Antonio Lanata

The Role of Nonlinear Dynamics in Affective Valence and Arousal Recognition

This paper reports on a new methodology for the automatic assessment of emotional responses. More specifically, emotions are elicited in agreement with a bidimensional spatial localization of affective states, that is, arousal and valence dimensions. A dedicated experimental protocol was designed and realized where specific affective states are suitably induced while three peripheral physiological signals, i.e., ElectroCardioGram (ECG), ElectroDermal Response (EDR), and ReSPiration activity (RSP), are simultaneously acquired. A group of 35 volunteers was presented with sets of images gathered from the International Affective Picture System (IAPS) having five levels of arousal and five levels of valence, including a neutral reference level in both. Standard methods as well as nonlinear dynamic techniques were used to extract sets of features from the collected signals. The goal of this paper is to implement an automatic multiclass arousal/valence classifier comparing performance when extracted features from nonlinear methods are used as an alternative to standard features. Results show that, when nonlinearly extracted features are used, the percentages of successful recognition dramatically increase. A good recognition accuracy (>;90 percent) after 40-fold cross-validation steps for both arousal and valence classes was achieved by using the Quadratic Discriminant Classifier (QDC).

SoC CMOS UWB Pulse Radar Sensor for Contactless Respiratory Rate Monitoring

An ultra wideband (UWB) system-on-chip radar sensor for respiratory rate monitoring has been realized in 90 nm CMOS technology and characterized experimentally. The radar testchip has been applied to the contactless detection of the respiration activity of adult and baby. The field operational tests demonstrate that the UWB radar sensor detects the respiratory rate of person under test (adult and baby) associated with sub-centimeter chest movements, allowing the continuous-time non-invasive monitoring of hospital patients and other people at risk of obstructive apneas such as babies in cot beds, or other respiratory diseases.

Wearable monitoring for mood recognition in bipolar disorder based on history-dependent long-term heart rate variability analysis.

Current clinical practice in diagnosing patients affected by psychiatric disorders such as bipolar disorder is based only on verbal interviews and scores from specific questionnaires, and no reliable and objective psycho-physiological markers are taken into account. In this paper, we propose to use a wearable system based on a comfortable t-shirt with integrated fabric electrodes and sensors able to acquire electrocardiogram, respirogram, and body posture information in order to detect a pattern of objective physiological parameters to support diagnosis. Moreover, we implemented a novel ad hoc methodology of advanced biosignal processing able to effectively recognize four possible clinical mood states in bipolar patients (i.e., depression, mixed state, hypomania, and euthymia) continuously monitored up to 18 h, using heart rate variability information exclusively. Mood assessment is intended as an intrasubject evaluation in which the patients states are modeled as a Markov chain, i.e., in the time domain, each mood state refers to the previous one. As validation, eight bipolar patients were monitored collecting and analyzing more than 400 h of autonomic and cardiovascular activity. Experimental results demonstrate that our novel concept of personalized and pervasive monitoring constitutes a viable and robust clinical decision support system for bipolar disorders recognizing mood states with a total classification accuracy up to 95.81%.

Revealing Real-Time Emotional Responses: a Personalized Assessment based on Heartbeat Dynamics

Emotion recognition through computational modeling and analysis of physiological signals has been widely investigated in the last decade. Most of the proposed emotion recognition systems require relatively long-time series of multivariate records and do not provide accurate real-time characterizations using short-time series. To overcome these limitations, we propose a novel personalized probabilistic framework able to characterize the emotional state of a subject through the analysis of heartbeat dynamics exclusively. The study includes thirty subjects presented with a set of standardized images gathered from the international affective picture system, alternating levels of arousal and valence. Due to the intrinsic nonlinearity and nonstationarity of the RR interval series, a specific point-process model was devised for instantaneous identification considering autoregressive nonlinearities up to the third-order according to the Wiener-Volterra representation, thus tracking very fast stimulus-response changes. Features from the instantaneous spectrum and bispectrum, as well as the dominant Lyapunov exponent, were extracted and considered as input features to a support vector machine for classification. Results, estimating emotions each 10 seconds, achieve an overall accuracy in recognizing four emotional states based on the circumplex model of affect of 79.29%, with 79.15% on the valence axis, and 83.55% on the arousal axis.

Comparative Evaluation of Susceptibility to Motion Artifact in Different Wearable Systems for Monitoring Respiratory Rate

The purpose of this study is to comparatively evaluate the performance of different wearable systems based on indirect breathing monitoring in terms of susceptibility to motion artifacts. These performances are compared with direct respiratory measurements using a spirometer, which is accurate, reliable, and less sensitive to movement artifacts, but cannot be integrated into truly wearable form. Experiments were carried out on four indirect methods implemented into wearable systems, inductive plethysmography, impedance plethysmography, piezoresistive pneumography, and piezoelectric pneumography, to ascertain the performance of each of them in terms of noise due to movement artifacts, as well as to study the effects of different movements or gestures during each test. A group of volunteers was asked to wear all of the breath monitoring systems simultaneously along with the face mask of the spirometer while carrying out four physical exercises in a gym under controlled conditions. Data are analyzed in the time and frequency domain to estimate the frequency respiration from each wearable system and compare it with those of the spirometer. Results confirmed that all the wearable systems are somehow affected by movement artifacts, but statistical investigation showed that for most of the physical exercises, three out of four, piezoelectric pneumography provided best performance in terms of robustness and reduced susceptibility to movement artifacts.

Dominant Lyapunov exponent and approximate entropy in heart rate variability during emotional visual elicitation

In this work we characterized the non-linear complexity of Heart Rate Variability (HRV) in short time series. The complexity of HRV signal was evaluated during emotional visual elicitation by using Dominant Lyapunov Exponents (DLEs) and Approximate Entropy (ApEn). We adopted a simplified model of emotion derived from the Circumplex Model of Affects (CMAs), in which emotional mechanisms are conceptualized in two dimensions by the terms of valence and arousal. Following CMA model, a set of standardized visual stimuli in terms of arousal and valence gathered from the International Affective Picture System (IAPS) was administered to a group of 35 healthy volunteers. Experimental session consisted of eight sessions alternating neutral images with high arousal content images. Several works can be found in the literature showing a chaotic dynamics of HRV during rest or relax conditions. The outcomes of this work showed a clear switching mechanism between regular and chaotic dynamics when switching from neutral to arousal elicitation. Accordingly, the mean ApEn decreased with statistical significance during arousal elicitation and the DLE became negative. Results showed a clear distinction between the neutral and the arousal elicitation and could be profitably exploited to improve the accuracy of emotion recognition systems based on HRV time series analysis.

Electrodermal activity in bipolar patients during affective elicitation.

Bipolar patients are characterized by a pathological unpredictable behavior, resulting in fluctuations between states of depression and episodes of mania or hypomania. In the current clinical practice, the psychiatric diagnosis is made through clinician-administered rating scales and questionnaires, disregarding the potential contribution provided by physiological signs. The aim of this paper is to investigate how changes in the autonomic nervous system activity can be correlated with clinical mood swings. More specifically, a group of ten bipolar patients underwent an emotional elicitation protocol to investigate the autonomic nervous system dynamics, through the electrodermal activity (EDA), among different mood states. In addition, a control group of ten healthy subjects were recruited and underwent the same protocol. Physiological signals were analyzed by applying the deconvolutive method to reconstruct EDA tonic and phasic components, from which several significant features were extracted to quantify the sympathetic activation. Experimental results performed on both the healthy subjects and the bipolar patients supported the hypothesis of a relationship between autonomic dysfunctions and pathological mood states.

Characterization of Depressive States in Bipolar Patients Using Wearable Textile Technology and Instantaneous Heart Rate Variability Assessment

The analysis of cognitive and autonomic responses to emotionally relevant stimuli could provide a viable solution for the automatic recognition of different mood states, both in normal and pathological conditions. In this study, we present a methodological application describing a novel system based on wearable textile technology and instantaneous nonlinear heart rate variability assessment, able to characterize the autonomic status of bipolar patients by considering only electrocardiogram recordings. As a proof of this concept, our study presents results obtained from eight bipolar patients during their normal daily activities and being elicited according to a specific emotional protocol through the presentation of emotionally relevant pictures. Linear and nonlinear features were computed using a novel point-process-based nonlinear autoregressive integrative model and compared with traditional algorithmic methods. The estimated indices were used as the input of a multilayer perceptron to discriminate the depressive from the euthymic status. Results show that our system achieves much higher accuracy than the traditional techniques. Moreover, the inclusion of instantaneous higher order spectra features significantly improves the accuracy in successfully recognizing depression from euthymia.

Oscillations of Heart Rate and Respiration Synchronize During Affective Visual Stimulation

The objective of this study is to investigate the synchronization between breathing patterns and heart rate during emotional visual elicitation, that is, using sets of images gathered from the international affective picture system having five levels of arousal and five levels of valence, including a neutral reference level. Thirty-five healthy volunteers were emotionally elicited in agreement with a bidimensional spatial localization of affective states, i.e., arousal/valence plane, while two peripheral physiological signals, ECG and Respiration activity, were acquired simultaneously. The synchronization was then quantified by applying the concept of phase synchronization of chaotic oscillators, i.e., the cardio-respiratory synchrogram. This technique allowed us to estimate the synchronization ratio m:n as the attendance of n heartbeats in each m respiratory cycle, even for noisy and nonstationary data. We found a stronger evidence of cardiorespiratory synchronization during arousal than during neutral states.

A novel EDA glove based on textile integrated electrodes for affective computing

This paper reports on performance evaluation of a preliminary system prototype based on a fabric glove, with integrated textile electrodes placed at the fingertips, able to acquire and process the electrodermal response (EDR) to discriminate affective states. First, textile electrodes have been characterized in terms of voltage–current characteristics and trans-surface electric impedance. Next, signal quality of EDR acquired simultaneously from textile and standard electrodes was comparatively evaluated. Finally, a dedicated experiment in which 35 subjects were enrolled, aiming at discriminating different affective states using only EDR was designed and realized. A new set of features extracted from non-linear methods were used, improving remarkably successful recognition rates. Results are, indeed, very satisfactory and promising in the field of affective computing.