Sanli S., Keleş H. O.
SENSORS, cilt.26, sa.8, ss.1-17, 2026 (SCI-Expanded, Scopus)
-
Yayın Türü:
Makale / Tam Makale
-
Cilt numarası:
26
Sayı:
8
-
Basım Tarihi:
2026
-
Doi Numarası:
10.3390/s26082388
-
Dergi Adı:
SENSORS
-
Derginin Tarandığı İndeksler:
Scopus, Science Citation Index Expanded (SCI-EXPANDED), Compendex, INSPEC, MEDLINE, Directory of Open Access Journals
-
Sayfa Sayıları:
ss.1-17
-
Ankara Üniversitesi Adresli:
Evet
Özet
Gentleness, defined as the ability to handle tissues delicately while minimizing unnecessary force, is a critical indicator of surgical proficiency. Objective and real-time assessment of gentleness in virtual reality (VR)-based training can improve the understanding of both psychomotor and cognitive components of surgical skill. This study evaluates and classifies participants’ gentleness during VR-based laparoscopic simulations using fNIRS-derived hemodynamic features. Twenty-three volunteers with no prior laparoscopic experience performed a VR-based double-grasper task while hemodynamic activity over frontal and motor cortical regions was recorded using eighteen fNIRS channels. In parallel, subjective workload (NASA-TLX), error counts, and gentleness performance score (GPS) were collected. Temporal features, including slope, root mean square, and standard deviation, were extracted from the fNIRS signals and used to train multiple machine learning classifiers. Performance labels were binarized into low and high groups using median splits of the gentleness performance score. Models were evaluated using stratified 5-fold cross-validation. Results revealed stronger right-frontal HbO activity and increased left-motor HbR responses in the low-performance group, suggesting higher cognitive effort and less efficient motor strategies. Across classifiers, slope-based features consistently outperformed variability- and amplitude-based metrics. The highest classification performance was achieved using HbR slope features with Random Forest classifiers (accuracy ≈ 0.85, AUC up to 0.93). These findings highlight the potential of fNIRS-based metrics for automated performance assessment in VR surgical training.