Publications

This work aims at corroborating the importance and efficacy of mutual learning in motor imagery (MI) brain–computer interface (BCI) by leveraging the insights obtained through our participation in the BCI race of the Cybathlon event. We hypothesized that, contrary to the popular trend of focusing mostly on the machine learning aspects of MI BCI training, a comprehensive mutual learning methodology that reinstates the three learning pillars (at the machine, subject, and application level) as equally significant could lead to a BCI–user symbiotic system able to succeed in real-world scenarios such as the Cybathlon event. Two severely impaired participants with chronic spinal cord injury (SCI), were trained following our mutual learning approach to control their avatar in a virtual BCI race game. The competition outcomes substantiate the effectiveness of this type of training. Most importantly, the present study is one among very few to provide multifaceted evidence on the efficacy of subject learning during BCI training. Learning correlates could be derived at all levels of the interface—application, BCI output, and electroencephalography (EEG) neuroimaging—with two end-users, sufficiently longitudinal evaluation, and, importantly, under real-world and even adverse conditions.

Objective: To evaluate the effectiveness of reinforced feedback in virtual environment (RFVE) treatment combined with conventional rehabilitation (CR) in comparison with CR alone, and to study whether changes are related to stroke etiology (ie, ischemic, hemorrhagic). Design: Randomized controlled trial. Setting: Hospital facility for intensive rehabilitation. Participants: Patients (N=136) within 1 year from onset of a single stroke (ischemic: n=78, hemorrhagic: n=58). Interventions: The experimental treatment was based on the combination of RFVE with CR, whereas control treatment was based on the same amount of CR. Both treatments lasted 2 hours daily, 5d/wk, for 4 weeks. Main Outcome Measures: Fugl-Meyer upper extremity scale (F-M UE) (primary outcome), FIM, National Institutes of Health Stroke Scale (NIHSS), and Edmonton Symptom Assessment Scale (ESAS) (secondary outcomes). Kinematic parameters of requested movements included duration (time), mean linear velocity (speed), and number of submovements (peak) (secondary outcomes). Results: Patients were randomized in 2 groups (RFVE with CR: n=68, CR: n=68) and stratified by stroke etiology (ischemic or hemorrhagic). Both groups improved after treatment, but the experimental group had better results than the control group (Mann-Whitney U test) for F-M UE (P<.001), FIM (P<.001), NIHSS (P≤.014), ESAS (P≤.022), time (P<.001), speed (P<.001), and peak (P<.001). Stroke etiology did not have significant effects on patient outcomes. Conclusions: The RFVE therapy combined with CR treatment promotes better outcomes for upper limb than the same amount of CR, regardless of stroke etiology.

Background: Dystroglycanopathy (α-DG) is a relatively common, clinically and genetically heterogeneous category of congenital forms of muscular dystrophy (CMD) and limb-girdle muscular dystrophy (LGMD) associated with hypoglycosylated α-dystroglycan. To date, mutations in at least 19 genes have been associated with α-DG. One of them, GMPPB, encoding the guanosine-diphosphate-mannose (GDP-mannose) pyrophosphorylase B protein, has recently been associated with a wide clinical spectrum ranging from severe Walker-Warburg syndrome to pseudo-metabolic myopathy and even congenital myasthenic syndromes. We re-sequenced the full set of known disease genes in 73 Italian patients with evidence of either reduced or nearly absent α-dystroglycan to assess genotype-phenotype correlations in this cohort. We used innovative bioinformatic tools to calculate the effects of all described GMPPB mutations on protein function and attempted to correlate them with phenotypic expressions. Results: We identified 13 additional cases from 12 families and defined seven novel mutations. Patients displayed variable phenotypes including less typical pictures, ranging from asymptomatic hyperCKemia, to arthrogryposis and congenital clubfoot at birth, and also showed neurodevelopmental comorbidities, such as seizures and ataxic gait, as well as autism-spectrum disorder, which is seldom described in clinical reports of dystroglycanopathies. We also demonstrated that few mutations recur in the Italian GMPPB-mutated population and that alterations of protein stability are the main effects of GMPPB missense variants. Conclusion: This work adds to the data on genotype-phenotype correlations in α-DG and offers new bionformatic tools to provide the conceptual framework needed to understand the complexity of these disorders.

Aims: Pompe disease is an autosomal recessive lysosomal storage disorder resulting from deficiency of acid α-glucosidase (GAA) enzyme. Histopathological hallmarks in skeletal muscle tissue are fibre vacuolization and autophagy. Since 2006, enzyme replacement therapy (ERT) is the only approved treatment with human recombinant GAA alglucosidase alfa. We designed a study to examine ERT-related skeletal muscle changes in 18 modestly to moderately affected late onset Pompe disease (LOPD) patients along with the relationship between morphological/biochemical changes and clinical outcomes. Treatment duration was short-to-long term. Methods: We examined muscle biopsies from 18 LOPD patients at both histopathological and biochemical level. All patients underwent two muscle biopsies, before and after ERT administration respectively. The study is partially retrospective because the first biopsies were taken before the study was designed, whereas the second biopsy was always performed after at least 6 months of ERT administration. Results: After ERT, 15 out of 18 patients showed improved 6-min walking test (6MWT; P = 0.0007) and most of them achieved respiratory stabilization. Pretreatment muscle biopsies disclosed marked histopathological variability, ranging from an almost normal pattern to a severe vacuolar myopathy. After treatment, we detected morphological improvement in 15 patients and worsening in three patients. Post-ERT GAA enzymatic activity was mildly increased compared with pretreatment levels in all patients. Protein levels of the mature enzyme increased in 14 of the 18 patients (mean increase = +35%; P < 0.05). Additional studies demonstrated an improved autophagic flux after ERT in some patients. Conclusions: ERT positively modified skeletal muscle pathology as well as motor and respiratory outcomes in the majority of LOPD patients.

The presence of major depressive symptoms is usually considered a negative long-term prognostic factor after an acute myocardial infarction (AMI); however, most of the supporting research was conducted before the era of immediate reperfusion by percutaneous coronary intervention. The aims of this study are to evaluate if depression still retains long-term prognostic significance in our era of immediate coronary reperfusion, and to study possible correlations with clinical parameters of physical performance. In 184 patients with recent ST-elevated AMI (STEMI), treated by immediate reperfusion, moderate or severe depressive symptoms (evaluated by Beck Depression Inventory version I) were present in 10 % of cases. Physical performance was evaluated by two 6-min walk tests and by a symptom-limited cardiopulmonary exercise test: somatic/affective (but not cognitive/affective) symptoms of depression and perceived quality of life (evaluated by the EuroQoL questionnaire) are worse in patients with lower levels of physical performance. Follow-up was performed after a median of 29 months by means of telephone interviews; 32 major adverse cardiovascular events (MACE) occurred. The presence of three vessels disease and low left ventricle ejection fraction are correlated with a greater incidence of MACE; only somatic/affective (but not cognitive/affective) symptoms of depression correlate with long-term outcomes. In patients with recent STEMI treated by immediate reperfusion, somatic/affective but not cognitive/affective symptoms of depression show prognostic value on long-term MACE. Depression symptoms are not predictors “per se” of adverse prognosis, but seem to express an underlying worse cardiac efficiency, clinically reflected by poorer physical performance.

Automated recognition of Activities of Daily Living allows to identify possible health problems and apply corrective strategies in Ambient Assisted Living (AAL). Activities of Daily Living analysis can provide very useful information for elder care and long-term care services. This paper presents an automated RGB-D video analysis system that recognises human ADLs activities, related to classical daily actions. The main goal is to predict the probability of an analysed subject action. Thus, abnormal behaviour can be detected. The activity detection and recognition is performed using an affordable RGB-D camera. Human activities, despite their unstructured nature, tend to have a natural hierarchical structure; for instance, generally making a coffee involves a three-step process of turning on the coffee machine, putting sugar in cup and opening the fridge for milk. Action sequence recognition is then handled using a discriminative Hidden Markov Model (HMM). RADiaL, a dataset with RGB-D images and 3D position of each person for training as well as evaluating the HMM, has been built and made publicly available.

This paper presents an RGB-D based human re-identification approach using novel biometrics features from the body’s volume. Existing work based on RGB images or skeleton features have some limitations for realworld robotic applications, most notably in dealing with occlusions and orientation of the user. Here, we propose novel features that allow performing re-identification when the person is facing side/backward or the person is partially occluded. The proposed approach has been tested for various scenarios including different views, occlusion and the public BIWI RGBD-ID dataset.

Double-layer hydroxide minerals are part of a very interesting group of natural and synthetic compounds with trigonal or hexagonal symmetry and a flexible layered crystal structure. They are formed extremely frequently in geologic, industrial, and synthetic processes. The ease of formation is related to the possibility of accommodating divalent and trivalent cations in the structure, together with a range of anionic species. Some compounds of the group, namely those based on hydrotalcite chemistry, are invariably found as products of the pozzolanic reaction between lime and clays in ancient mortars and modern binders that serve as alternatives to Portland clinker. The present review wishes to relate the structural properties of hydrotalcitetype compounds to the crystal-chemical mechanisms taking place during long-term pozzolanic processes. The kinetics of CO3 exchange between the hydroxide and the atmosphere has important negative consequences for the radiocarbon dating of ancient mortars.

Human detection and tracking are essential aspects to be considered in service robotics, as the robot often shares its workspace and interacts closely with humans. This paper presents an online learning framework for human classification in 3D LiDAR scans, taking advantage of robust multi-target tracking to avoid the need for data annotation by a human expert. The system learns iteratively by retraining a classifier online with the samples collected by the robot over time. A novel aspect of our approach is that errors in training data can be corrected using the information provided by the 3D LiDAR-based tracking. In order to do this, an efficient 3D cluster detector of potential human targets has been implemented. We evaluate the framework using a new 3D LiDAR dataset of people moving in a large indoor public space, which is made available to the research community. The experiments analyse the real-time performance of the cluster detector and show that our online learned human classifier matches and in some cases outperforms its offline version.

Stress remains a significant social problem for individuals in modern societies. This paper presents a machine learning approach for the automatic detection of stress of people in a social situation by combining two sensor systems that capture physiological and social responses. We compare the performance using different classifiers including support vector machine, AdaBoost, and k-nearest neighbor. Our experimental results show that by combining the measurements from both sensor systems, we could accurately discriminate between stressful and neutral situations during a controlled Trier social stress test (TSST). Moreover, this paper assesses the discriminative ability of each sensor modality individually and considers their suitability for real-time stress detection. Finally, we present an study of the most discriminative features for stress detection.