Change blindness refers to the substantial difficulty of observers to detect changes between successive presentations of images that are separated by a brief blank. Using the EEG technique, the main goal of our study was to investigate the neural correlates of change blindness. Adding a five level confidence rating scale to the task, our hypothesis was that the variation in the level of confidence reflects different neural change related responses. The behavioural data showed that slower reaction times (RT's) accounts for low confidence level rather than implicit detection for missed changes. Conducting the classic components analysis we identified three distinct effects: an early negative deflection, a posterior negative peak and the P3 component. The first event-related potential identified on the Pz electrode, an early negative 70-130ms component marks a significant difference between no-change condition and both missed and detected changes. Also, between 250-350ms (N2pc) a significant difference was registered on PO7 and PO8 electrodes which separated both low confidence missed changes and detected changes from no-change and high confidence missed changes. We identified the P3 component between 500-700ms on almost all the electrodes chosen for this analysis. A significant difference between low confidence and both no-change and high confidence missed changes marked this component. The cluster analysis we performed revealed different map activation between both no-change and low confidence missed changes compared to detected changes and low confidence missed changes concomitant with the N2pc (250-350ms component). Also simultaneously with the N2pc, the source localisation by LAURA revealed that both detected changes and low confidence missed changes relative to high confidence missed changes indicated right occipital and right parietal neural sources. The N2pc which is directly linked to the detection of a change can also be elicited by low confidence missed trials suggesting that participants had some awareness of the change because the attention shift drew their gaze on the change. As in the case of the N2pc, we have identified a similar P3 wave deflection for this condition as for detected changes. Following the results of the different analyses presented here we can draw the conclusion that observers can be aware of the change and still say they did not see the change because of a low confidence level.