Microcontroller student projects

Lecture Details

Several popular methods of measuring brain activity are electroencephalography (EEG), magnetic resonance imaging (MRI), positron emission topography (PET), computed axial tomography (CAT), magnetoencephalography, and near-infrared spectroscopy (NIRS). Of these methods, only EEG and NIRS are within any measure of feasibility given our time and budget for this project. MRI uses a powerful magnetic field, PET uses radiations from a nuclear tracer, CAT uses X-rays, and MEG uses unobtainable superconducting magnetometers. EEG is based on picking up electric fields resulting from the electric pulses of the underlying neurons. NIRS, on the other hand, works by shining infrared light into the scalp and measuring the amount of reflected light with a series of phototransistors. Since blood reflects infrared light depending on its level of oxygen concentration and activity within a region of the brain corresponds to increased blood flow to that area, the reflected infrared lights is a measure of local brain activity. The EEG is able provide a lot of information with high temporal resolution at the price of less-trivial sensor implementation than NIRS. Without amplification potentials measured at the scalp are on the order of microvolts whereas initial NIRS signals can be as high as the 10 mV range. Since we want high sensor density, the lower cost and easier implementation of an NIRS system are very desirable.