We propose a novel approach for the reconstruction of functional networks representing brain dynamics based on the idea that the coparticipation of two brain regions in a common cognitive task should result in a drop in their identifiability, or in the uniqueness of their dynamics. This identifiability is estimated through the score obtained by deep learning models in supervised classification tasks and therefore requires no a priori assumptions about the nature of such coparticipation. The method is tested on EEG recordings obtained from Alzheimer’s and Parkinson’s disease patients, and matched healthy volunteers, for eyes-open and eyes-closed resting–state conditions, and the resulting functional networks are analysed through standard topological metrics. Both groups of patients are characterised by a reduction in the identifiability of the corresponding EEG signals, and by differences in the patterns that support such identifiability. Resulting functional networks are similar, but not identical to those reconstructed by using a correlation metric. Differences between control subjects and patients can be observed in network metrics like the clustering coefficient and the assortativity in different frequency bands. Differences are also observed between eyes open and closed conditions, especially for Parkinson’s disease patients. Author Summary Functional networks are becoming a standard tool in neuroscience, thanks to their ability to extract the interactions between brain regions and representing them as simple mathematical objects. Many metrics for reconstructing these networks have so far been proposed, based on different assumptions about how those interactions manifest in the recorded data, and consequently depicting a limited part of the picture. We here introduce a more general approach, based on the idea that brain regions should have a unique dynamic, which becomes more similar when they coparticipate in a cognitive task, and on the quantification of such uniqueness through deep learning models.