Determining the Significant Kinematic Features for Characterizing Stress during Surgical Tasks Using Spatial...

Area of Science:

• Medical simulation and surgical training
• Surgical skill assessment
• Human factors in surgery

Background:

• Intraoperative stress negatively affects surgical skills, particularly in novice surgeons.
• Stress increases instrument tip velocity, acceleration, and jerk, leading to less smooth movements.
• The most sensitive kinematic marker for stress detection remains unclear.

Purpose of the Study:

• To identify the most significant kinematic feature affected by intraoperative stress.
• To develop a classifier to distinguish between normal and stressed surgical movements.
• To analyze the contribution of kinematic features to stress classification.

Main Methods:

• Utilized a spatial attention-based Long-Short-Term-Memory (LSTM) classifier.
• Collected kinematic data from novice surgeons performing a peg transfer task under normal and stressed conditions.
• Employed Leave-One-User-Out (LOUO) cross-validation for model testing.

Main Results:

• The LSTM classifier achieved 77.11% accuracy in distinguishing normal from stressed movements.
• Velocity and acceleration showed higher attention for normal movements.
• Jerk on the non-dominant hand demonstrated the most significant attention increase for stressed movements.

Conclusions:

• Jerk on the non-dominant hand is an effective indicator of stress in novice surgeons.
• Kinematic analysis, particularly jerk, can objectively assess surgical stress.
• This finding aids in developing stress-detection tools for surgical training.