Abstracts

RATIONALE: To satisfy the exacting needs of a sophisticated EMU, we have developed our own digital video recording system. Our implementation was designed to address the following needs not ordinarily available: 1 Full frame rate (30/sec) recording. 2 absolute frame by frame synchronization with EEG. 3 Smooth playback in forward or reverse, at accelerated or slow-motion speed. 4 Playback to computer screen, conventional NTSC monitor, or VCR. 5 Immunity to network or server failure. 6 Immediate network access to recorded data. METHODS: We have built our own digital video acquisition and review capability from the ground up in order to control all aspects of the process, employing: 1 Custom-written MPEG1 driver to permit time-stamping of every frame. 2 Acquisition interface cards for both EEG and video on same hardware bus permitting perfectly controlled synchronization. 3 Single computer acquisition of both EEG and video. No server or second computer required. 4 Peer to peer networking. Previously recorded data immediately served from recording system to any review station on network without interrupting acquisition. 5 Operator control of compression level. 6 Hardware decoding on replay to limit CPU resource requirement and permit video display to computer monitor or external NTSC monitor. RESULTS: We have developed a digital video solution customized for multitasking EEG/video. A single computer handles EEG acquisition up to 128 channels simultaneous with full frame rate video. Satisfactory performance is achieved at a data rate of 1.15 Mbit/sec. A 100 Mbit Ethernet connection is required. Time correlation to the ms (ie per frame) throughout the recording period has been achieved. The implementation permits independent (blind) or continuously correlated EEG and video review. Two way automatic seek, ie synchronization at any time from EEG to video or video to EEG, is available. CONCLUSIONS: The design criteria have been achieved, providing the advantages of digital video without sacrificing analog-type features (eg jog shuttle). Full frame rate video is critical for evaluating small facial movements. Subjectively the video quality is as good as MPEG2 but requires 1/4 the network speed and storage.