Electroencephalographic characterization in epilepsy patients.
- c2019.
- 87p.
Thesis Report.
SUMMARY:
Background: Epilepsy is a chronic brain disorder that poses challenge both to clinicians and to patients. It is more common in low socioeconomic countries like Nepal. Its comorbidities are often neglected and go untreated. Electroencephalography is an important tool for understanding brain function. Its clinical use is largely for the diagnosis of epilepsy. However, it suffers from being low in sensitivity for the diagnosis, may not reflect the severity of the condition, and has inter-observer variability. Over the past half a century, its shortcomings have also been tried to improve. For this purpose, one of the methods is quantitative computation of EEG signals. Thus, the apparently normal EEG can be further analyzed to find the potential marker of disease. EEG, being composite wave, can be spectrally analyzed for different frequency waves that it is composed of. Also being a chaotic biological signal, it can be studied (possibly better) by nonlinear analysis tools such as entropy values. In our study, we have used both the parameters -power values of signals as linear variable, and Shannon spectral entropy as nonlinear one, in each frequency band. We also assessed the presence of depression and anxiety (generalized anxiety disorder) in each individual by standard questionnaires. The same procedure was done in healthy volunteers and each parameter was compared with those of patients. We hypothesized that background EEG in epilepsy is significantly different than EEG of control group.
Objectives: Primary objective: To study the characteristics of abnormal EEG pattern and its background activities in clinically diagnosed epilepsy cases. 8 Secondary objective: To estimate the proportion of depression and anxiety associated with epilepsy. Material and Methods: This was a comparative cross-sectional study done at EEG lab, department of Physiology on epileptic patients and healthy controls. Ethical clearance was taken from Institutional Review Committee of BPKIHS before start of study. Inclusion criteria applied for patients were subjects with clinically diagnosed epilepsy (suspected epileptic who had positive features of epilepsy in EEG), age 16 years and above, and abstinence of 24 hours from substance abuse and 4 hours from smoking. Exclusion criteria were subjects with known diagnosis of neuropsychiatric or any chronic disorders, use of drugs as medication or as abuse within past 24 hours, and left handed individuals. We relied upon history of subjects for the criteria assessment. Apparently healthy volunteers were taken with similar criteria as cases but with normal EEG. Mood and anxiety disorders were assessed by PHQ-2 (for depression) and GAD2 (for anxiety) questionnaires. EEG of subjects was recorded using standard 10- 20 protocol of scalp EEG recording. Settings used for recording: high cut 70 Hz, time constant 0.3, notch filter at 50 Hz and common average as reference method. Each EEG was coded and data was kept anonymous. No personal data was revealed outside the project. To analyze the signals, 5 sec epochs were selected on clean background EEG on both case and control groups. Additionally, 2-5 sec of epoch was also taken immediately before epileptiform potential in patients. These signals were filtered in 1 to 30 Hz band, channels of interest were selected and power spectrum 9 analysis was done by Fourier transform. Absolute power was obtained in each frequency. Relative power and Shannon spectral entropy were calculated as per their definition. The data was presented in frequency bands of 1-4 Hz (delta), 5-7 Hz (theta), 8-13 Hz (alpha) and 14-30 Hz (beta). Data of the two groups were compared by statistical analysis using Mann Whitney test and Wilcoxon signedrank test, taking p value <0.05 as statistically significant.
Results: Both groups were comparable in terms of cardiorespiratory and anthropometric variables. The epileptiform discharges in patients mainly consisted of spikes, sharp waves and slow waves, often in combination. Depression was a common comorbidity, present in 22.22% of epileptic patients. In quantitative EEG comparison, there was higher power in patients compared to controls, which was distributed in all scalp sites studied; and was relatively more concentrated in theta band. Such distribution was not much affected by presence of depression in patients. In our data, no subject groups showed any asymmetries in any side at alpha band. Finally, the background EEG of patients was significantly different before or away from epileptiform discharges, but without any specific trend.
Conclusion: The seemingly normal background of epileptic EEG is different from that of healthy controls. The power value is high in most frequency range in all scalp sites, which is relatively higher in theta band. Depression is a common associated comorbidity in epilepsy. Depression may not affect the EEG in spectral analysis.