Figure 1.
The presented images (images 1–3), saliency maps, and heat maps.
A–C: the three CT images presented to subjects, D–F: saliency maps, G–I: heat maps in neurologists, J–L: heat maps in controls. Presented CT images are the normal brain (A: image 1), cerebral hemorrhage from the putamen (B: image 2), and cerebral infarction due to embolism (C: image 3). Saliency maps reveal that the most outstanding areas are the ventricles and cistern (D), the large hemorrhagic area (E), and the region of physiological calcification (F). Heat maps in neurologists and controls reveal that the most frequently fixated areas are similar between the two groups in images 1 and 2 but not in image 3; specifically, they are the ventricles and cistern in image 1 (G, J), the large hemorrhagic area in image 2 (H, K), and the ACA infarction area for neurologists and the region of physiological calcification for controls in image 3 (I, L).
Figure 2.
The presented images (images 4–6), saliency maps, and heat maps.
A–C: the three CT images presented to subjects, D–F: saliency maps, G–I: heat maps in neurologists, J–L: heat maps in controls. Presented CT images are lacunar infarction (A: image 4), hyperacute cerebral infarction with old infarctions (B: image 5), and subarachnoid hemorrhage with acute subdural hemorrhage (C: image 6). Saliency maps reveal that the most outstanding areas are the ventricles (D, E, F). Heat maps reveal that neurologists gaze more frequently at the clinically important lesions than controls do in all images; these are the lacunar infarction area in image 4 (G), the hyperacute MCA infarction area in image 5 (H), and the acute subdural hemorrhagic area in image 6 (I).
Figure 3.
Cerebral hemorrhage from putamen (image 2).
A: Region of interest (ROI) surrounding the large hemorrhagic area, B: dwell time in the ROI (A), C: ROI surrounding the intra-ventricular hemorrhage, D: dwell time in the ROI (C). The dwell time in the large hemorrhagic area did not significantly differ between the two groups (B), whereas the dwell time in the intra-ventricular hemorrhage was significantly longer for neurologists than for controls (D).
Figure 4.
Cerebral infarction due to embolism (image 3).
A: ROI surrounding the large infarction area, B: dwell time in the ROI (A), C: ROI surrounding the ACA infarction area, D: dwell time in the ROI (C). The dwell time in the large infarction area did not significantly differ between the two groups (B), whereas the dwell time in the ACA infarction area was significantly longer for neurologists than for controls (D).