Abstract:Human listeners can reliably recognize speech in complex listening environments. The underlying neural mechanisms, however, remain unclear and cannot yet be emulated by any artificial system. In this dissertation, we study how speech is represented in the human auditory cortex and how the neural representation contributes to reliable speech recognition. Cortical activity from normal hearing human subjects is noninvasively recorded using magnetoencephalography, during natural speech listening. It is first demonstrated that neural activity from auditory cortex is precisely synchronized to the slow temporal modulations of speech, when the speech signal is presented in a quiet listening environment. How this neural representation is affected by acoustic interference is then investigated. Acoustic interference degrades speech perception via two mechanisms, informational masking and energetic masking, which are addressed respectively by using a competing speech stream and a stationary noise as the interfering sound. When two speech streams are presented simultaneously, cortical activity is predominantly synchronized to the speech stream the listener attends to, even if the unattended, competing speech stream is 8 dB more intense. When speech is presented together with spectrally matched stationary noise, cortical activity remains precisely synchronized to the temporal modulations of speech until the noise is 9 dB more intense. Critically, the accuracy of neural synchronization to speech predicts how well individual listeners can understand speech in noise. Further analysis reveals that two neural sources contribute to speech synchronized cortical activity, one with a shorter response latency of about 50 ms and the other with a longer response latency of about 100 ms. The longer-latency component, but not the shorter-latency component, shows selectivity to the attended speech and invariance to background noise, indicating a transition from encoding the acoustic scene to encoding the behaviorally important auditory object, in auditory cortex. Taken together, we have demonstrated that during natural speech comprehension, neural activity in the human auditory cortex is precisely synchronized to the slow temporal modulations of speech. This neural synchronization is robust to acoustic interference, whether speech or noise, and therefore provides a strong candidate for the neural basis of acoustic background invariant speech recognition.

On Cortical Coding of Vocal Communication Sounds in Primates

Temporal coding of speech in human auditory cortex

Cortical Neural Coding of Speech in Simple and Complex Auditory Scenes.

Neural coding strategies in auditory cortex.

Spectral and temporal integration in the auditory cortex of awake primates and its implications for coding of species-specific communication sounds

Cortical Coding of Auditory Features.

Dynamics of auditory-vocal interaction in monkey auditory cortex.

Spectral and Temporal Integration in the Auditory Cortex of Alert Primates

Vocalization Induced Cfos Expression in Marmoset Cortex

Vocal processing networks in the human and marmoset brain

Two organizing principles of vocal production: Implications for nonhuman and human primates

Differential Representation of Species-Specific Primate Vocalizations in the Auditory Cortices of Marmoset and Cat.

Recent advances in understanding the auditory cortex

Frontal cortex activity underlying the production of diverse vocal signals during social communication in marmoset monkeys

Cortical Representations of Pitch: Theories and Experiments

Cortical Representations of Pitch in Monkeys and Humans

Frontal cortex activity during the production of diverse social communication calls in marmoset monkeys

Cortical representations of pitch in monkeys and humans.

Harmonic Template Neurons In Primate Auditory Cortex Underlying Complex Sound Processing

Representing the dynamics of natural marmoset vocal behaviors in frontal cortex

Neural representation of pitch in auditory cortex