Studies On Working Memory And Training Effects Of Abacus-based Mental Calculation With Neuroimaging Techniques

This study explored the neurophysiological measures of individual differences in working memory (WM), as well as the influences of abacus-based mental calculation (AMC) training on WM and the underlying neural mechanism. Combined with various experimental paradigms,high temporal-resolution Event Related Potential (ERP) and Event Related Synchronization/Desynchronization (ERS/ERD) techniques were used to monitor the individual differences of WM. ERP can help to reflect the time course of cognitive operations. The analyses of ERP include ocular artifacts correction, baseline correction, low-pass filter and signal averaging.With regard to ERS/ERD,it refers to spectral oscillations induced by particular cognitive operations and can provide information about the dynamics of functional network formation.The ERS/ERD analyses concern performing the discrete Fourier transform on segmented signals,and computing the relative power variations within specific frequency band on brain regions ofinterest.Furthermore, a longitudinal design was conducted with the high spatial-resolution functional Magnetic Resonance Imaging (fMRI) technique to reveal the AMC training effect on WM at the neural level. Regularly, the preprocessing of fMRI data involves slice-timing,motion-correction,registration and smoothing. Further statistical analyses were then performedusing a general linear model. With respect to the network analyses, graph theory was applied to characterize the functional network through different topological properties.Study 1 investigated how individual differences of WM were reflected in neurophysiological responses with the electroencephalography (EEG) approach. A modified digit span (MODS) test and an n-back task were adopted in this study. The behavioral results revealed that participants with lower WM capacities were more easily affected by task loads.EEG results showed prominent differences among participants with variant WM capacities during cognitive operations both at the ERP and ERS/ERD patterns. Participants with greater WM exhibited larger P300 amplitude, greater frontal 0 ERS as well as lower parietal a ERD relative to low WM participants. The results suggest that there are qualitative neural differences among individuals with different WM when performing cognitive operations. Individuals with high WM capacities may make more efficient use of neural resources to keep their attention focused on the task-relevant information.Study 2 examined the impact of AMC training on verbal and visuospatial WM of young adults as well as the underlying neural basis using the task-state fMRI approach. A memory span test and a visuospatial n-back task were employed in this study. The results showed that the AMC trained participants achieved superior verbal WM capacities relative to their pears. Also,the performance in the visuospatial WM task was significantly improved after AMC training,and the performance gains were observed to be correlated with the activation decreases in the right frontoparietal circuitry and left occipitotemporal junction. These findings suggest that the AMC training effect could transfer to WM functions, and such transfer might attribute to the functional plasticity of visuospatial neural substrates.Combined with the rest-state fMRI approach and the graph theory analysis, study 3 explored whether long-term AMC training had any influences on WM functions and related spontaneous network properties in children. N-back paradigms with different types of stimuli were adopted in this study. The results reveal that AMC training helps to improve the development of children's WM extensively, as well as enhance the small-worldness of WM network and the efficiency of information communication within it.The current study reveals the behavioral performance and neurophysiological responses of individuals with different WM in cognitive operations (particularly in high WM load tasks),filling the blank of this issue in the field of EEG research. As the first longitudinal neuroimaging research on AMC training, this study demonstrate that the advantages of AMC in calculating abilities could transfer to domain-general WM functions, extending the findings of previous AMC researches. In addition, the current results provide strong supports for the functional plasticity of WM network, and may shed lights on the neural mechanism underlying far transfer effects.