Marcello Bracale

Nonlinear heart rate variability features for real-life stress detection. Case study : students under stress due to university examination

BackgroundThis study investigates the variations of Heart Rate Variability (HRV) due to a real-life stressor and proposes a classifier based on nonlinear features of HRV for automatic stress detection.Methods42 students volunteered to participate to the study about HRV and stress. For each student, two recordings were performed: one during an on-going university examination, assumed as a real-life stressor, and one after holidays. Nonlinear analysis of HRV was performed by using Poincaré Plot, Approximate Entropy, Correlation dimension, Detrended Fluctuation Analysis, Recurrence Plot. For statistical comparison, we adopted the Wilcoxon Signed Rank test and for development of a classifier we adopted the Linear Discriminant Analysis (LDA).ResultsAlmost all HRV features measuring heart rate complexity were significantly decreased in the stress session. LDA generated a simple classifier based on the two Poincaré Plot parameters and Approximate Entropy, which enables stress detection with a total classification accuracy, a sensitivity and a specificity rate of 90%, 86%, and 95% respectively.ConclusionsThe results of the current study suggest that nonlinear HRV analysis using short term ECG recording could be effective in automatically detecting real-life stress condition, such as a university examination.

Remote Health Monitoring of Heart Failure With Data Mining via CART Method on HRV Features

Disease management programs, which use no advanced information and computer technology, are as effective as telemedicine but more efficient because less costly. We proposed a platform to enhance effectiveness and efficiency of home monitoring using data mining for early detection of any worsening in patients condition. These worsenings could require more complex and expensive care if not recognized. In this letter, we briefly describe the remote health monitoring platform we designed and realized, which supports heart failure (HF) severity assessment offering functions of data mining based on the classification and regression tree method. The system developed achieved accuracy and a precision of 96.39% and 100.00% in detecting HF and of 79.31% and 82.35% in distinguishing severe versus mild HF, respectively. These preliminary results were achieved on public databases of signals to improve their reproducibility. Clinical trials involving local patients are still running and will require longer experimentation.

Discrimination Power of Short-Term Heart Rate Variability Measures for CHF Assessment

In this study, we investigated the discrimination power of short-term heart rate variability (HRV) for discriminating normal subjects versus chronic heart failure (CHF) patients. We analyzed 1914.40 h of ECG of 83 patients of which 54 are normal and 29 are suffering from CHF with New York Heart Association (NYHA) classification I, II, and III, extracted by public databases. Following guidelines, we performed time and frequency analysis in order to measure HRV features. To assess the discrimination power of HRV features, we designed a classifier based on the classification and regression tree (CART) method, which is a nonparametric statistical technique, strongly effective on nonnormal medical data mining. The best subset of features for subject classification includes square root of the mean of the sum of the squares of differences between adjacent NN intervals (RMSSD), total power, high-frequencies power, and the ratio between low- and high-frequencies power (LF/HF). The classifier we developed achieved sensitivity and specificity values of 79.3% and 100 %, respectively. Moreover, we demonstrated that it is possible to achieve sensitivity and specificity of 89.7% and 100 %, respectively, by introducing two nonstandard features ΔAVNN and ΔLF/HF, which account, respectively, for variation over the 24 h of the average of consecutive normal intervals (AVNN) and LF/HF. Our results are comparable with other similar studies, but the method we used is particularly valuable because it allows a fully human-understandable description of classification procedures, in terms of intelligible “if ... then ...” rules.

Study of the control strategy of the quadriceps muscles in anterior knee pain

Anterior knee pain (AKP) is a common pathological condition, particularly among young people and athletes, associated to an abnormal motion of the patella during the bending of the knee and possibly dependent on a muscular or structural imbalance. A lack of synergy in the quadriceps muscles results in a dynamic misalignment of the patella, which in turn produces pain. AKP rehabilitative therapy consists of conservative treatment whose main objective is to strengthen the Vastus Medialis. The aim of this article is to study the quadriceps muscle control strategy in AKP patients during an isokinetic exercise. Analysis of the muscle activation strategy is important for an objective measurement of the knee functionality in that it helps to diagnose and monitor the rehabilitative treatment. Surface electromyography (EMG) from the three superficial muscles of the femoral quadriceps during a concentric isokinetic exercise has been analyzed along with the signals of knee joint position and torque. A group of 12 AKP patients has been compared with a group of 30 normal subjects. Analysis of the grand ensemble average of the EMG linear envelopes in AKP patients reveals significant modifications in Vastus Medialis activity compared to the other quadriceps muscles. In order to study the synergy of the muscles, temporal identifiers have been associated to the EMG linear envelopes. To this end, EMG linear envelope decomposition in Gaussian pulses turned out to be effective and the results highlight an appreciable delay in the activation of the Vastus Medialis in AKP patients. This muscular unbalance can explain the abnormal motion of the patella.

Totally laparoscopic gastrectomy for gastric cancer: Meta-analysis of short-term outcomes

Abstract Introduction: We present a review of the literature, together with a meta-analysis of short-term outcomes of totally laparoscopic gastrectomy (TLG) compared with open gastrectomy (OG). Material & methods: We carried out a search in the Pubmed and Cochrane databases from September 2003 to May 2009. Controlled studies on early outcomes were included, both prospective and retrospective, randomized and non-randomized. Results: We found nine eligible studies, one of which was a randomized controlled trial (RCT), while eight were series of patients (three consecutive). The study group consisted of 1,492 patients, 828 of whom had been treated with TLG and 664 treated with OG. TLG for gastric cancer shows a 32.5% (p < 0.001) longer operative time than OG, whereas TLG demonstrated a 44% (p < 0.001) reduction in blood loss, a 34% (p < 0.001) reduction time to first flatus and a 33.7% reduced (p < 0.001) hospital stay. No notable differences were registered regarding morbidity and mortality rates, and no significant difference was observed between the two groups regarding the extent of the lymphadenectomy. Conclusions: Despite a longer operative time for TLG, with a gastrointestinal recovery rate faster than the OG one for gastric cancer results, no notable differences were recorded between the two techniques for the morbidity and mortality rates and in the spread of the lymphadenectomy.

Automatic recognition of vertebral landmarks in fluoroscopic sequences for analysis of intervertebral kinematics

Intervertebral kinematics closely relates to the functionality of the spinal segments. Direct measurement of the intervertebral kinematics in vivo is very problematic. The use of a fluoroscopic device can provide continuous screening of the lumbar tract during patient spontaneous motion, with an acceptable, low X-ray dose. The kinematic analysis is intended to be limited to planar motion. Kinematic parameters are computed from vertebral landmarks on each frame of the image sequence. Landmarks are normally selected manually in spite of the fact that this is subjective, tedious to perform and regarded as one of the major contributors to errors in the computed kinematic parameters. The aim of this work is to present an innovative method for the automatic recognition of vertebral landmarks throughout a fluoroscopic image sequence to provide an objective and more precise quantification of intervertebral kinematics. The recognition procedure is based upon comparing vertebral features in two adjacent frames by means of a cross-correlation index, which is also robust despite the low signal-to-noise ratio of the lumbar fluoroscopic images. To provide a quantitative assessment of this method a calibration model was used which consisted of two lumbar vertebrae linked by a universal joint. The reliability and accuracy of the kinematic measurements have been investigated. The errors are of the order of a millimetre for the localisation of the intervertebral centre of rotation and tenths of a degree for the intervertebral angle. Error analysis suggests that this method improves the accuracy of the intervertebral kinematic calculations and has the potential to automate the selection of anatomical landmarks.

Estimation of out-of-plane vertebra rotations on radiographic projections using CT data: A simulation study

This study extends previous research concerning in vivo intervertebral motion by means of single-plane fluoroscopy in an attempt to overcome 2D analysis limitations. Knowledge of out-of-plane vertebra rotations will extend the results provided by planar kinematic studies, which is particularly important for lateral bending investigation where axial rotation accompanies side bending, but is also valuable in sagittal analysis (e.g. indicating an absence of coupled axial rotation). Combining a fluoroscopic projection of a vertebra with volumetric information provided by CT data, vertebra 3D position can be estimated. Out-of-plane vertebral rotations are estimated by comparing Digitally Reconstructed Radiographs (DRRs) in different orientations with a reference fluoroscopic projection, maximising the image cross-correlation index. DRRs have been computed from CT-data using a ray-casting algorithm. In this work a feasibility study of the method was performed by means of a computer simulation. To this end the CT volume (vertebra L4, segmented) provided by the Visible Human Project was utilised and reference fluoroscopic projections were simulated in different orientations adding various levels of noise. Accuracy and precision of the proposed method was determined. Error analysis reveals that an accuracy of less than 1 degree can be achieved in computation of out-of-plane vertebral angles.

Eye movement baseline oscillation and variability of eye position during foveation in congenital nystagmus

Despite the inability to maintain steady fixation, congenital nystagmus does not necessarily reduce visual acuity, that can be achieved during the foveation periods. The duration of the foveation period, but also the cycle-to-cycle variability of eye position and velocity during foveations play an important role. A quantitative relationship that relates visual acuity with foveation time and cycle-to-cycle variability of eye position during foveation has been previously proposed. In many infrared-oculographic and electro-oculographic eye position recordings of our database, a sinusoidal-like oscillation of the baseline was observed, on which the nystagmus waveforms lay. This oscillation may contribute to increase cycle-to-cycle variability during foveations. The aim of this work is to extract the baseline oscillation from the recordings and to verify its relationship with eye position variability during foveation. On the basis of the observations, the baseline oscillation was assumed to be sinusoidal, and was estimated (using a least mean square technique) from eye movement signals recorded during fixation intervals, at different gaze positions, from 20 patients affected by congenital nystagmus with low visual acuity. The average baseline oscillation amplitude was 1.31°, while the average frequency was 0.34 Hz. Baseline oscillation amplitude was well correlated (with a coefficient of 0.66) to the standard deviations of eye-position during foveation, which in turn is connected to visual acuity.

Quadriceps muscles activation in anterior knee pain during isokinetic exercise

The aim of this paper is to study the femoral quadriceps muscles activation in patients suffering from anterior knee pain (AKP) during an isokinetic exercise. AKP is a common pathological condition, associated to an abnormal motion of the patella. It possibly depends on a muscular or structural imbalance. A lack of synergy in the quadriceps muscles results in a dynamic misalignment of the patella, which in turn produces pain. A quantitative analysis of the muscle activation strategy is important for an objective measurement of the knee functionality in that it helps to diagnose and monitor the rehabilitative treatment. Surface electromyography (EMG) from the three superficial muscles of the femoral quadriceps during a concentric isokinetic exercise has been analysed. A group of AKP patients has been compared with a control group of healthy subjects. Ensemble average of the EMG linear envelopes reveals significant modifications in Vastus Medialis activity in AKP patients. In order to quantify the synergy of the muscles, different parameters have been associated to EMG linear envelopes significant features. A decomposition in gaussian pulses and an asymmetry coefficient have been utilised. The results relative to these concise parameters highlight an appreciable delay and a modification in the activation of the Vastus Medialis in AKP patients.

An expectation-maximisation approach for simultaneous pixel classification and tracer kinetic modelling in dynamic contrast enhanced-magnetic resonance imaging

Traditionally, tracer kinetic modelling and pixel classification of DCE-MRI studies are accomplished separately, although they could greatly benefit from each other. In this article, we propose an expectation-maximisation scheme for simultaneous pixel classification and compartmental modelling of DCE-MRI studies. The key point in the proposed scheme is the estimation of the kinetic parameters (Ktrans and Kep) of the two-compartmental model. Typically, they are estimated via nonlinear least-squares fitting. In our scheme, by exploiting the iterative nature of the EM algorithm, we use instead a Taylor expansion of the modelling equation. We developed the theoretical framework for the particular case of two classes and evaluated the performances of the algorithm by means of simulations. Results indicate that the accuracy of the proposed method supersedes the traditional pixel-by-pixel scheme and approaches the theoretical lower bound imposed by the Cramer–Rao theorem. Preliminary results on real data were also reported.