Multiple sclerosis (MS) represents the most common acquired demyelinating disorder of the central nervous system (CNS). Its pathogenesis, in parallel with the well-established role of mechanisms pertaining to autoimmunity, involves several key functions of immune, glial and nerve cells. The disease’s natural history is complex, heterogeneous and may evolve over a relapsing-remitting (RRMS) or progressive (PPMS/SPMS) course. Acute inflammation, driven by infiltration of peripheral cells in the CNS, is thought to be the most relevant process during the earliest phases and in RRMS, while disruption in glial and neural cells of pathways pertaining to energy metabolism, survival cascades, synaptic and ionic homeostasis are thought to be mostly relevant in long-standing disease, such as in progressive forms. In this complex scenario, many mechanisms originally thought to be distinctive of neurodegenerative disorders are being increasingly recognized as crucial from the beginning of the disease. The present review aims at highlighting mechanisms in common between MS, autoimmune diseases and biology of neurodegenerative disorders. In fact, there is an unmet need to explore new targets that might be involved as master regulators of autoimmunity, inflammation and survival of nerve cells.

The use of electroencephalography (EEG) has recently grown as a means to diagnose neurodegenerative pathologies such as Alzheimer’s disease (AD). AD recognition can benefit from machine learning methods that, compared with traditional manual diagnosis methods, have higher reliability and improved recognition accuracy, being able to manage large amounts of data. Nevertheless, machine learning methods may exhibit lower accuracies when faced with incomplete, corrupted, or otherwise missing data, so it is important do develop robust pre-processing techniques do deal with incomplete data. The aim of this paper is to develop an automatic classification method that can still work well with EEG data affected by artifacts, as can arise during the collection with, e.g., a wireless system that can lose packets. We show that a recurrent neural network (RNN) can operate successfully even in the case of significantly corrupted data, when it is pre-filtered by the robust principal component analysis (RPCA) algorithm. RPCA was selected because of its stated ability to remove outliers from the signal. To demonstrate this idea, we first develop an RNN which operates on EEG data, properly processed through traditional PCA; then, we use corrupted data as input and process them with RPCA to filter outlier components, showing that even with data corruption causing up to 20% erasures, the RPCA was able to increase the detection accuracy by about 5% with respect to the baseline PCA.

There is strong evidence that homocysteine is a risk factor not only for cerebrovascular diseases but also for degenerative dementias. A recent consensus statement renewed the importance and the role of high levels of homocysteine in cognitive decline in several forms of degenerative dementia, such as Alzheimer’s disease. Although the molecular mechanisms by which homocysteine causes cell dysfunction are known, both the impact of homocysteine on specific cognitive functions and the relationship between homocysteine level and non-Alzheimer dementias have been poorly investigated. Most of the studies addressing the impact of hyperhomocysteinemia on dementias have not examined the profile of performance across different cognitive domains, and have only relied on screening tests, which provide a very general and coarse-grained picture of the cognitive status of the patients. Yet, trying to understand whether hyperhomocysteinemia is associated with the impairment of specific cognitive functions would be crucial, as it would be, in parallel, learning whether some brain circuits are particularly susceptible to the damage caused by hyperhomocysteinemia. These steps would allow one to (i) understand the actual role of homocysteine in the pathogenesis of cognitive decline and (ii) improve the diagnostic accuracy, differential diagnosis and prognostic implications. This review is aimed at exploring and revising the state of the art of these two strictly related domains. Suggestions for future research are provided.

Objective The applause sign, originally reported as a specific sign of progressive supranuclear palsy (PSP), has recently been found in several parkinsonian disorders. Its nature is still uncertain. It has been interpreted as a motor perseveration or a form of apraxia. The present study aims to: (a) verify the specificity of the applause sign for parkinsonian disorders, examining the presence of the applause sign in cortical dementias which should be error free and (b) clarify the nature of the applause sign (resulting or not from apraxia). Methods 77 subjects were included: 10 PSP, 15 frontotemporal dementia (FTD), 29 Alzheimer's disease (AD) and 23 normal controls. The presence of apraxia was an exclusion criterion. All patients underwent a detailed neuropsychological examination, and cognitive performance was correlated to the applause sign. Results All patient groups showed the applause sign and differed significantly from normal subjects who were error free. No difference was found when comparing PSP with FTD and FTD with AD. AD differed significantly from PSP but they were not error free (31% of patients with AD showed the applause sign). The only correlation with background neuropsychology was found for measures of executive functions. Conclusions The presence of the applause sign in cortical dementia does not confirm the specificity of the applause sign for parkinsonian disorders. The applause sign should be interpreted as a sign of frontal lobe dysfunction rather than a form of apraxia, and can likely be detected in any kind of disease which involves frontal lobe structures to some extent.