Presentation
Search Abstracts | Symposia | Slide Sessions | Poster Sessions | Lightning Talks
Microstructural underpinnings of top-down signals in cortical speech tracking
Poster D76 in Poster Session D, Wednesday, October 25, 4:45 - 6:30 pm CEST, Espace Vieux-Port
Saskia Helbling1,2; 1Ernst Strüngmann Institute (ESI) for Neuroscience, Frankfurt am Main, Germany, 2Max Planck Institute for Human Cognitive and Brain Sciences, Leipzig, Germany
Low-frequency oscillatory activity in the auditory cortex synchronizes to rhythmic patterns in the envelope of the acoustic input (Lakatos et al., 2005). Speech is quasi-rhythmic and neural activity in auditory cortex follows rhythmic features in the speech envelope, with prominent tracking occurring in the theta frequency range (4-8 Hz) for syllables rates and in the delta frequency range (0.5-4 Hz) for prosodic rhythm (Ahissar et al. 2001; Gross et al., 2013). Previous studies have investigated how speech tracking is modulated by interactions with oscillatory activity in other brain areas, revealing that the synchronisation of low-frequency oscillations and speech envelope is under the influence of top-down control from (pre-) motor and orbito-frontal areas (Park et al., 2015; Keitel et al., 2017). Here we aim to investigate the microstructural foundations of top-down control in speech comprehension by relating MRI-based estimates of local cortical myelination, (as a proxy of local cell density) to the strength of speech tracking across subjects. We hypothesise that higher myelination in areas linked to top-down control of speech tracking indicates stronger top-down signals and thus predicts stronger speech tracking in earlier auditory areas. We acquired MEG data from 45 healthy participants who listened passively to continuous speech from a short story (Das Schwimmteam, Miranda July, 11 min 23 sec). For each participant we also acquired high-resolution (0.8 mm isotropic) quantitative MRI maps using a multi-parameter map protocol (Weiskopf et al., 2013) on a 3T Connectome scanner. The hMRI toolbox in SPM was used to generate quantitative R1, R2* and MT maps (https://github.com/hMRI-group/hMRI-toolbox). Longitudinal relaxation (R1) values were sampled at 50% cortical depth, as R1 has been demonstrated to be a marker for cortical myelination. Speech tracking was assessed at the sensor-level by calculating the cerebro-acoustic coherence between the speech envelope and pre-processed brain signals converted to BIDS format, spanning a frequency range from 0.4 to 20 Hz. Consistent with previous findings, cluster-based permutation statistics in sensor space revealed significant speech tracking in the delta and theta frequency ranges, predominately at fronto-temporal and temporal sensors, respectively. We found that theta frequency band speech tracking during story listening was significantly correlated with cortical R1 values at bilateral early and associative auditory cortex, left inferior parietal cortex (IPC) and right hippocampus. Delta band sensor coherence correlated negatively with cortical R1 values in bilateral cingulate motor area, left inferior frontal cortex, including Broca area, as well as anterior cingulate areas. We suggest that the positive correlations at inferior parietal cortex for the theta band speech tracking may reflect the role of IPC in providing semantic constraints during language comprehension (Seghier et al., 2013), while the negative correlations at inferior frontal regions, bilateral cingulate motor and anterior cingulate areas may be related to motor predictions in speech perception (Morillon and Baillet, 2017). The unexpected direction of these negative correlations can potentially be explained by the lower cell density at mid cortical layers in cortical areas associated with top-down signals (Goulas et al., 2018).
Topic Areas: Speech Perception,