PET imaging of the normal human auditory system: responses to speech in quiet and in background noise.

The neural mechanisms involved in listening to sentences, and then detecting and verbalizing a specific word are poorly understood, but most likely involve complex neural networks. We used positron emission tomography to identify the areas of the human brain that are activated when young, normal hearing males and females were asked to listen to a sentence and repeat the last word from the Speech in Noise (SPIN) test. Listening conditions were (1) Quiet, (2) Speech, (3) Noise, and (4) SPIN with stimuli presented monaurally to either the left ear or the right ear. The least difficult listening task, Speech, resulted in bilateral activation of superior and middle temporal gyrus and pre-central gyrus. The Noise and SPIN conditions activated many of the same regions as Speech alone plus additional sites within the cerebellum, thalamus and superior/middle frontal gyri. Comparison of the SPIN condition versus Speech revealed additional activation in the right anterior lobe of the cerebellum and right medial frontal gyrus, near the cingulate. None of the left ear-right ear stimulus comparison revealed any significant differences except for the SPIN condition that showed greater activation in the left superior temporal gyrus for stimuli presented to the right ear. No gender differences were observed. These results demonstrate that repeating the last word in a sentence activates mainly auditory and motor areas of the brain when Speech is presented, whereas more difficult tasks, such as SPIN or multi-talker Noise, activate linguistic, attentional, cognitive, working memory, and motor planning areas.