer (75). However, the combination of spectral and state-space methods proved to be highly sensitive in the early recognition of cognitive decline (62).

New methods derived from the mathematical framework of graph theory have recently become popular in EEG analysis. Graph theory is used to describe network connectivity from EEG data. Nodes in the network often correspond to electrodes, and some measures of connectivity between electrode pairs are used to characterize edges (90). The nodes which receive more inputs and have more connections are referred to as hubs. Human cognition seems to be strongly related to the efficacy of the integration of different brain nodes. The normal human large-scale functional connectivity network can be described as a small-world network, with numerous local connections between adjacent nodes and a few but prominent connections between distant regions, which together lend high clustering and short path lengths to the network (91). Recent studies revealed that small-world like network properties of are replaced by random-world characteristics in AD. Studies demonstrated that AD patients (n=18) have a prominent decrease in the characteristic path length, degree correlation and mean clustering coefficient of delta and gamma band activity indicating severe loss in the local and global connectivity parameters (86, 90). These findings were corroborated by MEG and fMRI studies as well including 18 (92), 20 (93), 21 (94) and 14 AD patients (95).

In summary, current EEG network analysis methods provide a promising new strategy for the diagnosis of AD and studying disease progression, and open a fresh view to cognitive decline from a network based perspective. State-space approaches seem to be also promising tools for the early recognition of MCI, especially when combined with spectral methods (Table 2).


3.3. Sleep-EEG studies in AD

Major neurocognitive disorders including Alzheimer’s disease as well as MCI are characterized with disturbed sleep. Growing evidence suggests thatsleepdisturbances precede the clinical onset ofcognitive decline in AD by years. A sleep-wake disturbance of clinical importance is found in up to half of the patients with dementia, and sun-down agitation is a frequent cause of institutionalization of demented patients (96). The circadian rhythm of dementia patients is disturbed with daytimesleepinessand disrupted nightsleep. Whereassleep changes may be severe in several types of dementia, a clinically significant sleep disorder usually develops only in the late phase ofAlzheimer's disease (97-98).

Sleep transforms in many ways during healthy aging. Changes include reduction in the proportion of slow wave sleep (SWS) and sleep slow wave activity (SWA), the number and amplitude ofsleepspindles as well as the density of rapid eye movements (REM) and the amplitude of circadian rhythms. With MCI there are further reductions of these parameters and all deteriorate further with the conversion to AD (99). Progressive changes inthe quality, architecture and neural regulation of sleep may contribute to cognitive decline (100). The sleep profiles of patients with dementia of diffuse Lewy body and AD are different; which may have diagnostic importance (101).

Sleep pathology appears to be essential component ofAD pathophysiology. Noticeably, sleep deprivation compromises cognitive and executive functions such as memory, attention, and response inhibition. In mice studies sleepdeprivation impaired the mice’s long-term and remote memory, even a month after the sleep deprivation session (102). The negative impact of sleep loss on memory has also been shown in patients with sleep disorders. Impairments of sleep-dependent memory consolidation for verbal and visual declarative information were found in patients with primary insomnia, for verbal declarative information in patients with obstructivesleepapnea, and for visual procedural skills in patients with narcolepsy-cataplexy (103). Conversely, increasing the amount of SWS and SWA by an anti-inflammatory agent in healthy humans improved memory consolidation (104). The positive cognitive effect of sleep slow waves was also shown in a transcranial slow oscillatory stimulation study where stimulation was applied duringthe afternoon naps of elderly individuals. The transcranial stimulation considerably increased frontal SWA significantly improving visualmemoryretention aftersleep, but not retention in the locationmemorysubtask and in the verbalmemorytask (105).

Changes in sleep microstructure have been observed as well in AD. Several studies have identified two types of sleep spindles: fast (13-15?Hz) centro-parietal and slow (11-13?Hz) frontal spindles. AD and MCI patients have shown a significant decrease in parietal fastspindledensity, which positively correlated with their loss in MMSE scores (106). Fast spindles are involved memory consolidation as enhancingsleepspindles with non-invasive brain stimulation in humans was found to significantly improve motor memory consolidation which correlated with the stimulation-induced increase of fast spindle activity (107).The impact of both slow and fast spindles on re-presentation learning has been shown in an odor re-exposure experiment (108).

Increased cerebrospinal fluid orexin levels were found causing sleep deterioration, which appeared to be associated with cognitive decline (109). The orexinergic system may be dysregulated in AD, while the role of orexin in memory function has remained controversial (110). Mice-experiments have shown an association of the amyloid-beta (Aß) peptide and mitochondrial dysfunction, supporting a primary role for mitochondrial Aß in AD pathology. Mitochondrial Aß peptide levels were strongly negatively associated to the scores of cognitive tasks in an AD transgenic mice model experiment, indicating that amyloid could compromise the cognitive functions via the altered the mitochondrial signaling system. The degree of cognitive impairment in AD transgenic mice can be linked to the extent of synaptic mitochondrial dysfunction and mitochondrial Aß peptide levels, suggesting that a mitochondrial signaling cascade induced by Aß may contribute to cognitive impairment (111). Further, chronic sleep deprivation caused mitochondrial dysfunction in the frontal cortex of mice and a significant mitochondria-related Aß increase in this cortical region suggesting that chronic sleep deprivation-induced mitochondrial dysfunction might be related to frontal mitochondria-related Aß accumulation, preceding Aß deposition in any other frontal cortical regions (112). These findings and the strong link of sleep-disruption in AD indicate a strong link between AD and sleep (113). It has been shown that insufficientsleepfacilitates the accumulation of Aß (114), potentially triggering earlier cognitive decline and conversion to AD (115). The sleep-wake cycle influences brain Aß levels, and sleep deprivation increases the concentration of soluble Aß leading to its accumulation, whereas sleep extension has the opposite effect (116). Furthermore, Aß accumulation leads to increased wakefulness. Individuals with still normal cognitive functions and early Aß deposition, report sleep abnormalities, similarly to mild dementia patients with incipient AD. Thus, sleep deprivation and AD may mutually amplify each other (117-118).

FMRI results suggest that the mechanism of the decline in short-termmemoryobserved after acutesleeprestriction is linked to the disruption of hippocampal-cortical connectivity (119). Age-related medial prefrontal cortex grey matter atrophy was associated with reduced non-REM SWA in older adults, correlating with the degree of the impairment of sleep-dependent memory retention. This memory impairment was associated with persistent hippocampal activation and reduced task-related hippocampal-prefrontal cortex functional connectivity, potentially causing compromised hippocampal-neocortical memory consolidation. Thus it seems that the age-related medial prefrontal cortex atrophy diminishes SWA, resulting in impaired long-term memory (120). In a human and monkey neocortical microelectrode array study, the state of SWS was associated with the highest coherence values in beta and gamma bands across the width of the neocortex, supporting the idea of the SWS-related memory consolidation (121). Aß-burden in the medial prefrontal cortex correlates with the impairment in non-REM SWA generation, which in turn is associated with sleep memory consolidation deficits during hippocampal-neocortical memory transfer. The association of the medial prefrontal cortex Aß pathology with a deficit in hippocampus-dependent memory consolidation was indirect as it depended on the intermediary factor of decreased non-REM SWA. These findings suggest that amyloid deposition could compromise the sleep-dependent memory consolidation via the disruption of the SWS (120).

In summary, changes in macrostructural parameters of sleep EEG indicate that alterations in sleep are prominent and early features of AD. However, further studies are required to examine the microstructural features including the sleep spindles, K-complexes, cyclic alternating patterns and high frequency oscillations. These microstructural features might reveal novel aspects of AD-related memory impairment and open new ways to understand how amyloid and tau could lead to cognitive decline (Figure 1).

https://www.bioscience.org/2018/v23/af/4587/fulltext.php?bframe=2.htm