Computational modeling of MEG event-related beta oscillations during auditory verb generation
Poster No:
1470
Submission Type:
Abstract Submission
Authors:
Minarose Ismail1, Davide Momi2, Zheng Wang3, Cathy Barr4, Randy McIntosh5, John Griffiths6, Darren Kadis4
Institutions:
1University of Toronto, Toronto, Ontario, 2CAMH, Toronto, Ontario, 3Centre for Addiction and Mental Health, Toronto, Ontario, 4Hospital for Sick Children, Toronto, Ontario, 5Simon Fraser University, Vancouver, BC, 6Centre for Addiction and Mental Health, Toronto, MT
First Author:
Co-Author(s):
Introduction:
Lateralization of low beta (13-23Hz) event-related desynchrony (ERD) and synchrony (ERS) during verb generation in MEG provides a robust assay of language hemispheric dominance1. In young children, low beta ERD and ERS are observed bilaterally, with lateralization emerging in early adolescence2,3. These task-related oscillatory changes could serve as neural signatures of language network maturation, as well as its potential for plasticity. However, the neural mechanisms underlying lateralization remain unclear. Exploring the emergence of macro-scale brain activity from anatomical network structure and micro-scale neuronal dynamics using biologically inspired neurocomputational models holds promise in addressing this gap.
Methods:
Individual connectome-based neural mass models were constructed for 10 healthy adolescents (15-18 years old) using the Jansen-Rit model4 and individual structural connectivity matrices derived from multi-shell diffusion-weighted MRI tractography, segmented with the Shen 200-node atlas5 . The auditory event was parameterized as a 40ms square-wave stimulus injected bilaterally into neural masses representing left and right Heschl's gyri. This model was used to fit trial-averaged MEG time series representing the early (-100-400ms) auditory evoked response in a verb generation task using the Whole-Brain Modelling in PyTorch (WhoBPyt)6 library, estimating local and global coupling parameters. For each individual, two sets of models were created: a verb model fitted with verb generation trials data, and a noise model fitted with speech-shaped noise trials data. These models were then used to simulate 1200ms epochs, and power spectral densities (PSDs) were computed using Welch's method for the 700-1200ms time window that, critically, was not used for fitting. Simulated and empirical PSDs in left and right frontal ROIs were then averaged across subjects and compared.
Results:
Brain network models for verb and noise conditions accurately captured auditory evoked field topographies at both the group average and the individual subject level (Figure 1). Interestingly, these early auditory-evoked models were also able to accurately predict the out-of-sample late (700-1200ms) effect of a left-lateralized low beta ERD and a right-lateralized low beta ERS (Figure 2). Consistent with empirical data, averaged PSDs from left frontal ROIs in the verb models correctly predicted a decrease in beta power relative to the noise models (ERD), while averaged PSDs of the right frontal ROIs predicted the observed increase in beta power relative to the noise models (ERS).
·Figure 1
·Figure 2
Conclusions:
We present the first individualized whole-brain model of auditory verb generation, demonstrating validity by reproducing the temporal dynamics of sensory-evoked activity, and its potential in predicting the lateralization of late oscillatory responses within higher-order language regions. Importantly, the model replicated late (700-1200ms) induced spectral effects after only being fit on early (-100-400ms) evoked response time series. Our findings suggests that language lateralization is encoded in the interaction between the structure and the dynamics of early sensory responses to language stimuli.
Language:
Language Other 2
Modeling and Analysis Methods:
Classification and Predictive Modeling 1
Connectivity (eg. functional, effective, structural)
Keywords:
Computational Neuroscience
Language
MEG
Modeling
1|2Indicates the priority used for review
Provide references using author date format
2. Sharma, V., Vannest, J., Greiner, H.M., Fujiwara, H., Tenney, J.R., Williamson, B. J., Kadis, D.S. (2021). ‘Beta synchrony for expressive language lateralizes to right hemisphere in development’, Scientific Reports 11, 3949. https://doi.org/10.1038/s41598-021- 83373-z.
3. Kadis, D. S., Pang, E. W., Mills, T., Taylor, M. J., McAndrews, M. P., & Smith, M. L. (2011). ‘Characterizing the normal developmental trajectory of expressive language lateralization using magnetoencephalography’. Journal of the International Neuropsychological Society, 17(5). doi:10.1017/S1355617711000932.
4. Jansen, B. H., & Rit, V. G. (1995). ‘Biological Cybernetics Electroencephalogram and visual evoked potential generation in a mathematical model of coupled cortical columns’. Biol. Cybern, 73. https://doi.org/10.1007/BF00199471.
5. Griffiths, J. D., Wang, Z., Ather, S. H., Momi, D., Rich, S., Diaconescu, A., McIntosh, A. R., & Shen, K. (2022). ‘Deep Learning-Based Parameter Estimation for Neurophysiological Models of Neuroimaging Data’, bioRxiv. https://doi.org/10.1101/2022.05.19.492664
6. Shen X, Tokoglu F, Papademetris X, Constable RT. (2013) ‘Groupwise whole-brain parcellation from resting-state fMRI data for network node identification’. Neuroimage, 82. https://doi.org/10.1016/j.neuroimage.2013.05.081.