Non-Invasive Detection of Hypoglycemic and Hyperglycemic Episodes in People with Diabetes using Electroencephalogram Signals

Patients with diabetes need to monitor their blood sugar levels frequently to ensure episodes of low or high sugar concentration can be treated timely. In this thesis, we aim at using brain waves to detect the onset of low blood sugar episodes without the need for taking blood. Novel computational intelligence algorithms were proposed to improve the correct detection rate of such dangerous episodes. Additionally, the innovation in using brain activity to detect low blood sugar was also extended to high blood sugar. The obtained results confirm the potential for using brain waves in detecting dangerous blood sugar conditions non-invasively.