EEG Data Collection

During each visual traffic-light detection session, the 31-channel EEG and 4-channel EOG using sintered Ag/AgCl electrodes with an unipolar reference at right earlobe were simultaneously recorded by the Scan NuAmps Express system (Compumedics Ltd., VIC, Australia). All the EEG/EOG channels were located based on a modified International 10-20 system based on the relationship between the location of an electrode and the underlying area of cerebral cortex. Before data acquisition, the contact impedance between EEG electrodes and scalp was calibrated to be less than 5kΩ. The EEG data were recorded with 16-bit quantization level at a sampling rate 1 KHz and down-sampled to 500 Hz for the simplicity of data processing. Then EEG data were preprocessed using a simple low-pass filter with a cut-off frequency of 50 Hz to remove the line noise (60 Hz and its harmonics) and other high-frequency noise for further analysis. Finally, we successfully collected more than 700 successful ERP events of one subject in a driving experiment. Fig. 3-2 shows an example of the collected time series of EEG signals of subject 1. The red/green/amber traffic-light stimuli were marked as red, green, and yellow lines, respectively. The subject’s correctly responded target responses were also observed followed the target stimuli, i.e. a blue line was observed about 300-ms fallen behind the red-light stimulus at 25.7 second, and a cyan line was also observed about 350-ms fallen behind the amber-light stimulus at 23.2 second. To further analyzing the relationship between the visual traffic-light stimuli and the subject’s corresponding response, the synchronously measured continuous EEG signals are separated into several epochs/trials where an epoch or a trial contains the sampled EEG data from –200 ms to 1000 ms with a light stimulus given at 0 ms and were connected together as shown in Fig. 3-3 for the analysis of traditional time-domain overlap-added averaged methods or ICA algorithm. The extracted single-trial epochs for the red light stimuli at Pz channel and their time-domain overlap-added average (black line) were shown in Fig. 3-4 (a). Note that the ERP,

P300, was clearly observed. Fig. 3-4 (b) shows the single-trial ERP-image plots of correctly responded target response data at Pz channel (occipital site) from a red-light visual stimulus.

The subject’s response time (black line) is very time-locked to the P300 ERP corresponding to red-light stimulus. Fig. 3-5 shows the time-domain overlap-added averaged ERP signals for three kinds of traffic-light stimuli in Pz channel. We can observe that the ERPs related to different traffic-lights have apparent differences. Although using the time-domain overlap-added averaged method can successfully observe the appealing differences between ERPs related to different stimuli, it costs much time to collect enough epochs (at least 30 trials) before performing such time-domain overlap-added averaged algorithm and can not be used for online applications. Therefore, in this thesis, we introduce a new single-trial analyzing method based on ICA algorithm to deal with the prior-average problem of time-domain overlap-added averaged method without loss any information of the original ERP signals.

Fig. 3-6 shows the scalp topography of the time series of averaged epochs for one stimulus (red light) of subject 1. The results demonstrated that the active brain responses to significant events or external stimuli involve synchronized oscillations in local field potentials in a number of brain regions as reported in previous studies [15-16]. These brain dynamic events appear to begin in the frontal cortex, implying they carry or channel top-down information about intention, including attentional focus, to sensorimotor brain areas [18]

triggering other dynamic events that carry or channel bottom-up information from sensory to response-selection areas [83]. The analyzing results in Figs. 3-2 to 3-6 calibrate the successful design of the visual traffic-light detection tasks.

Red Light Green Light Amber Light Right_Button Left_Button

Time (sec)

EEG Channe l In dex

Red Light Green Light Amber Light Right_Button Left_Button

Time (sec)

EEG Channe l In dex

Figure 3-2. An example of the recorded raw data of EEG signals with synchronous onset times of three kinds of the traffic-light stimuli and two kinds of subject’s responses. The onset time of the red/green/amber traffic lights are presented as red, green, and yellow lines, and the subject’s responses, pressing a right button for a red light and a left button for an amber light, are presented as blue and cyan lines, respectively.

Stimulus Response

Epoch (trial) Index: -200ms~1000ms

EEG Channel Index

Stimulus Response

Epoch (trial) Index: -200ms~1000ms

EEG Channel Index

Figure 3-3. Extracted epochs (dashed intervals) for one stimulus (red right) and subject’s response (right button).

(a)

(b)

Figure 3-4. (a) Observed epochs (trials) for the red light stimuli at Pz channels and their time-domain overlap-added average (black line). Note that the ERPs, P300, was clearly

Figure 3-5. Time-domain overlap-added averaged ERP signals for three kinds of traffic-light stimuli in Pz channel.

Figure 3-6. Scalp topography of the time series of an averaged epoch for one stimulus (red light). These results show that active brain responses to significant events or external stimuli involve synchronized oscillations in local field potentials in a number of brain regions [28-29].

These brain dynamic events appear to begin in the frontal cortex, implying they carry or channel top-down information about intention, including attentional focus, to sensorimotor brain areas triggering other dynamic events that carry or channel bottom-up information from sensory to response-selection areas.