Mass Spectrometry as a Quantitative Proteomic Analysis Tool for the Search for Temporal Lobe Epilepsy Biomarkers: A Systematic Review.

Temporal lobe epilepsy (TLE) is the most common form of epilepsy in adults. Tissue reorganization at the site of the epileptogenic focus is accompanied by changes in the expression patterns of protein molecules. The study of mRNA and its corresponding proteins is crucial for understanding the pathogenesis of the disease. Protein expression profiles do not always directly correlate with the levels of their transcripts; therefore, it is protein profiling that is no less important for understanding the molecular mechanisms and biological processes of TLE. The study and annotation of proteins that are statistically significantly different in patients with TLE is an approach to search for biomarkers of this disease, various stages of its development, as well as a method for searching for specific targets for the development of a further therapeutic strategy. When writing a systematic review, the following aggregators of scientific journals were used: MDPI, PubMed, ScienceDirect, Springer, and Web of Science. Scientific articles were searched using the following keywords: "proteomic", "mass-spectrometry", "protein expression", "temporal lobe epilepsy", and "biomarkers". Publications from 2003 to the present have been analyzed. Studies of brain tissues, experimental models of epilepsy, as well as biological fluids, were analyzed. For each of the groups, aberrantly expressed proteins found in various studies were isolated. Most of the studies omitted important characteristics of the studied patients, such as: duration of illness, type and response to therapy, gender, etc. Proteins that overlap across different tissue types and different studies have been highlighted: DPYSL, SYT1, STMN1, APOE, NME1, and others. The most common biological processes for them were the positive regulation of neurofibrillary tangle assembly, the regulation of amyloid fibril formation, lipoprotein catabolic process, the positive regulation of vesicle fusion, the positive regulation of oxidative stress-induced intrinsic apoptotic signaling pathway, removal of superoxide radicals, axon extension, and the regulation of actin filament depolymerization. MS-based proteomic profiling for a relevant study must accept a number of limitations, the most important of which is the need to compare different types of neurological and, in particular, epileptic disorders. Such a criterion could increase the specificity of the search work and, in the future, lead to the discovery of biomarkers for a particular disease.

Extreme value theory inspires explainable machine learning approach for seizure detection.

Epilepsy is one of the brightest manifestations of extreme behavior in living systems. Extreme epileptic events are seizures, that arise suddenly and unpredictably. Usually, treatment strategies start by analyzing brain activity during the seizures revealing their type and onset mechanisms. This approach requires collecting data for a representative number of events which is only possible during the continuous EEG monitoring over several days. A big part of the further analysis is searching for seizures on these recordings. An experienced medical specialist spends hours checking the data of a single patient and needs assistance from the automative systems for seizure detection. Machine learning methods typically address this issue in a supervised fashion and exhibit a lack of generalization. The extreme value theory allows addressing this issue with the unsupervised machine learning methods of outlier detection. Here, we make the first step toward using this approach for the seizure detection. Based on our recent work, we specified the EEG features showing extreme behavior during seizures and loaded them to the one-class SVM, a popular outlier detection algorithm. Testing the proposed approach on 83 patients, we reported 77% sensitivity and 12% precision. In 60 patients, sensitivity was 100%. In the rest 23 subjects, we observed deviations from the extreme behavior. The one-class SVM used a single subject's data for training; therefore, it was stable against between-subject variability. Our results demonstrate an effective convergence between the extreme value theory, a physical concept, and the outlier detection algorithms, a machine learning concept, toward solving the meaningful task of medicine.