Speech processing around the auditory cortex: the distinct contribution of gray and white matter
Poster No:
1041
Submission Type:
Abstract Submission
Authors:
Qiuhui Bi1,2, Peipei Qin1, Gaolang Gong1
Institutions:
1State Key Laboratory of Cognitive Neuroscience and Learning & IDG/McGovern Institute for Brain Resea, BeijIng, China, 2School of Artificial Intelligence, Beijing Normal University, BeijIng, China
First Author:
Qiuhui Bi
State Key Laboratory of Cognitive Neuroscience and Learning & IDG/McGovern Institute for Brain Resea|School of Artificial Intelligence, Beijing Normal University
BeijIng, China|BeijIng, China
State Key Laboratory of Cognitive Neuroscience and Learning & IDG/McGovern Institute for Brain Resea|School of Artificial Intelligence, Beijing Normal University
BeijIng, China|BeijIng, China
Co-Author(s):
Peipei Qin
State Key Laboratory of Cognitive Neuroscience and Learning & IDG/McGovern Institute for Brain Resea
BeijIng, China
State Key Laboratory of Cognitive Neuroscience and Learning & IDG/McGovern Institute for Brain Resea
BeijIng, China
Gaolang Gong
State Key Laboratory of Cognitive Neuroscience and Learning & IDG/McGovern Institute for Brain Resea
BeijIng, China
State Key Laboratory of Cognitive Neuroscience and Learning & IDG/McGovern Institute for Brain Resea
BeijIng, China
Introduction:
Understanding the structural substrates of functional activity is one of the major challenges in neuroscience. Putatively, the microcircuit within gray matter (GM) serves as the basis of local neural activity, and axons within white matter (WM) supports rapid transfer of information across remote brain regions. Previous studies have revealed distinct contribution of GM/WM structural properties to specific functional activation, depending on brain regions and hemispheres [1,2,3]. In this study, we aim to ascertain how the underlying GM/WM contribute to functional activity of the planum temporale (PT) and Heschl's gyrus (HG) during auditory speech tasks, with respect to different hemispheres and speech processing component.
Methods:
Subjects. 782 right-handed participants with high-quality T1w structural MRI, diffusion MRI and language fMRI from the human connectome project (S1200 release) were included.
ROI delineation. The ROIs of PT/HG (Fig 1A) were manually delineated following a well-established procedure by Altarelli et al [4].
Functional measures. We applied a widely-used LI toolbox [5] approach to quantify functional activation of left and right ROIs and the asymmetry indexes (AIs). For three contrasts in the language processing task, factor analysis was performed on the PT/HG activation left, right, and AI across the three contrasts, separately, resulting two main factors: speech perception and comprehension.
Structure measures. For GM, the total surface area, average thickness, myelin content, neurite density index (NDI) and orientation dispersion index (ODI) across all vertices were calculated for each ROI. For WM, tractography of the long and posterior segments of the arcuate fasciculus (AF) was conducted using MRtrix3 [6] with multiple ROIs (Fig 1B, C), separately for PT and HG. Average fibre length, NDI and ODI were calculated for each tract. The AI of structural measure was computed as (L – R) / (L + R)
Statistical analysis. General linear models were applied to each functional activation of PT/HG, with left/right activation and functional AI as the dependent variable, and structural left/right measures and AIs as the independent variables. Dominance analyses were performed to evaluate the contribution (general dominance weight, GDW) of structural measures to the variance of functional measures. To determine whether the degree of GM/WM measures' contribution to functional measure differ between the left and right hemisphere, as well as between the ipsilateral PT and HG, permutation tests were then performed.
ROI delineation. The ROIs of PT/HG (Fig 1A) were manually delineated following a well-established procedure by Altarelli et al [4].
Functional measures. We applied a widely-used LI toolbox [5] approach to quantify functional activation of left and right ROIs and the asymmetry indexes (AIs). For three contrasts in the language processing task, factor analysis was performed on the PT/HG activation left, right, and AI across the three contrasts, separately, resulting two main factors: speech perception and comprehension.
Structure measures. For GM, the total surface area, average thickness, myelin content, neurite density index (NDI) and orientation dispersion index (ODI) across all vertices were calculated for each ROI. For WM, tractography of the long and posterior segments of the arcuate fasciculus (AF) was conducted using MRtrix3 [6] with multiple ROIs (Fig 1B, C), separately for PT and HG. Average fibre length, NDI and ODI were calculated for each tract. The AI of structural measure was computed as (L – R) / (L + R)
Statistical analysis. General linear models were applied to each functional activation of PT/HG, with left/right activation and functional AI as the dependent variable, and structural left/right measures and AIs as the independent variables. Dominance analyses were performed to evaluate the contribution (general dominance weight, GDW) of structural measures to the variance of functional measures. To determine whether the degree of GM/WM measures' contribution to functional measure differ between the left and right hemisphere, as well as between the ipsilateral PT and HG, permutation tests were then performed.
Results:
The explained variance for each structural measure is illustrated in Figure 2A. As shown, GM measures of the left ROIs consistently explained a larger proportion of the variance for the speech perception, but a less proportion of the variance for the speech comprehension. Particularly, the model of speech comprehension was dominated by the fibre length of long AF segment in both left PT and HG, accounting for 41.9% and 47.6% of the all explained variance separately. In contrast, GM and WM measures of the right ROIs explained similar amount of variance for both speech perception and comprehension.
For the speech comprehension, WM measures explained significantly greater proportion of variance of the left ROI than the right ROI (Fig 2B): PT, L-R=0.07, P<0.001; HG, L-R=0.03, P<0.01. The premutation test further revealed greater explained variance of WM measures in the left PT, compared with the left HG for both speech perception (PT-HG=0.04, P=0.01) and comprehension (PT-HG=0.05, P=0.01). In addition, PT WM asymmetric measures captured a higher proportion of the variance of speech comprehension lateralization (PT-HG=0.03, P=0.03), compared with the HG (Fig 2B).
For the speech comprehension, WM measures explained significantly greater proportion of variance of the left ROI than the right ROI (Fig 2B): PT, L-R=0.07, P<0.001; HG, L-R=0.03, P<0.01. The premutation test further revealed greater explained variance of WM measures in the left PT, compared with the left HG for both speech perception (PT-HG=0.04, P=0.01) and comprehension (PT-HG=0.05, P=0.01). In addition, PT WM asymmetric measures captured a higher proportion of the variance of speech comprehension lateralization (PT-HG=0.03, P=0.03), compared with the HG (Fig 2B).
Conclusions:
The present study demonstrated distinct contribution of GM and WM structural measures to speech processing activity around the auditory cortex, depending on the selected region of interest, hemisphere, and functional components.
Language:
Speech Perception 1
Language Other 2
Novel Imaging Acquisition Methods:
Anatomical MRI
Diffusion MRI
Non-BOLD fMRI
Keywords:
Hemispheric Specialization
MRI
Other - Speech processing; Planum temporale (PT); Arcuate fasciculus (AF)
1|2Indicates the priority used for review
Provide references using author date format
[2] Ocklenburg, S. et al. (2018), ‘Neurite architecture of the planum temporale predicts neurophysiological processing of auditory speech’, Science Advances, vol. 4, no. 7, Art. number 7
[3] Hickok, G. et al. (2007), ‘The cortical organization of speech processing’, Nature Reviews Neuroscience, vol. 8, no. 5, pp. 393-402
[4] Altarelli, I. et al. (2014), ‘Planum Temporale Asymmetry in Developmental Dyslexia: Revisiting an Old Question’, Human Brain Mapping, vol. 35, no. 12, pp. 5717–5735
[5] Wilke, M. et al. (2007), ‘LI-tool: A new toolbox to assess lateralization in functional MR-data’, Journal of Neuroscience Methods, vol. 163, no. 1, pp. 128–136
[6] Tournier, J.-D. et al. (2019), ‘MRtrix3: A fast, flexible and open software framework for medical image processing and visualisation’, NeuroImage, vol. 202, pp. 116137