Extracting brain behavior change in patients with migraine by quantitative analysis of electroencephalogram signal of patients compared to healthy people

Background. Migraine disease is the second most common cause of headaches. Despite the high prevalence, the exact etiology of migraine is yet unknown. In this study, to evaluate the behavior change of electroencephalography (EEG) signals in migraine patients, various features of the EEG signals of migraine patients and healthy controls (HCs) were extracted and compared. Methods. This cross-sectional analytical study was conducted on 21 HCs and 18 migraine patients. Various features, such as fractal dimension (FD), approximate entropy (ApEn)


Background
Migraine disease (MD) is the second most common cause of headaches.Physiologically, different theories such as vascular theory, cortical spreading depression (CSD), migraine generating network (MGN), genetic theory, etc. have been stated for the cause of MD, but the exact cause of the disease is still unknown.In addition to headaches, MD also includes a wide range of symptoms such as vomiting, photophobia, osmophobia, and phonophobia.Brain signals or images are not a tool for MD diagnosis.MD diagnosis is only based on the patient's description of the disease and the doctor's decision.The studies of brain signals of MD patients report an increase in the number of spikes in the EEG signals of MD, an increase in the power of the theta frequency sub-band, an increase in the connections between neurons in the frontal region, and in general, differences in brain function due to MD.

Methods
This cross-sectional analytical study included 21 healthy controls (HCs) and 18 migraine patients.There were five males in each group.The age range of the subjects was 19 to 54.To investigate the EEG signals and better understand the cause of headaches, different features of the EEG signals were extracted and compared between HCs and migraine patients.For this purpose, EEG signals recorded by Carnegie Mellon University were used.In the first stage, the EEG signals were re-referenced, and after removing the DC offset, they were filtered by a 0.1 Hz highpass filter.The electricity noise was also removed by a 60 Hz notch filter.Then, for each signal, the noisefree segments were chosen.To calculate the energy of different frequency sub-bands, including delta (0 to 3 Hz), theta (4 to 8 Hz), alpha (8 to 12 Hz), beta (13 to 30 Hz), and gamma (more than 30 Hz), first these sub-bands were extracted using the wavelet transform, then their energy was calculated.
On the other hand, the complex features as well as fractal dimension (FD), approximate entropy (ApEn), and largest lyapunov exponent (LLE) were calculated from the EEG signals of both migraine patients and HCs.The extracted features from both groups were compared using the T-test or Kruskal-Wallis test, and the channels with significant differences (P<0.05) were reported.

Results
According to the results, the amount of ApEn in migraine patients was higher than that of HC in 120 channels, and only in eight channels did these values exhibit different performance.Based on statistical tests, three channels had significant differences.If we consider EEG behavior as a response of a dynamic system, an increase in ApEn means an increase in random behavior, indicating that the neuronal behavior of the brain deviates from the precise behavior of HCs, and tends to random behavior.Based on the results, the majority of channels had a lower value of LLE, with a significant difference observed in six channels.It can be said that the larger LLE shows the complexity of the dynamic system.In general, the reduction of LLE means that a migraine patient's brain moves away from specific behaviors such as deterministic chaos toward more random behaviors.These results are complementary to the changes in ApEn behavior.In these results, FD increased in most of the channels.The higher FD value means that higher degrees are needed to model the system, leading to a weaker relationship between different parts of the signal.This can indicate the role of the frontal region in the path of pain process in migraine patients.The increase in energy in theta waves can be due to anxiety and stress, which is prevalent in migraine patients.Based on the results, significant differences in delta and theta frequency sub-bands mostly occur in the occipital channels.Although significant differences in beta and gamma frequency sub-bands mostly occur in the frontal channels, in alpha waves, an increase in energy can be seen in 127 channels, but none of these differences are significant in this sub-band.Beta and gamma waves usually indicate activities such as problem-solving and mental activities.Energy increase in these sub-bands can be expressed in the way that the brain activity of migraine patients is normally the same as the brain activity during solving problems.Individuals in good health can readily assess environmental changes, while migraine sufferers find that doing so is like to solving a puzzle.In addition, the gamma frequency sub-band can correspond to dysfunction in pain-processing areas.

Conclusion
After observing the EEG behavior in general, it can be said that the mentioned differences are comprehensive, and some channels do not follow these changes.Furthermore, because significant changes are seen in a smaller number of these channels, it is simply not possible to distinguish healthy people from migraine patients in a headachefree state.However, if we consider EEG behavior as the response of a dynamic system, it can be said that although migraine patients do not exhibit a distinguishable difference from healthy people, the functioning mechanism of their brains has changed.Migraine patients' brains leave the function that is close to deterministic chaos and move toward random behaviors.