We measured musicians and non-musicians by using structural magnetic resonance imaging to investigate relationships between cortical features of the left planum temporale (PT) and the categorization of consonant-vowel (CV) syllables and their reduced-spectrum analogues. The present work is based on previous functional studies consistently showing that the left PT is particularly responsive to transient acoustic features in CV syllables and their reduced-spectrum analogues, and on striking evidence pointing to structural alterations of the left PT as a function of musicianship. By combining these two observations, we hypothesized to find that differences in cortical surface area (SA) and cortical thickness (CT) of the left PT in musicians may facilitate the categorization of fast-changing phonetic cues. Behavioural results indicated that musicians and non-musicians achieved a comparable performance in the categorization of CV syllables, whereas the musicians performed significantly better than the controls in the more demanding reduced-spectrum condition. This better behavioural performance corresponds to an increased cortical SA of the left PT in musicians compared to non-musicians. No differences in CT of the left PT were found between groups. In line with our predictions, we revealed a positive correlation between cortical SA of the left PT in musicians and the behavioural performance during the acoustically more demanding reduced-spectrum condition. Hence, we provide first evidence for a relationship between musical expertise, cortical SA of the left PT, and the processing of fast-changing phonetic cues.