Tracking information flow during cognitive tasks using time-resolved transfer entropy
Poster No:
1525
Submission Type:
Abstract Submission
Authors:
Chetan Gohil1, Oliver Cliff1, Ben Fulcher2, James Shine3, Joseph Lizier3
Institutions:
1University of Sydney, Sydney, NSW, 2University of Sydney, Sydney, Australia, 3The University of Sydney, Sydney, NSW
First Author:
Co-Author(s):
Introduction:
Modern theories of cognitive neuroscience rely on concepts related to information processing. Quantifying the directional pathway of information and relating this to cognitive task behaviour is an important step in understanding how the brain supports cognition. However, the capacity to quantify, and hence formalise, information flow has been relatively underdeveloped. Here, we make progress towards this.
First, we tackle a key methodological challenge that arises when applying an information-theoretic approach to functional magnetic resonance imaging (fMRI) data, namely the estimation of the underlying probability distribution. Using a working memory task from the Human Connectome (HCP) fMRI dataset [1], we show how this approach allows us to quantify information processing during task. Furthermore, we show how spatial patterns of information processing are dynamically reconfigured in different cognitive states. This approach provides a novel method of understanding brain function.
First, we tackle a key methodological challenge that arises when applying an information-theoretic approach to functional magnetic resonance imaging (fMRI) data, namely the estimation of the underlying probability distribution. Using a working memory task from the Human Connectome (HCP) fMRI dataset [1], we show how this approach allows us to quantify information processing during task. Furthermore, we show how spatial patterns of information processing are dynamically reconfigured in different cognitive states. This approach provides a novel method of understanding brain function.
Methods:
We characterise information processing using two measures [2,3]: active information storage (AIS), which quantifies how much past information is being retained at a particular brain region, and transfer entropy (TE), which quantifies the directional flow of information between pairs of regions. See Figure 1A for definitions.
An implicit assumption that is often overlooked in calculating these measures is that the observed data provides a good estimate of the underlying probability distribution (Figure 1A). Deciding how to estimate this probability distribution is a key challenge, particularly in task fMRI. This is because task data contains a limited number of samples and is highly non-stationary, switching from one task condition to the next.
We propose a novel solution to overcome this issue by pooling across all the data we have for a subject, including both task and rest (Figure 1B). This approach provides us a means to dynamically track information processing in response to task. Such a description is currently absent and severely needed to understand the link between information processing and cognition.
We apply these measures to a visual image N-back task. We studied 100 subjects. First, the data was preprocessed with bias field and motion correction, then parcellated to 333 regions of interest [4]. Next, we normalised, detrended and bandpass filtered the data (0.01-0.08 Hz). Finally, we deconvolved the haemodynamic response [5] from the data to minimise any time-lag effects.
An implicit assumption that is often overlooked in calculating these measures is that the observed data provides a good estimate of the underlying probability distribution (Figure 1A). Deciding how to estimate this probability distribution is a key challenge, particularly in task fMRI. This is because task data contains a limited number of samples and is highly non-stationary, switching from one task condition to the next.
We propose a novel solution to overcome this issue by pooling across all the data we have for a subject, including both task and rest (Figure 1B). This approach provides us a means to dynamically track information processing in response to task. Such a description is currently absent and severely needed to understand the link between information processing and cognition.
We apply these measures to a visual image N-back task. We studied 100 subjects. First, the data was preprocessed with bias field and motion correction, then parcellated to 333 regions of interest [4]. Next, we normalised, detrended and bandpass filtered the data (0.01-0.08 Hz). Finally, we deconvolved the haemodynamic response [5] from the data to minimise any time-lag effects.
Results:
We found the N-back task altered information processing (AIS and TE) in the brain. Figure 1C (left) shows the changes from rest (blue) to task (red). Each dot is a brain region averaged over subjects. There is a clear global shift in information processing that has been induced by the task. We see an increase in TE in and out of each region and a decrease in AIS, suggesting in task information is flowing more between regions than being stored locally.
Furthermore, contrasting the 2-back vs 0-back conditions (Figure 1C right), we find the spatial patterns of information processing are altered in going from a low to high cognitive load. We see AIS is decreased, and TE has increased, particularly in higher-order brain regions associated with working memory [6].
Furthermore, contrasting the 2-back vs 0-back conditions (Figure 1C right), we find the spatial patterns of information processing are altered in going from a low to high cognitive load. We see AIS is decreased, and TE has increased, particularly in higher-order brain regions associated with working memory [6].
Conclusions:
We introduce a new method for calculating information-theoretic measures that is more suited to the analysis of task neuroimaging data. This method allows us to identify the regions involved in information processing relevant to a task and dynamically quantify the information flows. We found the spatial patterns of information processing (AIS, TE) dynamically reconfigure in a working memory task and in response to different cognitive loads. By providing a dynamic task-relevant characterisation of information processing, the approach outlined in this work offers a powerful new methodology for studying in the brain.
Modeling and Analysis Methods:
Activation (eg. BOLD task-fMRI) 2
Methods Development 1
Keywords:
Cognition
Computational Neuroscience
Data analysis
FUNCTIONAL MRI
Modeling
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:
Provide references using APA citation style.
[2] Cover, T. M. (1999). Elements of information theory. John Wiley & Sons.
[3] Bossomaier, T., Barnett, L., Harré, M., Lizier, J. T., Bossomaier, T., Barnett, L., ... & Lizier, J. T. (2016). Transfer entropy (pp. 65-95). Springer International Publishing.
[4] Gordon, E. M., Laumann, T. O., Adeyemo, B., Huckins, J. F., Kelley, W. M., & Petersen, S. E. (2016). Generation and evaluation of a cortical area parcellation from resting-state correlations. Cerebral cortex, 26(1), 288-303.
[5] Wu, G. R., Liao, W., Stramaglia, S., Ding, J. R., Chen, H., & Marinazzo, D. (2013). A blind deconvolution approach to recover effective connectivity brain networks from resting state fMRI data. Medical image analysis, 17(3), 365-374.
[6] Chai, W. J., Abd Hamid, A. I., & Abdullah, J. M. (2018). Working memory from the psychological and neurosciences perspectives: a review. Frontiers in psychology, 9, 401.