Honghua Chen

NeuralRNN unifies RNN methods for cognitive neuroscience tasks by providing a single framework for model implementation, comparison, and extension.

This paper identifies two distinct temporal timescales organizing human vocal behavior, showing they are present from infancy and support different functions in vocal production and synchronization.

The paper shows that the auditory cortex uses efficient coding to encode time intervals in speech, with neural responses adapting to duration statistics and exhibiting compressive nonlinearity.

The paper shows low-frequency neural activity reflects context-dependent parsing of word sequences, tracking sentences rather than individual words.

The paper investigates how the auditory system segregates speech from echoes and encodes them as separate streams, showing that this process restores missing temporal modulations in echoic speech.