Poster No:
792
Submission Type:
Abstract Submission
Authors:
Jacob Paul1, Gilles de Hollander2
Institutions:
1University of Melbourne, Parkville, Victoria, 2University of Zurich, Zurich, ZH
First Author:
Jacob Paul
University of Melbourne
Parkville, Victoria
Co-Author:
Introduction:
Numerical cognition relies on being able to enumerate and compare the numerosity of multiple sets of objects. Human intraparietal cortices are known to encode numerosity of single sets through topographically organised (i.e., numerotopic) maps (Harvey, 2017; Paul, 2022). Here we aimed to explore how the human brain encodes the numerosity of multiple sets presented simultaneously. We hypothesised that the numerosity of concurrent stimulus arrays might be represented separately and kept neurally distinct with visuospatial (i.e., retinotopic) coding (Groen, 2021). We test this hypothesis using ultra-high-field (7T) fMRI together with a tailored retinotopic mapping and numerical decision-making task and Bayesian decoding approach. If the numerosity of concurrent stimulus arrays can indeed be decoded from parietal cortex based on their retinotopic position, this will advance our understanding of how the human brain flexibly integrates higher-order stimulus features from multiple relevant stimuli.
Methods:
Eight healthy adults (4 female) each completed 16 runs (32 trials each run, 512 trials total) of a numerical decision-making task. On each trial, two sets of black/white intermixed dots (n = 5-25) were shown on the left and right sides of a grey screen for 0.6s. A blank screen followed for 4s during which time participants were instructed to maintain central fixation. A leftward/rightward-pointed arrow post-cued participants to report the numerosity of either set. Responses were collected on a number line with feedback provided after each trial.
Data were acquired on a 7T Siemans Magnetom with a 32-channel head coil. T1-weighted anatomical scans (MP2RAGE) had 0.75mm isotropic resolution. T2*-weighted functional scans (2D EPI) had 1.6mm isotropic resolution resulting in 84 slices, 130x130 voxels, TR=800ms, TE= 22ms, and flip-angle=45 degrees. The first four functional scans were discarded to ensure a steady signal state. Reverse phase top-up scans were collected for distortion correction.
Population receptive field (pRF) models were estimated from task BOLD time-series (Dumoulin, 2008), with a focus on parietal maps (NPC). Bayesian inversion of these encoding models (García-Barretto, 2023; de Hollander, 2024) was used to decode the maximum likelihood numerosity representation at each voxel given the observed BOLD response.
Results:
Participants were accurate overall, with an average absolute error for left-cued stimuli of 3.1 (SD=0.62) and right-cued stimuli of 3.16 (SD=0.62). Left stimuli were underestimated (-0.52, SD=1.0) while right stimuli were overestimated (0.48, SD=0.73), t(7)=-3.00, p=0.013. If we fit two 1D pRFs simultaneously, preferred numerosities correlated across voxels (t(7)=7.86, p<0.001 for left NPC, t(7)=3.89, p=0.006 for right NPC). This is evidence for a model where numerosity tuning is the same but there is more/less tuning for left/right. The weight for numerosities on the left is slightly higher in right NPC (50%, SD=8.3%) compared to left NPC (44%, SD=9.9%), t(7)=-2.05, p=0.039, one-sided.
When we invert the PRF model we can decode the left numerosity and right numerosity at the same time above chance, and although the average correlation is modest these correlations are reliably different from 0 (p-values are one-sided): right NPC decoding left stimuli (r=0.11, SD=0.10, t(7)=3.35, p=0.006) and left NPC decoding right stimuli (r=0.10, SD=0.12, t(7)=2.43, p=0.023). Lastly, decoding accuracy in right NPC predicts behavioural accuracy of left-cued stimuli (r(8)= -0.84, p=0.004), and similarly left NPC predicts right-cued stimuli (r(8)= -0.61, p=0.055).
Conclusions:
Our findings show how retinotopic and numerotopic coding in human association cortices support flexible representation of multiple distinct numerosities. They also suggest numerical choice options in decision-making tasks can be presented concurrently, allowing for more efficient experimental designs. Shifts in visual attention during numerical decisions could be further considered.
Higher Cognitive Functions:
Space, Time and Number Coding 1
Modeling and Analysis Methods:
Activation (eg. BOLD task-fMRI)
Bayesian Modeling 2
Novel Imaging Acquisition Methods:
BOLD fMRI
Perception, Attention and Motor Behavior:
Perception: Visual
Keywords:
Cognition
FUNCTIONAL MRI
HIGH FIELD MR
Perception
Vision
1|2Indicates the priority used for review
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.
I accept
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:
I do not want to participate in the reproducibility challenge.
Please indicate below if your study was a "resting state" or "task-activation” study.
Task-activation
Healthy subjects only or patients (note that patient studies may also involve healthy subjects):
Healthy subjects
Was this research conducted in the United States?
No
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.
Yes
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.
Not applicable
Please indicate which methods were used in your research:
Functional MRI
Behavior
Computational modeling
For human MRI, what field strength scanner do you use?
7T
Which processing packages did you use for your study?
AFNI
Free Surfer
Other, Please list
-
fMRIPrep
Provide references using APA citation style.
Barretto-García, M. (2023). Individual risk attitudes arise from noise in neurocognitive magnitude representations. Nature Human Behaviour, 7(9), 1551-1567.
de Hollander, G. (2024). Rapid Changes in Risk Preferences Originate from Bayesian Inference on Parietal Magnitude Representations. bioRxiv, 2024-08. doi: https://doi.org/10.1101/2024.08.23.609296
Dumoulin, S. O. (2008). Population receptive field estimates in human visual cortex. Neuroimage, 39(2), 647-660.
Groen, I. I. (2022). Visuospatial coding as ubiquitous scaffolding for human cognition. Trends in Cognitive Sciences, 26(1), 81-96.
Harvey, B.M. (2017). A network of topographic numerosity maps in human association cortex. Nature Human Behaviour, 1(2), 0036.
Paul, J. M. (2022). Numerosity tuning in human association cortices and local image contrast representations in early visual cortex. Nature Communications, 13(1), 1340.
No