EARLY DETECTION OF VASCULAR DEMENTIA USING QUANTITATIVE ANALYSIS OF EEG SIGNALS

• Main Author: Hadiyoso, Sugondo
• Format: Dissertations
• Language: Indonesia
• Online Access: https://digilib.itb.ac.id/gdl/view/71426
• Institution: Institut Teknologi Bandung

Dementia is a common symptom of neurological disorders that describe a decline in cognitive function in the brain. One of the most common forms of dementia after Alzheimer's is Vascular Dementia (DVa) in post-stroke patients. Post-stroke vascular dementia is caused by degenerative cerebrovascular disease. Dementia can affect attention, memory, and other cognitive functions. If not treated quickly and appropriately, the decline will take place continuously. The diagnosis mechanism can be carried out by genetic profile analysis, MRI, CT, and PET imaging modalities, but the complexity and cost of these methods are invasive and tends to be expensive. One of the medical modalities that can be an alternative for evaluating and even diagnosing dementia is Electroencephalograph (EEG). In the last few decades, research on the characterization of EEG signals in cases of dementia has been carried out using both conventional and quantitative approaches. Due to the limitations of conventional methods, namely high subjectivity and difficulty in being applied in large populations, quantitative EEG or QEEG is strongly recommended to overcome this problem. In this study, a method for characterizing EEG waves in post-stroke patients with mild cognitive impairment and dementia is proposed by calculating and analyzing QEEG parameters. This research proposes linear and non-linear QEEG methods for characterization through spectral, coherence, and signal complexity analysis approaches. The characterization results showed that post-stroke patients with cognitive impairment had a relatively lower beta wave power than the normal group. Meanwhile, the relative strength of delta waves in cognitively impaired patients tends to be higher than normal. Then there is the relationship between the strength of the EEG signal and the severity of dementia. Investigation of coherence in this study showed a pattern of decreased coherence scores in post-stroke patients with cognitive impairment compared to the normal elderly group. Significantly decreased coherence was found in the temporal lobe, which plays an important role in hearing, language, and memory. Meanwhile, signal complexity analysis showed that signal complexity in post-stroke patients with cognitive impairment tended to be lower than in the normal group. EEG signal analysis on memory activity recordings showed a significant difference (p<0.05) for all observed EEG electrodes between dementia, mild cognitive impairment, and normal groups. Another result found is a relationship between the degree of complexity of the EEG signal and the severity of cognitive impairment. The result of this characterization calculation becomes a feature vector to be validated using the classification method. Simulations were conducted to classify normal elderly, post-stroke patients with mild cognitive impairment, and poststroke dementia subjects. Support Vector Machine (SVM) and k-Nearest Neighbor (k-NN) are used to evaluate the performance of the proposed feature extraction method. The results of the classification simulation show that the highest accuracy of 96% is achieved using Gaussian SVM. The findings of this study indicate that QEEG analysis can be a method to investigate and evaluate the severity of dementia in post-stroke patients. This method is expected to be used for early detection or detection of mild cognitive impairment. The proposed method can simplify the detection process with an automatic classification algorithm as additional validation in clinical diagnosis.