Aristotle University Of Thessaloniki (AUTH)
Μελέτη στατιστικών παραμέτρων ηλεκτροεγκεφαλογραφικού φάσματος ασθενών με ήπια νοητική διαταραχή
Abstract
dc:descriptionMild Cognitive Impairment (MCI) is a major medical-social problem as it is, in some cases, the precursor of dementia impairments. This thesis fits into the overall research on studying the MCI disease. This thesis has been designed to study the electroencephalogram (EEG) combined with Transcranial Magnetic Stimulation (TMS) as one of the most promising methods for early and reliable diagnosis of MCI. It is a neurophysiological technique presented in the mid 90's; it allowed for the first time non-invasive investigation of cortical excitability of the human brain. However, over the years, it became clear that this method is associated with numerous methodological problems that severely restricted its penetration in the daily clinical practice. The purpose of this study was to resolve the major methodological problems arising from the implementation of this complex process and to execute a pilot clinical trial in patients with MCI. Actual recordings from the TMS-EEG recording system of the Clinical Neurophysiology Laboratory in 3rd Neurological Clinic of the Thessaloniki Aristotle University were used as pilot data. Utilizing data from the clinical trial, the study dealt with the specific objectives set out: (a) removing the TMS and eye blink related artifacts by applying advanced signal analysis statistical methods, (b) exploring the possibility to implement the newly proposed infinite reference and benchmarking against the widely used average reference, (c) evaluating these references by studying event related coherence as a measure of inter-hemispheric connectivity, and (d) identifying differences in a set of electro-physiological parameters between the control group and patients with MCI. In the first phase of the thesis, a protocol was designed in order to describe in detail the process, techniques and tools needed to address the most important challenges that one faces when one tries to implement the highly complex process of EEG combined with TMS. The protocol describes in detail the steps of the process, including recording, pre-processing and analysis of EEG signal, focusing on methodological, technical and practical problems of the process. One of the major issues referenced in the literature and documented in the daily clinical practice is the problem of “contamination” introduced to EEG recordings due to TMS pulse. The intensity, range and general characteristics of the artifact make its removal extremely difficult and conceal important information of the brain's response to the stimulation, thus complicating the diagnosis of MCI. The strong distortion of the signal, the inability to make a reliable measurement of critical factors such as waves P30, N100, N15, and the inability to reveal connectivity between regions of the brain are important side effects of the adoption of magnetic stimulation. We studied the artifact characteristics and its physiology and, based on evaluation of methods proposed by international literature, a TMS artifact correction model was adopted. In our survey, we conclude that the selected correction model makes it possible to completely remove the TMS artifact appearing in EEG recordings in the first 10 ms after the magnetic stimulation. The model was applied to responses from magnetic stimulation of the left and right primary motor cortex (M1 area) and in the prefrontal cortex. Based on this method, a model that produces these responses and relates them to the stimulation intensity is proposed. The second phase of this study involved finding a reference which both is suitable for use with TMS and exhibits improved performance over the EEG signal quality. The reference point properties, whether determined by its physical location or by its computational characteristics, have fundamental influence on the signal. Traditionally it was thought that there is no electrically neutral reference point, as any reference point taken on the event related potential (ERP) waveform, reflects the contribution of both the active area and the reference point. We investigated the applicability of the newly proposed REST method - an IR reconstruction method - in order to approximately transform EEG data recorded using a reference point at infinity. The infinite reference is comparatively evaluated against the widely used reference average. For this purpose, stimulation of the left primary motor cortex (M1 region) was applied on a patient with complete agenesis callosum of congenital type, while recording with 60-channel TMS compatible EEG device (eXimia, Nexstim Ltd, Finland). Using eight (8) - shaped coil, we applied magnetic pulses with amplitude on the 120% of the motor cortex stimulation threshold. Before analysis, EEG data were re-referenced using IR and AR references. This study confirmed recent theoretical simulations and empirical observations, thus highlighting the infinite reference as a valid method for investigating connectivity, spectral and topographic analysis in neurophysiological studies. It has also been confirmed that the introduction of infinite reference significantly reduces the rate of false positive interhemispherical connectivity and therefore excels compared to the widely used reference average. This finding is consistent with prior studies using both simulated and actual data. In the third phase of this study, the aim was to identify differences in a set of electro-physiological parameters between the control group and patients with MCI. EEG data analysis has been used in order to identify differences between the control group and patients. In consistency with prior studies, we observed that early separation between the healthy group and the group of patients (MCI) is achieved using measures of sensitivity. Although the sample size was small, the two groups were distinctly separated, but the statistical significance of these comparisons is quite weak and requires further investigation with a larger sample size. Although the mechanism affecting P30 cannot be revealed, the coordinates of P30 peak seem to be related to cognitive decline. The object group (healthy, MCI) seems to have a strong effect on P30 average value for both hemispheres. Therefore, the value of P30 as a diagnostic measure for the cognitive function decline recognition has been confirmed; it exhibited greater efficiency compared to N100 height, which does not correlate significantly with the MMSE memory decline scale. In conclusion, this study establishes a methodology which suitable for the implementation of the complex TMS-EEG method, and demonstrates its applicability for the investigation of patients with Mild Cognitive Impairment and other disorders of the central nervous system in general.
Degree
thesis:*- Grantor dc:publisher
- Aristotle University Of Thessaloniki (AUTH)
- Year dc:date
- 2012
Author and committee
dc:creator, dc:contributor.*- Authors dc:creator
-
- Pseftogianni, Dimitra
- Ψευτογιάννη, Δήμητρα
Subjects
dc:subject × 12Rights
- Language dc:language
- gre
Identifiers
dc:identifier.*- Identifier
- 10.12681/eadd/35002