The proposed study will investigate the sensorimotor transformation between auditory and articulatory neural representations of speech during phonological short-term memory. This transformation, which is also essential in speech feedback control for error correction, speech development, and simple repetition, is thought to occur in area Spt (Sylvian-parietal-temporal) of the dominant posterior planum temporale, however the detailed neural mechanisms have not been fully elucidated and a recent claim for bi-hemispheric sensorimotor integration as been made. In this study, high-density direct cortical recordings, or electrocorticography (ECoG), will be obtained from the peri-Sylvian frontal, parietal, and temporal lobes in awake, behaving humans. These subjects are undergoing surgical treatment for medically refractory epilepsy with electrodes implanted for seizure foci localization. By employing multiple behavioral and computational approaches, this study will 1) localize sensorimotor transformations in phonological short-term memory and 2) determine the mechanism of these phonological sensorimotor transformations. During a repetition task, subjects will be acoustically presented with words and pseudo-words of variable load (ranging from 1 to 9 syllables), or spectrally matched non-speech sounds. After a delay period of variable length (range 2 to 15 seconds), subjects will be cued to overtly repeat the stimulus. Spectrotemporal analysis of lime-locked ECoG signals will identify cortical areas involved and tier sequence of activation. Functional connectivity and directed transfer function analysis will explore interactions between cortical areas and test the hypothesis that a `phonological loop' underlies short- term maintenance of phonologic information. Neurobehavioral correlations between behavioral parameters (performance), experimental parameters (stimulus load, delay duration, word versus pseudo-word versus non- speech), and neural parameters (activation strength, cortical network connectivity, information flow directionality) will identify those cortial areas involved in maintenance of phonologic information. By analyzing neural state-space trajectories with high temporal resolution we will test a hypothesized mechanism for phonological sensorimotor transformation. Elucidating the neural mechanisms of sensorimotor integration in the context of the phonological loop will add to our knowledge of how and where the brain performs this vital transformation. This will lead to a better understanding of disorders such as conductive aphasia and developmental stuttering, which are thought to be disorders of sensorimotor integration. Additionally, understanding this transformation may lead to development of better speech-based brain-computer interfaces.
Speech involves the perception and extraction of phonological information from an acoustic signal, the production of dynamic and tightly coordinated articulatory gestures, and a transformation between these auditory and articulatory neural representations. The proposed study will investigate how the brain performs this sensorimotor integration, and what its role is in short term phonological memory. The knowledge gained about speech processing from these experiments will advance our understanding of speech development, communication disorders and will have implications for speech-based brain-computer interfaces
| Breshears, Jonathan D; Hamilton, Liberty S; Chang, Edward F (2018) Spontaneous Neural Activity in the Superior Temporal Gyrus Recapitulates Tuning for Speech Features. Front Hum Neurosci 12:360 |
