It's not all in your feet: Improving penalty kick performance with human-avatar interaction and machine learning.

Penalty kicks are increasingly decisive in major international football competitions. Yet, over 30% of shootout kicks are missed. The outcome of the kick often relies on the ability of the penalty taker to exploit anticipatory movements of the goalkeeper to redirect the kick toward the open side of the goal. Unfortunately, this ability is difficult to train using classical methods. We used an augmented reality simulator displaying an holographic goalkeeper to test and train penalty kick performance with 13 young elite players. Machine learning algorithms were used to optimize the learning rate by maintaining an optimal level of training difficulty. Ten training sessions of 20 kicks reduced the redirection threshold by 120 ms, which constituted a 28% reduction with respect to the baseline threshold. Importantly, redirection threshold reduction was observed for all trained players, and all things being equal, it corresponded to an estimated 35% improvement of the success rate.

Preventing falls: the use of machine learning for the prediction of future falls in individuals without history of fall.

Nowadays, it becomes of paramount societal importance to support many frail-prone groups in our society (elderly, patients with neurodegenerative diseases, etc.) to remain socially and physically active, maintain their quality of life, and avoid their loss of autonomy. Once older people enter the prefrail stage, they are already likely to experience falls whose consequences may accelerate the deterioration of their quality of life (injuries, fear of falling, reduction of physical activity). In that context, detecting frailty and high risk of fall at an early stage is the first line of defense against the detrimental consequences of fall. The second line of defense would be to develop original protocols to detect future fallers before any fall occur. This paper briefly summarizes the current advancements and perspectives that may arise from the combination of affordable and easy-to-use non-wearable systems (force platforms, 3D tracking motion systems), wearable systems (accelerometers, gyroscopes, inertial measurement units-IMUs) with appropriate machine learning analytics, as well as the efforts to address these challenges.

Influence of mental workload on motion perception: A direct comparison of luminance-based and contrast-based stimuli.

In order to study the impact of increased mental workload on motion detection, twenty-four observers performed a motion discrimination task in which they had to detect odd moving patches. Two types of moving patches were used, namely luminance-based and contrast-based patches. For both types of patches, the motion discrimination task was performed with and without an additional N-Back task aimed at increasing the mental workload. The dual task decreased discrimination performance for both types of patches, but the difference was significantly larger for contrast-based patches, i.e., for second-order motion stimuli, both as an absolute and relative increment. This suggests that motion discrimination requires larger cognitive resources for contrast-based than for luminance-based stimuli, thereby hinting at the higher complexity of the cognitive mechanisms underlying second-order motion detection.

Model Based or Model Free? Comparing Adaptive Methods for Estimating Thresholds in Neuroscience.

The quantification of human perception through the study of psychometric functions Ψ is one of the pillars of experimental psychophysics. In particular, the evaluation of the threshold is at the heart of many neuroscience and cognitive psychology studies, and a wide range of adaptive procedures has been developed to improve its estimation. However, these procedures are often implicitly based on different mathematical assumptions on the psychometric function, and unfortunately, these assumptions cannot always be validated prior to data collection. This raises questions about the accuracy of the estimator produced using the different procedures. In the study we examine in this letter, we compare five adaptive procedures commonly used in psychophysics to estimate the threshold: Dichotomous Optimistic Search (DOS), Staircase, PsiMethod, Gaussian Processes, and QuestPlus. These procedures range from model-based methods, such as the PsiMethod, which relies on strong assumptions regarding the shape of Ψ, to model-free methods, such as DOS, for which assumptions are minimal. The comparisons are performed using simulations of multiple experiments, with psychometric functions of various complexity. The results show that while model-based methods perform well when Ψ is an ideal psychometric function, model-free methods rapidly outshine them when Ψ deviates from this model, as, for instance, when Ψ is a beta cumulative distribution function. Our results highlight the importance of carefully choosing the most appropriate method depending on the context.

A review of center of pressure (COP) variables to quantify standing balance in elderly people: Algorithms and open-access code.

Postural control is often quantified by recording the trajectory of the center of pressure (COP)-also called stabilogram-during human quiet standing. This quantification has many important applications, such as the early detection of balance degradation to prevent falls, a crucial task whose relevance increases with the aging of the population. Due to the complexity of the quantification process, the analyses of sway patterns have been performed empirically using a number of variables, such as ellipse confidence area or mean velocity. This study reviews and compares a wide range of state-of-the-art variables that are used to assess the risk of fall in elderly from a stabilogram. When appropriate, we discuss the hypothesis and mathematical assumptions that underlie these variables, and we propose a reproducible method to compute each of them. Additionally, we provide a statistical description of their behavior on two datasets recorded in two elderly populations and with different protocols, to hint at typical values of these variables. First, the balance of 133 elderly individuals, including 32 fallers, was measured on a relatively inexpensive, portable force platform (Wii Balance Board, Nintendo) with a 25-s open-eyes protocol. Second, the recordings of 76 elderly individuals, from an open access database commonly used to test static balance analyses, were used to compute the values of the variables on 60-s eyes-open recordings with a research laboratory standard force platform.

The 5P program, personalized and participatory primary prevention pathway: Rational and design of a clinical trial in general practice.

The aging of the population is leading to an increase in the number of people with loss of autonomy, placing a strain on the health care system. Its prevention at early stages such as the frailty stage would allow an improvement in the quality of life of seniors while limiting health care expenses. The "Atout Age" prevention program set up by the health public authorities of Reunion Island for retired people and the new frailty assessment tools based on mathematical machine learning algorithms could improve the ambulatory care of senior citizens. At present, referral care remains hospital with comprehensive geriatric assessment and there is a lack of evidence of the effectiveness of a prevention pathway for loss of autonomy in primary care. For these reasons, the 5P program "Personalized and Participative Primary Prevention Pathway" has been started in order to obtain scientific evidence. In this article, we present the objectives, design and first results, used in the 5P program up to the implementation of a clinical trial in general practice. The program is articulated in 3 phases. A first phase to evaluate the acceptability of innovative screening tools for frailty. A second pilot phase evaluates the feasibility of a large-scale ambulatory clinical trial in general practice. The last phase described in this article, is a multisite, pseudo-randomized, controlled clinical trial measuring the impact of the "Atout Age" workshops on the physical performance and the quality of life of seniors compared with their usual ambulatory follow-up.

A Langevin-Based Model With Moving Posturographic Target to Quantify Postural Control.

Falls are a major concern of public health, particularly for older adults, as the consequences of falls include serious injuries and death. Therefore, the understanding and evaluation of postural control is considered key, as its deterioration is an important risk factor predisposing to falls. In this work we introduce a new Langevin-based model, local recall, that integrates the information from both the center of pressure (CoP) and the center of mass (CoM) trajectories, and compare its accuracy to a previously proposed model that only uses the CoP. Nine healthy young participants were studied under quiet bipedal standing conditions with eyes either open or closed, while standing on either a rigid surface or a foam. We show that the local recall model produces significantly more accurate prediction than its counterpart, regardless of the eyes and surface conditions, and we replicate these results using another publicly available human dataset. Additionally, we show that parameters estimated using the local recall model are correlated with the quality of postural control, providing a promising method to evaluate static balance. These results suggest that this approach might be interesting to further extend our understanding of the underlying mechanisms of postural control in quiet stance.

An opinion paper on the maintenance of robustness: Towards a multimodal and intergenerational approach using digital twins.

The increasing number of frail elderly people in our aging society is becoming problematic: about 11% of community-dwelling older persons are frail and another 42% are pre-frail. Consequently, a major challenge in the coming years will be to test people over the age of 60 years to detect pre-frailty at the earliest stage and to return them to robustness using the targeted interventions that are becoming increasingly available. This challenge requires individual longitudinal monitoring (ILM) or follow-up of community-dwelling older persons using quantitative approaches. This paper briefly describes an effort to tackle this challenge. Extending the detection of the pre-frail stages to other population groups is also suggested. Appropriate algorithms have been used to begin the tracing of faint physiological signals in order to detect transitions from robustness to pre-frailty states and from pre-frailty to frailty states. It is hoped that these studies will allow older adults to receive preventive treatment at the correct institutions and by the appropriate professionals as early as possible, which will prevent loss of autonomy. Altogether, ILM is conceived as an emerging property of databases ("digital twins") and not the reverse. Furthermore, ILM should facilitate a coordinated set of actions by the caregivers, which is a complex challenge in itself. This approach should be gradually extended to all ages, because frailty has no age, as is testified by overwork, burnout, and post-traumatic syndrome.

Apprenticeship Learning for a Predictive State Representation of Anesthesia.

OBJECTIVE: In this paper, we present an original decision support algorithm to assist the anesthesiologists delivery of drugs to maintain the optimal Depth of Anesthesia (DoA). METHODS: Derived from a Transform Predictive State Representation algorithm (TPSR), our model learned by observing anesthesiologists in practice. This framework, known as apprenticeship learning, is particularly useful in the medical field as it is not based on an exploratory process - a prohibitive behavior in healthcare. The model only relied on the four commonly monitored variables: Heart Rate (HR), the Mean Blood Pressure (MBP), the Respiratory Rate (RR) and the concentration of anesthetic drug (AAFi). RESULTS: Thirty-one patients have been included. The performances of the model is analyzed with metrics derived from the Hamming distance and cross entropy. They demonstrated that low rank dynamical system had the best performances on both predictions and simulations. Then, a confrontation of our agent to a panel of six real anesthesiologists demonstrated that 95.7% of the actions were valid. CONCLUSION: These results strongly support the hypothesis that TPSR based models convincingly embed the behavior of anesthesiologists including only four variables that are commonly assessed to predict the DoA. SIGNIFICANCE: The proposed novel approach could be of great help for clinicians by improving the fine tuning of the DoA. Furthermore, the possibility to predict the evolutions of the variables would help preventing side effects such as low blood pressure. A tool that could autonomously help the anesthesiologist would thus improve safety-level in the surgical room.

NurseNet: Monitoring Elderly Levels of Activity with a Piezoelectric Floor.

Monitoring the activity of elderly individuals in nursing homes is key, as it has been shown that physical activity leads to significant health improvement. In this work, we introduce NurseNet, a system that combines an unobtrusive, affordable, and robust piezoelectric floor sensor with a convolutional neural network algorithm, which aims at measuring elderly physical activity. Our algorithm is trained using signal embedding based on atoms of a pre-learned dictionary and focuses the network's attention on step-related signals. We show that NurseNet is able to avoid the main limitation of floor sensors by recognizing relevant signals (i.e., signals produced by patients) and ignoring events related to the medical staff, offering a new tool to monitor elderly activity in nursing homes efficiently.

Estimating Center of Mass Trajectory in Quiet Standing: a Review.

As falls prevalence increases with the aging of the population, early detection of balance degradation is of great importance for efficient prevention and treatment. This work compares a wide range of state-of-the-art methods to estimate the trajectory of the center of mass - a key aspect of postural control quantification. This comparison is done through multiple complementary metrics over a large dataset, highlighting the pros and cons of each method as well as the key influence of the data preprocessing. Additionally, we introduce several improvements and parameter tunings for these methods that increase their accuracy or reduce contextual information requirement.

On the importance of local dynamics in statokinesigram: A multivariate approach for postural control evaluation in elderly.

The fact that almost one third of population >65 years-old has at least one fall per year, makes the risk-of-fall assessment through easy-to-use measurements an important issue in current clinical practice. A common way to evaluate posture is through the recording of the center-of-pressure (CoP) displacement (statokinesigram) with force platforms. Most of the previous studies, assuming homogeneous statokinesigrams in quiet standing, used global parameters in order to characterize the statokinesigrams. However the latter analysis provides little information about local characteristics of statokinesigrams. In this study, we propose a multidimensional scoring approach which locally characterizes statokinesigrams on small time-periods, or blocks, while highlighting those which are more indicative to the general individual's class (faller/non-faller). Moreover, this information can be used to provide a global score in order to evaluate the postural control and classify fallers/non-fallers. We evaluate our approach using the statokinesigram of 126 community-dwelling elderly (78.5 ± 7.7 years). Participants were recorded with eyes open and eyes closed (25 seconds each acquisition) and information about previous falls was collected. The performance of our findings are assessed using the receiver operating characteristics (ROC) analysis and the area under the curve (AUC). The results show that global scores provided by splitting statokinesigrams in smaller blocks and analyzing them locally, classify fallers/non-fallers more effectively (AUC = 0.77 ± 0.09 instead of AUC = 0.63 ± 0.12 for global analysis when splitting is not used). These promising results indicate that such methodology might provide supplementary information about the risk of fall of an individual and be of major usefulness in assessment of balance-related diseases such as Parkinson's disease.

A Non Linear Scoring Approach for Evaluating Balance: Classification of Elderly as Fallers and Non-Fallers.

Almost one third of population 65 years-old and older faces at least one fall per year. An accurate evaluation of the risk of fall through simple and easy-to-use measurements is an important issue in current clinic. A common way to evaluate balance in posturography is through the recording of the centre-of-pressure (CoP) displacement (statokinesigram) with force platforms. A variety of indices have been proposed to differentiate fallers from non fallers. However, no agreement has been reached whether these analyses alone can explain sufficiently the complex synergies of postural control. In this work, we study the statokinesigrams of 84 elderly subjects (80.3+- 6.4 years old), which had no impairment related to balance control. Each subject was recorded 25 seconds with eyes open and 25 seconds with eyes closed and information pertaining to the presence of problems of balance, such as fall, in the last six months, was collected. Five descriptors of the statokinesigrams were computed for each record, and a Ranking Forest algorithm was used to combine those features in order to evaluate each subject's balance with a score. A classical train-test split approach was used to evaluate the performance of the method through ROC analysis. ROC analysis showed that the performance of each descriptor separately was close to a random classifier (AUC between 0.49 and 0.54). On the other hand, the score obtained by our method reached an AUC of 0.75 on the test set, consistent over multiple train-test split. This non linear multi-dimensional approach seems appropriate in evaluating complex postural control.

Preprocessing the Nintendo Wii Board Signal to Derive More Accurate Descriptors of Statokinesigrams.

During the past few years, the Nintendo Wii Balance Board (WBB) has been used in postural control research as an affordable but less reliable replacement for laboratory grade force platforms. However, the WBB suffers some limitations, such as a lower accuracy and an inconsistent sampling rate. In this study, we focus on the latter, namely the non uniform acquisition frequency. We show that this problem, combined with the poor signal to noise ratio of the WBB, can drastically decrease the quality of the obtained information if not handled properly. We propose a new resampling method, Sliding Window Average with Relevance Interval Interpolation (SWARII), specifically designed with the WBB in mind, for which we provide an open source implementation. We compare it with several existing methods commonly used in postural control, both on synthetic and experimental data. The results show that some methods, such as linear and piecewise constant interpolations should definitely be avoided, particularly when the resulting signal is differentiated, which is necessary to estimate speed, an important feature in postural control. Other methods, such as averaging on sliding windows or SWARII, perform significantly better on synthetic dataset, and produce results more similar to the laboratory-grade AMTI force plate (AFP) during experiments. Those methods should be preferred when resampling data collected from a WBB.