Abstracts

Rationale: Improvement of digital brainwave recorder in high resolution of channels and sampling rates, and long-term recording although provides accuracy in diagnosis of epileptic seizures, or localizes ictal focus/foci, does not change time span in reading EEG. In this study, we tried to develop intelligent systems to assist medical professionals to find ictal onset in an easy way. The first system is an EEG specialization visualization system (ESVS) that transforms channel signals into a color map. The second intelligent system is an intelligent EEG epileptic seizure prediction system (IEESPS). Methods: In the ESVS, the extracted original data from long-term electrocorticography (ECoG) were converted into a scaled-color map which demonstrated the consequent changing of standard deviation degree in comparison with upstream epoch in each channel. In the IEESPS, we computed the short-term Lyapunov exponent max (STLmax) from original EEG data and then converted the STLmax of each channel into T-index for seizure prediction. Both systems were implemented by LabVIEW (Laboratory Virtual Instrument Engineering Workbench) and MatLab. The users can tune setting of parameters in the control panel to adjust in various sampling and window size conditions. Results: In the ESVS, the transformed color map was able to display outstanding features in the channels that corresponded to the recognized ECoG channels with ictal activities. In the IEESPS, two threshold value of the epileptic seizure, T1 and T2, 5.25 and 4.75 are calculated respectively based on the findings of Sackellares. The result in the following figure shows an example how seizure is warning. An epileptic seizure data is read in this system, the system will then automatically calculate the T-index value at every time span. At the time of T-index value drops below both T1 and T2, the system will display the warning sign (by changing the color of warning button) and predict the occurrence of seizure. Conclusions: Preliminarily, our systems did provide an automatic and quick way to assist clinicians to review the ictal ECoGs. Temporal changing of ECoG voltage reflected on our system underlines the electrophysiology of epileptic seizures. In the future, more data to correct the systems and more blinded exams to improve the accuracy are necessary.