Influence Of Abacus-based Mental Calculation Training On Children’s Numerical Processing With ERP Evidence

This study examined whether long-term abacus-based mental calculation (AMC) training improved numerical processing efficiency and at what stage of information processing the effect appeared. Different experimental paradigms and Event Related Potential (ERP) recording techniques were used to monitor the temporal dynamics during numerical processing.Study 1 examined the efficiency of calculation between the AMC group and the control group and whether this difference was transferred to other capabilities in the IQ test.Study 2 explored whether there was difference in sensitivity to numbers between abacus and the control group, using the Rapid Serial Visual Presentation (RSVP) paradigm. The results indicated that there was no difference in the effect of Attentional Blink between the two groups, suggesting numbers are not any special stimuli to the abacus children.Study 3 explored whether there was difference in the efficiency of accessing numerical representation between the two groups, utilizing the number compare paradigm. The behavioral results showed that the Distance Effect (DE) was smaller in the abacus group than the control group. The ERP results showed significant DE in bilateral parietal sites in abacus group, both in N1 and P2p amplitudes, while the control group only showed DE in left parietal site in P2p amplitude. These results showed a higher efficiency and larger involvement of parietal cortex in extracting numerical representation for the abacus children.Study 4 adopted a numerical Stroop task that required to determine the numerical magnitude (NC task) or the physical size (PC task) of two numbers presented simultaneously. In the NC task, the abacus group showed faster response times but similar accuracy compared to the control group. In the PC task, the two groups exhibited the same speed and accuracy. With regards to ERP results, the abacus group displayed congruity effects both in the earlier (N1) and later (N2 and LPC (late positive component) time domain, while the control group only displayed congruity effects for LPC. In the left parietal region, LPC amplitudes were larger for the abacus than the control group. Individual differences for LPC amplitudes over left parietal area showed a positive correlation with RTs in the NC task.Study 5 explored whether there were differences between the alerting and executive control network between the two groups, utilizing the ANT paradigm. The results showed there were no difference existed in the two networks of attention between the two groups.Our results suggested that the influence of AMC training on numerical processing mainly happened in the stage of accessing numerical representation. In the perception and more general executive stages, the training did not show significant effects.