Georgios Rigas

A User Independent, Biosignal Based, Emotion Recognition Method

A physiological signal based emotion recognition method, for the assessment of three emotional classes: happiness, disgustand fear, is presented. Our approach consists of four steps: (i) biosignal acquisition, (ii) biosignal preprocessing and feature extraction, (iii) feature selection and (iv) classification. The input signals are facial electromyograms, the electrocardiogram, the respiration and the electrodermal skin response. We have constructed a dataset which consists of 9 healthy subjects. Moreover we present preliminary results which indicate on average, accuracy rates of 0.48,0.68 and 0.69 for recognition of happiness, disgust and fear emotions, respectively.

Automated Levodopa-induced dyskinesia assessment

An automated methodology for Levodopa-induced dyskinesia (LID) assessment is presented in this paper. The methodology is based on the analysis of the signals recorded from accelerometers and gyroscopes, which are placed on certain positions on the subjects body. The obtained signals are analyzed and several features are extracted. Based on these features a classification technique is used for LID detection and classification of its severity. The method has been evaluated using a group of 10 subjects. Results are presented related to each individual sensor as well as for various sensor combinations. The obtained results indicate high classification ability (93.73% classification accuracy).

Extraction of consensus protein patterns in regions containing non-proline cis peptide bonds and their functional assessment

BackgroundIn peptides and proteins, only a small percentile of peptide bonds adopts the cis configuration. Especially in the case of amide peptide bonds, the amount of cis conformations is quite limited thus hampering systematic studies, until recently. However, lately the emerging population of databases with more 3D structures of proteins has produced a considerable number of sequences containing non-proline cis formations (cis-nonPro).ResultsIn our work, we extract regular expression-type patterns that are descriptive of regions surrounding the cis-nonPro formations. For this purpose, three types of pattern discovery are performed: i) exact pattern discovery, ii) pattern discovery using a chemical equivalency set, and iii) pattern discovery using a structural equivalency set. Afterwards, using each pattern as predicate, we search the Eukaryotic Linear Motif (ELM) resource to identify potential functional implications of regions with cis-nonPro peptide bonds. The patterns extracted from each type of pattern discovery are further employed, in order to formulate a pattern-based classifier, which is used to discriminate between cis-nonPro and trans-nonPro formations.ConclusionsIn terms of functional implications, we observe a significant association of cis-nonPro peptide bonds towards ligand/binding functionalities. As for the pattern-based classification scheme, the highest results were obtained using the structural equivalency set, which yielded 70% accuracy, 77% sensitivity and 63% specificity.

A Multiscale Approach for Modeling Atherosclerosis Progression

Progression of atherosclerotic process constitutes a serious and quite common condition due to accumulation of fatty materials in the arterial wall, consequently posing serious cardiovascular complications. In this paper, we assemble and analyze a multitude of heterogeneous data in order to model the progression of atherosclerosis (ATS) in coronary vessels. The patients medical record, biochemical analytes, monocyte information, adhesion molecules, and therapy-related data comprise the input for the subsequent analysis. As indicator of coronary lesion progression, two consecutive coronary computed tomography angiographies have been evaluated in the same patient. To this end, a set of 39 patients is studied using a twofold approach, namely, baseline analysis and temporal analysis. The former approach employs baseline information in order to predict the future state of the patient (in terms of progression of ATS). The latter is based on an approach encompassing dynamic Bayesian networks whereby snapshots of the patients status over the follow-up are analyzed in order to model the evolvement of ATS, taking into account the temporal dimension of the disease. The quantitative assessment of our work has resulted in 93.3% accuracy for the case of baseline analysis, and 83% overall accuracy for the temporal analysis, in terms of modeling and predicting the evolvement of ATS. It should be noted that the application of the SMOTE algorithm for handling class imbalance and the subsequent evaluation procedure might have introduced an overestimation of the performance metrics, due to the employment of synthesized instances. The most prominent features found to play a substantial role in the progression of the disease are: diabetes, cholesterol and cholesterol/HDL. Among novel markers, the CD11b marker of leukocyte integrin complex is associated with coronary plaque progression.

On Assessing Motor Disorders in Parkinson's Disease

In this paper we propose an automated method for assessing motor symptoms in Parkinson’s disease. Levodopa-induced dyskinesia (LID) and Freezing of Gait (FoG) are detected based on the analysis of signals recorded from wearable devices, i.e. accelerometers and gyroscopes, which are placed on certain positions on the patient’s body. The signals are initially pre-processed and then analyzed, using a moving window, in order to extract features from them. These features are used for LID and FoG assessment. Two classification techniques are employed, decision trees and random forests. The method has been evaluated using a group of patients and the obtained results indicate high classification ability, being 96.11% classification accuracy for FoG detection and 92.59% for LID severity assessment.

A multilevel and multiscale approach for the prediction of oral cancer reoccurrence

Oral cancer is the predominant neoplasm of the head and neck. Annually, more than 0.5 million new patients are diagnosed with oral cancer, worldwide. After the initial treatment and patient remission, reoccurrence rates still remain quite high. Early identification of such relapses is of crucial significance. Up to now, several approaches have been proposed for this purpose yielding however, unsatisfactory results. This is mainly attributed to the non-unified nature of these studies which focus only on a subset of the factors involved in the development and reoccurrence of oral cancer. Here we propose an orchestrated approach based on Dynamic Bayesian Networks (DBNs) for the prediction of a potential relapse after the disease has reached remission. A broad range of heterogeneous data sources featuring clinical, imaging and genomic information are assembled and analyzed during a predefined time-span, in order to decipher new and informative feature groups that correlate significantly with the progression of the disease and identify early potential relapses (local or metastatic) of the disease.

Predicting rapid progression of Parkinson's Disease at baseline patients evaluation

The rate of Parkinsons Disease (PD) progression in the initial post-diagnosis years can vary significantly. In this work, a methodology for the extraction of the most informative features for predicting rapid progression of the disease is proposed, using public data from the Parkinsons Progression Markers Initiative (PPMI) and machine learning techniques. The aim is to determine if a patient is at risk of expressing rapid progression of PD symptoms from the baseline evaluation and as close to diagnosis as possible. By examining the records of 409 patients from the PPMI dataset, the features with the best predictive value at baseline patient evaluation are found to be sleep problems, daytime sleepiness and fatigue, motor symptoms at legs, cognition impairment, early axial and facial symptoms and in the most rapidly advanced cases speech issues, loss of smell and affected leg muscle reflexes.

Mining motor symptoms UPDRS data of Parkinson's disease patients for the development of Hoehn and Yahr estimation decision support system

In this work a decision support system (DSS) for the conversion of Unified Parkinsons Disease Rating Scale (UPDRS) motor symptoms into a Hoehn & Yahr stage representation is proposed. Accurate estimation of a Parkinsons Disease patients Hoehn & Yahr stage is of great importance since this single value is enough to represent condition, severity of symptoms and localization and disease progression. For the first time data mining techniques are used to enhance Hoehn & Yahr stage estimation performance in a DSS. In its core a classification algorithm is trained using motor evaluation UPDRS data and new instances can then be automatically classified to provide suggestions and facilitate the clinicians final decision. Different classification methods and feature evaluation approaches are evaluated using public UPDRS data from the Parkinsons Progression Markers Initiative (PPMI). Overall, the Hoehn & Yahr stage classification accuracy reaches 87%.

Sequential vs concurrent epirubicin and docetaxel as adjuvant chemotherapy for high-risk, node-negative, early breast cancer: an interim analysis of a randomised phase III study from the Hellenic Oncology Research Group

Background:Sequential anthracyclines and taxanes are standard adjuvant chemotherapy for patients with high-risk axillary node-positive breast cancer. We compared a sequential to a concurrent regimen in high-risk node-negative early breast cancer.Methods:Patients were eligible if they had tumours >2 cm or T1c with two of the following characteristics: no oestrogen receptor (ER) and progesterone receptor (PR) expression, histological grade III, Ki67 >40% and vascular, lymphovascular or perineural invasion. They were randomised to receive four cycles of epirubicin 90 mg m−2 followed by four cycles of docetaxel 75 mg m−2 (sequential regimen) or six cycles of epirubicin 75 mg m−2 plus docetaxel 75 mg m−2 (concurrent regimen). All chemotherapy cycles were administered every 21 days with G-CSF prophylaxis only for the concurrent arm. The primary endpoint was disease-free survival (DFS).Results:Between 2001 and 2013, 658 women received the sequential (n=329) or the concurrent (n=329) regimen. The median age was 53 years, 43.9% of the patients were premenopausal and of the tumours 44.2% were ⩽2 cm, 52.7% histological grade 3 and 35.3% hormone receptor-negative. After a median follow-up of 70.5 months, there were 29 (8.8%) vs 42 (12.8%) disease relapses (P=0.102) and 11 (3.3%) vs 19 (5.8%) deaths (P=0.135), in the sequential and concurrent arm, respectively. The 5-year DFS rates were 92.6% vs 88.2% for sequential and concurrent arm, respectively (hazard ratio (HR): 1.591; 95% confidence interval (CI): 0.990–2.556; P=0.055). Toxicity included grade 2–4 neutropenia in 54% vs 41% (P=0.001), febrile neutropenia 2.7% vs 6.1% (P=0.06), nausea/vomiting 18.5% vs 12.4% (P=0.03) of patients in the sequential and concurrent arm. There were no toxic deaths.Conclusions:Sequential compared with the concurrent administration of anthracyclines and taxanes is associated with a non-significant but possibly clinically meaningful improvement in DFS. In the era of molecular selection of patients for adjuvant chemotherapy, this study offers valuable information for the optimal administration of anthracyclines and taxanes in patients with node-negative disease.

A visualization system for histological image annotation and 3D reconstruction of parametric geometries of the inner ear

The human cochlea is responsible for transducing mechanical stimuli to electrical signals. However, there are aspects of cochlear physiology that are not clearly understood and computational modeling has been recently used for this purpose. The computational approaches are usually based on artificial cochlea geometries, which present the cochlea as a box model. In this work, we present a system for the annotation of realistic histological images of the inner ear to reconstruct the cochlea structures using a realistic centerline. Using this approach the user can reconstruct as many structures are visible at the histological images or simplified geometries. The system gives the capabilities to modify the shape, size and length of the basilar membrane and scalae structures. Also, a visualization platform is provided. The approach can directly be used for modeling of inner ear mechanics implementing parametric design to understand accurately the mechanism of basilar membrane excitation from the mechanical stimuli of the pressure at the level of the stapes footplate.