Exploring the transmission of cognitive task information through optimal brain pathways
Poster No:
1058
Submission Type:
Abstract Submission
Authors:
Zhengdong Wang1, Dazhi Yin2, yifeixue yang2, Ziyi Huang2, Kaiqiang Su2, Hengcheng Zhu3, Wanyun Zhao2
Institutions:
1East China Normal University, Shanghai, ShangHai, 2East China Normal University, Shanghai, Shanghai, 3University of Minnesota Twin Cities, Twin Cities, MN
First Author:
Co-Author(s):
Introduction:
Understanding the large-scale information processing is the central goal of cognitive neuroscience. Growing evidence shows that functions originate from intricate interactions among neural elements in the brain from micro- to macro-scale (Laughlin & Sejnowski, 2003; Pessoa, 2023). Emerging activity flow models demonstrate that cognitive task information is transferred by interregional functional or structural connectivity (Cole et al., 2016; Yan et al., 2021; Zhao et al., 2024). In contrast, graph-theory-based models typically assume that neural communication occurs via the shortest path of brain networks (Seguin et al., 2018). Whether the shortest path is the optimal route for cognitive information transmission remains unclear. Although brain network is devoted to the transmission of task information (Sydnor et al., 2021), empirical evidence is still lacking currently. Based on activity flow modeling, in this study, we tested the shortest path or direct path is the optimal route for empirical cognitive information transmission (Figure 1). Furthermore, a model with spatial embeddings was proposed to improve prediction performance.
Methods:
Dataset: The MRI data used for this study were collected from the HCP (https://www.humanconnectome.org). We selected 100 unrelated participants with both resting-state and task-state fMRI.
Activity flow modeling: Based on the model of activity flow mapping(Cole et al., 2016), cognitive task activation of a given brain region can be predicted by linearly summing the activation of all other regions, weighted by resting-state functional connectivity (FC). The formulation of this model is expressed as follows:
E_j=∑(i≠j∈V)A_i*F_ij
where Ej is the predicted activation for the target region j in a given task, Ai denotes the actual activation of region i in a given task, and Fij is the weight of the FC between regions i and j. The accuracy of activity flow prediction was evaluated using a similarity measure based on Pearson's correlation. Here, all types of activity flow routes were derived from sparse FC networks (i.e., 2%, 5%, 10%, 15%, 25%). The following statistical analyses were performed on the area under the curve (AUC) of the prediction accuracy (r) across all density values.
Shortest path calculation: A commonly used transformation from network connectivity to topological distance is Tuv = 1/Fuv, where Fuv represents the connectivity weight between nodes u and v, and Tuv denotes the topological distance. Dijkstra's algorithm was implemented to compute the shortest path length (SPL) based on a topological distance matrix for each network density. SPLwei and SPLbin indicates that SPL calculation was based on the weighted network and the binary network respectively.
Spatial embedding of activity routes: For a given parcellation template, the distance matrix, D\in R^{N\times N}, calculated by spatial coordinates of network nodes. Then, spatial embedding can be incorporated into the classical activity flow model as follows:
E_j=sum{i~=j in V}{A_iF_{ij}}{1}*{D_{ij}}
Activity flow modeling: Based on the model of activity flow mapping(Cole et al., 2016), cognitive task activation of a given brain region can be predicted by linearly summing the activation of all other regions, weighted by resting-state functional connectivity (FC). The formulation of this model is expressed as follows:
E_j=∑(i≠j∈V)A_i*F_ij
where Ej is the predicted activation for the target region j in a given task, Ai denotes the actual activation of region i in a given task, and Fij is the weight of the FC between regions i and j. The accuracy of activity flow prediction was evaluated using a similarity measure based on Pearson's correlation. Here, all types of activity flow routes were derived from sparse FC networks (i.e., 2%, 5%, 10%, 15%, 25%). The following statistical analyses were performed on the area under the curve (AUC) of the prediction accuracy (r) across all density values.
Shortest path calculation: A commonly used transformation from network connectivity to topological distance is Tuv = 1/Fuv, where Fuv represents the connectivity weight between nodes u and v, and Tuv denotes the topological distance. Dijkstra's algorithm was implemented to compute the shortest path length (SPL) based on a topological distance matrix for each network density. SPLwei and SPLbin indicates that SPL calculation was based on the weighted network and the binary network respectively.
Spatial embedding of activity routes: For a given parcellation template, the distance matrix, D\in R^{N\times N}, calculated by spatial coordinates of network nodes. Then, spatial embedding can be incorporated into the classical activity flow model as follows:
E_j=sum{i~=j in V}{A_iF_{ij}}{1}*{D_{ij}}
Results:
(1) We found that the AUC values of the prediction accuracy based on the shortest path (SPLwei and SPLbin) were significantly lower than those based on the direct path (Figure 2a).This finding suggests that the direct path outperformed the shortest path in cognitive information transmission.
(2) The accuracy of activity flow prediction was significantly enhanced (all ps < 0.05, Bonferroni-corrected) for both the direct and shortest paths when the physical distance was considered (Figure 2b-c). This result indicates that the spatial embedding of routes may play a crucial role in information transmission.
(2) The accuracy of activity flow prediction was significantly enhanced (all ps < 0.05, Bonferroni-corrected) for both the direct and shortest paths when the physical distance was considered (Figure 2b-c). This result indicates that the spatial embedding of routes may play a crucial role in information transmission.
Conclusions:
Our findings not only challenge the shortest path assumption through the use of empirical network models, but also suggest that routed information transmission might be modulated by spatial geometry. This study sheds light on the mechanistic relationships between cognitive task activation, resting-state network topology and spatial geometry.
Modeling and Analysis Methods:
Activation (eg. BOLD task-fMRI) 1
Classification and Predictive Modeling 2
Connectivity (eg. functional, effective, structural)
Methods Development
Keywords:
Computational Neuroscience
Computing
FUNCTIONAL MRI
Modeling
MRI
1|2Indicates the priority used for review
Abstract Information
By submitting your proposal, you grant permission for the Organization for Human Brain Mapping (OHBM) to distribute your work in any format, including video, audio print and electronic text through OHBM OnDemand, social media channels, the OHBM website, or other electronic publications and media.
The Open Science Special Interest Group (OSSIG) is introducing a reproducibility challenge for OHBM 2025. This new initiative aims to enhance the reproducibility of scientific results and foster collaborations between labs. Teams will consist of a “source” party and a “reproducing” party, and will be evaluated on the success of their replication, the openness of the source work, and additional deliverables. Click here for more information. Propose your OHBM abstract(s) as source work for future OHBM meetings by selecting one of the following options:
Please indicate below if your study was a "resting state" or "task-activation” study.
Healthy subjects only or patients (note that patient studies may also involve healthy subjects):
Was this research conducted in the United States?
Were any human subjects research approved by the relevant Institutional Review Board or ethics panel? NOTE: Any human subjects studies without IRB approval will be automatically rejected.
Were any animal research approved by the relevant IACUC or other animal research panel? NOTE: Any animal studies without IACUC approval will be automatically rejected.
Please indicate which methods were used in your research:
For human MRI, what field strength scanner do you use?
Provide references using APA citation style.
Laughlin, S. B., & Sejnowski, T. J. (2003). Communication in neuronal networks. Science, 301(5641), 1870-1874. https://doi.org/https://doi.org/10.1126/science.1089662
Pessoa, L. (2023). The Entangled Brain. Journal of Cognitive Neuroscience, 35(3), 349-360. https://doi.org/https://doi.org/10.1162/jocn_a_01908
Seguin, C., van den Heuvel, M. P., & Zalesky, A. (2018). Navigation of brain networks. Proceedings of the National Academy of Sciences of the United States of America, 115(24), 6297-6302. https://doi.org/https://doi.org/10.1073/pnas.1801351115
Sydnor, V. J., Larsen, B., Bassett, D. S., Alexander-Bloch, A., Fair, D. A., Liston, C., Mackey, A. P., Milham, M. P., Pines, A., Roalf, D. R., Seidlitz, J., Xu, T., Raznahan, A., & Satterthwaite, T. D. (2021). Neurodevelopment of the association cortices: Patterns, mechanisms, and implications for psychopathology. Neuron, 109(18), 2820-2846. https://doi.org/https://doi.org/10.1016/j.neuron.2021.06.016
Yan, T. Y., Liu, T. T., Ai, J., Shi, Z. Y., Zhang, J., Pei, G. Y., & Wu, J. L. (2021). Task-induced activation transmitted by structural connectivity is associated with behavioral performance. Brain Structure & Function, 226(5), 1437-1452. https://doi.org/https://doi.org/10.1007/s00429-021-02249-0
Zhao, W., Su, K., Zhu, H., Kaiser, M., Fan, M., Zou, Y., Li, T., & Yin, D. (2024). Activity flow under the manipulation of cognitive load and training. NeuroImage, 297, 120761. https://doi.org/https://doi.org/10.1016/j.neuroimage.2024.120761