Objective：To examine changes in whole-brain neuromagnetic networks in children with self-limited epilepsy with centrotemporal spikes（SeLECTS）and assess the association between these network alterations and cognitive performance. Methods： Magnetoencephalography（MEG）data were collected from 63 unmedicated children with SeLECTS and 30 healthy controls（HC）. Corrected amplitude envelope correlation（AEC-c）analysis was performed to estimate the whole-brain oscillatory functional connectivity（FC）. Cognitive function was assessed in all children using the fourth edition of the Wechsler Intelligence Scale for Children（WISC-Ⅳ）. Spearman’s correlation analysis was applied to determine the relationship between AEC-c values and WISC-Ⅳ indices. Results：Compared to the HC group，children with SeLECTS exhibited reduced FC in several networks，including the left entorhinal cortex and the left rostral anterior cingulate in the theta band，the right lateral orbitofrontal cortex and the right pericalcarine in the beta band，and the right fusiform gyrus and the left inferior parietal gyrus in the gamma2 band（all P ＜ 0.05）. Conversely， increased FC was observed between the right fusiform gyrus and the left parahippocampal gyrus，the left insular gyrus and the left rostral anterior cingulate in the theta band，and the left postcentral gyrus and the right cuneus in the gamma1 band（P ＜ 0.05）Regarding the relationship between FC and cognitive performance，in the theta band，WMI scores showed a positive correlation with the AEC-c value between the right fusiform gyrus and the left parahippocampal gyrus（r=0.255，P ＜ 0.05）. Additionally，in the theta band， PSI scores were positively correlated with the AEC-c value between the left insular gyrus and the left rostral anterior cingulate（r= 0.258；P ＜ 0.05）. In the gamma1 band，a negative correlation was found between PRI scores and the AEC-c value of the left postcentral gyrus and the right cuneus（r=- 0.266，P ＜ 0.05）. Conclusion：The FC networks of children with SeLECTS exhibited significant differences in theta，beta，and gamma bands compared to the HC group. These altered brain regions were located in key nodes of cognition-related networks，suggesting early mechanistic insights into cognitive dysfunction. The AEC-c value of enhanced FC network showed linear correlation with cognitive-related metrics，highlighting MEG’s potential for assessing cognitive impairment.