Abstracts

Simultaneous EEG-fMRI recording is becoming an increasingly important tool for epileptogenic tissue localization. Reduction of the noise in EEG data recorded during fMRI acquisition is crucial for accurate epileptogenic tissue localization. The main sources of the noise include the RF pulse artifact, the gradient artifact, and the cardiac pulse artifact. Although the RF pulse artifact is relatively small in amplitude, its amplitude fluctuates randomly, making it more salient than the gradient artifact after post processing for noise reduction. This research investigates the cause of the RF pulse artifact. We hypothesize that the RF pulse artifact is caused by a rectification effect at the carbon-metal interfaces that exist between carbon wires and other electronic components. To test this hypothesis, two small wire loops, one made of a carbon wire and the other made of a copper wire, were attached to a spherical phantom that was placed in a 3T MR scanner (Siemens Trio). The carbon wire loop was connected to the rest of measurement circuit using copper crimps. The outputs from the both loops were low-pass filtered at 10.7 MHz and were recorded by an analog oscilloscope. Two types of MR sequences were used: an echo-planar imaging (EPI) sequence (TR=100 milliseconds (ms), TE=34 ms, FA=130) and a sequence that consisted only of RF pulses with a rectangular envelope (TR=101 ms, pulse duration = 1ms, FA=180). The oscilloscope recorded DC shifts that resembled the envelope of the RF pulse (i.e. envelope of a sync function for the EPI, a rectangular function for the RF-only sequence) from the carbon wire loop. No artifact was recorded from the copper loop corresponding to the RF pulses. The average amplitude of the RF pulse artifact from the carbon wire depended on the condition of the carbon-metal junction: the firmer the both ends of the carbon wires were crimped, the smaller the amplitude was. The amplitude of the RF pulse artifact fluctuated randomly and even changed polarities. In the firmly crimped condition, the mean absolute amplitude of the artifact was 34.8 microvolts (standard deviation = 24.4 microvolts). The waveform of the RF pulse artifact and the dependence of the amplitude on the carbon-metal junctions indicate that the rectification effect of the carbon-metal interface causes the RF pulse artifact. Use of metal wires and better carbon-metal connection methods are possible solutions to the problem. (Supported by NIH R01-NS38467.)