This paper aims to present that biofeedback based on electromyography (EMG) can help transtibial amputees to improve their capability to manipulate powered prosthetic ankles using Direct Electromyography (dEMG) control. First, we constructed a novel haptic based EMG biofeedback system composed of a HD haptic vest and an encoder that mapped EMG magnitude of residual shank muscles to tactile vibration patterns. Then we recruited six transtibial amputees and six non-disabled participants to conduct a position matching task, in which they conducted dEMG control on a powered prosthetic ankle with/without the biofeedback system after a brief acclimation. Impact of the EMG biofeedback system was evaluated based on their accuracy in matching specific targets in the position matching task. Our results demonstrated that 1) the EMG biofeedback can improve participants’ target matching accuracy; 2) the biofeedback led to more separable EMG signals among amputee participants when they tried to reach different targets; 3) for amputees, the improvement is more clearly observed when matching towards dorsiflexion targets; and 4) high performance variation among amputees is observed. Our work also demonstrated the feasibility of providing effective proprioceptive feedback based on muscle contraction level without involving any surgical procedure on lower limb amputees.

Supersaturated designs investigate more factors than there are runs and are often constructed under a criterion measuring a design’s proximity to an unattainable orthogonal design. The most popular analysis identifies active factors by inspecting the solution path of a penalized estimator, such as the lasso. Recent criteria encouraging positive correlations between factors have been shown to produce designs with more definitive solution paths so long as the active factors have positive effects. Two open problems affecting the understanding and practicality of supersaturated designs are: (1) do optimal designs under existing criteria maximize support recovery probability across an estimator’s solution path and (2) why do designs with positively correlated columns produce more definitive solution paths when the active factors have positive sign effects? To answer these questions, we develop criteria maximizing the lasso’s sign recovery probability. We prove that an orthogonal design is an ideal structure when the signs of the active factors are unknown, and a design with constant, small, positive correlations is ideal when the signs are assumed known. A computationally efficient design search algorithm is proposed that first filters through optimal designs under new heuristic criteria to select the one that maximizes the lasso sign recovery probability.

The rise of internet-based services and products in the late 1990's brought about an unprecedented opportunity for online businesses to engage in large scale data-driven decision making. Over the past two decades, organizations such as Airbnb, Alibaba, Amazon, Baidu, Booking, Alphabet's Google, LinkedIn, Lyft, Meta's Facebook, Microsoft, Netflix, Twitter, Uber, and Yandex have invested tremendous resources in online controlled experiments (OCEs) to assess the impact of innovation on their customers and businesses. Running OCEs at scale has presented a host of challenges requiring solutions from many domains. In this paper we review challenges that require new statistical methodologies to address them. In particular, we discuss the practice and culture of online experimentation, as well as its statistics literature, placing the current methodologies within their relevant statistical lineages and providing illustrative examples of OCE applications. Our goal is to raise academic statisticians' awareness of these new research opportunities to increase collaboration between academia and the online industry.

The analysis of screening experiments is often done in two stages, starting with factor selection via an analysis under a main effects model. The success of this first stage is influenced by three components: (1) main effect estimators' variances and (2) bias, and (3) the estimate of the noise variance. Component (3) has only recently been given attention with design techniques that ensure an unbiased estimate of the noise variance. In this paper, we propose a design criterion based on expected confidence intervals of the first stage analysis that balances all three components. To address model misspecification, we propose a computationally-efficient all-subsets analysis and a corresponding constrained design criterion based on lack-of-fit. Scenarios found in existing design literature are revisited with our criteria and new designs are provided that improve upon existing methods.

A screening experiment attempts to identify a subset of important effects using a relatively small number of experimental runs. Given the limited run size and a large number of possible effects, penalized regression is a popular tool used to analyze screening designs. In particular, an automated implementation of the Gauss-Dantzig selector has been widely recommended to compare screening design construction methods. Here, we illustrate potential reproducibility issues that arise when comparing screening designs via simulation, and recommend a graphical method, based on screening probabilities, which compares designs by evaluating them along the penalized regression solution path. This method can be implemented using simulation, or, in the case of lasso, by using exact local lasso sign recovery probabilities. Our approach circumvents the need to specify tuning parameters associated with regularization methods, leading to more reliable design comparisons. This article contains supplementary materials including code to implement the proposed methods.

The tuning parameter selection strategy for penalized estimation is crucial to identify a model that is both interpretable and predictive. However, popular strategies (e.g., minimizing average squared prediction error via cross-validation) tend to select models with more predictors than necessary. A simple yet powerful cross validation strategy is proposed which is based on maximizing the squared correlation between the observed and predicted values, rather than minimizing squared error loss for the purposes of support recovery. The strategy can be applied to all penalized least-squares estimators and, under certain conditions, the metric implicitly performs a bias adjustment named the α-modification. When applied to the Lasso estimator, the α-modification is closely related to the relaxed Lasso estimator. The approach is demonstrated on a functional variable selection problem to identify optimal placement of surface electromyogram sensors to control a robotic hand prosthesis.

In practice, optimal screening designs for arbitrary run sizes are traditionally generated using the D-criterion with factor settings fixed at +/- 1, even when considering continuous factors with levels in [-1, 1]. This paper identifies cases of undesirable estimation variance properties for such D-optimal designs and argues that generally A-optimal designs tend to push variances closer to their minimum possible value. New insights about the behavior of the criteria are found through a study of their respective coordinate-exchange formulas. The study confirms the existence of D-optimal designs comprised only of settings +/- 1 for both main effect and interaction models for blocked and un-blocked experiments. Scenarios are also identified for which arbitrary manipulation of a coordinate between [-1, 1] leads to infinitely many D-optimal designs each having different variance properties. For the same conditions, the A-criterion is shown to have a unique optimal coordinate value for improvement. We also compare Bayesian version of the A- and D-criteria in how they balance minimization of estimation variance and bias. Multiple examples of screening designs are considered for various models under Bayesian and non-Bayesian versions of the A- and D-criteria.

Automatically personalizing complex control of robotic prostheses to improve gait performance, such as gait symmetry, is challenging. Recently, human-in-the-loop (HIL) optimization and reinforcement learning (RL) have shown promise in achieving optimized control of wearable robots for each individual user. However, HIL optimization methods lack scalability for high-dimensional space, while RL has mostly focused on optimizing robot kinematic performance. Thus, we propose a novel hierarchical framework to personalize robotic knee prosthesis control and improve overall gait performance. Specifically, in this study the framework was implemented to simultaneously design target knee kinematics and tune 12 impedance control parameters for improved symmetry of propulsive impulse in walking. In our proposed framework, HIL optimization is used to identify an optimal target knee kinematics with respect to symmetry improvement, while RL is leveraged to yield an optimal policy for tuning impedance parameters in high-dimensional space to match the kinematics target. The proposed framework was validated on human subjects, walking with robotic knee prosthesis. The results showed that our design successfully shaped the target knee kinematics as well as configured 12 impedance control parameters to improve propulsive impulse symmetry of the human users. The knee kinematics that yielded best propulsion symmetry did not preserve the normative knee kinematics profile observed in non-disabled individuals, suggesting that restoration of normative joint biomechanics in walking does not necessarily optimize the gait performance of human-prosthesis systems. This new framework for prosthesis control personalization may be extended to other wearable devices or different gait performance optimization goals in the future.

Weighted optimality criteria allow an experimenter to express hierarchical interest across estimable functions through a concise weighting system. We show how such criteria can be implicitly influenced by the estimable functions' minimum variances, leading to nonintuitive variance properties of the optimal designs. To address this, we propose a new optimality and evaluation approach that incorporates these minimum variances. A modified c-optimality criterion is introduced to calculate an estimable function's minimum variance while requiring estimability of all other functions of interest. These minimum variances are then incorporated into a standardized weighted A-criterion that has an intuitive weighting system. We argue that optimal designs under this criterion tend to satisfy the conditions of a new design property we call weight adherence that sets appropriate expectations for how a given weighting system will influence variance properties. A practical, exploratory approach is then described for weighted optimal design generation and evaluation. Examples of the exploratory approach and weight adherence are provided for two types of factorial experiments.

Chemical agents released into the environment can induce oxidative stress in organisms, which is detrimental for health. Although environmental exposures typically incorporate multiple chemicals, organismal studies on oxidative stress derived from chemical agents commonly study exposures to individual compounds. In this work, we explore how chemical mixtures drive the oxidative stress response under various conditions in the nematode C. elegans, by quantitatively assessing levels of gst-4 expression. Our results indicate that naphthoquinone mixtures drive responses differently than individual components, and that altering environmental conditions, such as increased heat and reduced food availability, result in dramatically different oxidative stress responses mounted by C. elegans. When exposed to heat, the oxidative stress response is diminished. Notably, when exposed to limited food, the oxidative stress response specific to juglone is significantly heightened, while identified antagonistic interactions between some naphthoquinone components in mixtures are abolished. This implies that organismal responses to xenobiotics is confounded by environment and stressor interactions. Given the high number of variables under study, and their potential combinations, a simplex centroid design was used to capture such non-trivial response over the design space. This makes the case for the adoption of Design of Experiments approaches as they can greatly expand the experimental space probed in noisy biological readouts, and in combinatorial experiments. Our results also reveal gaps in our current knowledge of the organismal oxidative stress response, which can be addressed by employing sophisticated design of experiments approaches to identify significant interactions.

Optimizing an expensive, black-box function f(·) is challenging when its input space is high-dimensional. Sequential design frameworks first model f(·) with a surrogate function and then optimize an acquisition function to determine input settings to evaluate next. Optimization of both f(·) and the acquisition function benefit from effective dimension reduction. Global variable selection detects and removes input variables that do not affect f(·) across the input space. Further dimension reduction may be possible if we consider local variable selection around the current optimum estimate. We develop a sequential design algorithm called sequential optimization in locally important dimensions (SOLID) that incorporates global and local variable selection to optimize a continuous, differentiable function. SOLID performs local variable selection by comparing the surrogate’s predictions in a localized region around the estimated optimum with the p alternative predictions made by removing each input variable. The search space of the acquisition function is further restricted to focus only on the variables that are deemed locally active, leading to greater emphasis on refining the surrogate model in locally active dimensions. A simulation study across multiple test functions and an application to the Sarcos robot dataset show that SOLID outperforms conventional approaches. Supplementary materials for this article are available online.

Articular cartilage has unique load-bearing properties but has minimal capacity for intrinsic repair. Here, we used three-dimensional weaving, additive manufacturing, and autologous mesenchymal stem cells to create a tissue-engineered, bicomponent implant to restore hip function in a canine hip osteoarthritis model. This resorbable implant was specifically designed to function mechanically from the time of repair and to biologically integrate with native tissues for long-term restoration. A massive osteochondral lesion was created in the hip of skeletally mature hounds and repaired with the implant or left empty (control). Longitudinal outcome measures over 6 months demonstrated that the implant dogs returned to normal preoperative values of pain and function. Anatomical structure and functional biomechanical properties were also restored in the implanted dogs. Control animals never returned to normal and exhibited structurally deficient repair. This study provides clinically relevant evidence that the bicomponent implant may be a potential therapy for moderate hip osteoarthritis.

In this article, we propose a new method for constructing supersaturated designs that is based on the Kronecker product of two carefully chosen matrices. The construction method leads to a partitioning of the factors of the design such that the factors within a group are correlated to the others within the same group, but are orthogonal to any factor in any other group. We refer to the resulting designs as group-orthogonal supersaturated designs. We leverage this group structure to obtain an unbiased estimate of the error variance, and to develop an effective, design-based model selection procedure. Simulation results show that the use of these designs, in conjunction with our model selection procedure enables the identification of larger numbers of active main effects than have previously been reported for supersaturated designs. The designs can also be used in group screening; however, unlike previous group-screening procedures, with our designs, main effects in a group are not confounded. Supplementary materials for this article are available online.

Despite the vast amount of literature on supersaturated designs (SSDs), there is a scant record of their use in practice. We contend this imbalance is due to conflicting recommendations regarding SSD use in the literature as well as the designs’ inabilities to meet practitioners’ analysis expectations. To address these issues, we first summarize practitioner concerns and expectations of SSDs as determined via an informal questionnaire. Next, we discuss and compare two recent SSDs that pair a design construction method with a particular analysis method. The choice of a design/analysis pairing is shown to depend on the screening objective. Group orthogonal SSDs, when paired with our new, modified analysis, are demonstrated to have high power even with many active factors. Constrained positive Var(s)-optimal designs, when paired with the Dantzig selector, are recommended when effect directions can be credibly specified in advance; this strategy reasonably controls Type I error rates while still identifying a high proportion of active factors.

Despite the vast amount of literature on supersaturated designs, there is a scant record of their use in practice. We contend this imbalance is due to the designs’ inabilities to meet practitioners’ analysis expectations and the existing literature’s lack of clearly stated expectations. To address this issue we discuss and compare two recent SSDs that pair a design construction method with a particular analysis method. Group orthogonal supersaturated designs (Jones et al. 2019), when paired with our new, modified analysis, are shown to have high power even with many active factors. V ar(s+) designs (Weese et al. 2017), when paired with the Dantzig selector, are recommended when effect directions can be reasonably specified in advance; this strategy reasonably controls type 1 error rates while still identifying a high proportion of active factors. The construction of both designs is less intuitive than classical supersaturated designs that focus almost solely on traditional measures of near-orthogonality, and forces reflection on current best practices.

Robotic hand prostheses require a controller to decode muscle contraction information, such as electromyogram (EMG) signals, into the user's desired hand movement. State-of-the-art decoders demand extensive training, require data from a large number of EMG sensors and are prone to poor predictions. Biomechanical models of a single movement degree-of-freedom tell us that relatively few muscles, and, hence, fewer EMG sensors are needed to predict movement. We propose a novel decoder based on a dynamic, functional linear model with velocity or acceleration as its response and the recent past EMG signals as functional covariates. The effect of each EMG signal varies with the recent position to account for biomechanical features of hand movement, increasing the predictive capability of a single EMG signal compared to existing decoders. The effects are estimated with a multistage, adaptive estimation procedure that we call Sequential Adaptive Functional Estimation (SAFE). Starting with 16 potential EMG sensors, our method correctly identifies the few EMG signals that are known to be important for an able-bodied subject. Furthermore, the estimated effects are interpretable and can significantly improve understanding and development of robotic hand prostheses.

Background Establishing gait symmetry is a major aim of amputee rehabilitation and may be more attainable with powered prostheses. Though, based on previous work, we postulate that users transfer a previously-learned motor pattern across devices, limiting the functionality of more advanced prostheses. The objective of this study was to preliminarily investigate the effect of increased stance time via visual feedback on amputees’ gait symmetry using powered and passive knee prostheses. Methods Five individuals with transfemoral amputation or knee disarticulation walked at their self-selected speed on a treadmill. Visual feedback was used to promote an increase in the amputated-limb stance time. Individuals were fit with a commercially-available powered prosthesis by a certified prosthetist and practiced walking during a prior visit. The same protocol was completed with a passive knee and powered knee prosthesis on separate days. We used repeated-measures, two-way ANOVA (alpha = 0.05) to test for significant effects of the feedback and device factors. Our main outcome measures were stance time asymmetry, peak anterior-posterior ground reaction forces, and peak anterior propulsion asymmetry. Results Increasing the amputated-limb stance time via visual feedback significantly improved the stance time symmetry ( p = 0.012) and peak propulsion symmetry ( p = 0.036) of individuals walking with both prostheses. With the powered knee prosthesis, the highest feedback target elicited 36% improvement in stance time symmetry, 22% increase in prosthesis-side peak propulsion, and 47% improvement in peak propulsion symmetry compared to a no feedback condition. The changes with feedback were not different with the passive prosthesis, and the main effects of device/ prosthesis type were not statistically different. However, subject by device interactions were significant, indicating individuals did not respond consistently with each device (e.g. prosthesis-side propulsion remained comparable to or was greater with the powered versus passive prosthesis for different subjects). Overall, prosthesis-side peak propulsion averaged across conditions was 31% greater with the powered prosthesis and peak propulsion asymmetry improved by 48% with the powered prosthesis. Conclusions Increasing prosthesis-side stance time via visual feedback favorably improved individuals’ temporal and propulsive symmetry. The powered prosthesis commonly enabled greater propulsion, but individuals adapted to each device with varying behavior, requiring further investigation.

The goal of this study was to implement powerful mixture design techniques, commonly used in process optimization, to investigate enhanced adverse effects upon co-exposure to environmental cyanotoxins. Exposure to cyanobacteria, which are found ubiquitously in environmental water reservoirs, have been linked to several neurodegenerative diseases. Despite the known co-occurrence of various cyanotoxins, the majority of studies investigating this link have focused on the investigation of a single cyanotoxin, a noncanonical amino acid called beta-methylamino-L-alanine (BMAA), which poorly recapitulates an actual environmental exposure. Interactions amongst cyanotoxic compounds is an area of great concern and remains poorly understood. To this end, we describe the use of a simplex axial mixture design to screen for interactive adverse effects of cyanotoxic mixtures. Using a combination of basic toxicity assays coupled with contemporary proteomic techniques, our results show the existence of a significant (p <= 0.01) interaction between BMAA and its isomers aminoethyl glycine (AEG) and 2,4-diaminobutyric acid (2,4DAB). Cyanotoxic mixtures significantly decreased cell viability by an average of 19% and increased caspases 3/7 activities by an average of 110% when compared to individual cyanotoxins (p <= 0.05). Cyanotoxic mixtures perturbed various biological pathways associated with neurodegeneration, including inhibition of protective autophagy and activation of mitochondrial dysfunction (z-score > vertical bar 2 vertical bar). Additionally, exposure to mixtures perturbed important upstream regulators involved in cellular dysfunction, morbidity, and development. Taken together, our results highlight: (1) the need to study combinations of cyanotoxins when investigating the link between cyanobacteria and neurodegenerative pathologies and (2) the application of design of experiment (DoE) as an efficient methodology to study mixtures of relevant environmental toxins.

OBJECTIVE:To evaluate accuracy of articular surfaces determined by use of 2 perpendicular CT orientations, micro-CT, and laser scanning.SAMPLE:23 cat cadavers.PROCEDURES:Images of antebrachia were obtained by use of CT (voxel size, 0.6 mm) in longitudinal orientation (CTLO images) and transverse orientation (CTTO images) and by use of micro-CT (voxel size, 0.024 mm) in a longitudinal orientation. Images were reconstructed. Craniocaudal and mediolateral length, radius of curvature, and deviation of the articular surface of the distal portion of the radius of 3-D renderings for CTLO, CTTO, and micro-CT images were compared with results of 3-D renderings acquired with a laser scanner (resolution, 0.025 mm).RESULTS:Measurement of CTLO and CTTO images overestimated craniocaudal and mediolateral length of the articular surface by 4% to 10%. Measurement of micro-CT images underestimated craniocaudal and mediolateral length by 1%. Measurement of CTLO and CTTO images underestimated mediolateral radius of curvature by 15% and overestimated craniocaudal radius of curvature by > 100%; use of micro-CT images underestimated them by 3% and 5%, respectively. Mean ± SD surface deviation was 0.26 ± 0.09 mm for CTLO images, 0.30 ± 0.28 mm for CTTO images, and 0.04 ± 0.02 mm for micro-CT images.CONCLUSIONS AND CLINICAL RELEVANCE:Articular surface models derived from CT images had dimensional errors that approximately matched the voxel size. Thus, CT cannot be used to plan conforming arthroplasties in small joints and could lack precision when used to plan the correction of a limb deformity or repair of a fracture.

The history of supersaturated designs is full of unrealized promise. Despite the vast amount of literature on supersaturated designs, there is a scant record of their use in practice. We argue this imbalance is due to the designs’ inability to meet practitioners’ analysis expectations and the existing literature’s lack of clearly stated expectations. To resolve this issue, we first summarize practitioner concerns and expectations of supersaturated designs gathered from a survey. Next we tie exploration and optimization, two different goals of experimentation, to tradeoffs between power and type 1 error. The best strategy, which integrates a design’s analysis with its construction, is shown to depend on the goal. Group orthogonal supersaturated designs (Jones et al. 2019), when paired with our new, modified analysis, are shown to have high power even with many active factors and so are recommended for more exploratory studies. V ar(s+) designs (Weese et al. 2017), when paired with the Dantzig selector, are recommended as the initial experiment for optimization studies, as they can identify many active factors with a low type 1 error. The construction of both designs is less intuitive than traditional supersaturated designs and forces reflection on current best practices.