Artificial intelligence in virtual standardized patients: Combining natural language understanding and rule based dialogue management to improve conversational fidelity.

INTRODUCTION: Advances in natural language understanding have facilitated the development of Virtual Standardized Patients (VSPs) that may soon rival human patients in conversational ability. We describe herein the development of an artificial intelligence (AI) system for VSPs enabling students to practice their history taking skills. METHODS: Our system consists of (1) Automated Speech Recognition (ASR), (2) hybrid AI for question identification, (3) classifier to choose between the two systems, and (4) automated speech generation. We analyzed the accuracy of the ASR, the two AI systems, the classifier, and student feedback with 620 first year medical students from 2018 to 2021. RESULTS: System accuracy improved from ∼75% in 2018 to ∼90% in 2021 as refinements in algorithms and additional training data were utilized. Student feedback was positive, and most students felt that practicing with the VSPs was a worthwhile experience. CONCLUSION: We have developed a novel hybrid dialogue system that enables artificially intelligent VSPs to correctly answer student questions at levels comparable with human SPs. This system allows trainees to practice and refine their history-taking skills before interacting with human patients.

Lower-Limb Muscle Strength, Anterior-Posterior and Inter-Limb Asymmetry in Professional, Elite Academy and Amateur Soccer Players.

Given the importance of the lower-limb strength and strength balance in soccer players and its relationship with injury prevention and performance, the present study compared quadriceps and hamstrings strength, the conventional (Hconc:Qconc), functional (Hecc:Qconc) hamstrings-to-quadriceps ratio and inter-limb strength asymmetry in professional, elite academy and amateur male soccer players. In this cross-sectional study, two hundred-six soccer players (professional = 75, elite academy = 68, amateurs = 63) volunteered to participate. Quadriceps and hamstrings isokinetic peak torque was investigated at 60° .s-1 in both the concentric and eccentric modality and at 300°.s-1 in the concentric modality. The conventional Hconc:Qconc, functional Hecc:Qconc ratio and quadriceps and hamstrings inter-limb strength asymmetry were then calculated. Professional players presented greater quadriceps and hamstrings strength than elite academy (effect size from small to moderate) and amateur players (moderate to very large). Both the conventional Hconc:Qconc and functional Hecc:Qconc ratio were greater in professional than elite academy and amateur players (small to moderate). Overall, quadriceps and hamstrings inter-limb strength asymmetry was greater in amateurs than professional (small to very large) and elite academy (trivial to large) players. The present findings provide coaches and medical staffs with normative lower-limb muscle strength data on professional, academy and amateur soccer players. Overall lower-limb muscle strength and inter-limb strength asymmetry could be used to evaluate possible inference on injury prevention and performance. The hamstrings-to-quadriceps ratio poorly differentiates between the soccer players background and offers limited prediction for injury prevention and performance.

Effects of Postactivation Potentiation After an Eccentric Overload Bout on Countermovement Jump and Lower-Limb Muscle Strength.

ABSTRACT: Beato, M, Stiff, A, and Coratella, G. Effects of postactivation potentiation after an eccentric overload bout on countermovement jump and lower-limb muscle strength. J Strength Cond Res 35(7): 1825-1832, 2021-This study aimed to evaluate the postactivation potentiation (PAP) effects of an eccentric overload (EOL) exercise on countermovement jump (CMJ) performance and isokinetic lower-limb muscle strength. Eighteen active men (mean ± SD, age 20.2 ± 1.4 years, body mass 71.6 ± 8 kg, and height 178 ± 7 cm) were involved in a randomized, crossover study. The subjects performed 3 sets per 6 repetitions of EOL half squats at maximal power using a flywheel ergometer. Postactivation potentiation using an EOL exercise was compared with a control condition (10-minute cycling at 1 W·kg-1). Countermovement jump height, peak power, impulse, and force were recorded at 15 seconds, 1, 3, 5, 7, and 9 minutes after an EOL exercise or control. Furthermore, quadriceps and hamstrings isokinetic strength were performed. Postactivation potentiation vs. control reported a meaningful difference for CMJ height after 3 minutes (effect size [ES] = 0.68, p = 0.002), 5 minutes (ES = 0.58, p = 0.008), 7 minutes (ES = 0.57, p = 0.022), and 9 minutes (ES = 0.61, p = 0.002), peak power after 1 minute (ES = 0.22, p = 0.040), 3 minutes (ES = 0.44, p = 0.009), 5 minutes (ES = 0.40, p = 0.002), 7 minutes (ES = 0.29, p = 0.011), and 9 minutes (ES = 0.30, p = 0.008), as well as quadriceps concentric, hamstrings concentric, and hamstrings eccentric peak torque (ES = 0.13, p = 0.001, ES = 0.24, p = 0.003, and ES = 0.22, p = 003, respectively) after 3-9 minutes of rest. In conclusion, the present outcomes highlight that PAP using an EOL bout improves height, peak power, impulse, and peak force during CMJ, as well as quadriceps and hamstrings isokinetic strength in male athletes. Moreover, the optimal time window for the PAP was found from 3 to 9 minutes.

The Validity and Between-Unit Variability of GNSS Units (STATSports Apex 10 and 18 Hz) for Measuring Distance and Peak Speed in Team Sports.

The aims of this study were (i) to investigate the criterion validity (vs. gold standard measurements) of the 10 and 18 Hz STATSports Apex units for measuring distances and peak speed (Vpeak) outcomes and (ii) to investigate the between-unit variability. Twenty university students were enrolled in the study (age 21 ± 2 years, weight 72 ± 6 kg, and height 1.76 ± 0.05 m). The criterion validity was tested by comparing the distances recorded by the units with ground truth reference (400-m trial, 128.5-m circuit, and 20-m trial). Vpeak values were compared with those determined by a gold standard criterion device (Stalker ATS Radar Gun) during a linear 20-m sprint. The distance biases for the Apex 10 Hz in the 400-m trial, 128.5-m circuit, and 20-m trial were 1.05 ± 0.87%, 2.3 ± 1.1%, and 1.11 ± 0.99%, respectively, while for the Apex 18 Hz the biases were 1.17 ± 0.73%, 2.11 ± 1.06%, and 1.15 ± 1.23%, respectively. Vpeak measured by the Apex 10 and 18 Hz were 26.5 ± 2.3 km h-1 and 26.5 ± 2.6 km h-1, respectively, with the criterion method reporting 26.3 ± 2.4 km h-1, with a bias of 2.36 ± 1.67% and 2.02 ± 1.24%, respectively. This study is the first to validate and compare the STATSports Apex 10 and 18 Hz. Between-analysis (t-test) for total distance and Vpeak reported non-significant differences. Apex units reported a small error of around 1-2% compared to the criterion distances during 400-m, 128.5-m circuit, 20-m trials, and Vpeak. In conclusion, both units could be used with confidence to measure these variables during training and match play.

Validity and Reliability of Global Positioning System Units (STATSports Viper) for Measuring Distance and Peak Speed in Sports.

Beato, M, Devereux, G, and Stiff, A. Validity and reliability of global positioning system units (STATSports Viper) for measuring distance and peak speed in sports. J Strength Cond Res 32(10): 2831-2837, 2018-Previous evidence has proven that large variability exists in the accuracy of different brands of global positioning systems (GPS). Therefore, any GPS model should be validated independently, and the results of a specific brand cannot be extended to others. The aim of this study is to assess the validity and reliability of GPS units (STATSports Viper) for measuring distance and peak speed in sports. Twenty participants were enrolled (age 21 ± 2 years [range 18 to 24 years], body mass 73 ± 5 kg, and height 1.78 ± 0.04 m). Global positioning system validity was evaluated by comparing the instantaneous values of speed (peak speed) determined by GPS (10 Hz, Viper Units; STATSports, Newry, Ireland) with those determined by a radar gun during a 20-m sprint. Data were analyzed using the Stalker (34.7 GHz, USA) ATS Version 5.0.3.0 software as gold standard. Distance recorded by GPS was also compared with a known circuit distance (400-m running, 128.5-m sports-specific circuit, and 20-m linear running). The distance bias in the 400-m trial, 128.5-m circuit, and 20-m trial was 1.99 ± 1.81%, 2.7 ± 1.2%, and 1.26 ± 1.04%, respectively. Peak speed measured by the GPS was 26.3 ± 2.4 km·h, and criterion was 26.1 ± 2.6 km·h, with a bias of 1.80 ± 1.93%. The major finding of this study was that GPS did not underestimate the criterion distance during a 400-m trial, 128.5-m circuit, and 20-m trial, as well as peak speed. Small errors (<5%, good) were found for peak speed and distances. This study supported the validity and reliability of this GPS model.