Healthy individuals differ in their ability to perform a variety of visual tasks. Observers also vary in their capability to improve visual skill with practice, a phenomenon called Visual Perceptual Learning (VPL) which represents an example of plasticity in the adult brain, and a core feature of visual cognition. Several lines of evidence indicate that initial performance and rate of learning tend to be inversely related. Thus, individuals who perform better initially tend to exhibit slower improvement. Indeed, in our previous study on learning a difficult shape identification task, we identified a behavioral descriptor of learning, called ‘task-fitness’, capturing initial performance, rate of learning, and time to reach the learning criterion. Specifically, we observed that subjects with good task-fitness exhibited high initial accuracy and learned the task in fewer blocks of practice but at a lower rate of improvement. Conversely, subjects with poor task-fitness showed lower initial accuracy took longer to reach criterion but their rate of improvement was higher. Critically, we found that the degree of task-fitness was positively correlated with the strength of pre-training resting state functional connectivity (FC) between stimulus-related visual regions such as good learners exhibited higher FC while bad learners showed lower degree of FC. More recently, in a subsequent work we employed repetitive transcranial magnetic stimulation (rTMS) over the visual occipital (i.e. right V2d/V3 and right Lateral Occipital, LO) and parietal (i.e. right posterior Intraparietal sulcus, pIPS) regions, previously shown to be modulated, to investigate their causal role in learning the shape identification task (i.e. the same experimental paradigm of Lewis and coll., 2009). We reported that interference with V2d/V3 and LO increased reaction times to learned stimuli as compared to both pIPS and Sham control conditions thus supporting the causal role of the visual network in the control of the perceptual learning. Grounded in this framework, in the present study, exploiting the high temporal resolution of electroencephalography (EEG), we want to evaluate how the neurophysiological correlates, both in the time domain (evoked potentials) and in the frequency domain (brain rhythms), are modified following an intensive training. Moreover, here we will investigate whether pre-training resting state can predict the individual attitude to learn a novel task by estimating several markers such as EEG microstates. For this purpose, we will record both spontaneous and task-evoked electrical brain activity at the beginning and at the end of the learning process and we will compare the respective markers extracted from the EEG signals.