2026-10-14 京都大学
洪水リスクの伝え方による脳の反応の違い（出典：Fujimi et al., 2026、CC BY 4.0、改変）
＜関連情報＞
- https://www.kyoto-u.ac.jp/ja/research-news/2026-10-14-0
- https://www.sciencedirect.com/science/article/pii/S1053811926005689?via%3Dihub
⾃然災害リスクの体験型情報と記述型情報により活性化される異なる神経ネットワーク） Distinct neural networks for experiential and descriptive representations of natural hazard risk
Toshio Fujimi, Takanori Kochiyama, Wataru Sato
NeuroImage Available online: 23 September 2026
DOI:https://doi.org/10.1016/j.neuroimage.2026.122253
Highlights
- Flood videos produced a greater flood-related increase in subjective fear than maps.
- Flood videos showed greater flood-related right amygdala responses than maps.
- Flood maps showed stronger flood-related frontoparietal responses than videos.
- Video-favoring effects centered on one perceptual–salience-related component.
- Map-favoring effects spanned multiple components, including control-related ones.
Abstract
Environmental risks are communicated through experiential formats, such as pictures, videos, and virtual reality simulations, and descriptive formats, such as hazard maps and graphs, yet the neural systems engaged by these formats remain unclear. We acquired functional magnetic resonance imaging data from 45 healthy adults while they viewed first-person three-dimensional computer-generated flood videos and official flood maps, along with matched non-flood controls, in a within-subject 2 × 2 factorial design. Participants rated subjective fear after each stimulus. Neural responses were characterized using whole-brain general linear model analyses, anatomically defined region-of-interest analyses, and group independent component analysis. Flood videos produced a greater flood-related increase in subjective fear than flood maps. Relative to maps, videos elicited stronger flood-related responses in bilateral occipito-temporal and temporo-parietal cortices, sensorimotor and cingulate regions, pulvinar nuclei, and the cerebellum; the region-of-interest analysis further revealed greater right amygdala activation. In contrast, flood maps showed greater flood-related responses in bilateral frontal and parietal regions, including dorsolateral prefrontal and intraparietal areas commonly implicated in cognitive control and symbolic–spatial processing. Independent component analysis indicated that the video-favoring interaction was predominantly expressed in a single component overlapping regions commonly associated with perceptual and salience processing, whereas the map-favoring interaction was distributed across multiple components, including components resembling frontoparietal control networks. Taken together, these regional and network-level differences were broadly consistent with the hypothesized distinction between experiential and descriptive risk processing. The two formats may therefore offer complementary routes for flood-risk communication.
