| Paper: | PS-1A.5 |
| Session: | Poster Session 1A |
| Location: | Symphony/Overture |
| Session Time: | Thursday, September 6, 16:30 - 18:30 |
| Presentation Time: | Thursday, September 6, 16:30 - 18:30 |
| Presentation: |
Poster
|
| Publication: |
2018 Conference on Cognitive Computational Neuroscience, 5-8 September 2018, Philadelphia, Pennsylvania |
| Paper Title: |
Voxel-wise Modeling with Spatial Regularization: Application to Semantic Mapping during Natural Listening |
| Manuscript: |
Click here to view manuscript |
| DOI: |
https://doi.org/10.32470/CCN.2018.1088-0 |
| Authors: |
ÖZGÜR YILMAZ, EMİN ÇELİK, TOLGA ÇUKUR, Bilkent University, Turkey |
| Abstract: |
Voxel-wise modeling (VM) is a powerful tool that is used to estimate responses of individual voxels evoked by features of complex natural stimuli. Still, VM discards spatial correlations across functional selectivities of neighboring voxels, and this can lead to decreased sensitivity during model estimation with noisy measurements. Here, we describe a spatially-informed voxel-wise modeling (SPIN-VM) method that utilizes response correlations in spatial neighborhoods of voxels. SPIN-VM performs regularization both across spatial neighborhoods of voxels and across model features of individual voxels. We compared the performance of SPIN-VM to regular VM on a dataset collected from a natural listening experiment. Compared to VM, SPIN-VM leads to higher prediction performances and better capturing of local coherence of semantic representations with SPIN-VM. |
