On the deconvolution analysis of electrodermal activity in bipolar patients

Dissecting clinical and biological heterogeneity in clinical states of bipolar disorder: a 10-year retrospective study from China

Objectives

To dissect clinical and biological heterogeneity in clinical states of bipolar disorder (BD), and investigate if neuropsychological symptomatology, comorbidity, vital signs, and blood laboratory indicators are predictors of distinct BD states.

Methods

A retrospective BD cohort was established with data extracted from a Chinese hospital's electronic medical records (EMR) between 2009 and 2018. Subjects were inpatients with a main discharge diagnosis of BD and were assessed for clinical state at hospitalization. We categorized all subjects into manic state, depressive state, and mixed state. Four machine learning classifiers were utilized to classify the subjects. A Shapley additive explanations (SHAP) algorithm was applied to the classifiers to aid in quantifying and visualizing the contributions of each feature that drive patient-specific classifications.

Results

A sample of 3,085 records was included (38.54% as manic, 56.69% as depressive, and 4.77% as mixed state). Mixed state showed more severe suicidal ideation and psychomotor abnormalities, while depressive state showed more common anxiety, sleep, and somatic-related symptoms and more comorbid conditions. Higher levels of body temperature, pulse, and systolic and diastolic blood pressures were present during manic episodes. Xgboost achieved the best AUC of 88.54% in manic/depressive states classification; Logistic regression and Random forest achieved the best AUCs of 75.5 and 75% in manic/mixed states and depressive/mixed states classifications, respectively. Myocardial enzymes and the non-enzymatic antioxidant uric acid and bilirubin contributed significantly to distinguish BD clinical states.

Conclusion

The observed novel biological associations with BD clinical states confirm that biological heterogeneity contributes to clinical heterogeneity of BD.

Point-process Nonlinear Autonomic Assessment of Depressive States in Bipolar Patients

Introduction: This article is part of the Focus Theme of Methods of Information in Medicine on “Biosignal Interpretation: Advanced Methods for Studying Cardiovascular and Respiratory Systems”.

Objectives: The goal of this work is to apply a computational methodology able to characterize mood states in bipolar patients through instantaneous analysis of heartbeat dynamics.

Methods: A Point-Process-based Nonlinear Autoregressive Integrative (NARI) model is applied to analyze data collected from five bipolar patients (two males and three females, age 42.4 ± 10.5 range 32−56) undergoing a dedicated affective elicitation protocol using images from the International Affective Picture System (IAPS) and Thematic Apperception Test (TAT). The study was designed within the European project PSYCHE (Personalised monitoring SYstems for Care in mental HEalth).

Results: Results demonstrate that the inclusion of instantaneous higher order spectral (HOS) features estimated from the NARI nonlinear assessment significantly improves the accuracy in successfully recognizing specific mood states such as euthymia and depression with respect to results using only linear indices. In particular, a specificity of 74.44% using the instantaneous linear features set, and 99.56% using also the nonlinear feature set were achieved. Moreover, IAPS emotional elicitation resulted in a more discriminant procedure with respect to the TAT elicitation protocol.

Conclusions: A significant pattern of instantaneous heartbeat features was found in depressive and euthymic states despite the inter-subject variability. The presented point-process Heart Rate Variability (HRV) nonlinear methodology provides a promising application in the field of mood assessment in bipolar patients.

Gender-specific velocity recognition of caress-like stimuli through nonlinear analysis of Heart Rate Variability

This study reports on the development of a gender-specific classification system able to discern between two levels of velocity of a caress-like stimulus, through information gathered from Autonomic Nervous System (ANS) linear and nonlinear dynamics. Specifically, caress-like stimuli were administered to 32 healthy volunteers (16 males) while monitoring electrocardiogram signal to extract Heart Rate Variability (HRV) series. Caressing stimuli were administered to the forearm at a fixed force level (6 N) and two levels of velocity, 9.4 mm/s and 37 mm/s. Standard HRV measures, defined in the time and frequency domain, as well as HRV nonlinear measures were extracted during the pre- and post-stimulus sessions, and given as an input to a Support Vector Machine (SVM) classifier implementing a leave-one-subject-out procedure. Results show an accuracy of velocity recognition of 70% for the men, and 84.38% for the women, when both standard and nonlinear HRV measures were taken into account. Conversely, non-significant results were achieved considering standard measures only, or a gender-aspecific classification. We can conclude that caress-like stimuli elicitation significantly affect HRV nonlinear dynamics with a highly specific gender dependency.

A Novel Algorithm for Movement Artifact Removal in ECG Signals Acquired from Wearable Systems Applied to Horses

This study reports on a novel method to detect and reduce the contribution of movement artifact (MA) in electrocardiogram (ECG) recordings gathered from horses in free movement conditions. We propose a model that integrates cardiovascular and movement information to estimate the MA contribution. Specifically, ECG and physical activity are continuously acquired from seven horses through a wearable system. Such a system employs completely integrated textile electrodes to monitor ECG and is also equipped with a triaxial accelerometer for movement monitoring. In the literature, the most used technique to remove movement artifacts, when noise bandwidth overlaps the primary source bandwidth, is the adaptive filter. In this study we propose a new algorithm, hereinafter called Stationary Wavelet Movement Artifact Reduction (SWMAR), where the Stationary Wavelet Transform (SWT) decomposition algorithm is employed to identify and remove movement artifacts from ECG signals in horses. A comparative analysis with the Normalized Least Mean Square Adaptive Filter technique (NLMSAF) is performed as well. Results achieved on seven hours of recordings showed a reduction greater than 40% of MA percentage (between before- and after- the application of the proposed algorithm). Moreover, the comparative analysis with the NLMSAF, applied to the same ECG recordings, showed a greater reduction of MA percentage in favour of SWMAR with a statistical significant difference (p-value < 0.0.5).

Electrodermal activity processing: a convex optimization approach

This paper reports on a novel model based on convex optimization methods for the analysis of the skin conductance (SC) as response of the electrodermal activity (EDA) to affective stimuli. Starting from previous assessed methodological approaches, this new model proposes a decomposition of SC into tonic and phasic components through the solution of a convex optimization problem. Previous knowledge about the physiology of the EDA is accounted for by means of an appropriate choice of constraints and regularizers. In order to test the effectiveness of the new approach, an experimental session in which 9 healthy subjects were stimulated using affective pictures gathered from the IAPS database was designed and carried out. The experimental session included series of negative-valence high-arousal images and series of neutral images, with an inter-stimulus interval of about 2 seconds for both neutral and high arousal pictures. Next, a statistical analysis was performed on a set of features extracted from the phasic driver and the tonic signal estimated by the model. Results showed that the phasic driver extracted from the model was able to strongly distinguish arousal sessions from neutral ones. Conversely, no significant difference was found for the tonic components. This experimental findings are consistent with the literature and confirm that the phasic component is strictly related to changes in the sympathetic activity of the autonomic nervous system. Although preliminary, these results are very encouraging and future work will progress to further validate the model through specific and controlled experiments.

cvxEDA: A Convex Optimization Approach to Electrodermal Activity Processing

GOAL

This paper reports on a novel algorithm for the analysis of electrodermal activity (EDA) using methods of convex optimization. EDA can be considered as one of the most common observation channels of sympathetic nervous system activity, and manifests itself as a change in electrical properties of the skin, such as skin conductance (SC).

METHODS

The proposed model describes SC as the sum of three terms: the phasic component, the tonic component, and an additive white Gaussian noise term incorporating model prediction errors as well as measurement errors and artifacts. This model is physiologically inspired and fully explains EDA through a rigorous methodology based on Bayesian statistics, mathematical convex optimization, and sparsity.

RESULTS

The algorithm was evaluated in three different experimental sessions to test its robustness to noise, its ability to separate and identify stimulus inputs, and its capability of properly describing the activity of the autonomic nervous system in response to strong affective stimulation.

SIGNIFICANCE

Results are very encouraging, showing good performance of the proposed method and suggesting promising future applicability, e.g., in the field of affective computing.

Electrodermal activity analysis during affective haptic elicitation

This paper investigates how the autonomic nervous system dynamics, quantified through the analysis of the electrodermal activity (EDA), is modulated according to affective haptic stimuli. Specifically, a haptic display able to convey caress-like stimuli is presented to 32 healthy subjects (16 female). Each stimulus is changed according to six combinations of three velocities and two forces levels of two motors stretching a strip of fabric. Subjects were also asked to score each stimulus in terms of arousal (high/low activation) and valence (pleasant/unpleasant), in agreement with the circumplex model of affect. EDA was processed using a deconvolutive method, separating tonic and phasic components. A statistical analysis was performed in order to identify significant differences in EDA features among force and velocity levels, as well as in their valence and arousal scores. Results show that the simulated caress induced by the haptic display significantly affects the EDA. In detail, the phasic component seems to be inversely related to the valence score. This finding is new and promising, since it can be used, e.g., as an additional cue for haptics design.

Can a Humanoid Face be Expressive? A Psychophysiological Investigation

Non-verbal signals expressed through body language play a crucial role in multi-modal human communication during social relations. Indeed, in all cultures, facial expressions are the most universal and direct signs to express innate emotional cues. A human face conveys important information in social interactions and helps us to better understand our social partners and establish empathic links. Latest researches show that humanoid and social robots are becoming increasingly similar to humans, both esthetically and expressively. However, their visual expressiveness is a crucial issue that must be improved to make these robots more realistic and intuitively perceivable by humans as not different from them. This study concerns the capability of a humanoid robot to exhibit emotions through facial expressions. More specifically, emotional signs performed by a humanoid robot have been compared with corresponding human facial expressions in terms of recognition rate and response time. The set of stimuli included standardized human expressions taken from an Ekman-based database and the same facial expressions performed by the robot. Furthermore, participants' psychophysiological responses have been explored to investigate whether there could be differences induced by interpreting robot or human emotional stimuli. Preliminary results show a trend to better recognize expressions performed by the robot than 2D photos or 3D models. Moreover, no significant differences in the subjects' psychophysiological state have been found during the discrimination of facial expressions performed by the robot in comparison with the same task performed with 2D photos and 3D models.

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.

Temporal trends of neuro-autonomic complexity during severe episodes of bipolar disorders

Bipolar disorder is a chronic psychiatric condition during which patients experience mood swings among depression, hypomania or mania, mixed state (depression-hypomania) and euthymia, i.e., good affective balance. Nowadays, an objective characterization of the temporal trends of the disease as a response to the pharmacological treatment through physiological signatures, especially during severe episodes, is still missing. In this study we show interesting findings relating neuro-autonomic complexity to severe pathological mood states. More specifically, we studied Sample Entropy (SampEn) measures on Heart Rate Variability series gathered from four bipolar patients recruited within the frame of the European project PSYCHE. Patients were monitored through long term ECG recordings from the first hospital admission until clinical remission, i.e., the euthymic state. We observed that a mood transition from mixed-state to euthymia passing through depression can be characterized by increased SampEn values, i.e. as the patient is going to recover, SampEn increases. These results are in agreement with the current literature reporting on the complexity dynamics of the cardiovascular system and can provide a promising and viable clinical decision support to objectify the diagnosis and improve the management of psychiatric disorders.

Inference of human affective states from psychophysiological measurements extracted under ecologically valid conditions

Compared to standard laboratory protocols, the measurement of psychophysiological signals in real world experiments poses technical and methodological challenges due to external factors that cannot be directly controlled. To address this problem, we propose a hybrid approach based on an immersive and human accessible space called the eXperience Induction Machine (XIM), that incorporates the advantages of a laboratory within a life-like setting. The XIM integrates unobtrusive wearable sensors for the acquisition of psychophysiological signals suitable for ambulatory emotion research. In this paper, we present results from two different studies conducted to validate the XIM as a general-purpose sensing infrastructure for the study of human affective states under ecologically valid conditions. In the first investigation, we recorded and classified signals from subjects exposed to pictorial stimuli corresponding to a range of arousal levels, while they were free to walk and gesticulate. In the second study, we designed an experiment that follows the classical conditioning paradigm, a well-known procedure in the behavioral sciences, with the additional feature that participants were free to move in the physical space, as opposed to similar studies measuring physiological signals in constrained laboratory settings. Our results indicate that, by using our sensing infrastructure, it is indeed possible to infer human event-elicited affective states through measurements of psychophysiological signals under ecological conditions.

Telemonitoring with respect to mood disorders and information and communication technologies: overview and presentation of the PSYCHE project

This paper reviews what we know about prediction in relation to mood disorders from the perspective of clinical, biological, and physiological markers. It then also presents how information and communication technologies have developed in the field of mood disorders, from the first steps, for example, the transition from paper and pencil to more sophisticated methods, to the development of ecological momentary assessment methods and, more recently, wearable systems. These recent developments have paved the way for the use of integrative approaches capable of assessing multiple variables. The PSYCHE project stands for Personalised monitoring SYstems for Care in mental HEalth.