Smart active particles learn and transcend bacterial foraging strategies.

Throughout evolution, bacteria and other microorganisms have learned efficient foraging strategies that exploit characteristic properties of their unknown environment. While much research has been devoted to the exploration of statistical models describing the dynamics of foraging bacteria and other (micro-) organisms, little is known, regarding the question of how good the learned strategies actually are. This knowledge gap is largely caused by the absence of methods allowing to systematically develop alternative foraging strategies to compare with. In the present work, we use deep reinforcement learning to show that a smart run-and-tumble agent, which strives to find nutrients for its survival, learns motion patterns that are remarkably similar to the trajectories of chemotactic bacteria. Strikingly, despite this similarity, we also find interesting differences between the learned tumble rate distribution and the one that is commonly assumed for the run and tumble model. We find that these differences equip the agent with significant advantages regarding its foraging and survival capabilities. Our results uncover a generic route to use deep reinforcement learning for discovering search and collection strategies that exploit characteristic but initially unknown features of the environment. These results can be used, e.g., to program future microswimmers, nanorobots, and smart active particles for tasks like searching for cancer cells, micro-waste collection, or environmental remediation.

An Ensemble Docking Approach for Analyzing and Designing Aptamer Heterodimers Targeting VEGF165.

Vascular endothelial growth factor 165 (VEGF165) is a prominent isoform of the VEGF-A protein that plays a crucial role in various angiogenesis-related diseases. It is homodimeric, and each of its monomers is composed of two domains connected by a flexible linker. DNA aptamers, which have emerged as potent therapeutic molecules for many proteins with high specificity and affinity, can also work for VEGF165. A DNA aptamer heterodimer composed of monomers of V7t1 and del5-1 connected by a flexible linker (V7t1:del5-1) exhibits a greater binding affinity with VEGF165 compared to either of the two monomers alone. Although the structure of the complex formed between the aptamer heterodimer and VEGF165 is unknown due to the highly flexible linkers, gaining structural information will still be valuable for future developments. Toward this end of accessing structural information, we adopt an ensemble docking approach here. We first obtain an ensemble of structures for both VEGF165 and the aptamer heterodimer by considering both small- and large-scale motions. We then proceed through an extraction process based on ensemble docking, molecular dynamics simulations, and binding free energy calculations to predict the structures of the VEGF165/V7t1:del5-1 complex. Through the same procedures, we reach a new aptamer heterodimer that bears a locked nucleic acid-modified counterpart of V7t1, namely RNV66:del5-1, which also binds well with VEGF165. We apply the same protocol to the monomeric units V7t1, RNV66, and del5-1 to target VEGF165. We observe that V7t1:del5-1 and RNV66:del5-1 show higher binding affinities with VEGF165 than any of the monomers, consistent with experiments that support the notion that aptamer heterodimers are more effective anti-VEGF165 aptamers than monomeric aptamers. Among the five different aptamers studied here, the newly designed RNV66:del5-1 shows the highest binding affinity with VEGF165. We expect that our ensemble docking approach can help in de novo designs of homo/heterodimeric anti-angiogenic drugs to target the homodimeric VEGF165.

A Wavelet-Based Approach for Motion Artifact Reduction in Ambulatory Seismocardiography.

Wearable sensing has become a vital approach to cardiac health monitoring, and seismocardiography (SCG) is emerging as a promising technology in this field. However, the applicability of SCG is hindered by motion artifacts, including those encountered in practice of which the strongest source is walking. This holds back the translation of SCG to clinical settings. We therefore investigated techniques to enhance the quality of SCG signals in the presence of motion artifacts. To simulate ambulant recordings, we corrupted a clean SCG dataset with real-walking-vibrational noise. We decomposed the signal using several empirical-mode-decomposition methods and the maximum overlap discrete wavelet transform (MODWT). By combining MODWT, time-frequency masking, and nonnegative matrix factorization, we developed a novel algorithm which leveraged the vertical axis accelerometer to reduce walking vibrations in dorsoventral SCG. The accuracy and applicability of our method was verified using heart rate estimation. We used an interactive selection approach to improve estimation accuracy. The best decomposition method for reduction of motion artifact noise was the MODWT. Our algorithm improved heart rate estimation from 0.1 to 0.8 r-squared at -15 dB signal-to-noise ratio (SNR). Our method reduces motion artifacts in SCG signals up to a SNR of -19 dB without requiring any external assistance from electrocardiography (ECG). Such a standalone solution is directly applicable to the usage of SCG in daily life, as a content-rich replacement for other wearables in clinical settings, and other continuous monitoring scenarios. In applications with higher noise levels, ECG may be incorporated to further enhance SCG and extend its usable range. This work addresses the challenges posed by motion artifacts, enabling SCG to offer reliable cardiovascular insights in more difficult scenarios, and thereby facilitating wearable monitoring in daily life and the clinic.

Study on the variable length simple pendulum oscillation based on the relative mode transfer method.

In this study, we employed the principle of Relative Mode Transfer Method (RMTM) to establish a model for a single pendulum subjected to sudden changes in its length. An experimental platform for image processing was constructed to accurately track the position of a moving ball, enabling experimental verification of the pendulum's motion under specific operating conditions. The experimental data demonstrated excellent agreement with simulated numerical results, validating the effectiveness of the proposed methodology. Furthermore, we performed simulations of a double obstacle pendulum system, investigating the effects of different parameters, including obstacle pin positions, quantities, and initial release angles, on the pendulum's motion through numerical simulations. This research provides novel insights into addressing the challenges associated with abrupt and continuous changes in pendulum length.

Chest Wall Motion Model of Cardiac Activity for Radar-Based Vital-Sign-Detection System.

An increasing number of studies on non-contact vital sign detection using radar are now beginning to turn to data-driven neural network approaches rather than traditional signal-processing methods. However, there are few radar datasets available for deep learning due to the difficulty of acquiring and labeling the data, which require specialized equipment and physician collaboration. This paper presents a new model of heartbeat-induced chest wall motion (CWM) with the goal of generating a large amount of simulation data to support deep learning methods. An in-depth analysis of published CWM data collected by the VICON Infrared (IR) motion capture system and continuous wave (CW) radar system during respiratory hold was used to summarize the motion characteristics of each stage within a cardiac cycle. In combination with the physiological properties of the heartbeat, appropriate mathematical functions were selected to describe these movement properties. The model produced simulation data that closely matched the measured data as evaluated by dynamic time warping (DTW) and the root-mean-squared error (RMSE). By adjusting the model parameters, the heartbeat signals of different individuals were simulated. This will accelerate the application of data-driven deep learning methods in radar-based non-contact vital sign detection research and further advance the field.

Patient's Healthy-Limb Motion Characteristic-Based Assist-As-Needed Control Strategy for Upper-Limb Rehabilitation Robots.

The implementation of a progressive rehabilitation training model to promote patients' motivation efforts can greatly restore damaged central nervous system function in patients. Patients' active engagement can be effectively stimulated by assist-as-needed (AAN) robot rehabilitation training. However, its application in robotic therapy has been hindered by a simple determination method of robot-assisted torque which focuses on the evaluation of only the affected limb's movement ability. Moreover, the expected effect of assistance depends on the designer and deviates from the patient's expectations, and its applicability to different patients is deficient. In this study, we propose a control method with personalized treatment features based on the idea of estimating and mapping the stiffness of the patient's healthy limb. This control method comprises an interactive control module in the task-oriented space based on the quantitative evaluation of motion needs and an inner-loop position control module for the pneumatic swing cylinder in the joint space. An upper-limb endpoint stiffness estimation model was constructed, and a parameter identification algorithm was designed. The upper limb endpoint stiffness which characterizes the patient's ability to complete training movements was obtained by collecting surface electromyographic (sEMG) signals and human-robot interaction forces during patient movement. Then, the motor needs of the affected limb when completing the same movement were quantified based on the performance of the healthy limb. A stiffness-mapping algorithm was designed to dynamically adjust the rehabilitation training trajectory and auxiliary force of the robot based on the actual movement ability of the affected limb, achieving AAN control. Experimental studies were conducted on a self-developed pneumatic upper limb rehabilitation robot, and the results showed that the proposed AAN control method could effectively estimate the patient's movement needs and achieve progressive rehabilitation training. This rehabilitation training robot that simulates the movement characteristics of the patient's healthy limb drives the affected limb, making the intensity of the rehabilitation training task more in line with the patient's pre-morbid limb-use habits and also beneficial for the consistency of bilateral limb movements.

Evaluation of the accuracy of a UWB tracker for in-home positioning for older adults.

The population of older adults is rapidly growing. In-home monitoring systems have been used to support aging-in-place. Ambient sensors or wearable localizers can be used but may be too low resolution, while camera systems are invasive to privacy. Ultra-wideband (UWB) localization offers precise positioning by placing anchors throughout the house and wearing a tag that is tracked by the anchors. In this study, the accuracy of UWB for indoor tracking was evaluated in a motion capture gait lab and in a mock condo in the Glenrose Rehabilitation Hospital. First, the configuration of UWB was tested, changing factors related to sampling time, anchor placement and line-of-sight. Comparing these factors to the configurations recommended by the manufacturer guidelines, accuracies remained within 14 cm. We then performed static and dynamic accuracy tests, with dynamic testing comprised of rolling and walking motions. In the motion capture lab, we found localization accuracies of 7.0 ± 11.1 cm while in the mock condo, we found accuracies of 27.3 ± 12.9 cm. Dynamic testing with rolling motions had an average of 19.1 ± 1.6 cm while walking was 20.5 ± 4.2 cm. The mean accuracy of UWB is within the 30 cm target for indoor localization.

3D far-field Lidar sensing and computational modeling for human identification.

3D sensors offer depth sensing that may be used for task-specific data processing and computational modeling. Many existing methods for human identification using 3D depth sensors primarily focus on Kinect data, where the range is very limited. This work considers a 3D long-range Lidar sensor for far-field imaging of human subjects in 3D Lidar full motion video (FMV) of "walking" action. 3D Lidar FMV data for human subjects are used to develop computational modeling for automated human silhouette and skeleton extraction followed by subject identification. We propose a matrix completion algorithm to handle missing data in 3D FMV due to self-occlusion and occlusion from other subjects for 3D skeleton extraction. We further study the effect of noise in the 3D low resolution far-field Lidar data in human silhouette extraction performance of the model. Moreover, this work addresses challenges associated with far-field 3D Lidar including learning with a limited amount of data and low resolution. Moreover, we evaluate the proposed computational algorithm using a gallery of 10 subjects for human identification and show that our method is competitive with the state-of-the-art OpenPose and V2VPose skeleton extraction models using the same dataset for human identification.

(Don't) look where you are going: Evidence for a travel direction signal in humans that is independent of head direction.

We often assume that travel direction is redundant with head direction, but from first principles, these two factors provide differing spatial information. Although head direction has been found to be a fundamental component of human navigation, it is unclear how self-motion signals for travel direction contribute to forming a travel trajectory. Employing a novel motion adaptation paradigm from visual neuroscience designed to preclude a contribution of head direction, we found high-level aftereffects of perceived travel direction, indicating that travel direction is a fundamental component of human navigation. Interestingly, we discovered a higher frequency of reporting perceived travel toward the adapted direction compared to a no-adapt control-an aftereffect that runs contrary to low-level motion aftereffects. This travel aftereffect was maintained after controlling for possible response biases and approaching effects, and it scaled with adaptation duration. These findings demonstrate the first evidence of how a pure travel direction signal might be represented in humans, independent of head direction. (PsycInfo Database Record (c) 2024 APA, all rights reserved).

Leadership dynamics in musical groups: Quantifying effects of musical structure on directionality of influence in concert performance videos.

Music ensemble performance provides an ecologically valid context for investigating leadership dynamics in small group interactions. Musical texture, specifically the relative salience of simultaneously sounding ensemble parts, is a feature that can potentially alter leadership dynamics by introducing hierarchical relationships between individual parts. The present study extended previous work on quantifying interpersonal coupling in musical ensembles by examining the relationship between musical texture and leader-follower relations, operationalised as directionality of influence between co-performers' body motion in concert video recordings. It was hypothesised that the directionality of influence, indexed by Granger Causality, would be greater for 'homophonic' textures with a clear distinction between melody and accompaniment parts than for 'polyphonic' textures with less distinction between melody and accompaniment. This hypothesis was tested by using pose estimation algorithms to track instrumentalists' body movements in a string quartet and a clarinet quintet, and then applying Granger Causality analysis to their head motion to estimate directional influence between instrumentalist pairs for sections of the pieces that varied in texture. It was found that Granger Causality values were generally higher (indicating greater directionality of influence) for homophonic than polyphonic textures. Furthermore, considering melody and accompaniment instrument roles revealed more evidence for the melody instrument influencing accompanying instruments than vice versa, plus a high degree of directionality among accompanying instruments, in homophonic textures. These observed patterns of directional information flow in co-performer body motion are consistent with changing leader-follower relations depending on hierarchical relations between ensemble parts in terms of the relative salience of melodic material in the musical texture. The finding that automatic pose estimation can detect modulations of leadership dynamics in standard video recordings under naturalistic performance conditions has implications for investigating interpersonal coordination in large-scale music video datasets representing different cultural traditions, and for exploring nonverbal communication in group activities more generally.

Brownian processes in human motor control support descending neural velocity commands.

The motor neuroscience literature suggests that the central nervous system may encode some motor commands in terms of velocity. In this work, we tackle the question: what consequences would velocity commands produce at the behavioral level? Considering the ubiquitous presence of noise in the neuromusculoskeletal system, we predict that velocity commands affected by stationary noise would produce "random walks", also known as Brownian processes, in position. Brownian motions are distinctively characterized by a linearly growing variance and a power spectral density that declines in inverse proportion to frequency. This work first shows that these Brownian processes are indeed observed in unbounded motion tasks e.g., rotating a crank. We further predict that such growing variance would still be present, but bounded, in tasks requiring a constant posture e.g., maintaining a static hand position or quietly standing. This hypothesis was also confirmed by experimental observations. A series of descriptive models are investigated to justify the observed behavior. Interestingly, one of the models capable of accounting for all the experimental results must feature forward-path velocity commands corrupted by stationary noise. The results of this work provide behavioral support for the hypothesis that humans plan the motion components of their actions in terms of velocity.

Chemical Cross-Linking Cellulose Aerogel-Based Triboelectric Nanogenerators for Energy Harvesting and Sensing Human Activities.

Zinc oxide (ZnO) is a widely employed material for enhancing the performance of cellulose-based triboelectric nanogenerators (C-TENGs). Our study provides a novel chemical interpretation for the improved output efficiency of ZnO in C-TENGs. C-TENGs exhibit excellent flexibility and integration, achieving a maximum open-circuit voltage (Voc) of 210 V. The peak power density is 54.4 µW/cm2 with a load resistance of 107 Ω, enabling the direct powering of 191 light-emitting diodes with the generated electrical output. Moreover, when deployed as self-powered sensors, C-TENGs exhibit prolonged operational viability and responsiveness, adeptly discerning bending and motion induced by human interaction. The device's sensitivity, flexibility, and stability position it as a promising candidate for a diverse array of energy-harvesting applications and advanced healthcare endeavors. Specifically, envisaging sensitized wearable sensors for human activities underscores the multifaceted potential of C-TENGs in enhancing both energy-harvesting technologies and healthcare practices.

Pelvic target volume inter-fractional motion during radiotherapy for cervical cancer with daily iterative cone beam computed tomography.

BACKGROUND: Tumor regression and organ movements indicate that a large margin is used to ensure target volume coverage during radiotherapy. This study aimed to quantify inter-fractional movements of the uterus and cervix in patients with cervical cancer undergoing radiotherapy and to evaluate the clinical target volume (CTV) coverage. METHODS: This study analyzed 303 iterative cone beam computed tomography (iCBCT) scans from 15 cervical cancer patients undergoing external beam radiotherapy. CTVs of the uterus (CTV-U) and cervix (CTV-C) contours were delineated based on each iCBCT image. CTV-U encompassed the uterus, while CTV-C included the cervix, vagina, and adjacent parametrial regions. Compared with the planning CTV, the movement of CTV-U and CTV-C in the anterior-posterior, superior-inferior, and lateral directions between iCBCT scans was measured. Uniform expansions were applied to the planning CTV to assess target coverage. RESULTS: The motion (mean ± standard deviation) in the CTV-U position was 8.3 ± 4.1 mm in the left, 9.8 ± 4.4 mm in the right, 12.6 ± 4.0 mm in the anterior, 8.8 ± 5.1 mm in the posterior, 5.7 ± 5.4 mm in the superior, and 3.0 ± 3.2 mm in the inferior direction. The mean CTV-C displacement was 7.3 ± 3.2 mm in the left, 8.6 ± 3.8 mm in the right, 9.0 ± 6.1 mm in the anterior, 8.4 ± 3.6 mm in the posterior, 5.0 ± 5.0 mm in the superior, and 3.0 ± 2.5 mm in the inferior direction. Compared with the other tumor (T) stages, CTV-U and CTV-C motion in stage T1 was larger. A uniform CTV planning treatment volume margin of 15 mm failed to encompass the CTV-U and CTV-C in 11.1% and 2.2% of all fractions, respectively. The mean volume change of CTV-U and CTV-C were 150% and 51%, respectively, compared with the planning CTV. CONCLUSIONS: Movements of the uterine corpus are larger than those of the cervix. The likelihood of missing the CTV is significantly increased due to inter-fractional motion when utilizing traditional planning margins. Early T stage may require larger margins. Personal radiotherapy margining is needed to improve treatment accuracy.

Giant honeybees (Apis dorsata) trade off defensiveness against periodic mass flight activity.

The giant honeybee Apis dorsata (Fabricius, 1793) is an evolutionarily ancient species that builds its nests in the open. The nest consists of a single honeycomb covered with the bee curtain which are several layers of worker bees that remain almost motionless with their heads up and abdomens down on the nest surface, except for the mouth area, the hub between inner- and outer-nest activities. A colony may change this semi-quiescence several times a day, depending on its reproductive state and ambient temperature, to enter the state of mass flight activity (MFA), in which nest organisation is restructured and defense ability is likely to be suppressed (predicted by the mass-flight-suspend-defensiveness hypothesis). For this study, three episode of MFA (mfa1-3) of a selected experimental nest were analysed in a case study with sequences of >60 000 images at 50 Hz, each comprise a short pre-MFA session, the MFA and the post-MFA phase of further 10 min. To test colony defensiveness under normative conditions, a dummy wasp was cyclically presented with a standardised motion programme (Pd) with intervening sessions without such a presentation (nPd). Motion activity at five selected surveillance zones (sz1-5) on the nest were analysed. In contrast to mfa1,2, in mfa3 the experimental regime started with the cyclic presentation of the dummy wasp only after the MFA had subsided. As a result, the MFA intensity in mfa3 was significantly lower than in mfa1-2, suggesting that a colony is able to perceive external threats during the MFA. Characteristic ripples appear in the motion profiles, which can be interpreted as a start signal for the transition to MFA. Because they are strongest in the mouth zone and shift to higher frequencies on their way to the nest periphery, it can be concluded that MFA starts earlier in the mouth zone than in the peripheral zones, also suggesting that the mouth zone is a control centre for the scheduling of MFA. In Pd phases of pre- and postMFA, the histogram-based motion spectra are biphasic, suggesting two cohorts in the process, one remaining at quiescence and the other involved in shimmering. Under MFA, nPd and Pd spectra were typically Gaussian, suggesting that the nest mates with a uniform workload shifted to higher motion activity. At the end of the MFA, the spectra shift back to the lower motion activities and the Pd spectra form a biphasic again. This happens a few minutes earlier in the peripheral zones than in the mouth zone. Using time profiles of the skewness of the Pd motion spectra, the mass-flight-suspend-defensiveness hypothesis is confirmed, whereby the inhibition of defense ability was found to increase progressively during the MFA. These sawtooth-like time profiles of skewness during MFA show that defense capability is recovered again quite quickly at the end of MFA. Finally, with the help of the Pd motion spectra, clear indications can be obtained that the giant honeybees engage in a decision in the sense of a tradeoff between MFA and collective defensiveness, especially in the regions in the periphery to the mouth zone.

Collective behavior from surprise minimization.

Collective motion is ubiquitous in nature; groups of animals, such as fish, birds, and ungulates appear to move as a whole, exhibiting a rich behavioral repertoire that ranges from directed movement to milling to disordered swarming. Typically, such macroscopic patterns arise from decentralized, local interactions among constituent components (e.g., individual fish in a school). Preeminent models of this process describe individuals as self-propelled particles, subject to self-generated motion and "social forces" such as short-range repulsion and long-range attraction or alignment. However, organisms are not particles; they are probabilistic decision-makers. Here, we introduce an approach to modeling collective behavior based on active inference. This cognitive framework casts behavior as the consequence of a single imperative: to minimize surprise. We demonstrate that many empirically observed collective phenomena, including cohesion, milling, and directed motion, emerge naturally when considering behavior as driven by active Bayesian inference-without explicitly building behavioral rules or goals into individual agents. Furthermore, we show that active inference can recover and generalize the classical notion of social forces as agents attempt to suppress prediction errors that conflict with their expectations. By exploring the parameter space of the belief-based model, we reveal nontrivial relationships between the individual beliefs and group properties like polarization and the tendency to visit different collective states. We also explore how individual beliefs about uncertainty determine collective decision-making accuracy. Finally, we show how agents can update their generative model over time, resulting in groups that are collectively more sensitive to external fluctuations and encode information more robustly.

Confinement of a Styryl Dye into Nanoporous Aluminophosphates: Channels vs. Cavities.

Styryl dyes are generally poor fluorescent molecules inherited from their flexible molecular structures. However, their emissive properties can be boosted by restricting their molecular motions. A tight confinement into inorganic molecular sieves is a good strategy to yield highly fluorescent hybrid systems. In this work, we compare the confinement effect of two Mg-aluminophosphate zeotypes with distinct pore systems (the AEL framework, a one-dimensional channeled structure with elliptical pores of 6.5 Å × 4.0 Å, and the CHA framework, composed of large cavities of 6.7 Å × 10.0 Å connected by eight-ring narrower windows) for the encapsulation of 4-DASPI styryl dye (trans-4-[4-(Dimethylamino)styryl]-1-methylpyridinium iodide). The resultant hybrid systems display significantly improved photophysical features compared to 4-DASPI in solution as a result of tight confinement in both host inorganic frameworks. Molecular simulations reveal a tighter confinement of 4-DASPI in the elliptical channels of AEL, explaining its excellent photophysical properties. On the other hand, a singular arrangement of 4-DASPI dye is found when confined within the cavity-based CHA framework, where the 4-DASPI molecule spans along two adjacent cavities, with each aromatic ring sitting on these adjacent cavities and the polymethine chain residing within the narrower eight-ring window. However, despite the singularity of this host-guest arrangement, it provides less tight confinement for 4-DASPI than AEL, resulting in a slightly lower quantum yield.

From an obliquely falling rod in a viscous fluid to the motion of suspended magnetic bead chains that are driven by a gradient magnetic field and that make an arbitrary angle with the magnetic force vector: A Stokes flow study.

In view of the growing role of magnetic particles under magnetic field influence in medical and other applications, and perforce the bead chaining, it is important to understand more generally the chain dynamics. As is well known, in the presence of a magnetic field, magnetic beads tend to form chains that are aligned with the magnetic field vector. In addition, if there is a magnetic field gradient, there will be a magnetic force acting on this chain. The main goal of the present research is to study the motion of a magnetic bead chain that makes an arbitrary angle with the magnetic force vector in the Stokes flow limit, that is, in the limit of zero Reynolds number. We used the public-domain computer program HYDRO++ to calculate the mobility matrix, which relates the magnetic force acting on the chain to the velocity of the chain, for a chain of N beads making an arbitrary angle with the magnetic force vector. Because of the presence of off-diagonal elements of the mobility matrix, as the chain is drawn in the direction of the magnetic force, it is also deflected to the side. We derived analytic solutions for this motion. Also, for bead chains moving in directions both parallel and perpendicular to their lengths, we fit three-parameter functions to solutions from HYDRO++. We found the fits to be excellent. Combining these results with the analytic solutions, we obtained expressions for the velocity components for the bead chains that provide excellent fits to HYDRO++ solutions for arbitrary angles. Finally, we apply the methodology used for the bead chain studies to the study of an obliquely falling rod in a viscous fluid and derive analytic solutions for the velocity components of the obliquely falling rod.

Estimating the mechanical cost of transport in human walking with a simple kinematic data-driven mechanical model.

This work utilizes a simplified, streamlined approach to study the mechanical cost of transport in human walking. Utilizing the kinematic motion data of the center of mass, velocities and accelerations are determined using kinematic analysis; the applied force is then obtained using inverse dynamics. We calculate the mechanical cost of transport per step from both synthetic and measured data, using a very simple mechanical model of walking. The approach studied can serve as an informative gait characteristic to monitor rehabilitation in human walking.

Simulating rigid head motion artifacts on brain magnitude MRI data-Outcome on image quality and segmentation of the cerebral cortex.

Magnetic Resonance Imaging (MRI) datasets from epidemiological studies often show a lower prevalence of motion artifacts than what is encountered in clinical practice. These artifacts can be unevenly distributed between subject groups and studies which introduces a bias that needs addressing when augmenting data for machine learning purposes. Since unreconstructed multi-channel k-space data is typically not available for population-based MRI datasets, motion simulations must be performed using signal magnitude data. There is thus a need to systematically evaluate how realistic such magnitude-based simulations are. We performed magnitude-based motion simulations on a dataset (MR-ART) from 148 subjects in which real motion-corrupted reference data was also available. The similarity of real and simulated motion was assessed by using image quality metrics (IQMs) including Coefficient of Joint Variation (CJV), Signal-to-Noise-Ratio (SNR), and Contrast-to-Noise-Ratio (CNR). An additional comparison was made by investigating the decrease in the Dice-Sørensen Coefficient (DSC) of automated segmentations with increasing motion severity. Segmentation of the cerebral cortex was performed with 6 freely available tools: FreeSurfer, BrainSuite, ANTs, SAMSEG, FastSurfer, and SynthSeg+. To better mimic the real subject motion, the original motion simulation within an existing data augmentation framework (TorchIO), was modified. This allowed a non-random motion paradigm and phase encoding direction. The mean difference in CJV/SNR/CNR between the real motion-corrupted images and our modified simulations (0.004±0.054/-0.7±1.8/-0.09±0.55) was lower than that of the original simulations (0.015±0.061/0.2±2.0/-0.29±0.62). Further, the mean difference in the DSC between the real motion-corrupted images was lower for our modified simulations (0.03±0.06) compared to the original simulations (-0.15±0.09). SynthSeg+ showed the highest robustness towards all forms of motion, real and simulated. In conclusion, reasonably realistic synthetic motion artifacts can be induced on a large-scale when only magnitude MR images are available to obtain unbiased data sets for the training of machine learning based models.