Humans detect and react to characteristics of timbre in speech and instrumental sounds, but the relationship between emotions conveyed by timbre in non-verbal vocalizations and those conveyed by electric guitar sounds is unknown. For this study, I created a series of sounds with varying timbre characteristics: non-verbal voice sounds and electric guitar sounds. Sounds were validated through categorization of emotion and ratings of intensity and believability. I found that sounds with low (slow onset) attack slope were most likely to be categorized as angry, while sounds with high (fast onset) attack slope were most likely to be categorized as happy. I collected EEG data from participants while they made judgements on the emotional similarity of guitar sounds (primes) when compared with vocal sounds (targets). I conducted Multifractal Detrended Fluctuation Analysis and MANOVAs, and I found systematic differences of the multifractal spectrum of EEG responses between conditions (emotion and sound type) that would be obscured by other forms of analysis. This information could be applied to development of more effective and psychologically healthy entertainment (music, film, etc.). It is also applicable in therapy situations in which there is a need to induce a certain neurological or emotional state.