Fig 1.
Hierarchical semantic decomposition of surgical activity.
e denotes that the segment can be performed using either of the robotic arms, b denotes that the segment is performed using both the robotic arms.
Fig 2.
Maneuver and gesture counts for overall task by skill level.
GRS = global rating score; E = expert; I = intermediate; N = novice. The error bars illustrate 95% bootstrap intervals. We observed statistically significant differences in the total count for maneuvers and the count for incomplete maneuvers between expert and novice trials by GRS-based definition but not using experience based-definition for surgical skill. We observed statistically significant differences in the total count for gestures and error gestures between expert and novice trials per both experience- and GRS-based definitions for surgical skill.
Fig 3.
Total gesture counts within individual maneuvers by skill level.
ST = suture throw; GPR = grasp-pull-run suture through tissue; OLK = one-loop knot; TLK = two-loop knot; IMS = inter-maneuver segment; E = expert; I = intermediate; N = novice. The error bars illustrate 95% bootstrap intervals. The Y-axis show the total number of gestures performed within each maneuver. The bar height represents the mean value, and the error bars represent 95% bootstrap intervals. We observed statistically significant differences in the total gesture counts for ST using a GRS-based skill definition, and for GPR and TLK using both experience- and GRS-based definitions for surgical skill.
Fig 4.
Counts for specific gestures within maneuvers by skill level.
DN = drive needle; GN = grasp needle; RN = rotate needle out of tissue; PN = pull needle; AD = adjustment gestures; LS = loop suture; GT = grasp suture tail through loop; PT = pull suture tail through loop; TK = pull ends of suture to tighten knot. The Y-axis shows the counts of a particular gesture (X-axis) performed within each maneuver. The dots in the Fig 4 represent mean values and the bars represent 95% bootstrap intervals. We observed statistically significant differences in the counts for ST_DN and OLK_TK between expert and novice trials by GRS-based definition for skill, for ST_GN, ST_RN, TLK_LS, TLK_GT, TLK_PT, TLK_TK, and TLK_AD between expert and novice trials by both experience- and GRS-based definitions for skill.
Fig 5.
State flow plots showing transitions among maneuvers in a surgeon’s knot and among gestures in its constituent maneuvers by skill level.
ST = suture throw; IMS = inter-maneuver segment; GPR = grasp-pull-run suture through tissue; OLK = one-loop knot; TLK = two-loop knot; GRS = global rating score; DN = drive needle; GN = grasp needle; RN = rotate needle out of tissue; PN = pull needle; AD = adjustment gestures; LS = loop suture; GT = grasp suture tail through loop; PT = pull suture tail through loop; TK = pull ends of suture to tighten knot. An arrow in the Fig 5 represents a transition from a maneuver shown at the tail to the maneuver shown at the head of the arrow. The thickness of the arrow corresponds to the fraction of total transitions arising out of the maneuver from which the arrow originates.