Neural sensitivity to changes in naturally produced speech sounds: a comparison of different stimuli presentation paradigms

Previous studies show that neural sensitivity to variability in synthetic speech, measured as change detection with the mismatch negativity (MMN), is similar across stimuli presentation paradigms that vary in duration and in how the speech memory trace is constructed. Since listeners perceive naturallyproduced and computer-synthesized speech differently, likely due to the complex characteristics of natural speech that are not captured synthetically, results may not apply to natural speech. We examined neural sensitivity to naturally produced Dutch vowels varying in speaker, sex, accent, and vowel category, and compared canonical hour-long MMN paradigms with a novel paradigms lasting 15 minutes. Results showed that MMN amplitudes across paradigms were virtually identical, indicating that shorter, more efficient MMN paradigms can be successfully adopted to examine natural speech perception. This result has implications for investigating populations (e.g., children and clinical populations) where task duration is an important factor.