3-D reconstruction of the human skeleton during motion

L’analisi del movimento umano ha come obiettivo la descrizione del movimento assoluto e relativo dei segmenti ossei del soggetto e, ove richiesto, dei relativi tessuti molli durante l’esecuzione di esercizi fisici. La bioingegneria mette a disposizione dell’analisi del movimento gli strumenti ed i metodi necessari per una valutazione quantitativa di efficacia, funzione e/o qualità del movimento umano, consentendo al clinico l’analisi di aspetti non individuabili con gli esami tradizionali. Tali valutazioni possono essere di ausilio all’analisi clinica di pazienti e, specialmente con riferimento a problemi ortopedici, richiedono una elevata accuratezza e precisione perché il loro uso sia valido. Il miglioramento della affidabilità dell’analisi del movimento ha quindi un impatto positivo sia sulla metodologia utilizzata, sia sulle ricadute cliniche della stessa. Per perseguire gli obiettivi scientifici descritti, è necessario effettuare una stima precisa ed accurata della posizione e orientamento nello spazio dei segmenti ossei in esame durante l’esecuzione di un qualsiasi atto motorio. Tale descrizione può essere ottenuta mediante la definizione di un modello della porzione del corpo sotto analisi e la misura di due tipi di informazione: una relativa al movimento ed una alla morfologia. L’obiettivo è quindi stimare il vettore posizione e la matrice di orientamento necessari a descrivere la collocazione nello spazio virtuale 3D di un osso utilizzando le posizioni di punti, definiti sulla superficie cutanea ottenute attraverso la stereofotogrammetria. Le traiettorie dei marker, così ottenute, vengono utilizzate per la ricostruzione della posizione e dell’orientamento istantaneo di un sistema di assi solidale con il segmento sotto esame (sistema tecnico) (Cappozzo et al. 2005). Tali traiettorie e conseguentemente i sistemi tecnici, sono affetti da due tipi di errore, uno associato allo strumento di misura e l’altro associato alla presenza di tessuti molli interposti tra osso e cute. La propagazione di quest’ultimo ai risultati finali è molto più distruttiva rispetto a quella dell’errore strumentale che è facilmente minimizzabile attraverso semplici tecniche di filtraggio (Chiari et al. 2005). In letteratura è stato evidenziato che l’errore dovuto alla deformabilità dei tessuti molli durante l’analisi del movimento umano provoca inaccuratezze tali da mettere a rischio l’utilizzabilità dei risultati. A tal proposito Andriacchi scrive: “attualmente, uno dei fattori critici che rallentano il progresso negli studi del movimento umano è la misura del movimento scheletrico partendo dai marcatori posti sulla cute” (Andriacchi et al. 2000). Relativamente alla morfologia, essa può essere acquisita, ad esempio, attraverso l’utilizzazione di tecniche per bioimmagini. Queste vengono fornite con riferimento a sistemi di assi locali in generale diversi dai sistemi tecnici. Per integrare i dati relativi al movimento con i dati morfologici occorre determinare l’operatore che consente la trasformazione tra questi due sistemi di assi (matrice di registrazione) e di conseguenza è fondamentale l’individuazione di particolari terne di riferimento, dette terne anatomiche. L’identificazione di queste terne richiede la localizzazione sul segmento osseo di particolari punti notevoli, detti repere anatomici, rispetto ad un sistema di riferimento solidale con l’osso sotto esame. Tale operazione prende il nome di calibrazione anatomica. Nella maggior parte dei laboratori di analisi del movimento viene implementata una calibrazione anatomica a “bassa risoluzione” che prevede la descrizione della morfologia dell’osso a partire dall’informazione relativa alla posizione di alcuni repere corrispondenti a prominenze ossee individuabili tramite palpazione. Attraverso la stereofotogrammetria è quindi possibile registrare la posizione di questi repere rispetto ad un sistema tecnico. Un diverso approccio di calibrazione anatomica può essere realizzato avvalendosi delle tecniche ad “alta risoluzione”, ovvero attraverso l’uso di bioimmagini. In questo caso è necessario disporre di una rappresentazione digitale dell’osso in un sistema di riferimento morfologico e localizzare i repere d’interesse attraverso palpazione in ambiente virtuale (Benedetti et al. 1994 ; Van Sint Jan et al. 2002; Van Sint Jan et al. 2003). Un simile approccio è difficilmente applicabile nella maggior parte dei laboratori di analisi del movimento, in quanto normalmente non si dispone della strumentazione necessaria per ottenere le bioimmagini; inoltre è noto che tale strumentazione in alcuni casi può essere invasiva. Per entrambe le calibrazioni anatomiche rimane da tenere in considerazione che, generalmente, i repere anatomici sono dei punti definiti arbitrariamente all’interno di un’area più vasta e irregolare che i manuali di anatomia definiscono essere il repere anatomico. L’identificazione dei repere attraverso una loro descrizione verbale è quindi povera in precisione e la difficoltà nella loro identificazione tramite palpazione manuale, a causa della presenza dei tessuti molli interposti, genera errori sia in precisione che in accuratezza. Tali errori si propagano alla stima della cinematica e della dinamica articolare (Ramakrishnan et al. 1991; Della Croce et al. 1999). Della Croce (Della Croce et al. 1999) ha inoltre evidenziato che gli errori che influenzano la collocazione nello spazio delle terne anatomiche non dipendono soltanto dalla precisione con cui vengono identificati i repere anatomici, ma anche dalle regole che si utilizzano per definire le terne. E’ infine necessario evidenziare che la palpazione manuale richiede tempo e può essere effettuata esclusivamente da personale altamente specializzato, risultando quindi molto onerosa (Simon 2004). La presente tesi prende lo spunto dai problemi sopra elencati e ha come obiettivo quello di migliorare la qualità delle informazioni necessarie alla ricostruzione della cinematica 3D dei segmenti ossei in esame affrontando i problemi posti dall’artefatto di tessuto molle e le limitazioni intrinseche nelle attuali procedure di calibrazione anatomica. I problemi sono stati affrontati sia mediante procedure di elaborazione dei dati, sia apportando modifiche ai protocolli sperimentali che consentano di conseguire tale obiettivo. Per quanto riguarda l’artefatto da tessuto molle, si è affrontato l’obiettivo di sviluppare un metodo di stima che fosse specifico per il soggetto e per l’atto motorio in esame e, conseguentemente, di elaborare un metodo che ne consentisse la minimizzazione. Il metodo di stima è non invasivo, non impone restrizione al movimento dei tessuti molli, utilizza la sola misura stereofotogrammetrica ed è basato sul principio della media correlata. Le prestazioni del metodo sono state valutate su dati ottenuti mediante una misura 3D stereofotogrammetrica e fluoroscopica sincrona (Stagni et al. 2005), (Stagni et al. 2005). La coerenza dei risultati raggiunti attraverso i due differenti metodi permette di considerare ragionevoli le stime dell’artefatto ottenute con il nuovo metodo. Tale metodo fornisce informazioni sull’artefatto di pelle in differenti porzioni della coscia del soggetto e durante diversi compiti motori, può quindi essere utilizzato come base per un piazzamento ottimo dei marcatori. Lo si è quindi utilizzato come punto di partenza per elaborare un metodo di compensazione dell’errore dovuto all’artefatto di pelle che lo modella come combinazione lineare degli angoli articolari di anca e ginocchio. Il metodo di compensazione è stato validato attraverso una procedura di simulazione sviluppata ad-hoc. Relativamente alla calibrazione anatomica si è ritenuto prioritario affrontare il problema associato all’identificazione dei repere anatomici perseguendo i seguenti obiettivi: 1. migliorare la precisione nell’identificazione dei repere e, di conseguenza, la ripetibilità dell’identificazione delle terne anatomiche e della cinematica articolare, 2. diminuire il tempo richiesto, 3. permettere che la procedura di identificazione possa essere eseguita anche da personale non specializzato. Il perseguimento di tali obiettivi ha portato alla implementazione dei seguenti metodi: • Inizialmente è stata sviluppata una procedura di palpazione virtuale automatica. Dato un osso digitale, la procedura identifica automaticamente i punti di repere più significativi, nella maniera più precisa possibile e senza l'ausilio di un operatore esperto, sulla base delle informazioni ricavabili da un osso digitale di riferimento (template), preliminarmente palpato manualmente. • E’ stato poi condotto uno studio volto ad indagare i fattori metodologici che influenzano le prestazioni del metodo funzionale nell’individuazione del centro articolare d’anca, come prerequisito fondamentale per migliorare la procedura di calibrazione anatomica. A tale scopo sono stati confrontati diversi algoritmi, diversi cluster di marcatori ed è stata valutata la prestazione del metodo in presenza di compensazione dell’artefatto di pelle. • E’stato infine proposto un metodo alternativo di calibrazione anatomica basato sull’individuazione di un insieme di punti non etichettati, giacenti sulla superficie dell’osso e ricostruiti rispetto ad un TF (UP-CAST). A partire dalla posizione di questi punti, misurati su pelvi coscia e gamba, la morfologia del relativo segmento osseo è stata stimata senza identificare i repere, bensì effettuando un’operazione di matching dei punti misurati con un modello digitale dell’osso in esame. La procedura di individuazione dei punti è stata eseguita da personale non specializzato nell’individuazione dei repere anatomici. Ai soggetti in esame è stato richiesto di effettuare dei cicli di cammino in modo tale da poter indagare gli effetti della nuova procedura di calibrazione anatomica sulla determinazione della cinematica articolare. I risultati ottenuti hanno mostrato, per quel che riguarda la identificazione dei repere, che il metodo proposto migliora sia la precisione inter- che intraoperatore, rispetto alla palpazione convenzionale (Della Croce et al. 1999). E’ stato inoltre riscontrato un notevole miglioramento, rispetto ad altri protocolli (Charlton et al. 2004; Schwartz et al. 2004), nella ripetibilità della cinematica 3D di anca e ginocchio. Bisogna inoltre evidenziare che il protocollo è stato applicato da operatori non specializzati nell’identificazione dei repere anatomici. Grazie a questo miglioramento, la presenza di diversi operatori nel laboratorio non genera una riduzione di ripetibilità. Infine, il tempo richiesto per la procedura è drasticamente diminuito. Per una analisi che include la pelvi e i due arti inferiori, ad esempio, l’identificazione dei 16 repere caratteristici usando la calibrazione convenzionale richiede circa 15 minuti, mentre col nuovo metodo tra i 5 e i 10 minuti.

Assimilation of Meteosat Second Generation (MSG) satellite data in a regional numerical weather prediction model using a one-dimensional variational approach

The quality of temperature and humidity retrievals from the infrared SEVIRI sensors on the geostationary Meteosat Second Generation (MSG) satellites is assessed by means of a one dimensional variational algorithm. The study is performed with the aim of improving the spatial and temporal resolution of available observations to feed analysis systems designed for high resolution regional scale numerical weather prediction (NWP) models. The non-hydrostatic forecast model COSMO (COnsortium for Small scale MOdelling) in the ARPA-SIM operational configuration is used to provide background fields. Only clear sky observations over sea are processed. An optimised 1D–VAR set-up comprising of the two water vapour and the three window channels is selected. It maximises the reduction of errors in the model backgrounds while ensuring ease of operational implementation through accurate bias correction procedures and correct radiative transfer simulations. The 1D–VAR retrieval quality is firstly quantified in relative terms employing statistics to estimate the reduction in the background model errors. Additionally the absolute retrieval accuracy is assessed comparing the analysis with independent radiosonde and satellite observations. The inclusion of satellite data brings a substantial reduction in the warm and dry biases present in the forecast model. Moreover it is shown that the retrieval profiles generated by the 1D–VAR are well correlated with the radiosonde measurements. Subsequently the 1D–VAR technique is applied to two three–dimensional case–studies: a false alarm case–study occurred in Friuli–Venezia–Giulia on the 8th of July 2004 and a heavy precipitation case occurred in Emilia–Romagna region between 9th and 12th of April 2005. The impact of satellite data for these two events is evaluated in terms of increments in the integrated water vapour and saturation water vapour over the column, in the 2 meters temperature and specific humidity and in the surface temperature. To improve the 1D–VAR technique a method to calculate flow–dependent model error covariance matrices is also assessed. The approach employs members from an ensemble forecast system generated by perturbing physical parameterisation schemes inside the model. The improved set–up applied to the case of 8th of July 2004 shows a substantial neutral impact.

Seizure prediction and control in epilepsy

The first part of my thesis presents an overview of the different approaches used in the past two decades in the attempt to forecast epileptic seizure on the basis of intracranial and scalp EEG. Past research could reveal some value of linear and nonlinear algorithms to detect EEG features changing over different phases of the epileptic cycle. However, their exact value for seizure prediction, in terms of sensitivity and specificity, is still discussed and has to be evaluated. In particular, the monitored EEG features may fluctuate with the vigilance state and lead to false alarms. Recently, such a dependency on vigilance states has been reported for some seizure prediction methods, suggesting a reduced reliability. An additional factor limiting application and validation of most seizure-prediction techniques is their computational load. For the first time, the reliability of permutation entropy [PE] was verified in seizure prediction on scalp EEG data, contemporarily controlling for its dependency on different vigilance states. PE was recently introduced as an extremely fast and robust complexity measure for chaotic time series and thus suitable for online application even in portable systems. The capability of PE to distinguish between preictal and interictal state has been demonstrated using Receiver Operating Characteristics (ROC) analysis. Correlation analysis was used to assess dependency of PE on vigilance states. Scalp EEG-Data from two right temporal epileptic lobe (RTLE) patients and from one patient with right frontal lobe epilepsy were analysed. The last patient was included only in the correlation analysis, since no datasets including seizures have been available for him. The ROC analysis showed a good separability of interictal and preictal phases for both RTLE patients, suggesting that PE could be sensitive to EEG modifications, not visible on visual inspection, that might occur well in advance respect to the EEG and clinical onset of seizures. However, the simultaneous assessment of the changes in vigilance showed that: a) all seizures occurred in association with the transition of vigilance states; b) PE was sensitive in detecting different vigilance states, independently of seizure occurrences. Due to the limitations of the datasets, these results cannot rule out the capability of PE to detect preictal states. However, the good separability between pre- and interictal phases might depend exclusively on the coincidence of epileptic seizure onset with a transition from a state of low vigilance to a state of increased vigilance. The finding of a dependency of PE on vigilance state is an original finding, not reported in literature, and suggesting the possibility to classify vigilance states by means of PE in an authomatic and objectic way. The second part of my thesis provides the description of a novel behavioral task based on motor imagery skills, firstly introduced (Bruzzo et al. 2007), in order to study mental simulation of biological and non-biological movement in paranoid schizophrenics (PS). Immediately after the presentation of a real movement, participants had to imagine or re-enact the very same movement. By key release and key press respectively, participants had to indicate when they started and ended the mental simulation or the re-enactment, making it feasible to measure the duration of the simulated or re-enacted movements. The proportional error between duration of the re-enacted/simulated movement and the template movement were compared between different conditions, as well as between PS and healthy subjects. Results revealed a double dissociation between the mechanisms of mental simulation involved in biological and non-biologial movement simulation. While for PS were found large errors for simulation of biological movements, while being more acurate than healthy subjects during simulation of non-biological movements. Healthy subjects showed the opposite relationship, making errors during simulation of non-biological movements, but being most accurate during simulation of non-biological movements. However, the good timing precision during re-enactment of the movements in all conditions and in both groups of participants suggests that perception, memory and attention, as well as motor control processes were not affected. Based upon a long history of literature reporting the existence of psychotic episodes in epileptic patients, a longitudinal study, using a slightly modified behavioral paradigm, was carried out with two RTLE patients, one patient with idiopathic generalized epilepsy and one patient with extratemporal lobe epilepsy. Results provide strong evidence for a possibility to predict upcoming seizures in RTLE patients behaviorally. In the last part of the thesis it has been validated a behavioural strategy based on neurobiofeedback training, to voluntarily control seizures and to reduce there frequency. Three epileptic patients were included in this study. The biofeedback was based on monitoring of slow cortical potentials (SCPs) extracted online from scalp EEG. Patients were trained to produce positive shifts of SCPs. After a training phase patients were monitored for 6 months in order to validate the ability of the learned strategy to reduce seizure frequency. Two of the three refractory epileptic patients recruited for this study showed improvements in self-management and reduction of ictal episodes, even six months after the last training session.

Machine learning methods for prediction of disulphide bonding states of cysteine residues in proteins

The goal of this thesis work is to develop a computational method based on machine learning techniques for predicting disulfide-bonding states of cysteine residues in proteins, which is a sub-problem of a bigger and yet unsolved problem of protein structure prediction. Improvement in the prediction of disulfide bonding states of cysteine residues will help in putting a constraint in the three dimensional (3D) space of the respective protein structure, and thus will eventually help in the prediction of 3D structure of proteins. Results of this work will have direct implications in site-directed mutational studies of proteins, proteins engineering and the problem of protein folding. We have used a combination of Artificial Neural Network (ANN) and Hidden Markov Model (HMM), the so-called Hidden Neural Network (HNN) as a machine learning technique to develop our prediction method. By using different global and local features of proteins (specifically profiles, parity of cysteine residues, average cysteine conservation, correlated mutation, sub-cellular localization, and signal peptide) as inputs and considering Eukaryotes and Prokaryotes separately we have reached to a remarkable accuracy of 94% on cysteine basis for both Eukaryotic and Prokaryotic datasets, and an accuracy of 90% and 93% on protein basis for Eukaryotic dataset and Prokaryotic dataset respectively. These accuracies are best so far ever reached by any existing prediction methods, and thus our prediction method has outperformed all the previously developed approaches and therefore is more reliable. Most interesting part of this thesis work is the differences in the prediction performances of Eukaryotes and Prokaryotes at the basic level of input coding when ‘profile’ information was given as input to our prediction method. And one of the reasons for this we discover is the difference in the amino acid composition of the local environment of bonded and free cysteine residues in Eukaryotes and Prokaryotes. Eukaryotic bonded cysteine examples have a ‘symmetric-cysteine-rich’ environment, where as Prokaryotic bonded examples lack it.

Combined Evaluation of ST-segment Elevation in lead AVR and ST-segment Depression in other Leads Enhances Prediction of in-hospital Cardiovascular Death in Patients with Non ST-segment Elevation Acute Coronary Syndrome

Objective: To investigate the prognostic significance of ST-segment elevation (STE) in aVR associated with ST-segment depression (STD) in other leads in patients with non-STE acute coronary syndrome (NSTE-ACS). Background: In NSTE-ACS patients, STD has been extensively associated with severe coronary lesions and poor outcomes. The prognostic role of STE in aVR is uncertain. Methods: We enrolled 888 consecutive patients with NSTE-ACS. They were divided into two groups according to the presence or not on admission ECG of aVR STE≥ 1mm and STD (defined as high risk ECG pattern). The primary and secondary endpoints were: in-hospital cardiovascular (CV) death and the rate of culprit left main disease (LMD). Results: Patients with high risk ECG pattern (n=121) disclosed a worse clinical profile compared to patients (n=575) without [median GRACE (Global-Registry-of-Acute-Coronary-Events) risk score =142 vs. 182, respectively]. A total of 75% of patients underwent coronary angiography. The rate of in-hospital CV death was 3.9%. On multivariable analysis patients who had the high risk ECG pattern showed an increased risk of CV death (OR=2.88, 95%CI 1.05-7.88) and culprit LMD (OR=4.67,95%CI 1.86-11.74) compared to patients who had not. The prognostic significance of the high risk ECG pattern was maintained even after adjustment for the GRACE risk score (OR = 2.28, 95%CI:1.06-4.93 and OR = 4.13, 95%CI:2.13-8.01, for primary and secondary endpoint, respectively). Conclusions: STE in aVR associated with STD in other leads predicts in-hospital CV death and culprit LMD. This pattern may add prognostic information in patients with NSTE-ACS on top of recommended scoring system.

The prediction of flash evaporation in superheated fuel injections for automotive applications

In gasoline Port Fuel Injection (PFI) and Direct Injection (GDI) internal combustion engines, the liquid fuel might be injected into a gaseous ambient in a superheated state, resulting in flash boiling of the fuel. The importance to investigate and predict such a process is due to the influence it has on the liquid fuel atomization and vaporization and thus on combustion, with direct implications on engine performances and exhaust gas emissions. The topic of the present PhD research involves the numerical analysis of the behaviour of the superheated fuel during the injection process, in high pressure injection systems like the ones equipping GDI engines. Particular emphasis is on the investigation of the effects of the fuel superheating degree on atomization dynamics and spray characteristics. The present work is a look at the flash evaporation and flash boiling modeling, from an engineering point of view, addressed to keep the complex physics involved as simple as possible, however capturing the main characteristics of a superheated fuel injection.

Changes in tumor stiffness for early prediction of tumor response to sorafenib: a proof-of-concept study with elastosonography in an animal model of Hepatocellular Carcinoma (HCC)

Background and aims: Sorafenib is the reference therapy for advanced Hepatocellular Carcinoma (HCC). No method exists to predict in the very early period subsequent individual response. Starting from the clinical experience in humans that subcutaneous metastases may rapidly change consistency under sorafenib and that elastosonography a new ultrasound based technique allows assessment of tissue stiffness, we investigated the role of elastonography in the very early prediction of tumor response to sorafenib in a HCC animal model. Methods: HCC (Huh7 cells) subcutaneous xenografting in mice was utilized. Mice were randomized to vehicle or treatment with sorafenib when tumor size was 5-10 mm. Elastosonography (Mylab 70XVG, Esaote, Genova, Italy) of the whole tumor mass on a sagittal plane with a 10 MHz linear transducer was performed at different time points from treatment start (day 0, +2, +4, +7 and +14) until mice were sacrified (day +14), with the operator blind to treatment. In order to overcome variability in absolute elasticity measurement when assessing changes over time, values were expressed in arbitrary units as relative stiffness of the tumor tissue in comparison to the stiffness of a standard reference stand-off pad lying on the skin over the tumor. Results: Sor-treated mice showed a smaller tumor size increase at day +14 in comparison to vehicle-treated (tumor volume increase +192.76% vs +747.56%, p=0.06). Among Sor-treated tumors, 6 mice showed a better response to treatment than the other 4 (increase in volume +177% vs +553%, p=0.011). At day +2, median tumor elasticity increased in Sor-treated group (+6.69%, range –30.17-+58.51%), while decreased in the vehicle group (-3.19%, range –53.32-+37.94%) leading to a significant difference in absolute values (p=0.034). From this time point onward, elasticity decreased in both groups, with similar speed over time, not being statistically different anymore. In Sor-treated mice all 6 best responders at day 14 showed an increase in elasticity at day +2 (ranging from +3.30% to +58.51%) in comparison to baseline, whereas 3 of the 4 poorer responders showed a decrease. Interestingly, these 3 tumours showed elasticity values higher than responder tumours at day 0. Conclusions: Elastosonography appears a promising non-invasive new technique for the early prediction of HCC tumor response to sorafenib. Indeed, we proved that responder tumours are characterized by an early increase in elasticity. The possibility to distinguish a priori between responders and non responders based on the higher elasticity of the latter needs to be validated in ad-hoc experiments as well as a confirmation of our results in humans is warranted.

Machine-learning methods for structure prediction of β-barrel membrane proteins

Different types of proteins exist with diverse functions that are essential for living organisms. An important class of proteins is represented by transmembrane proteins which are specifically designed to be inserted into biological membranes and devised to perform very important functions in the cell such as cell communication and active transport across the membrane. Transmembrane β-barrels (TMBBs) are a sub-class of membrane proteins largely under-represented in structure databases because of the extreme difficulty in experimental structure determination. For this reason, computational tools that are able to predict the structure of TMBBs are needed. In this thesis, two computational problems related to TMBBs were addressed: the detection of TMBBs in large datasets of proteins and the prediction of the topology of TMBB proteins. Firstly, a method for TMBB detection was presented based on a novel neural network framework for variable-length sequence classification. The proposed approach was validated on a non-redundant dataset of proteins. Furthermore, we carried-out genome-wide detection using the entire Escherichia coli proteome. In both experiments, the method significantly outperformed other existing state-of-the-art approaches, reaching very high PPV (92%) and MCC (0.82). Secondly, a method was also introduced for TMBB topology prediction. The proposed approach is based on grammatical modelling and probabilistic discriminative models for sequence data labeling. The method was evaluated using a newly generated dataset of 38 TMBB proteins obtained from high-resolution data in the PDB. Results have shown that the model is able to correctly predict topologies of 25 out of 38 protein chains in the dataset. When tested on previously released datasets, the performances of the proposed approach were measured as comparable or superior to the current state-of-the-art of TMBB topology prediction.

Subject-specific musculoskeletal models of the lower limbs for the prediction of skeletal loads during motion

The determination of skeletal loading conditions in vivo and their relationship to the health of bone tissues, remain an open question. Computational modeling of the musculoskeletal system is the only practicable method providing a valuable approach to muscle and joint loading analyses, although crucial shortcomings limit the translation process of computational methods into the orthopedic and neurological practice. A growing attention focused on subject-specific modeling, particularly when pathological musculoskeletal conditions need to be studied. Nevertheless, subject-specific data cannot be always collected in the research and clinical practice, and there is a lack of efficient methods and frameworks for building models and incorporating them in simulations of motion. The overall aim of the present PhD thesis was to introduce improvements to the state-of-the-art musculoskeletal modeling for the prediction of physiological muscle and joint loads during motion. A threefold goal was articulated as follows: (i) develop state-of-the art subject-specific models and analyze skeletal load predictions; (ii) analyze the sensitivity of model predictions to relevant musculotendon model parameters and kinematic uncertainties; (iii) design an efficient software framework simplifying the effort-intensive phases of subject-specific modeling pre-processing. The first goal underlined the relevance of subject-specific musculoskeletal modeling to determine physiological skeletal loads during gait, corroborating the choice of full subject-specific modeling for the analyses of pathological conditions. The second goal characterized the sensitivity of skeletal load predictions to major musculotendon parameters and kinematic uncertainties, and robust probabilistic methods were applied for methodological and clinical purposes. The last goal created an efficient software framework for subject-specific modeling and simulation, which is practical, user friendly and effort effective. Future research development aims at the implementation of more accurate models describing lower-limb joint mechanics and musculotendon paths, and the assessment of an overall scenario of the crucial model parameters affecting the skeletal load predictions through probabilistic modeling.

Learning with Kernels on Graphs: DAG-based kernels, data streams and RNA function prediction.

In many application domains data can be naturally represented as graphs. When the application of analytical solutions for a given problem is unfeasible, machine learning techniques could be a viable way to solve the problem. Classical machine learning techniques are defined for data represented in a vectorial form. Recently some of them have been extended to deal directly with structured data. Among those techniques, kernel methods have shown promising results both from the computational complexity and the predictive performance point of view. Kernel methods allow to avoid an explicit mapping in a vectorial form relying on kernel functions, which informally are functions calculating a similarity measure between two entities. However, the definition of good kernels for graphs is a challenging problem because of the difficulty to find a good tradeoff between computational complexity and expressiveness. Another problem we face is learning on data streams, where a potentially unbounded sequence of data is generated by some sources. There are three main contributions in this thesis. The first contribution is the definition of a new family of kernels for graphs based on Directed Acyclic Graphs (DAGs). We analyzed two kernels from this family, achieving state-of-the-art results from both the computational and the classification point of view on real-world datasets. The second contribution consists in making the application of learning algorithms for streams of graphs feasible. Moreover,we defined a principled way for the memory management. The third contribution is the application of machine learning techniques for structured data to non-coding RNA function prediction. In this setting, the secondary structure is thought to carry relevant information. However, existing methods considering the secondary structure have prohibitively high computational complexity. We propose to apply kernel methods on this domain, obtaining state-of-the-art results.

A New Prediction Model for Slope Stability Analysis

The instability of river bank can result in considerable human and land losses. The Po river is the most important in Italy, characterized by main banks of significant and constantly increasing height. This study presents multilayer perceptron of artificial neural network (ANN) to construct prediction models for the stability analysis of river banks along the Po River, under various river and groundwater boundary conditions. For this aim, a number of networks of threshold logic unit are tested using different combinations of the input parameters. Factor of safety (FS), as an index of slope stability, is formulated in terms of several influencing geometrical and geotechnical parameters. In order to obtain a comprehensive geotechnical database, several cone penetration tests from the study site have been interpreted. The proposed models are developed upon stability analyses using finite element code over different representative sections of river embankments. For the validity verification, the ANN models are employed to predict the FS values of a part of the database beyond the calibration data domain. The results indicate that the proposed ANN models are effective tools for evaluating the slope stability. The ANN models notably outperform the derived multiple linear regression models.