Preparation for speech perception and production functionally dissociates ventral, dorsal and articulatory pathways in the human brain

Introduction: Current thinking divides speech processing in speech perception and speech recognition. A dual-stream model of speech processing proposes a left-lateralized dorsal stream mapping acoustic speech signals to frontal articulatory networks aiding in speech perception. The bilaterally organized ventral stream is supposed to subserve speech recognition (Hickok and Poeppel, Nature Reviews Neuroscience 2007). Yet, overt articulation does not engage the dorsal stream, at least for overt reading (Kell et al., Brain 2009 and Kell et al., Cerebral Cortex 2010). We set out to disentangle the functional roles of the dorsal and ventral speech processing stream and their contribution to lateralization by comparing preparation for and execution of speech processing with speech production.

Methods: We performed a functional MRI experiment on 28 healthy right-handed participants. Our cue-target paradigm involved four different conditions: reading sentences overtly or covertly, or observing consonant or symbol strings that had a sentence-like structure. Three different speech levels were dissociable and contrasted against symbols as visual control condition: Consonant strings were used as unutterable pseudo-words (pre-lexical speech processing). To target speech comprehension, German sentences were presented. In the overt speech condition, these sentences had to be articulated. The cue informed subjects about the upcoming task and implicitly contained information about the expected stimulus type. Subjects were familiarized with the stimuli and tasks to allow cognitive association of a cue with a specific stimulus type and task. The variable instruction delay (2–4s) allowed for a temporal dissociation of task preparation from execution. The EPI images were preprocessed and contrastes in SPM8 and results family wise error corrected for multiple comparisons at p<0.05.

Results and discussion: Covert reading targets the bilateral ventral and the left-lateralized dorsal language streams, while overt speech production involves a third anterior bilateral articulatory network of prefrontal, temporal and central regions. This argues against a mere articulatory role for the dorsal stream but implies that speech comprehension takes advantage of pre-articulatory speech representations. While sensorimotor integration seems to be key in lateralization of speech production, our data suggest a more linguistic mechanism in dorsal stream lateralization during speech recognition.