Assessment of competence in surgical skills using functional magnetic resonance imaging: a feasibility study

Patient safety is fundamental to modern medical practice; safe surgery saves lives. Ensuring surgical competence is becoming more difficult at a time when surgeons are being trained in fewer hours. Accurate objective assessment of technical skills ability is lacking in standardization. Functional magnetic resonance imaging (fMRI) has a long history in neuroscience, psychiatry, and cognitive studies. Many studies have explored levels of perceived expertise in sports and musical ability. Little has been published on actual rather than perceived motor skills. This study sought to assess the feasibility of utilizing a novel assessment method by measuring blood oxygen level-dependent signal changes (BOLD) in specific brain regions via fMRI during a surgical skills task.

Images were acquired using fMRI in a pilot study of 9 subjects (3 experts, 3 intermediates, and 3 novices) when performing and imagining performing a basic surgical procedure: hand tying of surgical knots. Level of expertise was based on years of experience and clinical grade. The quality and quantity of knots were assessed objectively by 2 experts who were independent of the study and blinded to the ability of the candidate. The effect of subject head motion caused by the task itself was assessed. The efficacy of fMRI data analyses in removing artifacts caused by this noise source in the data was explored.

Shifts of less than 1 voxel (3 × 3 × 3.55 mm(3)) were recorded in all participants and were successfully corrected in all cases in the fMRI preprocessing step. Decreased BOLD activity was observed in experts compared to novices when "knot tying" was compared with the control "finger tap." Increased BOLD activity was observed in experts compared with novices when "imagining a task" in the primary visual cortex, an area important in perceptual learning. Experts and intermediates performed consistently with 100% square knots. Novices had an average of 2 slip knots. Regarding knot quantity, the number of knots ranged from 14 to 26 in novices, 38 to 47 in intermediates, and 54 to 58 in experts. A Kruskal-Wallis rank sum test revealed that the difference between the 3 groups was statistically significant in the quantity of square knots tied (p = 0.147). Specific regions of interest identified concurred with findings of previous studies and included the left supramarginal, left rolandic operculum, and left postcentral regions.

We found that fMRI is a feasible method of exploring actual and perceived motor skill abilities. Head motion during performance of a motor skill does not preclude the attainment of meaningful data. Larger numbers are needed to further investigate these early findings.