Neural representations of relational demands emerge in higher-order networks' alpha and beta rhythms
Poster No:
786
Submission Type:
Abstract Submission
Authors:
Conor Robinson1, Luca Cocchi1, Takuya Ito2, Luke Hearne1
Institutions:
1QIMR Berghofer Medical Research Institute, Herston, Queensland, 2T.J. Watson Research Center, IBM Research, Yorktown Heights, NY
First Author:
Co-Author(s):
Introduction:
Relational complexity (RC) theory quantifies problem complexity by the number of relations processed in a single-step simultaneously (Halford et al., 1998). While previous research shows increased neural activity in higher-order networks during relational reasoning (Hearne et al., 2015), how the brain represents varying relational demands remains unknown. Using representational similarity analysis (RSA) on two independent functional magnetitic resonance imaging (fMRI) and electroencephalography (EEG) datasets collected during a visuospatial reasoning task, we aimed to disambiguate RC-specific representations from those reflecting general cognitive effort (CE), determine where and when these representations emerge in the brain, and examine how they relate to individual differences in performance (Robinson et al., 2024).
Methods:
Data from 85 healthy participants (18-35 years) collected across two datasets completed the Latin Square Task ((Birney et al., 2006), Figure 1A) were included in the final analysis: 7T fMRI (n=40) (Hearne et al., 2017) and 64-channel EEG (n=45). fMRI data were preprocessed using fMRIPrep and FreeSurfer, with puzzle-specific activations estimated using whole-brain general linear modeling. EEG data were preprocessed using EEGLAB and filtered into frequency bands (theta: 4-8 Hz, alpha: 8-12 Hz, beta: 13-30 Hz, gamma: 30-45 Hz). To compare datasets, neural activity was transformed into representational dissimilarity matrices (RDMs, Figure 1A) (Kriegeskorte et al., 2008): fMRI data were parcellated into 376 regions (Glasser et al., 2016) grouped into thirteen functional networks (Ji et al., 2019), while EEG data were analysed at the sensor-level in 100 ms segments. These RDMs were analysed individually and combined to identify RC and CE representations and their relationship to performance Figure 1B.
Results:
LST Task Performance
Both datasets showed robust complexity effects, with significant declines in accuracy (fMRI: F(2,78) = 80.61, p < .001; EEG: F(2,88) = 101.60, p < .001) and increases in response time (fMRI: F(2,78) = 51.09, p < .001; EEG: F(2,88) = 108.00, p < .001). Performance measures showed high correspondence between datasets (accuracy: rs = .95, p < .001; response time: rs = .83, p < .001).
Multimodal and Cross-modal Evidence for Neural Representations encoding RC
fMRI RSA showed RC was encoded across the brain (Figure 2A, left 367/376 regions, pFDR < .05), with peak representations in prefrontal and parietal areas. EEG RSA revealed that RC representations emerged late in the trial period (2.1 - 4.2s, pFDR < .05), specifically in alpha and beta frequency bands (Figure 2A, right).
Using model-based EEG-fMRI fusion analysis (Figure 2B), we found that higher-order networks showed the strongest cross-modal representational similarity, with peak commonality emerging around 3.6-3.8s in alpha and beta bands, where RC accounted for most of the shared variance between imaging modalities.
RC neural representations are linked to task performance
Exploratory analyses revealed that better task performance was associated with weaker RC representations in both frontoparietal and cingulo-opercular networks and reduced peak beta-band representation (Figure 2C).
Both datasets showed robust complexity effects, with significant declines in accuracy (fMRI: F(2,78) = 80.61, p < .001; EEG: F(2,88) = 101.60, p < .001) and increases in response time (fMRI: F(2,78) = 51.09, p < .001; EEG: F(2,88) = 108.00, p < .001). Performance measures showed high correspondence between datasets (accuracy: rs = .95, p < .001; response time: rs = .83, p < .001).
Multimodal and Cross-modal Evidence for Neural Representations encoding RC
fMRI RSA showed RC was encoded across the brain (Figure 2A, left 367/376 regions, pFDR < .05), with peak representations in prefrontal and parietal areas. EEG RSA revealed that RC representations emerged late in the trial period (2.1 - 4.2s, pFDR < .05), specifically in alpha and beta frequency bands (Figure 2A, right).
Using model-based EEG-fMRI fusion analysis (Figure 2B), we found that higher-order networks showed the strongest cross-modal representational similarity, with peak commonality emerging around 3.6-3.8s in alpha and beta bands, where RC accounted for most of the shared variance between imaging modalities.
RC neural representations are linked to task performance
Exploratory analyses revealed that better task performance was associated with weaker RC representations in both frontoparietal and cingulo-opercular networks and reduced peak beta-band representation (Figure 2C).
Conclusions:
Here we demonstrate that RC is explicitly encoded in neural representations across both space and time, with strongest representations in higher-order networks in alpha and beta frequency bands. We found these neural representations were better explained by RC than general CE and were linked to individual differences in performance. This suggests that during the LST, the brain primarily organises information according to RC rather than general task difficulty.
Higher Cognitive Functions:
Reasoning and Problem Solving 1
Higher Cognitive Functions Other
Modeling and Analysis Methods:
Activation (eg. BOLD task-fMRI) 2
EEG/MEG Modeling and Analysis
Keywords:
Cognition
Electroencephaolography (EEG)
FUNCTIONAL MRI
Other - complexity; reasoning; representational similarity analysis; frontoparietal network
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?
Which processing packages did you use for your study?
Provide references using APA citation style.
Glasser, M. F., et al (2016). A multi-modal parcellation of human cerebral cortex. Nature, 536(7615), 171-178. https://doi.org/10.1038/nature18933
Halford, G. S., et al (1998). Processing capacity defined by relational complexity: implications for comparative, developmental, and cognitive psychology. Behavior Brain Science, 21(6), 803-831; 831-864. https://doi.org/10.1017/s0140525x98001769
Hearne, L., et al (2015). Interactions between default mode and control networks as a function of increasing cognitive reasoning complexity. Human Brain Mapping, 36(7), 2719-2731. https://doi.org/https://doi.org/10.1002/hbm.22802
Hearne, L. J., et al (2017). Reconfiguration of Brain Network Architectures between Resting-State and Complexity-Dependent Cognitive Reasoning. Journal of Neuroscience, 37(35), 8399-8411. https://doi.org/10.1523/JNEUROSCI.0485-17.2017
Ji, J. L., et al (2019). Mapping the human brain's cortical-subcortical functional network organization. Neuroimage, 185, 35-57. https://doi.org/10.1016/j.neuroimage.2018.10.006
Kriegeskorte, N., et al (2008). Representational similarity analysis - connecting the branches of systems neuroscience. Frontiers in Systems Neuroscience, 2. https://doi.org/10.3389/neuro.06.004.2008
Robinson, C., et al (2024). Relational integration demands are tracked by temporally delayed neural representations in alpha and beta rhythms within higher-order cortical networks. bioRxiv, 2024.2010.2016.618779. https://doi.org/10.1101/2024.10.16.618779