Epilepsy is a chronic neurological disorder caused by abnormal neuronal activity that is diagnosed visually by analyzing electroencephalography (EEG) signals. Background: Surgical operations are the only option for epilepsy treatment when patients are refractory to treatment, which highlights the role of classifying focal and generalized epilepsy syndrome. Therefore, developing a model to be used for diagnosing focal and generalized epilepsy automatically is important. Methods: A classification model based on longitudinal bipolar montage (LB), discrete wavelet transform (DWT), feature extraction techniques, and statistical analysis in feature selection for RNN combined with long short-term memory (LSTM) is proposed in this work for identifying epilepsy. Initially, normal and epileptic LB channels were decomposed into three levels, and 15 various features were extracted. The selected features were extracted from each segment of the signals and fed into LSTM for the classification approach. Results: The proposed algorithm achieved a 96.1% accuracy, a 96.8% sensitivity, and a 97.4% specificity in distinguishing normal subjects from subjects with epilepsy. This optimal model was used to analyze the channels of subjects with focal and generalized epilepsy for diagnosing purposes, relying on statistical parameters. Conclusions: The proposed approach is promising, as it can be used to detect epilepsy with satisfactory classification performance and diagnose focal and generalized epilepsy.

Purpose: Routine electroencephalogram (EEG) examinations uses intermittent photic stimulation (IPS) for investigation of the visual cortex EEG responses during resting time. This study aimed to discover brain dynamics effects of IPS in 28 generalized epilepsy patients and 28 healthy subjects. Methodology: Signal processing techniques were used in feature extraction by Fast Fourier transform (FFT), feature dimension reduction by t-test (significant, p<0.05) and classification by nearest neighbor (k-NN) and support vector machine (SVM). Results: The epilepsy group had higher level of amplitude in Theta waves compared to the healthy group. The Alpha waves in the resting time and for all IPS frequencies were observed with lower level of amplitude in healthy subjects compared to the epilepsy group. The k-NN (85.7% accuracy) classifier had the best discrimination of epilepsy from healthy group for resting time versus during IPS at 18 Hz IPS. However, using SVM (75.0% accuracy), IPS at 25 Hz yielded the best discrimination between resting time versus IPS in epilepsy where the healthy group responded similarly in all IPS frequencies. Conclusions: This study shows that IPS at 18 Hz and 25 Hz are suitable IPS frequencies for k-NN and SVM, respectively, to discriminate non-photosensitive generalized epilepsy from normal subjects during interictal.

Obsessive-compulsive disorder (OCD) is a mental illness causing patients to suffer from recurring undesirable thoughts (obsessions) conducting to do affairs repetitively (compulsions). Brain signals recorded by Electroencephalogram (EEG) can be analyzed in order to present a diagnostic procedure considering the localization approach. In this study, the signals acquired by EEG have been recorded from three groups; two case groups; patients with severe obsessive symptoms and patients with severe compulsive symptoms, and one healthy control group. Brain signal processing techniques have been applied on the signals emitted from frontal and parieto-occipital regions to discover the features leading to the best discrimination between case groups and healthy controls. In this regard, after preprocessing, the features of time and frequency domains presenting the significant meaningful relation were nominated for classification by linear discrimination analysis (LDA). Although the parieto-occipital region performed better in the diagnosis for both obsessive and compulsive groups, the features gained from the frontal cortex resulted in better discrimination for only the compulsive group. In addition, time domain features had a more significant influence in diagnosis rather than frequency domain for both case groups. The study presented particular characteristics of brain signals in two dimensions of OCD in specific brain regions leading to more accurate presurgical assessments in the studies between the affected brain regions and behavioral issues.

No abstract

No abstract