Artificial Neural Network for Laparoscopic Skills Classification Using Motion Signals from Apple Watch

Objective and automated assessment of surgical technical skills with IoT systems: A systematic literature review

“…The main purpose for using sensors in a surgical procedure is to track tool motion (59 different articles); this is done by incorporating sensors that capture motion in surgical instruments, such as catheters, needles, graspers or scissors [31][55] [70][75] [80]. Sensors are also used to track motion in the surgeon's body during a surgical procedure (25 different articles); this is done by incorporating sensors which capture motion typically on the surgeon's hands, wrists, arms or head [86][91] [94][99] [103]. In addition, sensors are also used to measure the cognitive load of the surgeon (13 different articles), incorporating sensors that collect activity on the surgeon's eyes, brain, muscles, heart or skin [8][87] [110].…”

“…It is important to note that some of the statistical methods and algorithms for classification and prediction rely on preliminary steps related to pre-processing the data collected. These include: 1) feature extraction (to collect the most important data and reduce the dimensionality of the data) using e.g., PCA (Principal Component Analysis) [26]; 2) data normalization (in case data are scaled differently) using e.g., Znormalization [86]; and 3) feature selection (to collect the most relevant features/indicators, excluding those which may cause noise), using e.g., mRMR (Minimum Redundancy Maximum Relevance) [21] or Linear Correlation Coefficients [26]. The pre-processing of the data, although may be of interest, is not the focus of this article.…”

“…The second most common indicator used is acceleration values in surgical instruments (22 articles) (including mean, maximum value, range, standard deviation, root-mean-square and total-sum-of-square). This indicator has shown significant differences between novice, intermediate and advanced surgeons when assessing the surgical technical skill EF [46] [102], specially non-linear acceleration parameters, such as kurtosis (a measure of the combined weight of the tails of a distribution relative to the center of the distribution) and power spectral density (a measure of the power content of the signal relative to the frequency) [86]. Indicators that have shown significant differences between novice and advanced surgeons beyond smoothness are: number of sub-movements obtained through acceleration (9 articles) when assessing EC [31][52] [102]; sub-movement duration obtained through acceleration (4 articles) when assessing EF [102]; and spectral arc length obtained through Fourier spectrum of acceleration (2 articles) when assessing EC [56].…”

“…In contrast, fixation duration, fixation rate, fixation number, saccade number, saccade duration, and saccade amplitude did not show very good results to classify levels of expertise [105]. In addition, other indicators that are not derived from data collected by sensors can be included in classification models, such as years of experience [8] or years of study of the surgeon [25], GRS real score [47], OSATS real score [24], GEARS real score [36], number of procedures made [28], and time and number of errors made [86]. Regarding prediction models and the indicators used, indicators can be combined with some weights to obtain scores about surgical technical skills.…”

Learning Enhancement in Higher Education with Wearable Technology

“…5: Various head-worn wearable have been used for educational purpose including A) Emotiv EPOC EEG system [47] B) Google Glasses in [49] and . C) Wearable wrist-worn like Apple Watch used in [60] and D)wearable chest-worn studied in [63] • . In [62] wrist-worn activity trackers equipped with biometric sensors were used in higher education regarding eHealth literacy acquisition.…”