PURPOSE:Anterior knee pain remains a leading cause of dissatisfaction following total knee arthroplasty (TKA), often without overt mechanical complications. The biomechanical consequences of altered patellar thickness remain incompletely understood. This study evaluated the effects of reconstructed patellar thickness on patellofemoral joint (PFJ) kinematics and kinetics using a cadaveric model. METHODS:Eleven knees from six unembalmed whole-lower-extremity cadaveric specimens were tested with a robotic simulator and optical motion capture. Patellar thickness was altered in three conditions: understuffed (-3 mm), neutral (native) and overstuffed (+4 mm). PFJ kinematics and contact forces were assessed during passive flexion and simulated stair descent. Total force and regional pressures at the medial facet C1, central ridge C2 and lateral facet C3 were recorded using a custom multi-sensor array. RESULTS:Overstuffing significantly altered PFJ kinematics, including increased posterior translation (p < 0.001), reduced flexion (p = 0.001) and altered valgus alignment (p = 0.020). Composite PFJ force increased with thickness during passive flexion (overstuffed 59.9 N; neutral 33.9 N; understuffed 23.5 N; p < 0.001) and stair descent (50.5 N, 27.5 N, 17.9 N; p < 0.001). Regional analysis showed a lateral shift in loading, with the lateral facet most affected (p < 0.001). Understuffing reduced joint pressures with minimal kinematic change. CONCLUSION:Small increases in reconstructed patellar thickness significantly alter PFJ motion and load distribution, particularly lateral facet forces. These results highlight the biomechanical sensitivity of the PFJ to thickness variation and support the need for thinner, modular or patient-specific components to enable finer intraoperative control. Restoring native geometry may help reduce anterior knee pain and dissatisfaction after TKA. LEVEL OF EVIDENCE:Level IV, controlled laboratory study.

Sex bias in cohorts occurs when one sex is overrepresented. This study evaluated sex bias in study participant cohorts among abstracts accepted for the International Society of Biomechanics (ISB) Congress in 2025. Analyses focused on participant inclusion and abstract characteristics by lead author affiliated country, research topic, and cohort type (human/animal/computational). Using NotebookLM Gemini 1.5 Pro and a predefined prompt, data were extracted and analysed from 1,394 accepted abstracts. Of the 1202 abstracts with study participants, 43% did not report sex distribution, 13% had very small samples sizes (<6 participants), 18% had an unbalanced cohort (<30% representation of one sex), and only 26% had a balanced cohort. Among unbalanced abstracts, 64% showed an overrepresentation of male participants, for which a scientific justification was provided in only 1% of cases; in female‑biased abstracts, justification was provided in 18% of cases. Variation in sex representation and reporting was observed across countries, topics, and cohort types, with applied domains such as Military and Sport biomechanics and Animal studies demonstrating lower levels of balance or transparency. Taken together, these findings indicate that sex distribution is frequently insufficiently reported or unevenly represented at the conference abstract stage, but these may be subsequently modified if the study is still in process.

Purpose Thumb phalangeal and metacarpal fractures are undergoing surgical intervention more frequently and can be stabilized using a range of modalities. Intramedullary screw (IMS) fixation has become more common in the management of these fractures. Violation of the articular cartilage and the extensor tendon is a concern when IMS fixation is used. The aim of this study was to assess the cartilaginous and soft tissue footprint resulting from retrograde IMS insertion in the proximal phalanx and metacarpal of the thumb. Methods Ten cadaveric thumbs underwent radiographic guided insertion of 2.8, 3.6, and 4.1 mm diameter headless compression screws. The width of the extensor tendon lesion was measured using a ruler, and the surface area of the defect within the articular cartilage using digital processing software. Results The mean footprint in the proximal phalanx following insertion of a 2.8 mm screw involved 15.58% of the extensor tendon width and 2.45% of the articular surface, increasing to 16.19% and 3.73%, respectively, when a 3.6 mm screw was inserted. The mean lesion in the metacarpal when a 3.6 mm screw was used was 14.79% of the extensor tendon width and 2.35% of the articular surface, with 9.29% tendon width and 4.10% articular surface disruption observed with the insertion of a 4.1 mm screw. Conclusions The use of retrograde intramedullary screw fixation in the proximal phalanx and metacarpal of the thumb results in violation of the articular cartilage and terminal extensor tendon that is comparable to prior reports assessing the lesser digits. Clinical relevance This study quantifies the disruption of the articular cartilage and terminal extensor tendon when intramedullary screws are inserted into thumb proximal phalanges and metacarpals in a retrograde fashion and finds comparable results to prior reports assessing the lesser digits. Quantification of articular and extensor tendon defects can be used with clinical outcomes to examine the benefits and detriments of intramedullary fixation of the thumb metacarpal and proximal phalanx.

Background: Surgical treatment of Rockwood grade V acromioclavicular (AC) joint injuries remains varied. We hypothesized that the addition of a second suspensory device between the clavicle and coracoid would yield superior biomechanical results over a single device. We also hypothesized that the addition of an internal brace across the AC joint to a suspensory device would yield superior results over the suspensory device in isolation. Methods: A total of 24 cadaveric shoulders were dissected and randomized to 4 groups with 4 different constructs implanted: group A, single AC TightRope (Arthrex Inc.); group B, double AC TightRope; group C, single Knotless AC TightRope (Arthrex Inc.); group D: single Knotless AC TightRope with AC InternalBrace ligament augmentation (Arthrex Inc.). These were then loaded in a robotic arm (SIMVITRO), where 250 cycles of 50 N of force in the superior plane was applied. Dynamic creep, displacement, translation, and stiffness were assessed. Results: Testing was successfully completed for all specimens. There were no failures due to fracture or translation of the clavicle >5 mm from the starting position. Reduction was maintained with a mean superior displacement of 1.7 mm (+/- 1.4 mm). The mean peak-to- peak displacement, superior and posterior translation, dynamic creep, and stiffness did not differ significantly between the construct groups. Conclusion: This study did not demonstrate any significant biomechanical differences between groups in terms of displacement, translation, creep, or stiffness.

PURPOSE:To compare unicortical and bicortical screw fixation for dorsal plating of extra-articular proximal phalangeal fractures. METHODS:Midshaft osteotomies were performed on 21 cadaveric proximal phalanges. The phalanges were fixed via dorsal plating using either unicortical or bicortical locking screws and subjected to a three-point bending test. Cyclical loading was performed, increasing displacement by 0.4 mm every block of five cycles, until construct failure. Clinical failure was defined as 2 mm of displacement. Data collected included maximal force at failure, force at clinical failure, number of cycles to failure, and displacement to failure. RESULTS:Similar values were noted in maximal force at clinical failure and biomechanical failure, number of cycles to failure, and displacement to failure between phalanges fixed using unicortical or bicortical locking screws. CONCLUSIONS:Both unicortical and bicortical locking screws provide sufficient stability to facilitate early motion without risk of clinically relevant displacement during plate fixation of midshaft proximal phalangeal fractures. CLINICAL RELEVANCE:Unicortical dorsal plate-screw constructs should be considered, where appropriate, to prevent tendon injury or complications.

Tendons are force-transmitting structures which facilitate musculoskeletal functioning. Characterizing variations between different anatomical tendons, regions within tendons, as well as between the sexes and with age can improve understanding of tendon physiology and pathology. A systematic search of the literature was conducted to identify and summarize microscopic structural (histological) variations in normal/healthy tendons in relation to these variables (Tendon, Region, Age, Sex, and Other). Regional differences within individual tendons have been investigated in numerous studies, however investigations comparing histological variations between a range of different tendons are sparse, with most focusing on a few select tendons. When injured, ageing tendons typically have a greater degree of pathological changes than younger tendons, but few studies have documented variations in tendon histology throughout typical (uninjured) ageing or across large age spans. Similarly, sex-related observations of tendon structure are underreported. This narrative review summarizes studies on these topics and explores interactions between these variables, as well as the implications of these in the context of selecting control samples for studies of tendon pathology. Future studies should endeavour to improve knowledge of tendon structural variations-specifically focusing on normal tendons-to facilitate understanding of tendon structure-function relationships, physiological mechanisms involved in tendon damage/healing, and to aid clinical research and practice.

PURPOSE:Phalangeal fractures are undergoing surgical intervention more frequently and can be stabilized via a range of modalities. Intramedullary screw (IMS) fixation has become popular in the management of metacarpal and phalangeal fractures, with promising short- to medium-term results. Violation of articular cartilage and the terminal extensor tendon is a concern when IMS fixation is used in the middle phalanx. The aim of this study was to assess the cartilaginous and soft tissue footprint resulting from retrograde IMS insertion in the middle phalanx. METHODS:Ten cadaveric hands underwent radiographic guided insertion of 1.7 and 2.2 mm diameter headless compression screws. The width of the extensor tendon lesion and the surface area of the surface defect within the articular cartilage were both measured using digital processing software. RESULTS:Using the 1.7 mm screw, the mean lesion involved 12.1% of the width of the extensor tendon and 1.8% of the total articular surface. Using the 2.2 mm screw, the mean lesion involved 20.4% of the width of the extensor tendon and 3.5% of the total articular surface. CONCLUSIONS:The use of retrograde intramedullary screw fixation in the middle phalanx results in minimal violation of the articular cartilage and terminal extensor tendon. CLINICAL RELEVANCE:This study finds minimal disruption of the articular cartilage and terminal extensor tendon when intramedullary screws are inserted into middle phalanges in a retrograde fashion. When combined with promising early- to mid-term clinical outcomes, these findings render intramedullary screw fixation a suitable therapeutic option in this cohort.

Introduction: Complications arising from the patellofemoral joint (PFJ) represent the third most common cause for revision in total knee arthroplasty (TKA). Previous in vitro biomechanical studies have altered the native attachments of muscles controlling the PFJ. The purpose of this study was to design an in vitro biomechanical setup that would allow testing of both native and arthroplasty knee joints, specifically the PFJ, without disturbing the native attachments of the quadriceps and hamstrings muscles. Methods: After finalising a prototype, a pelvis-to-toe human cadaver specimen was tested. The simVITRO platform was used to simulate movement and control force trajectories. A motion capture system was used to capture the motion of the bones and to measure knee flexion angle and patellar movement with respect to the femur. The forces applied in the PFJ were measured using a custom patella sensor. Results: Displacement of the reflective cluster attached to the femur was measured during compression loading at different flexion angles, passive flexion and stairs descent trajectory. The femur showed less than 1 mm and 3 mm displacement with respect to the femur clamp in passive flexion and stairs descent. The most translation of 8.37 mm (<2% average femur length) was observed at 90° flexion which occurred at 483 N simulated compression force. Conclusion: This novel design provides a methodology for studying the biomechanics of the PFJ in vitro that preserves the soft tissues influencing the behaviour of the joint. This setup provides a biomechanics model that can be utilised to better understand and study the PFJ in vitro.

Purpose (the aim of the study): Osteoarthritis (OA) affects joints throughout the body with higher incidence in knee (20%), hand (17%) and hip (6%). Although the aetiology of OA is multifactorial, the primary risk factors may differ between joints (e.g. ankle OA is predominantly post-traumatic). The local environment has a critical role and the mechanical properties of tissues are important for maintaining joint structural integrity and function. While cartilage mechanics have been extensively investigated, results are commonly derived from a single joint and/or focus on the most commonly affected joints (e.g.

Tendon allograft and xenograft processing often involves one or more steps of freezing and thawing. As failure strength is an important graft consideration, this study aimed to evaluate effects on failure properties when varying freeze-thaw conditions. Kangaroo tendons, a potential xenograft source, were used to evaluate changes in ultimate tensile strength (UTS), failure strain and elastic modulus after exposure to different freezer-storage temperatures (-20 degrees C vs. -80 degrees C), storage durations (1, 3, 6, 9, or 12 months), number of freeze-thaw cycles (1, 2, 3, 4, 5, or 10), or freeze-thaw temperature ranges (including freezing in liquid nitrogen to thawing at 37 degrees C). Tendons stored for 6 or more months had significantly increased UTS and elastic modulus compared with 1 or 3 months of storage. This increase occurred irrespective of the freezing temperature (-20 degrees C vs. -80 degrees C) or the number of freeze-thaw cycles (1 vs. 10). In contrast, UTS, failure strain and the elastic modulus were no different between storage temperatures, number of freeze-thaw cycles and multiple freeze-thaw cycles across a range of freeze and thaw temperatures. Common freeze-thaw protocols did not negatively affect failure properties, providing flexibility for graft testing, storage, transportation and decellularisation procedures. However, the change in properties with the overall storage duration has implications for assessing the consistent performance of grafts stored for short versus extended periods of time (<6 months vs. >6 months), and the interpretation of data obtained from tissues of varying or unknown storage durations.

Purpose To provide a biomechanical comparison of dorsal plating, lateral plating and intramedullary screw [IMS] fixation for extra-articular proximal phalangeal fractures. Methods Midshaft osteotomies were performed on 36 cadaveric proximal phalanges. The phalanges were fixed by dorsal plating, lateral plating or IMS fixation, and subjected to a fourpoint bending force. Force was applied to achieve displacement of 1 mm/s, until construct failure or to a maximum of 10 mm of displacement. Clinical failure was defined as 2 mm of displacement, and force required to result in 1 mm and 2 mm of displacement was recorded, as was mode of failure. Results Dorsal plating [127.5 N +/- 52.6; 46.51-229.17] and lateral plating [77.1 N +/- 25.1; 48.3-113.8] required significantly greater force to achieve 1 mm of displacement when compared to IMS [41.2 N +/- 12.4; 20.6-62.3]. Dorsal plating [339.2 N +/- 91.8; 158.5-538.6] required significantly greater force than lateral plating [154.5 N +/- 33.8; 99.0-204.4] and intramedullary screw fixation [110.0 +/- 38.6; 51.1-189.3] to result in 2 mm of displacement. Lateral and dorsal plating constructs failed through plate bending, screw cut-out or plate failure, whilst IMS failed via implant deformity. All three constructs required greater force to result in even 1 mm of displacement than what is likely subjected through rehabilitation via active motion. Conclusions Lateral plating and IMS fixation offer sufficient stiffness to withstand the likely forces subjected via early active motion without displacement. Clinical relevance Dorsal plating required significantly greater force than lateral plating and intramedullary screw fixation to achieve 1 mm of displacement when used in extra-articular proximal phalangeal fractures in an in vitro setting. However, all three modalities confer enough stability to likely withstand the forces associated with active range of motion (J Hand Surg Am. 2024;49(3):247-252. Copyright (c) 2024 by the American Society for Surgery of the Hand. All rights reserved.)

BACKGROUND:Methodological heterogeneity hinders data comparisons across isolated studies of tendon and ligament properties, limiting clinical understanding and affecting the development and evaluation of replacement materials. PURPOSE:To create an open-access data set on the morphological, biomechanical, and biochemical properties of clinically important tendons and ligaments of the lower limb, using consistent methodologies, to enable direct tendon/ligament comparisons. STUDY DESIGN:Descriptive laboratory study. METHODS:Nineteen distinct lower limb tendons and ligaments were retrieved from 8 fresh-frozen human cadavers (5 male, 3 female; aged 49-65 years) including Achilles, tibialis posterior, tibialis anterior, fibularis (peroneus) longus, fibularis (peroneus) brevis, flexor hallucis longus, extensor hallucis longus, plantaris, flexor digitorum longus, quadriceps, patellar, semitendinosus, and gracilis tendons; anterior cruciate, posterior cruciate, medial collateral, and lateral collateral ligaments; and 10 mm-wide grafts from the contralateral quadriceps and patellar tendons. Outcomes included morphology (tissue length, ultrasound-quantified cross-sectional area [CSAUS], and major and minor axes), biomechanics (failure load, ultimate tensile strength [UTS], failure strain, and elastic modulus), and biochemistry (sulfated glycosaminoglycan [sGAG] and hydroxyproline contents). Tissue differences were analyzed using mixed-model regression. RESULTS:There was a range of similarities and differences between tendons and ligaments across outcomes. A key finding relating to potential graft tissue suitability was the comparable failure loads, UTS, CSAUS, sGAG, and hydroxyproline present between hamstring tendons (a standard graft source) and 5 tendons not typically used for grafting: fibularis (peroneus) longus and brevis, flexor and extensor hallucis longus, and flexor digitorum longus tendons. CONCLUSION:This study of lower limb tendons and ligaments has enabled direct comparison of morphological, biomechanical, and biochemical human tissue properties-key factors in the selection of suitable graft tissues. This analysis has identified 6 potential new donor tissues with properties comparable to currently used grafts. CLINICAL RELEVANCE:This extensive data set reduces the need to utilize data from incompatible sources, which may aid surgical decisions (eg, evidence to expand the range of tendons considered suitable for use as grafts) and may provide congruent design inputs for new biomaterials and computational models. The complete data set has been provided to facilitate further investigations, with the capacity to expand the resource to include additional outcomes and tissues.

Background: Spinal cord injury presents a significant burden globally, with traumatic spinal cord injury being the predominant cause historically. However, nontraumatic spinal cord injury (NTSCI) is emerging as a significant contributor, particularly in devel-oped nations, yet it remains poorly understood due to underreporting and misclassifi-cation. NTSCI, spanning various etiologies such as bony growths, vascular conditions, infections, neoplastic conditions, and immune disorders, poses unique challenges in diagnosis and treatment, often resulting in lifelong morbidity. This study addresses the lack of suitable animal models for NTSCI research, especially in neonatal animals. Methods: Utilizing a solid spacer approach, we developed a compression NTSCI model applicable to both neonatal and adult Sprague-Dawley rats. Results: Through anatomical measurements and in vivo experiments, we confirmed the feasibility and safety of the spacer insertion procedure and observed no acute off- target effects. Conclusion: The versatility of this model lies in its adaptability to different ages of rats, offering a cost-effective and reproducible means to induce graded injuries. Moreover, behavioral assessments demonstrated observable hindlimb function, validating the model's utility for studying functional outcomes. Although challenges persist, par-ticularly in accounting for spinal column growth in neonatal animals, this model fills a crucial gap in pediatric NTSCI research. By providing a platform to investigate patho-physiological mechanisms and test potential treatments, it offers promising avenues for advancing our understanding and management of pediatric NTSCI

Background: The use of allograft tendons has increased for primary and revision anterior cruciate ligament reconstruction, but allograft supply is currently limited to a narrow range of tendons and donors up to the age of 65 years. Expanding the range of donors and tendons could help offset an increasing clinical demand. Purpose: To investigate the effects of donor age, sex, height, and specific tendon on the mechanical properties of a range of human lower leg tendons. Study Design: Descriptive laboratory study. Methods: Nine tendons were retrieved from 39 fresh-frozen human cadaveric lower legs (35 donors [13 female, 22 male]; age, 49-99 years; height, 57-85 inches [145-216 cm]) including: Achilles tendon, tibialis posterior and anterior, fibularis longus and brevis, flexor and extensor hallucis longus, plantaris, and flexor digitorum longus. Tendons underwent tensile loading to failure measuring cross-sectional area (CSA), maximum load, strain at failure, ultimate tensile strength, and elastic modulus. Results from 332 tendons were analyzed using mixed-effects linear regression, accounting for donor age, sex, height, and weight. Results: Mechanical properties were significantly different among tendons and were substantially greater than the effects of donor characteristics. Significant effects of donor sex, age, and height were limited to specific tendons: Achilles tendon, tibialis posterior, and tibialis anterior. All other tendons were unaffected. The Achilles tendon was most influenced by donor variables: greater CSA in men (β = 15.45 mm2; Šidák adjusted P < .0001), decreased maximum load with each year of increased age (β = −17.20 N per year; adjusted P = .0253), and increased CSA (β = 1.92 mm2 per inch; adjusted P < .0001) and maximum load (β = 86.40 N per inch; adjusted P < .0001) with each inch of increased height. Conclusion: Mechanical properties vary significantly across different human tendons. The effects of donor age, sex, and height are relatively small, are limited to specific tendons, and affect different tendons uniquely. The findings indicate that age negatively affected only the Achilles tendon (maximum load) and challenge the exclusion of donors aged >65 years across all tendon grafts. Clinical Relevance: The findings support including a broader range of tendons for use as allografts for anterior cruciate ligament reconstruction and reviewing the current exclusion criterion of donors aged >65 years.

Climate change and fishing have impacted marine species richness ( R ) at global and local scales. It has previously been shown that R of the fish community in the North Sea has increased since the early 1980’s. Over the same period, ocean temperature has increased, and fishing mortality has decreased in the North Sea. Because these are confounded over time, either trend could plausibly explain the increase in R . Therefore, a logic-based approach was used to disentangle the effects of temperature and fishing on R , using spatio-temporal models fitted to survey data. To investigate the effect of temperature on R , fish species were subset by thermal affinity, as either Lusitanian (warm) or Boreal (cold) species. To investigate the effect of fishing mortality on R , species were subset by management category as either quota (assumed to be targeted) or non-quota species. Trends in these subsets were plotted separately to investigate which subsets of the fish community have contributed to the overall R increase. Over three decades, fish R increased by an average of 2.5 species per haul. These increases were predominantly of Lusitanian non-quota species (1.9). A small increase was observed in quota species (0.6); however, this increase was driven by quota-Lusitanian species (0.4). Our results suggest that temperature rather than fishing mortality was the driver of R increase in the North Sea since 1991 and highlight the importance of long-term monitoring in detecting ecological responses to climate change at the community level.

This paper presents a labeling methodology for marine life data using a weakly supervised learning framework. The methodology iteratively trains a deep learning model using non-expert labels obtained from crowdsourcing. This approach enables us to converge on a labeled image dataset through multiple training and production loops that leverage crowdsourcing interfaces. We present our algorithm and its results on two separate sets of image data collected using the Seabed autonomous underwater vehicle. The first dataset consists of 10,505 images that were point annotated by NOAA biologists. This dataset allows us to validate the accuracy of our labeling process. We also apply our algorithm and methodology to a second dataset consisting of 3,968 completely unlabeled images. These image categories are challenging to label, such as sponges. Qualitatively, our results indicate that training with a tiny subset and iterating on those results allows us to converge to a large, highly annotated dataset with a small number of iterations. To demonstrate the effectiveness of our methodology quantitatively, we tabulate the mean average precision (mAP) of the model as the number of iterations increases.

Purpose: Significant abnormalities in the synovial T-cell profile of osteoarthritis (OA) patients have been found implicating a crucial role for proinflammatory CD4+ T-helper cell subsets, such as Th1 and Th17. Interleukin (IL)12 is not only directly involved in the polarization of Th1 cells, but also promotes polarization of Th17 cells via a shared p40 subunit with IL23, making the IL12 pathway a promising target for immunomodulation. Although evidence exists that highlights the importance of this pathway during the development of rheumatoid arthritis to date the role of this pathway in OA is unknown. We investigated the cellular and histopathological response following IL12 pathway modulation in an unilateral anterior-cruciate-ligament (ACL) rupture model of post-traumatic (pt)OA.

Objectives: Histological scoring remains the gold-standard for quantifying post-traumatic osteoarthritis (ptOA) in animal models, allowing concurrent evaluation of numerous joint tissues. Available systems require scoring multiple sections/joint making analysis laborious and expensive. We investigated if a single section allowed equivalent quantitation of pathology in different joint tissues and disease stages, in three ptOA models.Method: Male 10-12-week-old C57BL/6 mice underwent surgical medial-meniscal-destabilization, anterior-cruciate-ligament (ACL) transection, non-invasive-ACL-rupture, or served as sham-surgical, non-invasive-ACL-strain, or na & iuml;ve/non-operated controls. Mice (n = 12/group) were harvested 1-, 4-, 8-, and 16-week post-intervention. Serial sagittal toluidine-blue/fast-green stained sections of the medial-femoro-tibial joint (n = 7/joint, 84 mu m apart) underwent blinded scoring of 40 histology-outcomes. We evaluated agreement between single-slide versus entire slide-set maximum or median scores (weighted-kappa), and sensitivity/specificity of single-slide versus median/maximum to detect OA pathology.Results: A single optimal mid-sagittal section showed excellent agreement with median (weighted-kappa 0.960) and maximum (weighted-kappa 0.926) scores. Agreement for individual histology-outcomes was high with only 19/240 median and 15/240 maximum scores having a weighted-kappa <= 0.4, the majority of these (16/19 and 11/15) in control groups. Statistically-significant histology-outcome differences between ptOA models and their controls detected with the entire slide-set were reliably reproduced using a single slide (sensitivity >93.15%, specificity >93.10%). The majority of false-negatives with single-slide scoring were meniscal and subchondral bone histology-outcomes (89%) and occurred in weeks 1-4 post-injury (84%).Conclusion: A single mid-sagittal slide reduced the time needed to score diverse histopathological changes by 87% without compromising the sensitivity or specificity of the analysis, across a variety of ptOA models and time-points.(c) 2023 The Author(s). Published by Elsevier Ltd on behalf of Osteoarthritis Research Society International. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/).