Human Brain Mapping


Florence, Italy



June 11-15, 2006

Ruey-Song Huang (1,2), Tzyy-Ping Jung (2), Jeng-Ren Duann (3), Scott Makeig (2) & Martin I. Sereno (1) , 1. Department of Cognitive Science, UCSD
2. Swartz Center for Computational Neuroscience, Institute for Neural Computation, University of California San Diego, La Jolla CA

Imaging event-related brain dynamics during continuous driving

INTRODUCTION: Event-related brain dynamics during continuous driving were studied using electroencephalographic (EEG) recording and functional magnetic resonance imaging (fMRI).

EEG METHODS: A virtual-reality scene was constructed to simulate driving on a highway at a constant speed. Each of five right-handed subjects participated in two 1-hour sessions during which 256-channel EEG signals and driving task parameters were recorded at 256 Hz. During 1-hour continuous driving, every 3 to 7 seconds, the car was linearly pulled towards the curb or into the opposite lane, with equal probability. Subjects were required to compensate for the drift by holding down an arrow key, and to release the key when the car was steered back into the center of the left lane. The extent of each drift event was measured by the absolute maximum deviation from the cruising position at deviation onset. EEG data were analyzed using independent component analysis and time-frequency analysis using the EEGLAB toolbox (sccn.ucsd.edu/eeglab).

fMRI METHODS: Subjects also participated in an fMRI driving session on a different day. The driving scene was projected onto a screen inside the scanner. Subjects steered the car using a two-key MR-compatible response box. Each subject participated in two 512-s periodic and two 1024-s random-ISI event-related scans. In the first two scans, the car drifted from lane center once every 16 sec. In the last two scans, the car drifted away every 5~10 sec. Functional data were acquired with an 8-channel phased-array head coil in a GE 3T scanner using a standard EPI sequence (TR=2 s, TE=30 ms, 64x64 voxels, 31 slices, 3.125x3.125x4 mm, 256 or 512 images). Data from the first two scans were analyzed using Fourier analysis. Significant activations plus their phases at the stimulus frequency (32 cycles/scan) were rendered onto inflated cortical surfaces using Freesurfer.

EEG RESULTS: A component with equivalent dipole sources located bilaterally in the intraparietal sulcus exhibited larger tonic alpha band power in large-error (drowsy) compared to low-error (alert) events. Alpha power was suppressed briefly after deviation onset, then increased strongly (~10 dB) just before the subject released the key. This transient (1.5~3 s) alpha rebound activity was consistently observed during all single events, regardless of alertness levels. Other components localized to premotor, somatomotor, posterior parietal and cingulate cortices also exhibited event-related brain dynamics in various frequency bands that were time-locked to different phases of the drift events.

fMRI RESULTS: Results showed that during each drift event, BOLD signals were activated, sequentially, in cingulate cortex, supplemental motor area (SMA), premotor, motor, somatosensory, and posterior parietal cortices. Data from the last two scans were analyzed using FMRLAB (sccn.ucsd.edu/fmrlab) and maps of maximally spatially independent BOLD process activities were rendered onto the inflated cortical surfaces. BOLD activities of several of these processes, active in middle temporal, posterior parietal, somatomotor, parietal-prefrontal, dorsal lateral prefrontal, medial prefrontal and cingulate cortices respectively, were strongly correlated with driving performance.

DISCUSSION: This study demonstrates how event-related brain dynamics during continuous driving may be explored using independent component analysis applied to data from multiple imaging modalities with differing spatial and temporal resolutions.

Return to Current Abstracts