A REAL-TIME EEG MONITORING SYSTEM

Ronald S. Hosek
Edward F. Owens

Life College
1269 Barclay Circle
Marietta, GA 30062

Presented at: Fourth Annual Conservative Health Science Research Conference
Palmer College of Chiropractic - West October 12 & 13, 1985

Abstract

While the primary medical application of the electroencephalogram (EEG) is the identification and localization of seizure disorders, brain damage or tumors, it is also useful for documenting states of consciousness. In this application, it is typically analyzed in terms of its frequency content, and a state of consciousness assigned as a function of how much of each frequency is present. The usual EEG frequency bands include Delta (0-4 Hz), Theta (4-8 Hz), and Alpha (8-12 Hz) and Beta ( greater than 12 Hz). Sleep has a preponderance of delta and Theta while waking states contain mostly Alpha and Beta. Relaxed aware states, such as meditation, generally contain largely; when visual imagery is present, there is often increased Theta. These latter types of states are the ones most often sought for therapeutic purposes, and may be "trained" using Alpha or Theta Biofeedback. The EEG is generally retrieved and turned into a number of tracings which are read ,off a moving chart and analyzed by an Encephalographer. In addition to being too costly, this approach is not acceptable for monitoring changes in real time.

The present work involves the design, construction and testing of a device which breaks the EEG into its frequency bands and plots the relative proportion of each frequency over time. Activity is taken from scalp electrodes in the normal fashion and amplified using a high-gain (500,000), low-noise (110 db CMRR) DC instrumentation amplifier. The amplified signal is led to a squaring circuit to provide leading and trailing pulse edges coincident in time with the zero-crossings of the signal. The leading pulse is used to start a precision voltage ramp generator; the trailing edge stops the generator and resets it. The voltage achieved by the ramp is proportional to the time between zero crossings and is taken to correspond to half the period of the current EEG signal. This time is compared to the half period times of the four frequency bands, a counter for the correct band is incremented, and a "band" light is flashed. At the end of a preset sampling time, generally 0.5 to one minute, the contents of the counters are transferred to an Apple computer through a parallel interface and plotted. The resultant graph shows the relative energy in each frequency band over time. The values are also stored on disk for later analysis. The system has been calibrated using a precision frequency generator and is accurate to within plus or minus 0.5 Hz. The total cost of the parts is under $100; a wire wrap schematic is available. Preliminary experiments with the unit have demonstrated definite changes with chiropractic adjustments. It has also been used in a biofeedback mode with the band lights providing feedback as to which frequencies are present.