Designing daily-life research combining experience sampling method with parallel data.

BACKGROUND: Ambulatory monitoring is gaining popularity in mental and somatic health care to capture an individual's wellbeing or treatment course in daily-life. Experience sampling method collects subjective time-series data of patients' experiences, behavior, and context. At the same time, digital devices allow for less intrusive collection of more objective time-series data with higher sampling frequencies and for prolonged sampling periods. We refer to these data as parallel data. Combining these two data types holds the promise to revolutionize health care. However, existing ambulatory monitoring guidelines are too specific to each data type, and lack overall directions on how to effectively combine them. METHODS: Literature and expert opinions were integrated to formulate relevant guiding principles. RESULTS: Experience sampling and parallel data must be approached as one holistic time series right from the start, at the study design stage. The fluctuation pattern and volatility of the different variables of interest must be well understood to ensure that these data are compatible. Data have to be collected and operationalized in a manner that the minimal common denominator is able to answer the research question with regard to temporal and disease severity resolution. Furthermore, recommendations are provided for device selection, data management, and analysis. Open science practices are also highlighted throughout. Finally, we provide a practical checklist with the delineated considerations and an open-source example demonstrating how to apply it. CONCLUSIONS: The provided considerations aim to structure and support researchers as they undertake the new challenges presented by this exciting multidisciplinary research field.

Early warning signals in psychopathology: what do they tell?

BACKGROUND: Despite the increasing understanding of factors that might underlie psychiatric disorders, prospectively detecting shifts from a healthy towards a symptomatic state has remained unattainable. A complex systems perspective on psychopathology implies that such symptom shifts may be foreseen by generic indicators of instability, or early warning signals (EWS). EWS include, for instance, increasing variability, covariance, and autocorrelation in momentary affective states-of which the latter was studied. The present study investigated if EWS predict (i) future worsening of symptoms as well as (ii) the type of symptoms that will develop, meaning that the association between EWS and future symptom shifts would be most pronounced for congruent affective states and psychopathological domains (e.g., feeling down and depression). METHODS: A registered general population cohort of adolescents (mean age 18 years, 36% male) provided ten daily ratings of their affective states for 6 consecutive days. The resulting time series were used to compute EWS in feeling down, listless, anxious, not relaxed, insecure, suspicious, and unwell. At baseline and 1-year follow-up, symptom severity was assessed by the Symptom Checklist-90 (SCL-90). We selected four subsamples of participants who reported an increase in one of the following SCL-90 domains: depression (N = 180), anxiety (N = 192), interpersonal sensitivity (N = 184), or somatic complaints (N = 166). RESULTS: Multilevel models showed that EWS in feeling suspicious anticipated increases in interpersonal sensitivity, as hypothesized. EWS were absent for other domains. While the association between EWS and symptom increases was restricted to the interpersonal sensitivity domain, post hoc analyses showed that symptom severity at baseline was related to heightened autocorrelations in congruent affective states for interpersonal sensitivity, depression, and anxiety. This pattern replicated in a second, independent dataset. CONCLUSIONS: The presence of EWS prior to symptom shifts may depend on the dynamics of the psychopathological domain under consideration: for depression, EWS may manifest only several weeks prior to a shift, while for interpersonal sensitivity, EWS may already occur 1 year in advance. Intensive longitudinal designs where EWS and symptoms are assessed in real-time are required in order to determine at what timescale and for what type of domain EWS are most informative of future psychopathology.

Classification of close binary stars using recurrence networks.

Close binary stars are binary stars where the component stars are close enough such that they can exchange mass and/or energy. They are subdivided into semidetached, overcontact, or ellipsoidal binary stars. A challenging problem in the context of close binary stars is their classification into these subclasses based solely on their light curves. Conventionally, this is done by observing subtle features in the light curves like the depths of adjacent minima, which is tedious when dealing with large datasets. In this work, we suggest the use of machine learning algorithms applied to quantifiers derived from recurrence networks to differentiate between classes of close binary stars. We show that overcontact binary stars occupy a region different from semidetached and ellipsoidal binary stars in a plane of characteristic path length and average clustering coefficient, computed from their recurrence networks. We use standard clustering algorithms and report that the clusters formed correspond to the standard classes with a high degree of accuracy.

A hybrid neural ordinary differential equation model of the cardiovascular system.

In the human cardiovascular system (CVS), the interaction between the left and right ventricles of the heart is influenced by the septum and the pericardium. Computational models of the CVS can capture this interaction, but this often involves approximating solutions to complex nonlinear equations numerically. As a result, numerous models have been proposed, where these nonlinear equations are either simplified, or ventricular interaction is ignored. In this work, we propose an alternative approach to modelling ventricular interaction, using a hybrid neural ordinary differential equation (ODE) structure. First, a lumped parameter ODE model of the CVS (including a Newton-Raphson procedure as the numerical solver) is simulated to generate synthetic time-series data. Next, a hybrid neural ODE based on the same model is constructed, where ventricular interaction is instead set to be governed by a neural network. We use a short range of the synthetic data (with various amounts of added measurement noise) to train the hybrid neural ODE model. Symbolic regression is used to convert the neural network into analytic expressions, resulting in a partially learned mechanistic model. This approach was able to recover parsimonious functions with good predictive capabilities and was robust to measurement noise.

A variable heart rate multi-compartmental coupled model of the cardiovascular and respiratory systems.

Current mathematical models of the cardiovascular system that are based on systems of ordinary differential equations are limited in their ability to mimic important features of measured patient data, such as variable heart rates (HR). Such limitations present a significant obstacle in the use of such models for clinical decision-making, as it is the variations in vital signs such as HR and systolic and diastolic blood pressure that are monitored and recorded in typical critical care bedside monitoring systems. In this paper, novel extensions to well-established multi-compartmental models of the cardiovascular and respiratory systems are proposed that permit the simulation of variable HR. Furthermore, a correction to current models is also proposed to stabilize the respiratory behaviour and enable realistic simulation of vital signs over the longer time scales required for clinical management. The results of the extended model developed here show better agreement with measured bio-signals, and these extensions provide an important first step towards estimating model parameters from patient data, using methods such as neural ordinary differential equations. The approach presented is generalizable to many other similar multi-compartmental models of the cardiovascular and respiratory systems.

Predicting recurrence of depression using cardiac complexity in individuals tapering antidepressants.

It is currently unknown whether the complexity and variability of cardiac dynamics predicts future depression and whether within-subject change herein precedes the recurrence of depression. We tested this in an innovative repeated single-subject study in individuals who had a history of depression and were tapering their antidepressants. In 50 individuals, electrocardiogram (ECG) derived Interbeat-interval (IBI) time-series data were collected for 5 min every morning and evening, for 4 months. Usable data were obtained from 14 participants who experienced a transition (i.e., a clinically significant increase in depressive symptoms) and 14 who did not. The mean, standard deviation, Higuchi dimension and multiscale entropy, calculated from IBIs, were examined for time trends. These quantifiers were also averaged over a baseline period and compared between the groups. No consistent trends were observed in any quantifier before increases in depressive symptoms within individuals. The entropy baseline levels significantly differed between the two groups (morning: P value < 0.001, Cohen's d = -2.185; evening: P value < 0.001, Cohen's d = -1.797) and predicted the recurrence of depressive symptoms, in the current sample. Moreover, higher mean IBIs and Higuchi dimensions were observed in individuals who experienced transitions. While we found little evidence to support the existence of within- individual warning signals in IBI time-series data preceding an upcoming depressive transition, our results indicate that individuals who taper antidepressants and showed lower entropy of cardiac dynamics exhibited a higher chance of recurrence of depression. Hence, entropy could be a potential digital phenotype for assessing the risk of recurrence of depression in the short term while tapering antidepressants.

Anticipating manic and depressive transitions in patients with bipolar disorder using early warning signals.

BACKGROUND: In bipolar disorder treatment, accurate episode prediction is paramount but remains difficult. A novel idiographic approach to prediction is to monitor generic early warning signals (EWS), which may manifest in symptom dynamics. EWS could thus form personalized alerts in clinical care. The present study investigated whether EWS can anticipate manic and depressive transitions in individual patients with bipolar disorder. METHODS: Twenty bipolar type I/II patients (with ≥ 2 episodes in the previous year) participated in ecological momentary assessment (EMA), completing five questionnaires a day for four months (Mean = 491 observations per person). Transitions were determined by weekly completed questionnaires on depressive (Quick Inventory for Depressive Symptomatology Self-Report) and manic (Altman Self-Rating Mania Scale) symptoms. EWS (rises in autocorrelation at lag-1 and standard deviation) were calculated in moving windows over 17 affective and symptomatic EMA states. Positive and negative predictive values were calculated to determine clinical utility. RESULTS: Eleven patients reported 1-2 transitions. The presence of EWS increased the probability of impending depressive and manic transitions from 32-36% to 46-48% (autocorrelation) and 29-41% (standard deviation). However, the absence of EWS could not be taken as a sign that no transition would occur in the near future. The momentary states that indicated nearby transitions most accurately (predictive values: 65-100%) were full of ideas, worry, and agitation. Large individual differences in the utility of EWS were found. CONCLUSIONS: EWS show theoretical promise in anticipating manic and depressive transitions in bipolar disorder, but the level of false positives and negatives, as well as the heterogeneity within and between individuals and preprocessing methods currently limit clinical utility.

Overnight affective dynamics and sleep characteristics as predictors of depression and its development in women.

STUDY OBJECTIVES: We examined (1) differences in overnight affective inertia (carry-over of evening affect to the next morning) for positive (PA) and negative affect (NA) between individuals with past, current, and no depression; (2) how sleep duration and quality influence overnight affective inertia in these groups, and (3) whether overnight affective inertia predicts depression development. METHODS: We used data of 579 women from the East-Flanders Prospective Twin Survey. For aim 1 and 2, individuals with past (n = 82), current (n = 26), and without (lifetime) depression (n = 471) at baseline were examined. For aim 3, we examined individuals who did (n = 58) and did not (n = 319) develop a depressive episode at 12-month follow-up. Momentary PA and NA were assessed 10 times a day for 5 days. Sleep was assessed daily with sleep diaries. Affective inertia was operationalized as the influence of evening affect on morning affect. Linear mixed-effect models were used to test the hypotheses. RESULTS: Overnight affective inertia for NA was significantly larger in the current compared to the non-depressed group, and daytime NA inertia was larger in the past compared to the non-depressed group. Overnight NA inertia was differently associated with shorter sleep duration in both depression groups and with lower sleep quality in the current compared to the non-depressed group. Overnight affective inertia did not predict depression development at 12-month follow-up. CONCLUSIONS: Current findings demonstrate the importance of studying complex affect dynamics such as overnight affective inertia in relation to depression and sleep characteristics. Replication of these findings, preferably with longer time-series, is needed.

Efficacy of early warning signals and spectral periodicity for predicting transitions in bipolar patients: An actigraphy study.

Early-warning signals (EWS) have been successfully employed to predict transitions in research fields such as biology, ecology, and psychiatry. The predictive properties of EWS might aid in foreseeing transitions in mood episodes (i.e. recurrent episodes of mania and depression) in bipolar disorder (BD) patients. We analyzed actigraphy data assessed during normal daily life to investigate the feasibility of using EWS to predict mood transitions in bipolar patients. Actigraphy data of 15 patients diagnosed with BD Type I collected continuously for 180 days were used. Our final sample included eight patients that experienced a mood episode, three manic episodes and five depressed episodes. Actigraphy data derived generic EWS (variance and kurtosis) and context-driven EWS (autocorrelation at lag-720) were used to determine if these were associated to upcoming bipolar episodes. Spectral analysis was used to predict changes in the periodicity of the sleep/wake cycle. The study procedures were pre-registered. Results indicated that in seven out of eight patients at least one of the EWS did show a significant change-up till four weeks before episode onset. For the generic EWS the direction of change was always in the expected direction, whereas for the context-driven EWS the observed effect was often in the direction opposite of what was expected. The actigraphy data derived EWS and spectral analysis showed promise for the prediction of upcoming transitions in mood episodes in bipolar patients. Further studies into false positive rates are suggested to improve effectiveness for EWS to identify upcoming bipolar episode onsets.