The new approach to brands has streamlined the creation of marketing strategies and customer relationships to achieve emotionally strong insights. Audio-branding is hardly used and is difficult to measure its help in creating customer connection with brands. To close this gap, we propose a deep learning model with mixed learning and central affinity, for the characterization of emotions in audio-branding by compressing electroencephalographic (EEG) signals obtained from the brain activation of an individual subjected to auditory brand strategies used for its branding. The results show that the proposed model allowed the emotional characterization of audio-branding for five brands (Microsoft Windows, Valve, MasterCard, THX, and XBOX) taking as a reference a set of 20 basic sounds, facilitating the iterative construction of identity and brand management, helping to strengthen or readjust brand marketing strategies. This neural model, given its capacity of adaptation and learning, can be applied in other spheres of action as an innovative solution to the characterization of emotions from auditory images.