Supplement
Video S1: Representative platelet retraction intensity (1, 2, 3). Citrate-anticoagulated whole blood with ADP-stimulated platelets was perfused. Representative retracting aggregate formations are shown in detail for the introduced intensity levels of 1, 2, and 3 in time laps. The original reacting times illustrated in the video were 3min (level 1) and 6min (level 2 + 3). The arrows indicate the direction of flow.
Video S2: Course of aggregate formation and embolization on different surfaces. Citrate-anticoagulated whole blood with ADP-stimulated platelets was perfused over four surfaces (glass, VWF, Carbothane, and Pelethane) at a flow rate 40ml/h. Surface-specific aggregate formation and embolization around the stagnation point is shown over 7min in time lapse.
Video S3: Dissolving aggregate formations after haltingADP inflow. Citrate-anticoagulated whole blood with ADP-stimulated platelets was perfused over a VWF-coated surface at a flow rate of 80ml/h for 10min. The time-lapse video starts when the ADP inflow was halted; the aggregates subsequently begin to dissolve and were carried away by the current.
Video S4: Course of aggregate formation and embolization with a washed blood cell suspension.A washed blood cell suspension with ADP-stimulated platelets was perfused over four surfaces (glass, VWF, Carbothane, and Pelethane) at a flow rate of 40ml/h (cf. video 1). Surface-specific aggregate formation and embolization around the stagnation point is shown over 7min in time lapse. Without plasma proteins in the fluid, the stability of the aggregates changed significantly compared to the experiments shown in video 1.
Figure S1: Image segmentation.The still images (12bit) from flow chamber experiments acquired with video-fluorescence microscopy show animage section at the stagnation point. The top right image shows the segmentation with an optimal threshold, whereas the images in the middle and bottom row illustratethe influence of thresholdvariations of +/-50 and +/- 100 grayscale values on the segmentation. As anoptimal grayscale threshold a compromise value was determined taking into account two factors: 1) the impact of bright shining larger aggregates and 2) including most of the small aggregates. The corona of bright shining larger aggregates should be excluded by raising the threshold. The arrows indicate examples of the variance of the dividing line at a bright corona with different thresholds. At the same time more small aggregates are detected by lowering the threshold. The situation shown with the thresholds of 1000 and 1200 represent the extreme segmentations. A threshold was selected between the both extremes. The bar in the top left image represents 50 µm.