Sex estimation from the tarsal bones in a Portuguese sample: a machine learning approach

A statistical evaluation of the sexual dimorphism of the acetabulum in an Iberian population

Sex estimation is essential for human identification within bioarchaeological and medico-legal contexts. Amongst the sexually dimorphic skeletal elements commonly utilised for this purpose, the pelvis is usually preferred because of its direct relationship with reproduction. Furthermore, the posterior part of the innominate bone has proven to have better preservation within degraded contexts. With the aim of investigating the potential of the vertical acetabular diameter as a sex marker, 668 documented individuals from three different Iberian skeletal collections were randomly divided into training and test samples and eventually analysed using different statistical approaches. Two traditional (Discriminant Function Analysis and Logistic Regression Analysis) and four Machine learning methodologies (Support Vector Classification, Decision Tree Classification, k Nearest Neighbour Classification, and Neural Networks) were performed and compared. Amongst these statistical modalities, Machine Learning methodologies yielded better accuracy outcomes, with DTC garnering highest accuracy percentages of 83.59% and 89.85% with the sex-pooled and female samples, respectively. With males, ANN yielded highest accuracy percentage of 87.70%, when compared to other statistical approaches. Higher accuracy obtained with ML, along with its minimal statistical assumptions, warrant these approaches to be increasingly utilised for further investigations involving sex estimation and human identification. In this line, the creation of a statistical platform with easier user interface can render such robust statistical modalities accessible to researchers and practitioners, effectively maximising its practical use. Future investigations should attempt to achieve this goal, alongside examining the influence of factors such as age, on the obtained accuracy outcomes.

Morphological Species Delimitation in The Western Pond Turtle (Actinemys): Can Machine Learning Methods Aid in Cryptic Species Identification?

As the discovery of cryptic species has increased in frequency, there has been an interest in whether geometric morphometric data can detect fine-scale patterns of variation that can be used to morphologically diagnose such species. We used a combination of geometric morphometric data and an ensemble of five supervised machine learning methods (MLMs) to investigate whether plastron shape can differentiate two putative cryptic turtle species, Actinemys marmorata and Actinemys pallida. Actinemys has been the focus of considerable research due to its biogeographic distribution and conservation status. Despite this work, reliable morphological diagnoses for its two species are still lacking. We validated our approach on two datasets, one consisting of eight morphologically disparate emydid species, the other consisting of two subspecies of Trachemys (T. scripta scripta, T. scripta elegans). The validation tests returned near-perfect classification rates, demonstrating that plastron shape is an effective means for distinguishing taxonomic groups of emydids via MLMs. In contrast, the same methods did not return high classification rates for a set of alternative phylogeographic and morphological binning schemes in Actinemys. All classification hypotheses performed poorly relative to the validation datasets and no single hypothesis was unequivocally supported for Actinemys. Two hypotheses had machine learning performance that was marginally better than our remaining hypotheses. In both cases, those hypotheses favored a two-species split between A. marmorata and A. pallida specimens, lending tentative morphological support to the hypothesis of two Actinemys species. However, the machine learning results also underscore that Actinemys as a whole has lower levels of plastral variation than other turtles within Emydidae, but the reason for this morphological conservatism is unclear.

Sex estimation from the hyoid bone measurements in an adult Eastern Turkish population using 3D CT images, discriminant function analysis, support vector machines, and artificial neural networks☆

The hyoid bone is one of the bones in the human body that shows sexual dimorphism. The anthropological and anthropometric characteristics that determine sexual dimorphism are influenced by demographic differences. The aim of this study was to investigate the rate of sexual dimorphism of the hyoid bone in the adult Eastern Turkish population from the examination of the 3D computed tomography images of 240 patients, using discriminant function analysis (DFA), support vector machines (SVM), and artificial neural networks (ANN). These evaluations were based on eight hyoid measurements that have been frequently used in previous CT studies. The results showed that all eight measurements were higher in males than in females (p = 0.000). It was determined that sex could be estimated accurately at up to 93.3 % using DFA, 93.8 % using SVM and 95.4 % using ANN. The maximum accuracy rate achieved to 94.2 % in males using SVM, and 95.8 % in females using ANN. These high rates of sexual dimorphism found using DFA, SVM, and ANN in this study indicate that characteristics of the hyoid bone can be utilized to determine sex in the Eastern Turkish population.

Assigned sex estimation with the clavicle and scapula: A study in a Portuguese reference sample

The estimation of biological sex is a critical step in the assessment of the biological profile of an anonymous skeletonized individual. In certain recovery circumstances, the most dimorphic skeletal areas, such as the pelvis, are absent or fragmented; in that case, other bones of the skeleton, including the clavicle and scapula, can be used to predict sex. The purpose of this research is to generate new models for the estimation of sex with clavicular and scapular measurements using a study-sample of 129 individuals with clavicle (65 males and 64 females) and 112 individuals with scapula (50 males and 62 females) from the Lisbon Identified Skeletal Collection (Portugal). A decision tree classifier (C4.5) and logistic regression (LR) were employed to create univariable and multivariable sex prediction models. Accuracy under cross-validation of the classification models is high (up to 93.8%), with minimal bias (<5%), particularly in the multivariable models. The proposed LR models facilitate the probabilistic estimation of biological sex, accounting for the significant overlap in the expression of sexual dimorphism.

A web application for sex and stature estimation from radiographic proximal femur for a Thai population

In both forensic and archaeological domains, the discovery of incomplete human remains is a frequent occurrence. Nevertheless, the estimation of biological profiles from such remains presents a challenge due to the absence of crucial skeletal elements, such as the skull and pelvis. This study aimed to assess the utility of the proximal femur in the forensic identification process by creating a web application for osteometric analysis of the proximal femur. The aim was to determine the sex and stature of an individual from radiographs of the left anteroposterior femur. To accomplish this, an automated method was developed for acquiring linear measurements from radiographic images of the proximal femur using Python tools. The application of Hough techniques and Canny edge detection was utilized to generate linear femoral dimensions from radiographs. A total of 354 left femora were radiographed and measured by the algorithm. The sex classification model employed in this study was the Naïve Bayes algorithm (accuracy = 91.2 %). Results indicated that Gaussian process regression (GPR) was the most effective method for estimating stature (mean error = 4.68 cm, SD = 3.93 cm). The proposed web application holds the potential to serve as a valuable asset in the realm of forensic investigations in Thailand, particularly in the estimation of biological profiles from fragmentary skeletal remains.

Hide and seek shark teeth in Random Forests: machine learning applied to Scyliorhinus canicula populations

Shark populations that are distributed alongside a latitudinal gradient often display body size differences at sexual maturity and vicariance patterns related to their number of tooth files. Previous works have demonstrated that Scyliorhinus canicula populations differ between the northeastern Atlantic Ocean and the Mediterranean Sea based on biological features and genetic analysis. In this study, we sample more than 3,000 teeth from 56 S. canicula specimens caught incidentally off Roscoff and Banyuls-sur-Mer. We investigate population differences based on tooth shape and form by using two approaches. Classification results show that the classical geometric morphometric framework is outperformed by an original Random Forests-based framework. Visually, both S. canicula populations share similar ontogenetic trends and timing of gynandric heterodonty emergence but the Atlantic population has bigger, blunter teeth, and less numerous accessory cusps than the Mediterranean population. According to the models, the populations are best differentiated based on their lateral tooth edges, which bear accessory cusps, and the tooth centroid sizes significantly improve classification performances. The differences observed are discussed in light of dietary and behavioural habits of the populations considered. The method proposed in this study could be further adapted to complement DNA analyses to identify shark species or populations based on tooth morphologies. This process would be of particular interest for fisheries management and identification of shark fossils.

The computational age-at-death estimation from 3D surface models of the adult pubic symphysis using data mining methods

Age-at-death estimation of adult skeletal remains is a key part of biological profile estimation, yet it remains problematic for several reasons. One of them may be the subjective nature of the evaluation of age-related changes, or the fact that the human eye is unable to detect all the relevant surface changes. We have several aims: (1) to validate already existing computer models for age estimation; (2) to propose our own expert system based on computational approaches to eliminate the factor of subjectivity and to use the full potential of surface changes on an articulation area; and (3) to determine what age range the pubic symphysis is useful for age estimation. A sample of 483 3D representations of the pubic symphyseal surfaces from the ossa coxae of adult individuals coming from four European (two from Portugal, one from Switzerland and Greece) and one Asian (Thailand) identified skeletal collections was used. A validation of published algorithms showed very high error in our dataset-the Mean Absolute Error (MAE) ranged from 16.2 and 25.1 years. Two completely new approaches were proposed in this paper: SASS (Simple Automated Symphyseal Surface-based) and AANNESS (Advanced Automated Neural Network-grounded Extended Symphyseal Surface-based), whose MAE values are 11.7 and 10.6 years, respectively. Lastly, it was demonstrated that our models could estimate the age-at-death using the pubic symphysis over the entire adult age range. The proposed models offer objective age estimates with low estimation error (compared to traditional visual methods) and are able to estimate age using the pubic symphysis across the entire adult age range.

Efficiency of the Adjusted Binary Classification (ABC) Approach in Osteometric Sex Estimation: A Comparative Study of Different Linear Machine Learning Algorithms and Training Sample Sizes

The adjusted binary classification (ABC) approach was proposed to assure that the binary classification model reaches a particular accuracy level. The present study evaluated the ABC for osteometric sex classification using multiple machine learning (ML) techniques: linear discriminant analysis (LDA), boosted generalized linear model (GLMB), support vector machine (SVM), and logistic regression (LR). We used 13 femoral measurements of 300 individuals from a modern Turkish population sample and split data into two sets: training (n = 240) and testing (n = 60). Then, the five best-performing measurements were selected for training univariate models, while pools of these variables were used for the multivariable models. ML classifier type did not affect the performance of unadjusted models. The accuracy of univariate models was 82−87%, while that of multivariate models was 89−90%. After applying ABC to the crossvalidation set, the accuracy and the positive and negative predictive values for uni- and multivariate models were ≥95%. Sex could be estimated for 28−75% of individuals using univariate models but with an obvious sexing bias, likely caused by different degrees of sexual dimorphism and between-group overlap. However, using multivariate models, we minimized the bias and properly classified 81−87% of individuals. A similar performance was also noted in the testing sample (except for FEB), with accuracies of 96−100%, and a proportion of classified individuals between 30% and 82% in univariate models, and between 90% and 91% in multivariate models. When considering different training sample sizes, we demonstrated that LR was the most sensitive with limited sample sizes (n < 150), while GLMB was the most stable classifier.

DXAGE 2.0 - adult age at death estimation using bone loss in the proximal femur and the second metacarpal

The accurate age at death assessment of unidentified adult skeletal individuals is a critical research task in forensic anthropology, being a key feature for the determination of biological profiles of individual skeletal remains. We have previously shown that the age-related decrease of bone mineral density (BMD) in the proximal femur could be used to assess age at death in women (Navega et al., J Forensic Sci 63:497-503, 2018). The present study aims to generate models for age estimation in both sexes through bone densitometry of the femur and radiogrammetry of the second metacarpal. The training sample comprised 224 adults (120 females, 104 males) from the "Coimbra Identified Skeletal Collection," and different models were generated through least squares regression and general regression neural networks (GRNN). The models were operationalized in a user-friendly online interface at https://osteomics.com/DXAGE2/ . The mean absolute difference between the known and estimated age at death ranges from 9.39 to 13.18 years among women and from 10.33 to 15.76 among men with the least squares regression models. For the GRNN models, the mean absolute difference between documented and projected age ranges from 8.44 to 12.58 years in women and from 10.56 to 16.18 years in men. DXAGE 2.0 enables age estimation in incomplete and/or fragmentary skeletal remains, using alternative skeletal regions, with reliable results.

Performance of the supervised learning algorithms in sex estimation of the proximal femur: A comparative study in contemporary Egyptian and Turkish samples

Sex estimation standards are population specific however, we argue that machine learning techniques (ML) may enhance the biological sex determination on trans-population application. Linear discriminant analysis (LDA) versus nine ML including quadratic discriminant analysis (QDA), support vector machine (SVM), Decision Tree (DT), Gaussian process (GPC), Naïve Bayesian (NBC), K-Nearest Neighbor (KNN), Random Forest (RFM) and Adaptive boosting (Adaboost) were compared. The experiments involve two contemporary populations: Turkish (n = 300) and Egyptian populations (n = 100) for training and validation, respectively. Base models were calibrated using isotonic and sigmoid calibration schemes. Results were analyzed at posterior probabilities (pp) thresholds >0.95 and >0.80. At pp = 0.5, ML algorithms yielded comparable accuracies in the training (90% to 97%) and test sets (81% to 88%) which are not modified after employing the calibration techniques. At pp >0.95, the raw RFM, LDA, QDA, and SVM models have shown the best performance however, calibration techniques improved the performance of various classifier especially NBC and Adaboost. By contrast, the performance of GPC, KNN, QDA models worsened by calibration. RFM has shown the best performance among all models at both thresholds whereas LDA benefited the best from using both calibration methods at pp >0.80. Complex ML models are not necessarily achieving better performance metrics. LDA and QDA remain the fastest and simplest classifiers. We demonstrated the capability of enhancing sex estimation using ML on an independent population sample however, differences in the underlying probability distribution generated by models were detected which warranted more cautious application by forensic practitioners.

Prognosis of sexual dimorphism with unfused hyoid bone: Artificial intelligence informed decision making with discriminant analysis

Depending on the metric and non-metric skeletal features of various bones, forensic experts proposed diverse sex identification methods. The main focus of the present study is to calculate sexual dimorphism in human unfused or disarticulated hyoid bone and compared it with studies conducted by different researchers. For this study, 293 unfused hyoid bones were accumulated and investigated from 173 male and 120 female cadavers of the northwest Indian population from the age of 15 to 80 years. Initially, discriminant analysis was performed on the dataset to predict sex and to get an idea for the crucial variables for sexual dimorphism. Later, significant variables predicted by the discriminant analysis were used for machine learning approaches to improve accuracy for sex determination. The standard scaler method is used for pre-processing of the data before machine learning analysis and to prevent overfitting and underfitting, 70 % of the whole dataset was utilized in the training of the model and the remaining data were used for testing the model. According to the discriminant analysis, body length (BL) and body height (BH) were found to be highly significant for the sex determination and predicted sex with 75.1 % accuracy. However, implementation of machine learning approaches such as the XG Boost classifier increased the accuracy to 83 % with sensitivity, and specificity scores of 0.81 and 0.84, respectively. Moreover, the ROC-AUC score achieved by the XG Boost classifier is 0.89; indicating machine learning investigation can improve the sex determination accuracy up to the appropriate standard.

Testing the validity of population-specific sex estimation equations: An evaluation based on talus and patella measurements

Sex estimation is essential for forensic scientists to identify human skeletal remains. However, the most sexually dimorphic elements like pelvis or skull are not always assessable. Osteometric analyses have proven useful in sex estimation, but also to be population specific. The main purpose of this study was to test the validity of contemporary Greek and Spanish discriminant functions for the talus and the patella, respectively, on a Swiss skeletal sample and to quantify the utility of the measurements as a novel approach in osteometric sex assessment. Four talus and three patella measurements on dry bone were obtained from 234 individuals of the modern cemetery SIMON Identified Skeletal Collection. The previously derived discriminant functions were applied, accuracies determined, the utility of the different measurements was assessed and new multivariable equations constructed. Accuracies varied between 67% and 86% for talus and 63% and 84% for patella, similar to those reported by the original studies. Multivariable equations should be preferred over equations based on single measurements and combining the most significant measurements rather than using several variables obtained the best possible accuracy. The new discriminant functions did not provide a substantial improvement to the original ones. The overall utility of talus and patella is limited, allowing sex estimation with sufficient certainty only in a small proportion of individuals. Discriminant functions developed in contemporary Greek or Spanish populations are in principle applicable also to Swiss contemporary populations. We recommend that at present existent studies of this type should be validated and tested rather than developing new formulas.

Sexing white 2D footprints using convolutional neural networks

Footprints are left, or obtained, in a variety of scenarios from crime scenes to anthropological investigations. Determining the sex of a footprint can be useful in screening such impressions and attempts have been made to do so using single or multi landmark distances, shape analyses and via the density of friction ridges. Here we explore the relative importance of different components in sexing two-dimensional foot impressions namely, size, shape and texture. We use a machine learning approach and compare this to more traditional methods of discrimination. Two datasets are used, a pilot data set collected from students at Bournemouth University (N = 196) and a larger data set collected by podiatrists at Sheffield NHS Teaching Hospital (N = 2677). Our convolutional neural network can sex a footprint with accuracy of around 90% on a test set of N = 267 footprint images using all image components, which is better than an expert can achieve. However, the quality of the impressions impacts on this success rate, but the results are promising and in time it may be possible to create an automated screening algorithm in which practitioners of whatever sort (medical or forensic) can obtain a first order sexing of a two-dimensional footprint.

Revisiting metric sex estimation of burnt human remains via supervised learning using a reference collection of modern identified cremated individuals (Knoxville, USA)

OBJECTIVES

This study aims to increase the rate of correctly sexed calcined individuals from archaeological and forensic contexts. This is achieved by evaluating sexual dimorphism of commonly used and new skeletal elements via uni- and multi-variate metric trait analyses.

MATERIALS AND METHODS

Twenty-two skeletal traits were evaluated in 86 individuals from the William M. Bass donated cremated collection of known sex and age-at-death. Four different predictive models, logistic regression, random forest, neural network, and calculation of population specific cut-off points, were used to determine the classification accuracy (CA) of each feature and several combinations thereof.

RESULTS

An overall CA of ≥ 80% was obtained for 12 out of 22 features (humerus trochlea max., and lunate length, humerus head vertical diameter, humerus head transverse diameter, radius head max., femur head vertical diameter, patella width, patella thickness, and talus trochlea length) using univariate analysis. Multivariate analysis showed an increase of CA (≥ 95%) for certain combinations and models (e.g., humerus trochlea max. and patella thickness). Our study shows metric sexual dimorphism to be well preserved in calcined human remains, despite the changes that occur during burning.

CONCLUSIONS

Our study demonstrated the potential of machine learning approaches, such as neural networks, for multivariate analyses. Using these statistical methods improves the rate of correct sex estimations in calcined human remains and can be applied to highly fragmented unburnt individuals from both archaeological and forensic contexts.

Sex classification of first molar teeth in cone beam computed tomography images using data mining

OBJECTIVE

The teeth have been used as a supplementary tool for sex differentiation as they are resistant to post-mortem degradation. The present study aimed to develop a new novel informatics framework for predicting sex from linear tooth dimension measurements achieved from cone beam computed tomography (CBCT) images.

METHOD AND MATERIALS

A clinical workflow using different machine learning methods was employed to predict the sex in the present study. The CBCT images of 485 subjects (245 men and 240 women) were evaluated for sex differentiation. Nine parameters were measured in both buccolingual and mesiodistal aspects of the teeth. We applied our dataset to Naïve Bayesian (NB), Random Forest (RF), and Support Vector Machine (SVM) as classifiers for prediction. Genetic feature selection was used to discover real features associated with sex classification.

RESULTS

The 10-fold cross-validation results indicated that NB had higher accuracy than SVM and RF for sex classification. The genetic algorithm (GA) indicated that the model could fit the data without using the enamel thickness and pulp height. The average classification accuracy of our clinical workflow was 92.31 %.

CONCLUSION

The results showed that NB was the best method for sex classification. The application of the first molar teeth in sex prediction indicated an acceptable level of sexual classification. Therefore, these odontometric parameters can be applied as an additional tool for sex determination in forensic anthropology.

Machine learning approaches for sex estimation using cranial measurements

The aim of the present study is to apply support vector machines (SVM) and artificial neural network (ANN) as sex classifiers and to generate useful classification models for sex estimation based on cranial measurements. Besides, the performance of the generated sub-symbolic machine learning models is compared with models developed through logistic regression (LR). The study was carried out on computed tomography images of 393 Bulgarian adults (169 males and 224 females). The three-dimensional coordinates of 47 landmarks were acquired and used for calculation of the cranial measurements. A total of 64 measurements (linear distances, angles, triangle areas and heights) and 22 indices were calculated. Two datasets were assembled including the linear measurements only and all measurements and index, respectively. An additional third dataset comprising all possible interlandmark distances between the landmarks was constructed. Two machine learning algorithms-SVM and ANN and a traditional statistical analysis LR-were applied to generate models for sex estimation. In addition, two advanced attribute selection techniques (Weka BestFirst and Weka GeneticSearch) were used. The classification accuracy of the models was evaluated by means of 10 × 10-fold cross-validation procedure. All three methods achieved accuracy results higher than 95%. The best accuracy (96.1 ± 0.5%) was obtained by SVM and it was statistically significantly higher than the best results achieved by ANN and LR. SVM and ANN reached higher accuracy by training on the full datasets than the selection datasets, except for the sample described by the interlandmark distances, where the reduction of attributes by the GeneticSearch algorithm improved the accuracy.

Sex assessment with the radius in Portuguese skeletal populations (late 19th - early to mid 20th centuries)

The assessment of sex is of immense relevance in the analysis of human skeletal remains, as other parameters of the biological profile are usually sex-specific (e.g., age at death or stature). The radius can be used to estimate sex when more dimorphic bones are not available or in the case of incomplete and fragmentary remains. Ten radius measurements collected in a sample of 364 individuals (166 females and 198 males) from the Coimbra Identified Skeletal Collection (late 19th - early 20th centuries) were employed to generate univariable and multivariable models for sex estimation. All models were evaluated with a 10-fold cross-validation method and an independent holdout sample from the Luís Lopes Collection (late 19th - mid 20th centuries) encompassing 50 individuals (25 females and 25 males). Univariable models show an accuracy ranging from 77.7% to 89.8% (cross-validation), and from 70% to 86% (test sample), while accuracy in the multivariable models varies from 88.7% to 93.4% (cross-validation), and 84.0% to 90.0% (test sample). Results suggest that measurements of the radius are useful to develop standard guidelines for sex estimation of anonymous skeletal remains.

New data about the 21st Century Identified Skeletal Collection (University of Coimbra, Portugal)

After the publication of the first article in 2014, 21st Century Identified Skeletal Collection, housed in the Laboratory of Forensic Anthropology, Department of Life Sciences at the University of Coimbra, Portugal, has been growing. Currently, the collection is composed of 302 complete adult skeletons of both sexes, which means that in 5 years it has doubled. The collection consists mostly of elderly individuals, with only 12.25% of the individuals aged less than 61 years old. All individuals are Portuguese nationals who died between 1982 and 2012. Ninety individuals exhibit prostheses, other medical devices and signs of surgical procedures. Moreover, a sub-collection of experimentally burned skeletons is under development, and currently includes 56 individuals (18.54% of the collection). The 21st Century Identified Skeletal Collection constitutes a fundamental tool for forensic anthropology research, including the development and validation studies of methods that focus on elderly individuals, as can be ascertained by the numerous scientific publications and academic scholarship that have been produced in previous years.

On the use of machine learning algorithms in forensic anthropology

The classification performance of the statistical methods binary logistic regression (BLR), multinomial and penalized multinomial logistic regression (MLR, pMLR), linear discriminant analysis (LDA), and the machine learning algorithms naïve Bayes classification (NBC), decision trees (DT), random forest (RF), artificial neural networks (ANN), support vector machines (linear, polynomial or radial) (SVM), multivariate adaptive regression splines (MARS), and extreme gradient boosting (XGB) is examined in skeletal sex/ancestry estimation. The datasets used to test the performance of these methods were obtained from a documented human skeletal collection, Athens Collection, and the Howells Craniometric data set. For their implementation, an R package has been written to search for the optimum tuning parameters under cross-validation and perform sex/ancestry classification. It was found that the classification performance may vary significantly depending on the problem. From the methods tested, LDA and the machine learning technique of linear SVM exhibit the best performance, with high prediction accuracy and relatively low bias in most of the tests. ANN and pMLR can generally be considered to give satisfactory predictions, whereas NBC when using metric traits and DT are the worst of the classification methods examined. The possibility of making the models developed via the machine learning algorithms applicable to other assemblages without the use of a training sample is also discussed.

Data mining for sex estimation based on cranial measurements

The aim of the present study is to develop effective and understandable classification models for sex estimation and to identify the most dimorphic linear measurements in adult crania by means of data mining techniques. Furthermore, machine learning models and models developed through logistic regression analysis are compared in terms of performance. Computed tomography scans of 393 adult individuals were used in the study. A landmark-based approach was applied to collect the metric data. The three-dimensional coordinates of 47 landmarks were acquired and used for calculation of linear measurements. Two datasets of cranial measurements were assembled, including 37standard measurements and 1081 interlandmark distances, respectively. Three data mining algorithms were applied: the rule induction algorithms JRIP and Ridor, and the decision tree algorithm J48. Two advanced attribute selection methods (Weka BestFirst and Weka GeneticSearch) were also used. The best accuracy result (91.9 %) was achieved by a set of rules learnt by the JRIP algorithm from the dataset constructed by application of the GeneticSearch selection algorithm to the dataset of standard cranial measurements. The set consisted of five rules including seven cranial measurements. Its accuracy was even better than the classification rates achieved by the logistic regression models. Concerning the second dataset of nonstandard measurements, the best accuracy (88.3 %) was obtained by using classification models learnt by two algorithms - JRIP with a dataset preprocessed by the BestFirst selection algorithm and Ridor with preprocessing by the GeneticSearch selection algorithm. Our experiments show that for the two datasets mentioned above the rule-based models contain smaller sets of rules with shorter lists of measurements and achieve better classification accuracy results in comparison with decision tree-based models.

Morphological variations of the human talus investigated using three-dimensional geometric morphometrics

INTRODUCTION

The shape of the talus determines the positional and kinematic features of the subtalar, talonavicular, and talocrural joints during walking. Thus, detailed knowledge of the pattern of sexual dimorphism of the human talus may be useful for revealing the pathogenetic mechanism of foot and knee disorders, which are more prevalent in females. The aim of this study was to characterize and visualize the three-dimensional shape variations of the talus in relation to sex and age using geometric morphometrics.

MATERIALS AND METHODS

Computed tomography images of 56 feet without talar injuries or disorders were used in this study. Thirty-seven anatomical landmarks were identified on a bone model of the talus to calculate principal components (PCs) of shape variations among specimens. PC scores were compared between sexes, and their correlations with age were also investigated.

RESULTS

The female talus had a longer neck and narrower head width than the male talus. The superior trochlea was tilted more laterally in the frontal plane in females. Furthermore, the female talar head was more twisted and was more elongated in the dorsoplantar direction.

CONCLUSIONS

Morphological features of the talus in females could alter the subtalar and talonavicular joint kinematics during walking and could be a structural factor in the pathogenetic mechanism underlying foot and knee disorders. This study contributes to the comprehensive understanding of shape variations in the human talus.

Effects of Hypertension, Diabetes, and Smoking on Age and Sex Prediction from Retinal Fundus Images

Retinal fundus images are used to detect organ damage from vascular diseases (e.g. diabetes mellitus and hypertension) and screen ocular diseases. We aimed to assess convolutional neural network (CNN) models that predict age and sex from retinal fundus images in normal participants and in participants with underlying systemic vascular-altered status. In addition, we also tried to investigate clues regarding differences between normal ageing and vascular pathologic changes using the CNN models. In this study, we developed CNN age and sex prediction models using 219,302 fundus images from normal participants without hypertension, diabetes mellitus (DM), and any smoking history. The trained models were assessed in four test-sets with 24,366 images from normal participants, 40,659 images from hypertension participants, 14,189 images from DM participants, and 113,510 images from smokers. The CNN model accurately predicted age in normal participants; the correlation between predicted age and chronologic age was R² = 0.92, and the mean absolute error (MAE) was 3.06 years. MAEs in test-sets with hypertension (3.46 years), DM (3.55 years), and smoking (2.65 years) were similar to that of normal participants; however, R² values were relatively low (hypertension, R² = 0.74; DM, R² = 0.75; smoking, R² = 0.86). In subgroups with participants over 60 years, the MAEs increased to above 4.0 years and the accuracies declined for all test-sets. Fundus-predicted sex demonstrated acceptable accuracy (area under curve > 0.96) in all test-sets. Retinal fundus images from participants with underlying vascular-altered conditions (hypertension, DM, or smoking) indicated similar MAEs and low coefficients of determination (R²) between the predicted age and chronologic age, thus suggesting that the ageing process and pathologic vascular changes exhibit different features. Our models demonstrate the most improved performance yet and provided clues to the relationship and difference between ageing and pathologic changes from underlying systemic vascular conditions. In the process of fundus change, systemic vascular diseases are thought to have a different effect from ageing. Research in context. Evidence before this study. The human retina and optic disc continuously change with ageing, and they share physiologic or pathologic characteristics with brain and systemic vascular status. As retinal fundus images provide high-resolution in-vivo images of retinal vessels and parenchyma without any invasive procedure, it has been used to screen ocular diseases and has attracted significant attention as a predictive biomarker for cerebral and systemic vascular diseases. Recently, deep neural networks have revolutionised the field of medical image analysis including retinal fundus images and shown reliable results in predicting age, sex, and presence of cardiovascular diseases. Added value of this study. This is the first study demonstrating how a convolutional neural network (CNN) trained using retinal fundus images from normal participants measures the age of participants with underlying vascular conditions such as hypertension, diabetes mellitus (DM), or history of smoking using a large database, SBRIA, which contains 412,026 retinal fundus images from 155,449 participants. Our results indicated that the model accurately predicted age in normal participants, while correlations (coefficient of determination, R²) in test-sets with hypertension, DM, and smoking were relatively low. Additionally, a subgroup analysis indicated that mean absolute errors (MAEs) increased and accuracies declined significantly in subgroups with participants over 60 years of age in both normal participants and participants with vascular-altered conditions. These results suggest that pathologic retinal vascular changes occurring in systemic vascular diseases are different form the changes in spontaneous ageing process, and the ageing process observed in retinal fundus images may saturate at age about 60 years. Implications of all available evidence. Based on this study and previous reports, the CNN could accurately and reliably predict age and sex using retinal fundus images. The fact that retinal changes caused by ageing and systemic vascular diseases occur differently motivates one to understand the retina deeper. Deep learning-based fundus image reading may be a more useful and beneficial tool for screening and diagnosing systemic and ocular diseases after further development.

Three-Dimensional Morphological Variations of the Human Calcaneus Investigated Using Geometric Morphometrics

The shape of the calcaneus determines the mechanical interaction of the foot with the ground during the heel-strike in human walking. Detailed knowledge of the pattern of sexual dimorphism of the human calcaneus could help to clarify the pathogenetic mechanism of foot and knee disorders, which are more prevalent in females. Therefore, the aim of this study was to characterize and visualize the three-dimensional shape variations of the calcaneus in relation to sex and age using geometric morphometrics. Computed tomography images of 56 feet without subtalar injuries or disorders were used in this study. Thirty-seven anatomical landmarks were identified on the bone model of the calcaneus to calculate principal components (PCs) of shape variations among specimens. The PC scores were compared between males and females, and their correlations with age were also analyzed. The female calcaneus was longer in length and shorter in height than that of males. The medial process of the calcaneal tuberosity in females was more inferiorly projected and the tuberosity was shifted more laterally. Also, the calcaneus was wider and the sustentaculum tali thickened with aging. Female structural features of the calcaneus alter the kinematics of the foot during walking and could be a structural factor in foot and knee disorders. This study contributes to a comprehensive understanding of shape variations in the human calcaneus. Clin. Anat., 33:751-758, 2020. © 2019 Wiley Periodicals, Inc.

Metric variation of the tibia in the Mediterranean: Implications in forensic identification

Ancestry estimation from skeletal remains is a challenging task, but essential for the creation of a complete biological profile. As such, the study of human variation between populations is important for the fields of biological and forensic anthropology, as well as medicine. Cranial and dental morphological variation have traditionally been linked to geographic affinity resulting in several methods of ancestry estimation, while the postcranial skeleton has been systematically neglected. The current study explores metric variation of the tibia in six Mediterranean populations and its validity in estimating ancestry in the Mediterranean. The study sample includes 909 individuals (470 males and 439 females) from Cyprus, Greece, Italy, Spain, Portugal and Turkey. The sample was divided in two subsamples: a reference and a validation sample. Multinomial regression models were created based on the reference sample and then applied to validation sample. The first model used three variables and resulted in 57% and 56% accuracy for the two samples respectively, while the second model (6 variables) resulted in 80% and 74% respectively. Classification between groups ranged from 28% to 95% for the reference sample and from 15% to 91% for the validation sample. The highest classification accuracy was noted for the Greek sample (95% and 90% for the reference and validation sample respectively), followed by the Turkish sample (74% and 78% respectively). The Spanish, Portuguese and Italian samples presented greater morphological overlap which resulted in lower classification accuracies. The results indicate that although the tibia presents considerable variation amongst neighbour populations it is not suitable as a sole skeletal element to separate all groups successfully. A combination of different skeletal elements may be required in order to achieve the levels of reliability required for forensic applications.

Sex estimation using the mandibular canine index components

The mandibular canine index (MCI) has been described as a suitable methodology for sex estimation in forensic scenarios but there are contradictory reports about its accuracy. Moreover, the two mandibular canine teeth must be available, which is not always a viable option. The aim of this study was to strip the MCI by analyzing the MCI itself and its components, in order to optimize its use for sex estimation. The mesiodistal dimensions of the mandibular canine crown and the mandibular canine arch width were measured in a sample of 120 cast models. Five predictor variables were considered in this study: the standard MCI, a variation of the MCI using the left canine, and MCI components (MD43, MD33 and D33-43). Multivariate binary logistic regression was performed using stepwise forward approach to select the most statistical relevant variables on the probability of a cast being from a female. The estimated probability was then analyzed with respect to performance in sex classification (ROC analysis and optimal cut-offs accuracy) and compared with the performance of the univariate variables. MCI43 and MCI33 presented the lowest performance (64.2% and 63.3% respectively), and the highest overall accuracy was attained using the MD43 and MD33 (85.8% in both cases). The multivariate logistic model obtained (using MD43 and MD33) exhibited the same accuracy as the logistic model based solely on MD43 (85.8%). Our results suggest that MD43 should be used instead of MCI for sex estimation.

Inferring locomotor behaviours in Miocene New World monkeys using finite element analysis, geometric morphometrics and machine-learning classification techniques applied to talar morphology

The talus is one of the most commonly preserved post-cranial elements in the platyrrhine fossil record. Talar morphology can provide information about postural adaptations because it is the anatomical structure responsible for transmitting body mass forces from the leg to the foot. The aim of this study is to test whether the locomotor behaviour of fossil Miocene platyrrhines could be inferred from their talus morphology. The extant sample was classified into three different locomotor categories and then talar strength was compared using finite-element analysis. Geometric morphometrics were used to quantify talar shape and to assess its association with biomechanical strength. Finally, several machine-learning (ML) algorithms were trained using both the biomechanical and morphometric data from the extant taxa to infer the possible locomotor behaviour of the Miocene fossil sample. The obtained results show that the different locomotor categories are distinguishable using either biomechanical or morphometric data. The ML algorithms categorized most of the fossil sample as arboreal quadrupeds. This study has shown that a combined approach can contribute to the understanding of platyrrhine talar morphology and its relationship with locomotion. This approach is likely to be beneficial for determining the locomotor habits in other fossil taxa.

Dimensions around the nutrient foramina of the tibia and fibula in the estimation of sex

Sex estimation from skeletal remains is one of the key components in establishing a biological profile and consequent identification of an individual in a forensic and medico-legal practice. The use of dimensions around the nutrient foramen in instances where long bones may be fragmented and damaged is of benefit due to the fact that the nutrient foramen is easily identifiable and may be preserved on the shaft of long bones. This study is an investigation of the usefulness of various measurements around the nutrient foramen of the tibia and fibula of South Africans in an attempt to develop osteometric standards for sex estimation. The sample included 206 tibiae and 204 fibulae of South African Africans (SAA) and South African whites (SAW) procured from the Raymond A. Dart Collection of Human Skeletons based at the University of the Witwatersrand. Sex was correctly classified for the tibia with an accuracy ranging between 79-82% in SAA and 84-88% in SAW, with the circumference at the level of the nutrient foramen as the single best predictor of sex in both populations. An accuracy ranging from 69 to 74% in SAA and 70-77% in SAW was observed for the combined measurements on the fibula. The current study confirms the usefulness of measurements around the nutrient foramen of the tibia in the assignment of sex. However functions of the fibula generally performed poorly and should be used with caution.

Sex estimation with the total area of the proximal femur: A densitometric approach

The estimation of sex is a central step to establish the biological profile of an anonymous skeletal individual. Imaging techniques, including bone densitometry, have been used to evaluate sex in remains incompletely skeletonized. In this paper, we present a technique for sex estimation using the total area (TA) of the proximal femur, a two-dimensional areal measurement determined through densitometry. TA was acquired from a training sample (112 females; 112 males) from the Coimbra Identified Skeletal Collection (University of Coimbra, Portugal). Logistic regression (LR), linear discriminant analysis (LDA), reduce error pruning trees (REPTree), and classification and regression trees (CART) were employed in order to obtain models that could predict sex in unidentified skeletal remains. Under cross-validation, the proposed models correctly estimated sex in 90.2-92.0% of cases (bias ranging from 1.8% to 4.5%). The models were evaluated in an independent test sample (30 females; 30 males) from the 21st Century Identified Skeletal Collection (University of Coimbra, Portugal), with a sex allocation accuracy ranging from 90.0% to 91.7% (bias from 3.3% to 10.0%). Overall, data mining classifiers, especially the REPTree, performed better than the traditional classifiers (LR and LDA), maximizing overall accuracy and minimizing bias. This study emphasizes the significant value of bone densitometry to estimate sex in cadaveric remains in diverse states of preservation and completeness, even human remains with soft tissues.

Estimation of stature from radiographic measurement of foot dimensions: Truncated foot length may be more reliable than full foot length

Foot length has been shown to be a reliable dimension for estimation of stature. However, phalanges of the foot are very small bones and their length may not be proportional to person's stature. Thus, we hypothesized that foot length measured excluding the phalanges, the truncated foot length, may be more reliable in stature estimation than full foot length. This study, therefore, aimed at comparing the accuracy of the regression equations derived from the truncated foot length and the full foot length. The study recruited a sample of 32 young adults (16 males and 16 females) aged from 20 to 35 years. Lateral radiographs of the right feet were obtained for each subject in a bilateral standing position while maintaining equal weight on both feet. Standing height of the participants was measured with a stadiometer. Truncated foot length and full foot length were measured on the lateral radiographs of the foot. Independent t-test was used to check for mean differences in the dimensions between genders. Linear regression analysis was used to determine the equations for stature estimation. Intra and inter-observer reliability were calculated from four precision estimates: absolute technical error of measurement (aTEM), relative technical error of measurement (rTEM), coefficient of reliability (Rr) and coefficient of variation (Cv). All the dimensions measured were significantly larger in males than females. Linear regression equations were derived for estimation of stature using both the truncated foot length and full foot length. The regression equations derived from truncated foot length have larger correlation coefficient, coefficient of determination, adjusted coefficient of determination as well as smaller standard error of estimation than those derived from full foot length. All the precision estimates showed that the measurement errors are within acceptable limits. This study suggests that even if the full foot length is available, excluding the phalanges may give more accurate stature estimation.

Sex estimation from the navicular bone in Spanish contemporary skeletal collections

Sexual estimation is fundamental to reconstruct the biological profile of individuals, but postdepositional factors can alter the resistance of the bones, thereby preventing accurate diagnosis especially when the skull and the pelvis are absent. Navicular bones are usually well preserved in archeological and forensic contexts and can a good alternative to discriminate sex. On the basis of these aspects, the present investigation analyzed the sexual dimorphism in 231 pairs of navicular bones from documented contemporary collections from Spain. Receiver operating characteristic (ROC) curve analysis and binary logistic regressions were carried out in eight replicable linear measurements of the navicular bone. Each of the eight variables showed a significant sexual dimorphism in our sample. The ROC curve results indicate that at least five out of the eight variables used have high ability for sex diagnosis, among which the maximum length of the cuneiform surface (LMAXCUN) showed a better performance (area under the curve value=0.86). Moreover, we introduced regression equations with combination of measurements that correctly allocated the skeletons with 80% or greater accuracy. The equation with high allocation accuracy rate (83.4%) included a combination of the maximum height of the navicular (HMAX), maximum length of the cuneiform surface (LMAXCUN), and maximum length of the talar facet (LMAXTAL). The regression equations presented here are useful for the Western Mediterranean populations and offer better alternatives than formulas based on other population groups.

A method for sex estimation using the proximal femur

The assessment of sex is crucial to the establishment of a biological profile of an unidentified skeletal individual. The best methods currently available for the sexual diagnosis of human skeletal remains generally rely on the presence of well-preserved pelvic bones, which is not always the case. Postcranial elements, including the femur, have been used to accurately estimate sex in skeletal remains from forensic and bioarcheological settings. In this study, we present an approach to estimate sex using two measurements (femoral neck width [FNW] and femoral neck axis length [FNAL]) of the proximal femur. FNW and FNAL were obtained in a training sample (114 females and 138 males) from the Luís Lopes Collection (National History Museum of Lisbon). Logistic regression and the C4.5 algorithm were used to develop models to predict sex in unknown individuals. Proposed cross-validated models correctly predicted sex in 82.5-85.7% of the cases. The models were also evaluated in a test sample (96 females and 96 males) from the Coimbra Identified Skeletal Collection (University of Coimbra), resulting in a sex allocation accuracy of 80.1-86.2%. This study supports the relative value of the proximal femur to estimate sex in skeletal remains, especially when other exceedingly dimorphic skeletal elements are not accessible for analysis.

A review of sex estimation techniques during examination of skeletal remains in forensic anthropology casework

Sex estimation is considered as one of the essential parameters in forensic anthropology casework, and requires foremost consideration in the examination of skeletal remains. Forensic anthropologists frequently employ morphologic and metric methods for sex estimation of human remains. These methods are still very imperative in identification process in spite of the advent and accomplishment of molecular techniques. A constant boost in the use of imaging techniques in forensic anthropology research has facilitated to derive as well as revise the available population data. These methods however, are less reliable owing to high variance and indistinct landmark details. The present review discusses the reliability and reproducibility of various analytical approaches; morphological, metric, molecular and radiographic methods in sex estimation of skeletal remains. Numerous studies have shown a higher reliability and reproducibility of measurements taken directly on the bones and hence, such direct methods of sex estimation are considered to be more reliable than the other methods. Geometric morphometric (GM) method and Diagnose Sexuelle Probabiliste (DSP) method are emerging as valid methods and widely used techniques in forensic anthropology in terms of accuracy and reliability. Besides, the newer 3D methods are shown to exhibit specific sexual dimorphism patterns not readily revealed by traditional methods. Development of newer and better methodologies for sex estimation as well as re-evaluation of the existing ones will continue in the endeavour of forensic researchers for more accurate results.

Possible application of CT morphometry of the calcaneus and talus in forensic anthropological identification

Computed tomography (CT) data provide information for volumetric and radiographic density analysis. The present study investigated the application of virtual CT volumetry of the tarsal bones to estimation of the sex, stature, and body weight using postmortem CT (PMCT) data of forensic autopsy cases. Three-dimensional (3D) images of the bilateral foot bones of intact Japanese subjects after adolescence (age ≥ 15 years, n = 179, 100 males and 79 females) were reconstructed on an automated CT image analyzer system. Measured parameters were mass volume, mean CT value (HU), and total CT value of the talus and calcaneus. Mean CT values of these bones showed age-dependent decreases in elderly subjects over 60 years of age for both sexes, with significant sex-related differences especially in the elderly. The mass volumes and total CT values of the talus and calcaneus showed significant sex-related differences, and also moderate correlations with body height and weight for bilateral bones in all cases (r = 0.58-0.78, p < 0.0001); however, the correlations of these parameters of the female talus with body weight were insufficient (r = 0.41-0.61, p < 0.0001). These observations indicate the applicability of virtual CT morphometry of the talus and calcaneus using an automated analyzer to estimate the sex and stature in forensic identification; however, greater variations should be considered in body weight estimations of females.