Cross-Subject ERSP Results

*

*

*

*

*

8~13 Hz

*

*

*

*

*

*

8~13 Hz

*

*

*

*

*

*

8~13 Hz

Fig. 4-21: The comparison of the averaged alpha power between the single tasks and dual tasks in the motor cluster. Left column (A, C, E): statistic test between dual-task cases and single math. Right column (B, D, F): statistic test between dual-task cases and single deviation. Note: the frequency response in motor area near the alpha band was not only influenced by the deviation task but also varied by button press.

4.2.3.2 Cross-Subject ERSP Results

The cross-subject averaged ERSP in the frontal cluster corresponding to the five cases were shown in Fig. 4-22. Significant power increases related to the math-task were observed in Fig. 4-22 (A), (B), (C) and (D).

We again demonstrated that the power increase in the frontal cluster is related to math-task. The theta power increase in the three dual-cases including case-1, case-2 and case-3 were slightly different to each other. Comparing to single-math task (Fig.

4-22 (A)), the power in dual-task cases were stronger. EEG theta increase was related to distracted effects in the literatures. Therefore, subjects distracted highest in the case which math presented at 400ms before deviation. The beta power increase which induced by mathematical equations in the literatures was appeared in the math-task and time-locked to mathematics onsets.

A

Fig. 4-22: The ERSP images of frontal cluster for five cases: case-4 (A), case-1 (B), case-2 (C), case-3 (D) and case-5 (E). The right column: the averaged scalp maps for the frontal component across 10 subjects. Color bars showed the magnitude of ERSPs. The middle column showed the onset sequences between two tasks. Pink dashed lines: the first event onset. Blue dashed lines: the mean of reaction time to deviation. Red dashed lines: the mean of reaction time to math. Black dashed lines: the averaged response time for car returning to the third lane. Red dot: the onset of math occurred. Blue dot: the onset of deviation presented. Note: the theta (5~7.8 Hz) and beta (12.2~17Hz) power were increased briefly after the math onset. The strongest power increase was observed in case-1. The shortest latency of theta band increase was appeared in case-1.

The comparison of the total power in the four cases with math-task was given in Fig. 4-23, which suggested that the amount of power increase in 5~7.8 Hz were different with different time of SOA. The most significant power increase occurred in case-1.

Fig. 4-23: The comparison of total power in cross-subject averaged ERSP images in frontal component between cases. The light blue bar were represented the total power in the theta (5~7.8 Hz) band. The dark blue bars were represented the total power in the beta (12.2~17 Hz) band. The bottom insets showed the onset sequences between two tasks. Note: the most significant power increase was occurred in case-1.

The comparison of the latency of ERSP time-locked to math onsets in the four cases with math-task was given in Fig. 4-24, which suggested that the latency of power increase in 5~7.8 Hz were different with different time of SOA. The shortest latency of power increase occurred in case-1 and the longest latency of power increase occurred in case-4.

0 Latency (aligning Math onset)

Latency (ms) Latency (aligning Math onset)

Latency (ms) Latency (aligning Math onset)

Latency (ms)

Fig. 4-24: Effects of distraction on onsets of theta and beta increases. Latencies were calculated from cross-subject averaged ERSP images of the frontal component.

Panel as Fig. 4-23. Note: the shortest latency of the theta increase was observed in case-1 and the longest latency of the theta increase was revealed in case-4.

Left

Fig. 4-25: The grand mean of cross-subject averaged ERSP images in left (left block:

8 subjects) and right (right block: 6 subjects) motor components between cases. Left block: left motor cluster. Right block: right motor cluster. Panels as Fig. 4-22. Note:

the alpha power was suppressed briefly after the first event onsets in all cases and the strongest alpha suppression was occurred in case-5.

The cross-subject average ERSP in the left and right motor clusters corresponding to the five cases were shown in Fig. 4-25. Significant power suppressions time-locked to event onsets were observed (case-1, case-2, case-3, case-4, and case-5). In case-4, the alpha suppression was observed continuously until the red dashed lines which were the subject response to the event with pressing a button. The alpha suppression continued after the black dashed lines (including case-1, case-2, case-3, and case-5), it maybe control the steering wheel again in the third lane for subjects.

Fig. 4-26: The comparison of total power in cross-subject averaged ERSP images in the left (left block: 8 subjects) and the right (right block: 6 subjects) motor components between cases. The light blue bars represented the total power of the alpha (8~13 Hz) band in individual cases. Note: the most significant power increases were occurred in case-5.

The comparison of the total power in the five cases was given in Fig. 4-26, which suggested that the amounts of power increase in alpha band were different with different time of SOA. The most significant power increase occurred in case-5 (the single-deviation task) and the smallest power increase occurred in case-4 (the single-math task).

0

Fig. 4-27: The distracted effects on latencies of alpha suppression measured averaged ERSP images of the motor components. A: the latency of alpha suppression in the left motor cluster across 8 subjects. B: the latency of alpha suppression in the right motor cluster across 6 subjects. Note: no apparently differences were observed among the five cases.

The comparison of the latency of ERSP time-lock to deviation-onset in the five cases was given in Fig. 4-27, which suggested that the latency of power increase in alpha band were not different with different time of SOA. The latency of power increase was following with the event onsets in both right and motor component.

D

Fig. 4-28: The grand mean of averaged ERSP images of the left (left block: 7 subjects) and right (right block: 5 subjects) occipital clusters for five cases. Panels as Fig. 4-22.

Note: the power at lower frequencies (0~8 Hz) was increased briefly after the onsets of the math presented. The most significant alpha power increase was occurred in case-4.

Since the experiment was using visual stimulus, then we could find the pattern inducing by visual stimulus (as shown in Fig. 4-28). In the occipital cluster, we found the lower frequency power which was increase time-locked to math onsets and the most significant alpha power increase which called rebound occurred in case-4. The lower frequency power was induced by P300 (event-related Potential (ERP)) amplitude, as shown in Fig. 4-29.

ERP (μV)

Time (ms)

ERP (μV)

Time (ms)

ERP (μV)

Time (ms)

ERP (μV)

Time (ms)

ERP (μV)

Time (ms)

ERP (μV)

Time (ms)

Fig. 4-29: The averaged ERPs of the left (left block: 7 subjects) and right (right block: 5 subjects) occipital clusters for five cases. Upper panels: the group averaged occipital independent component. Red line: case-1, blue line: case-2, green line:

case-3, black line: case-4, pink line: case-5. The bottom panels showed all ERP traces. The brown arrows indicated the location of the P300. Note the peak of P300 was time-locked to the onset of math presented.

Our results showed that independent component processes in the frontal cortex exhibited theta (5~7.8 Hz) and beta (12.2~17 Hz) increase that were consistent within subjects. Compared dual-task to single task, the total power in theta (5 ~ 7.8 Hz) band of dual-task was higher than single task. Similarly, the case which presented at 400ms before deviation was highest than that of the other dual-task cases. There was a time sifting at the onset of theta increase in the cases with math events. The beta (12.2~17

Hz) increase was induced by math. In occipital component, we found the pattern of inducing by visual stimulus and rebound which induced by pressing a button. In motor component there was all alpha suppression time-locked to the event onsets.