Distinct Top-down Induction Mechanisms Of Auditory Neural Representations

Sensory representation can be reactivated without any external stimulation,such as in mental imagery.Previous MEG and fMRI studies have found that imagined speaking(imagine oneself speaking)and imagined hearing(imagine oneself hearing)are two ways to internally reconstruct the auditory representations,separately indicating the mechanisms of motor-simulation-based and memory-retrieval-based routes in brains.In this study,we employed articulation imagery(AI)and hearing imagery(HI)in two experiments to further investigate the behavioral and neural representations of the proposed distinct neural pathways.Experiment 1 investigated the behavioral aftereffects of imagery tasks on speech perception.Seven-level /ba/-/da/ continuum was used as sound stimuli.Participants needed to classify the sounds they heard to either /ba/ or /da/ after the imagery tasks.The results showed that participants biased their judgements to the syllable in the AI condition more than that in the HI condition.Experiment 2 used two types of silent video clips to selectively induce the two proposed top-down induction mechanisms.fMRI results showed that AI induced greater activity in cortical areas related to speech production,including the left precentral and triangular part of the inferior frontal gyrus.Auditory cortices containing bilateral superior temporal gyrus,bilateral middle temporal gyrus and left temporal pole were also more activated in the AI condition,suggesting that motor-simulation-based processes produced stronger auditory reactivation.Whereas a distributed frontal-parietal neural network that had been implicated in memory retrieval was observed during HI,including the left medial superior frontal gyrus,bilateral middle frontal gyrus,bilateral superior parietal gyrus,left inferior parietal gyrus,left supramarginal gyrus and precuneus.These consistent behavioral and neuroimaging results suggested that motorsimulation-based and memory-retrieval-based pathways can internally induce auditory representations.However,they are implemented in different functional neural networks and can distinctly interact with the bottom-up information to shape auditory perception.