The Autonomic Craving Signature: physiological signals as a daily-life biomarker of craving

The Autonomic Craving Signature: physiological signals as a daily-life biomarker of craving | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Article The Autonomic Craving Signature: physiological signals as a daily-life biomarker of craving Emmanuelle Baillet, Fuschia Serre, Hakim Si-Mohammed, Cassandre Romao, and 5 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-4432651/v1 This work is licensed under a CC BY 4.0 License Status: Posted Version 1 posted You are reading this latest preprint version Abstract Craving variations have been shown to be prospectively associated to relapse and are a target for treatment in substance use disorders (SUD). We show that a set of physiological features captured in daily life can distinguish within-day craving from no-craving states, among SUD participants in a 14-day multimodal study. Craving was assessed Signal-contingent and event-contingent. Continuous recording of physiological signals was conducted daily. After labeling samples as “craving” and “no-craving” using a double-check verification, features of each signal were extracted, normalized and standardized and a principal component analysis was applied. Using Support Vector Machines and Linear Discriminant Analysis classifiers, physiological signals allowed to discriminate craving from no-craving episodes with a high cross-validation accuracy (73.6% and 72.2% respectively). This finding is an important breakthrough that paves the way to a biomarker of craving in daily life. Health sciences/Biomarkers/Diagnostic markers Health sciences/Diseases/Psychiatric disorders/Addiction Health sciences/Biomarkers/Prognostic markers Craving Substance Use Disorder Biomarker Ecological Momentary Assessment Wearable Sensor Machine Learning Classification Heart Rate Electrodermal Activity Accelerometer Skin Temperature Figures Figure 1 Figure 2 INTRODUCTION Craving, an intense desire to use, is a key contributor to substance use disorders (SUD) 1–3 . Added as a SUD diagnostic criterion in the Diagnostic and Statistical Manual of Mental Disorders , fifth edition (DSM-5) 4,5 , it has been demonstrated to be the most discriminant and specific criterion in Item Responses Theory analysis 6 and the most central criterion in networks of DSM-5 SUD diagnostic criteria 7 . Craving may also vary in intensity and frequency within hours or days in the same individual, under the influence of internal and environmental factors 8 . Versatile emotional states have been a challenge for research but are now particularly well captured by Ecological Momentary Assessment (EMA) 9 that allows a real-time measurement of individual momentary states in the natural environment through repeated assessments overtime. Use of EMA in craving research fueled a generic clinical model of addiction, the cues – craving – use model 10 in which cues – conditional stimuli associated with past experiences of use 11 – were prospectively associated to craving, which in turn was prospectively associated with use in the following hours 12 . In a meta-analysis pooling 51,788 substance-using participants from 237 studies, craving was reported to prospectively predict substance use and relapse, suggesting it could play a causative role 3 . Recently, a study based on a functional magnetic resonance imagery (fMRI) cue-reactivity protocol reported a neurobiological craving signature (NCS) common to food and substances, that predicted craving and distinguished substance users from non-users 13 . Another study identified a neurobiological signature of craving based on functional connectivity from resting-state high-density electroencephalography (EEG) 14 . This represents a potential major step forward in the discovery of biomarkers of craving and provides a window into understanding its neurobiological mechanisms. However, fMRI or EEG are not yet usable in daily-life, limiting the transitioning of NCS to everyday clinical practice. Yet, neurobiological changes specific to craving could be captured by changes in the autonomic nervous system (ANS) 15,16 . Heart rate (HR), skin temperature (TEMP), breathing rates or electrodermal activity (EDA) are commonly used as indicators of psychological arousal and can be easily captured in daily-life by ambulatory monitoring devices. Several studies have shown a concomitance between signs of physiological arousal such as a significant increase in HR, EDA, TEMP or accelerometer (ACC) and the appearance of craving 17–22 . The magnitude of craving was correlated with these physiological activity changes suggesting that the latter could actually represent underlying mechanisms of craving or, their physiological expression 16,23,24 , and thereby be a potential objective marker implementable in the natural context of participants. Some studies have examined the physiological correlates of craving in human lab settings 18,25,26 or in daily-life 27–32 with limitations in model development, accuracy and reproducibility 33 . Firstly, the use of a single physiological parameter (i.e., HR 27,29 ) may not be specific and reliable enough to identify such a complex phenomenon as craving is likely to be. A pattern, or combination, of physiological measures (i.e., combination of HR, EDA and TEMP) may be more appropriate. Secondly, some measures of craving may not be craving specific, which means there was no guarantee that the phenomenon observed and analyzed was really craving 34 and not the association of different phenomena, such as stress co-occurring with craving 27–29 , especially because stress and craving share some physiological similarities 15 . Thirdly, to identify physiological correlates of craving, the time window explored to date ranged from 40 to 180 minutes 27–29 , while craving duration could be shorter (5 to 30 minutes) 35–37 . Finally, craving needs to be explored with its intensity fluctuations over time, requiring EMA and adapted length of assessment 38 . To overcome these limitations, our aim was to identify a transdiagnostic and reliable pattern of physiological signals of ANS (heart rate, electrodermal activity, skin temperature and accelerometry) associated with craving captured in daily-life using a multimodal method (EMA and wearable sensors) among participants with SUDs. We expected that craving vs. no-craving episodes would be distinguished based on physiological and EMA data. METHODS Participants Participants were included from the ADDICTAQUI Cohort 39 or from not-in-treatment population. The ADDICTAQUI cohort includes individuals seeking treatment for a DSM-5 use disorder in outpatient addiction treatment clinics in Aquitaine, France. Participants from the not-in-treatment population were recruited through social networks. To be eligible for the study, participants in treatment and not-in-treatment had to be adults (> 18 years), with at least one DSM-5 SUD diagnosis for tobacco, alcohol or cannabis and reported craving in the past 30 days. Exclusion criteria were pregnancy, cardiovascular disease, non-dominant upper extremity limitation, or psychiatric conditions incompatible with understanding protocol. The study was approved by the French Regulation and ethical committee (CPP EST III, approval registration 2021-A01835-36) and all participants provided written informed consent. Experimental Design The study was a 14-day prospective daily-life data collection with EMA and wearable sensors. During the enrollment visit, clinical assessment was provided with the Addiction Severity Index (ASI) 40 and Mini International Neuropsychiatric Interview (MINI) 41 to collect demographic and historical data about SUD and other psychiatric disorders. A brief training was performed with devices. Participants were asked to wear the sensor on their non-dominant wrist all day for the duration of the study, and to remove it only to sleep. EMA surveys, presented 4 times/day at random, approximately every 3h30, were adapted to the primary substance of addiction according to ASI and in case of poly-addiction, the choice was made with the participant. Main outcomes measures Physiological parameters: The E4® sensor (Empatica Inc., Cambridge MA, USA) is a research device with good signal quality compared to a reference device 42–44 . It is a non-invasive wearable which allows the continuous acquisition of physiological data in real-time: blood volume pulse (BVP) (in beats/min (bpm)), electrodermal activity (EDA) (in microSiemens (µS)), skin temperature (TEMP) (in Celsius degree (°C)) and 3-axis accelerometers (ACC) (in gravity (g)). Craving measurements: Craving was assessed signal-contingently on a 7-point numerical scale in EMA from “1” (not at all) to “7” (extreme craving) in response to the question "Since the last questionnaire, have you felt a desire to use [primary substance] ?", and event-contingently, by button pressing (“TAG”) on the sensor when they felt craving for their primary substance, regardless of intensity. Analysis strategy To label classes of craving and no-craving, a double-check selection procedure was used: “No-craving” class was defined when no craving was reported both on the sensor (no-TAG) and on the corresponding EMA survey (intensity = 1); and “Craving” class was defined when craving was reported both on the sensor (TAG) and on the corresponding EMA (intensity > 4 corresponding to the average intensity of the craving scale). Periods when the participant mentioned stress (> 1 on the EMA) were excluded. Then, based on the literature exploring craving duration 35–37,45−47 , we used epochs of 20 minutes (centered around the TAG for “craving” or randomly for “no-craving”), in which several features were extracted for each physiological signal recorded to reflect their real values (Supplementary Table 1). In order to be able to run statistical analyses without risking over-fitting of the model on the data, we included only the participants for whom at least 20 samples of each class could be extracted. If the classes were unbalanced in time of occurrence, we selected the samples that were the closer from the samples selected in the other class. Then we used Principal Component Analysis (PCA) 48 to reduce data dimensionality by identifying Principal Components (PCs) that capture the maximum variance in physiological data, regardless of class 49 . Student t-tests were performed on each feature (“craving” vs. “no-craving”) in order to identify those that contributed the most to the first PCs. Finally, Support vector machines (SVM) 50 , minimum distance to mean (MDM) 51 , linear discriminant analysis (LDA) 52 classifiers were used. Their performance was assessed through a k-fold cross-validation (k = 10) 53 . For each fold, 80% of the samples were randomly assigned to the training test while the other 20% were used as a testing set. Our models were evaluated with standard metrics in which craving samples were the positive class: specificity, recall, precision, accuracy and area under the curve (AUC) of the receiver operating characteristic (ROC) curve 54 . The theoretical chance levels were estimated at 65% 55 . Finally, to estimate statistical significance of the cross-validation classifier accuracy for an α risk of 5%, we performed permutation tests with 100 randomized cross-validation iterations. The proposed model is detailed in the Supplement and was implemented using Python (3.12.0) and R Studio (4.1.0) software. The code used is available on request to the first and corresponding authors. RESULTS From April 2022 to July 2023, 43 (95.6%) of 45 participants included had worn the sensor and were considered as the study sample for analysis. Participants of the study sample were aged 32.9 years (standard deviation [SD] = 9.4), and 55% (n = 24) were women and 45% were men (n = 19). The most frequent primary substance was tobacco (46.5%, n = 20), followed by cannabis (32.5%, n = 14) and alcohol (20.9%, n = 9). On average participants had 7.3 DSM-5 SUD diagnostic criteria (SD = 2.1) and a mean of 5.4 on the ASI severity score for the primary substance (SD = 1.3) (Table 1 ). Table 1 Characteristics of the sample and description of daily-life variables. Samples characteristics Full sample (N = 43) Sample of the ACS (N = 5) % (N) M (SD) Min Max % (N) M (SD) Min Max Gender (woman) 55.8 ( 24 ) 60.0 ( 3 ) Employed 76.2 ( 32 ) 80.0 ( 4 ) Age 32.9 (9.4) 22 60 39.2 (10.1) 30 52 Education (years) 14.5 (1.9) 9 17 12.6 (2.6) 9 16 Primary substance Tobacco 46.5 ( 20 ) 40.0 ( 2 ) Cannabis 32.5 ( 14 ) 40.0 ( 2 ) Alcohol 20.9 ( 9 ) 20.0 ( 1 ) ASI severity score for primary substance 5.4 (1.3) 0 7 5.4 (1.8) 0 7 MINI SUD severity 7.3 (2.1) 0 11 7.4 (2.6) 0 11 MINI current diagnosis Mood disorder 1 20.9 ( 9 ) 25.0 ( 1 ) Anxiety disorder 1 20.9 ( 9 ) 40.0 ( 2 ) Signal-contingent EMA 85.8%, observations = 2 017 2 91.4%, observations = 256 2 Use of primary substance 51.4 (1037) 23.8 ( 61 ) Craving intensity ( 1 – 7 ) 4.1 (2.1) 0 7 3.5 (2.4) 0 7 Craving episodes (intensity > 1) 77.2 (1556) 59.5 (152) Event-contingent sensor (> 0) 1 654 38.5 (33.4) 2 146 295 59.0 (25.8) 29 93 1 Mood disorders include major depressive episode, dysthymia and hypomania; anxiety disorders include social phobia, generalized anxiety disorder and panic disorder. 2 Frequencies, percentages and means are based on the total number of valid electronic surveys during the assessment period. ASI = Addiction Severity Index; MINI = Mini International Neuropsychiatric Disorder; EMA = Ecological Momentary Assessment. [Table 1 ] EMA completion rate was 85.8% (n = 2,017 assessments completed) and the sensor was used on average 8.9 hours/day (SD = 2.9). In total, 5,512 hours of physiological data were captured over the course of EMA-days. The total number of craving TAG was 1,654 TAGs with an average of 38.5 TAG (SD = 33.4) per participant and the total number of EMA craving reports was 1,556 with an average craving intensity of 4.1/7 (SD = 2.08). For sample labeling, 1883 samples were labeled as “no-craving”, 1109 samples were labeled as “craving”, and 4296 samples were excluded because they did not meet our double-check selection criteria (Supplementary Fig. 1). From these selected samples, 190 features were extracted, normalized and standardized after which 98 were again rejected for missing values, high correlation or low standard deviation (Supplementary Table 2). Consequently, each “craving” and “no-craving” samples was composed of a vector of 92 standardized and normalized features. To maintain the balance between classes “craving” and “no-craving” and avoid overfitting, we only retained participants with at least 20 samples/classes. Thus, 192 samples of “craving” and 192 of “no-craving” were analyzed by PCA and these were issued from 5 participants. As a result, 31 PCs, which explained 98.8% of variance data, were kept for the classification model (Supplementary Fig. 2). The first component (PC1) explained the majority of the total variance (73.8%) and was mainly composed by EDA features while the second component (PC2) explained 7.9% of the total variance. Five features of PC1 with greater weights (59.43% of PC1) were significantly different between our two classes: “EDA Tonic Min” (p < .0001), “EDA Tonic Mean” (p < .0001), “EDA Tonic Max” (p < .0001), “EDA Wavelet 0 RMS” (p < .0001), and “EDA Skew” (p = .020) (Table 2 ). In PC2, among five features with greater weights (42.27% of PC2), “HRV SDNNI5” (p = .001), “HRV SDANN5” (p = .017), “HRV SDANN2” (p < .0001) and “HRV SDANN1” (p < .0001) were significantly different between our two classes, but not “HRV DFA α 2” (p = 0.128) (Table 2 ). Table 2 Description of the main features of the first PCs applied to craving/no-craving samples. PC Features M (SD)[Craving] M (SD)[No-craving] t-ratio dof Prob.>ItI PC1 EDA Tonic Min 2.59 (9.52) 11.65 (21.75) -5.28 261.67 < .0001 PC1 EDA Tonic Mean 0.51 (0.89) 1.20 (1.36) -5.86 329.5 < .0001 PC1 EDA Tonic Max 0.74 (1.01) 1.46 (1.38) -5.82 349.8 < .0001 PC1 EDA Wavelet 0 RMS 0.20 (0.70) 0.82 (1.16) -6.27 314.6 < .0001 PC1 EDA Skew 1.28 (3.05) 0.59 (2.78) 2.33 378.9 0.020 PC2 HRV SDNNI5 1.82 (4.89) 0.64 (1.38) 3.22 221.3 0.001 PC2 HRV SDNN5 0.98 (4.64) 0.16 (0.85) 2.41 203.8 0.017 PC2 HRV SDANN2 1.22 (2.62) 0.29 (1.65) 4.16 322.6 < .0001 PC2 HRV SDANN1 1.08 (1.92) 0.22 (1.24) 5.22 327.6 < .0001 PC2 HRV DFA α 2 -0.20 (2.01) 0.04 (0.83) -1.53 254.5 0.128 Unit of all EDA features was microSiemens (µS); second (s) for HRV SDANNI5, HRV SDANN5, HRV SDANN2 and HRV SDANN1 features; HRV DFAα2 had no unit. Bold values indicate statistically significant P values. [Table 2 ] Then, the 31 PCs from the training set were used to perform SVM, MDM and LDA classifications, and performance in discriminating craving from no-craving was estimated on the test set. For the MDM, the sensitivity was 66.2% and the specificity was 51.0%; the AUC of the model’s ROC was 0.58 (Fig. 1 ) and the model’s accuracy and precision were 58.3% and 55.5% respectively. For the SVM, the sensitivity was 74.5% and the specificity was 72.8%; the AUC of the model’s ROC was 0.83 (Fig. 1 ) and the model’s accuracy and precision were 73.6% and 73.1% respectively. For the LDA, the sensitivity was 76.8% and the specificity was 67.7%; the AUC of the model’s ROC was 0.83 (Fig. 1 ) and the model’s accuracy and precision were 72.2% and 69.8% respectively. For all measures and classifiers, none of the permutation samples performed (general accuracy score: 0.55) as well as the observed results (MDM: p = 0.02; SVM: p = 0.01; LDA: p = 0.01) at the 95% confidence level (Fig. 2 ). [FIGURE 1 ] [FIGURE 2 ] DISCUSSION The aim of the present work was to identify craving with physiological signals of ANS in daily-life among participants with SUDs. Using SVM and LDA classifiers, physiological signals allowed us to discriminate episodes of craving vs. no-craving with a high cross-validation accuracy (73.6% and 72.2% respectively). Taking into account the limitations of previous research, to our knowledge this is the first study to develop a pipeline to detect craving from physiological signals across substances in daily-life by multimodal methods. The application of PCA provided significant advantages for pattern recognition of craving 56 and allowed us to calculate PCs, a linear combination of physiological features, through which classifiers were able to distinguish between craving and no-craving episodes. These features were mainly extracted from EDA and HR signals. The “EDA Skew” feature, derived from the phasic component of EDA, corresponds to skewness of EDA distribution and more positive values indicate a peak of activity 57 , which was found in craving compared to no-craving episodes. This is potentially important, especially for the interaction with other features that were extracted from the tonic component, generally used to reflect the overall arousal level of an individual in the absence of specific stimuli, while phasic component refers to faster changes, e.g. caused by sudden emotional arousal or external stimuli 58 . In line with our results, craving might induce an activity peak reflected by the phasic component, but there seems to be an association with the tonic component. This result provides both information on the general state of arousal and on a specific response to craving. Our hypothesis is that craving could be reflected in basal levels of autonomic arousal and even in the absence of acute episodes of perceived/reported craving, individuals with SUD may have a higher basic autonomic arousal (tonic component of EDA), which is increased in case of an acute episode of reported craving (phasic component of EDA). This phenomenon is particularly common in anxiety disorder characterized by a high level of HR and EDA at rest 59 . Craving overall could be the expression of an altered activation of the autonomic system at rest, i.e., a basal elevation of some parameters such as HR and EDA in the case of no-craving, and a high arousal in case of craving. This basal level, which is already high in individuals with SUD, increases during craving episodes, making it possible to distinguish it from no-craving episodes. This set of features forms a physiological pattern, which we label the autonomous craving signature (ACS) , with good specificity 34 and consistency 60 . As this was a study in daily-life, the individuals were not in the same places and did not encounter these 2 conditions at the same times (time within-day or between-days) which reinforces the consistency of ACS under any condition 60 . Moreover, the double-check selection of episodes with the EMA signal contingent (short-term retrospective measure) and the participant-initiated event contingent (immediate measure) enabled us to guarantee that craving episodes reported by the individual were really craving (and not an unintentional press on the sensor, for example) and to approach the validity of physiologic responses of these episodes in everyday life. Also, as each individual was their own control, our model considers both within- and between-individual variability. An important point was that we detected a difference between craving and no-craving episodes with ACS, i.e., features that composed the ACS had a significant difference between our 2 conditions (craving vs. no-craving). This model allowed us to take into account the natural variation of the autonomic system and excluded the hypothesis that the ACS could only result from these variations. In fact, the ACS pattern makes it possible to overcome Laws of Autonomic Constraint 60 . While we do not rule-out the possibility that some of the features included in the ACS pattern were linked to other phenomena we cannot control, our hypothesis is that the combination of these markers makes the set specific and provides information on the physiological expression of craving, paving the way to better understanding the underlying mechanisms. Important limitations have to be acknowledged. One limitation was our choice of epoch length that was of 20 minutes, centered on the participant self-initiated TAG. Despite our significant results, the temporality of craving and its duration remains to be explored. The main difficulty was balancing data: some participants had either not enough episodes of no-craving or of craving. Consequently, much data had to be discarded and finally the ACS was identified from 5 participants out of 43 and requires to be replicated. While this may not seem like much, it is encouraging, since we have reliable results. Furthermore, when compared to the full sample, the 5 participants were not different based on their descriptive characteristics, including craving reports. In our models, we chose low thresholds (correlation, standard deviation) and our results indicate a classification of around 75%, suggesting that the detection of craving was both possible as shown by our findings and could be improved by setting higher restrictive thresholds. The next step would be to test the generalization of this model on a new dataset and to explore, as suggested in previous research, how the ACS may be associate with the NCS 13 . CONCLUSION The identification of craving from physiological data could be an important step in the development of prognostic biomarkers of relapse in SUD. In addition to facilitate the long-term monitoring of craving without the need for the patient to engage in an effortful awareness that is difficult to sustain overtime, the use of ACS may also be helpful for individuals with low insight of their substance-seeking behaviors or with difficulty in reporting craving, which is associated to poor treatment outcome 2 . The ACS paves the way to the development, in daily-life, of protocols of craving regulation that could build a sustainable warning system that would be activated when this pattern is present to allow a relay towards care if needed or towards a just-in-time intervention 61 such as biofeedback or mindfulness 62 . The regulation of craving is key to better prevention of relapse and more successful treatment outcomes 63 . Declarations DATA SHARING STATEMENT Original data are available on request to Marc Auriacombe. CONFLICT OF INTEREST: We report no conflicts of interest related to this work. The researchers confirm their independence from funders and sponsors. AUTHOR CONTRIBUTION: MA was the overall principal investigator and study supervisor, obtained funding and access to participants. MA, EB, FS developed the study design and methods. EB set up the study, included participants, conducted analyses, interpreted the data, and drafted the manuscript. TM contributed to analysis and data management; CM and AB to inclusion of participants. HSM and CJK helped in the conceptualization of analysis, in architecture of classification model and in the interpretation of data. All authors undertook the critical revision of the manuscript for important intellectual content and all authors significantly contributed to the manuscript and approved the final version. ACKNOWLEDGEMENTS This study received financial support from the French government in the framework of the University of Bordeaux's IdEx "Investments for the Future" program / GPR BRAIN_2030 and by French Cancer and Public Health Institutes (INCA-IRESP-2020-169) PhD doctoral grant to Emmanuelle Baillet. Funding for the ADDICTAQUI cohort data collection was provided by Research Grant PHRC (2006–2014) from the French Ministry of Health, French Government Addiction Agency MILDT/MILDECA grant 2010 and 2016 to Marc Auriacombe. Funding for the EMA study was provided by Research Grant AAP-19-ADDICTIONS-16 from the public health research institute (IReSP) and Aviesan Alliance as part of the call for research projects to combat addiction to psychoactive substances to Marc Auriacombe and Fuschia Serre. Office and staff support was provided by CH Charles Perrens hospital. The authors express theirs thanks to all participants for their contribution and are grateful to all the interviewers of the ADDICTAQUI team, especially to Chloé VACHER, Arthur BRUNEAU, Axel ALLACHE and Cassandre ROMAO and to the medical team that managed patients during the time of the study, Saman Sarram, Jacques Dubernet, Bérengère Gelot. Marc Auriacombe and Fuschia Serre had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis. References Auriacombe, M., Serre, F., Denis, C. & Fatseas, M. in The Routledge Handbook of the Philosophy and Science of Addiction (eds H Pickard & S.H. Ahmed) Ch. 10, 132–144 (2018). Cavicchioli, M., Vassena, G., Movalli, M. & Maffei, C. Is craving a risk factor for substance use among treatment-seeking individuals with alcohol and other drugs use disorders? A meta-analytic review. Drug Alcohol Depend 212, 108002, doi: 10.1016/j.drugalcdep.2020.108002 (2020). Vafaie, N. & Kober, H. Association of Drug Cues and Craving With Drug Use and Relapse: A Systematic Review and Meta-analysis. JAMA Psychiatry, doi: 10.1001/jamapsychiatry.2022.1240 (2022). American Psychiatric Association. Diagnostic and Statistical Manual of Mental Disorders (DSM-5) . (American Psychiatric Pub, 2013). Hasin, D. S. et al. DSM-5 Criteria for Substance Use Disorders: Recommendations and Rationale. American Journal of Psychiatry 170, 834–851, doi: 10.1176/appi.ajp.2013.12060782 (2013). Kervran, C. et al. Item Response Theory analyses of DSM-5 substance use disorder criteria in French outpatient addiction clinic participants. How much is craving special? Drug and Alcohol Dependence 212, 108036, doi: 10.1016/j.drugalcdep.2020.108036 (2020). Gauld, C. et al. The Centrality of Craving in Network Analysis of Five Substance Use Disorders. Drug and Alcohol Dependence, 109828, doi: https://doi.org/10.1016/j.drugalcdep.2023.109828 (2023). Sayette, M. A. et al. The measurement of drug craving. Addiction 95 Suppl 2, S189–210 (2000). Stone, A. A. & Shiffman, S. Ecological Momentary Assessment (Ema) in Behavioral Medicine. Annals of Behavioral Medicine 16, 199–202, doi: 10.1093/abm/16.3.199 (1994). Fatseas, M. et al. Craving and substance use among patients with alcohol, tobacco, cannabis or heroin addiction: a comparison of substance- and person-specific cues: Cues, craving and substance use. Addiction 110, 1035–1042, doi: 10.1111/add.12882 (2015). Carter, B. L. & Tiffany, S. T. Meta-analysis of cue-reactivity in addiction research. Addiction (Abingdon, England) 94, 327–340 (1999). Serre, F., Fatseas, M., Denis, C., Swendsen, J. & Auriacombe, M. Predictors of craving and substance use among patients with alcohol, tobacco, cannabis or opiate addictions: Commonalities and specificities across substances. Addictive Behaviors 83, 123–129, doi: 10.1016/j.addbeh.2018.01.041 (2018). Koban, L., Wager, T. D. & Kober, H. A neuromarker for drug and food craving distinguishes drug users from non-users. Nature Neuroscience 26, 316–325, doi: 10.1038/s41593-022-01228-w (2023). Tian, W. et al. An electroencephalographic signature predicts craving for methamphetamine. Cell Reports Medicine 5 (2024). Sinha, R. et al. Enhanced negative emotion and alcohol craving, and altered physiological responses following stress and cue exposure in alcohol dependent individuals. Neuropsychopharmacology: Official Publication of the American College of Neuropsychopharmacology 34, 1198–1208, doi: 10.1038/npp.2008.78 (2009). Wang, W. et al. Cue-elicited craving, thalamic activity, and physiological arousal in adult non-dependent drinkers. J Psychiatr Res 116, 74–82, doi: 10.1016/j.jpsychires.2019.06.005 (2019). Carter, B. L. et al. Generalized craving, self-report of arousal, and cue reactivity after brief abstinence. Nicotine Tob Res 11, 823–826, doi: 10.1093/ntr/ntp068 (2009). Zhao, M. et al. Cue-induced craving and physiological reactions in recently and long-abstinent heroin-dependent patients. Addict Behav 37, 393–398, doi: 10.1016/j.addbeh.2011.11.030 (2012). Li, P. et al. Incubation of alcohol craving during abstinence in patients with alcohol dependence. Addict Biol 20, 513–522, doi: 10.1111/adb.12140 (2015). Carter, B. L. & Tiffany, S. T. Meta-analysis of cue-reactivity in addiction research. Addiction 94, 327–340 (1999). Sinha, R. et al. Enhanced negative emotion and alcohol craving, and altered physiological responses following stress and cue exposure in alcohol dependent individuals. Neuropsychopharmacology 34, 1198–1208, doi: 10.1038/npp.2008.78 (2009). Trotzke, P., Starcke, K., Pedersen, A. & Brand, M. Cue-induced craving in pathological buying: empirical evidence and clinical implications. Psychosom Med 76, 694–700, doi: 10.1097/psy.0000000000000126 (2014). Murphy, A. et al. Time-dependent neuronal changes associated with craving in opioid dependence: an fMRI study. Addiction Biology 23, 1168–1178, doi: https://doi.org/10.1111/adb.12554 (2018). Volkow, N. D. et al. Cocaine cues and dopamine in dorsal striatum: mechanism of craving in cocaine addiction. J Neurosci 26, 6583–6588, doi: 10.1523/jneurosci.1544-06.2006 (2006). Gullapalli, B. T. et al. On-body Sensing of Cocaine Craving, Euphoria and Drug-Seeking Behavior Using Cardiac and Respiratory Signals. Proc. ACM Interact. Mob. Wearable Ubiquitous Technol. 3, Article 46, doi: 10.1145/3328917 (2019). Fatseas, M. et al. Cue-Induced Reactivity, Cortisol Response and Substance Use Outcome in Treated Heroin Dependent Individuals. Biological Psychiatry 70, 720–727, doi: 10.1016/j.biopsych.2011.05.015 (2011). van Lier, H. G. et al. An ideographic study into physiology, alcohol craving and lapses during one hundred days of daily life monitoring. Addict Behav Rep 16, 100443, doi: 10.1016/j.abrep.2022.100443 (2022). Carreiro, S. et al. Wearable sensor-based detection of stress and craving in patients during treatment for substance use disorder: A mixed methods pilot study. Drug and Alcohol Dependence 209, 107929, doi: 10.1016/j.drugalcdep.2020.107929 (2020). Kennedy, A. P. et al. Continuous in-the-field measurement of heart rate: Correlates of drug use, craving, stress, and mood in polydrug users. Drug and Alcohol Dependence 151, 159–166, doi: 10.1016/j.drugalcdep.2015.03.024 (2015). Chatterjee, S. et al. mCrave: Continuous Estimation of Craving During Smoking Cessation. Proc ACM Int Conf Ubiquitous Comput 2016, 863–874, doi: 10.1145/2971648.2971672 (2016). Boyer, E. W. et al. Preliminary efforts directed toward the detection of craving of illicit substances: the iHeal project. J Med Toxicol 8, 5–9, doi: 10.1007/s13181-011-0200-4 (2012). Shi, R. et al. in 2015 IEEE 39th Annual Computer Software and Applications Conference. 282–287. Tackett, J. L., Brandes, C. M., King, K. M. & Markon, K. E. Psychology’s replication crisis and clinical psychological science. Annual Review of Clinical Psychology 15, 579–604, doi: 10.1146/annurev-clinpsy-050718-095710 (2019). Quigley, K. S. & Barrett, L. F. Is there consistency and specificity of autonomic changes during emotional episodes? Guidance from the Conceptual Act Theory and psychophysiology. Biol Psychol 98, 82–94, doi: 10.1016/j.biopsycho.2013.12.013 (2014). Shiffman, S. et al. Efficacy of acute administration of nicotine gum in relief of cue-provoked cigarette craving. Psychopharmacology (Berl) 166, 343–350, doi: 10.1007/s00213-002-1338-1 (2003). Niaura, R. et al. Comparative efficacy of rapid-release nicotine gum versus nicotine polacrilex gum in relieving smoking cue-provoked craving. Addiction 100, 1720–1730, doi: 10.1111/j.1360-0443.2005.01218.x (2005). Heishman, S. J., Lee, D. C., Taylor, R. C. & Singleton, E. G. Prolonged duration of craving, mood, and autonomic responses elicited by cues and imagery in smokers: Effects of tobacco deprivation and sex. Exp Clin Psychopharmacol 18, 245–256, doi: 10.1037/a0019401 (2010). De Calheiros Velozo, J. et al. Designing daily-life research combining experience sampling method with parallel data. Psychological Medicine, 1–10, doi: 10.1017/S0033291722002367 (2022). Auriacombe, M. ADDICTAQUI - Aquitaine Addiction Cohort: Trajectories of people with addiction (substances or behaviour) in contact with health-care system. Medical, neurobiological, sociological and psychological characteristics. Prospective multicentric, multidisciplinary study < https://epidemiologie-france.aviesan.fr/en/epidemiology/records/cohorte-addiction-aquitaine-trajectoires-de-personnes-presentant-une-addiction-aux-substances-ou-une-addiction-comportementale-en-contact-avec-le-dispositif-de-soins.-caracteristiques-medicales-neurobiologiques-sociologiques-et-psychologiques.-etude#tab_1> (2017). Denis, C. et al. Usefulness and validity of the modified Addiction Severity Index: A focus on alcohol, drugs, tobacco, and gambling. Substance Abuse 37, 168–175, doi: 10.1080/08897077.2015.1036334 (2016). Sheehan, D. V. et al. The Mini-International Neuropsychiatric Interview (M.I.N.I.): the development and validation of a structured diagnostic psychiatric interview for DSM-IV and ICD-10. The Journal of Clinical Psychiatry 59 Suppl 20, 22–33;quiz 34–57 (1998). Schuurmans, A. A. T. et al. Validity of the Empatica E4 Wristband to Measure Heart Rate Variability (HRV) Parameters: a Comparison to Electrocardiography (ECG). Journal of Medical Systems 44, 190, doi: 10.1007/s10916-020-01648-w (2020). Enewoldsen, N. Analysis of the quality of electrodermal activity and heart rate data recorded in daily life over a period of one week with an E4 wristband , University of Twente, (2016). van Lier, H. G. et al. A standardized validity assessment protocol for physiological signals from wearable technology: Methodological underpinnings and an application to the E4 biosensor. Behavior Research Methods 52, 607–629, doi: 10.3758/s13428-019-01263-9 (2020). Heishman, S. J., Saha, S. & Singleton, E. G. Imagery-induced tobacco craving: duration and lack of assessment reactivity bias. Psychol Addict Behav 18, 284–288, doi: 10.1037/0893-164x.18.3.284 (2004). Zhong, F., Wu, L. Z. & Han, J. S. Suppression of cue-induced heroin craving and cue-reactivity by single-trial transcutaneous electrical nerve stimulation at 2 Hz. Addict Biol 11, 184–189, doi: 10.1111/j.1369-1600.2006.00020.x (2006). Heishman, S. J., Singleton, E. G. & Moolchan, E. T. Tobacco Craving Questionnaire: Reliability and validity of a new multifactorial instrument. Nicotine & Tobacco Research 5, 645–654, doi: 10.1080/1462220031000158681 (2003). Abdi, H. & Williams, L. J. Principal component analysis. WIREs Computational Statistics 2, 433–459, doi: https://doi.org/10.1002/wics.101 (2010). Migenda, N., Möller, R. & Schenck, W. Adaptive dimensionality reduction for neural network-based online principal component analysis. PLOS ONE 16, e0248896, doi: 10.1371/journal.pone.0248896 (2021). Salcedo-Sanz, S., Rojo-Álvarez, J. L., Martínez-Ramón, M. & Camps-Valls, G. Support vector machines in engineering: an overview. WIREs Data Mining and Knowledge Discovery 4, 234–267, doi: https://doi.org/10.1002/widm.1125 (2014). Vibhute, A. D. et al. in Proceedings of the Second International Conference on Computer and Communication Technologies: IC3T 2015, Volume 2. 413–427 (Springer). Graf, R., Zeldovich, M. & Friedrich, S. Comparing linear discriminant analysis and supervised learning algorithms for binary classification—A method comparison study. Biometrical Journal n/a, doi: https://doi.org/10.1002/bimj.202200098 (2022). Wong, T. T. & Yeh, P. Y. Reliable Accuracy Estimates from k-Fold Cross Validation. IEEE Transactions on Knowledge and Data Engineering 32, 1586–1594, doi: 10.1109/TKDE.2019.2912815 (2020). Baratloo, A., Hosseini, M., Negida, A. & El Ashal, G. Part 1: Simple Definition and Calculation of Accuracy, Sensitivity and Specificity. Emerg (Tehran) 3, 48–49 (2015). Müller-Putz, G., Scherer, R., Brunner, C., Leeb, R. & Pfurtscheller, G. Better than random: a closer look on BCI results. International journal of bioelectromagnetism 10, 52–55 (2008). Prasad, S. & Bruce, L. M. Limitations of Principal Components Analysis for Hyperspectral Target Recognition. IEEE Geoscience and Remote Sensing Letters 5, 625–629, doi: 10.1109/LGRS.2008.2001282 (2008). Romine, W., Schroeder, N., Banerjee, T. & Graft, J. Toward Mental Effort Measurement Using Electrodermal Activity Features. Sensors 22, 7363 (2022). Boucsein, W. et al. Publication recommendations for electrodermal measurements. Psychophysiology 49, 1017–1034, doi: 10.1111/j.1469-8986.2012.01384.x (2012). Pruneti, C., Saccò, M., Cosentino, C. & Sgromo, D. Relevance of autonomic arousal in the stress response in psychopathology. J. Basic Appl. Sci 12, 176–184 (2016). Quigley, K. S. & Barrett, L. F. Is there consistency and specificity of autonomic changes during emotional episodes? Guidance from the Conceptual Act Theory and psychophysiology. Biological Psychology 98, 82–94, doi: https://doi.org/10.1016/j.biopsycho.2013.12.013 (2014). Serre, F. et al. The Craving-Manager smartphone app designed to diagnose substance use/addictive disorders, and manage craving and individual predictors of relapse: a study protocol for a multicenter randomized controlled trial. Front Psychiatry 14, 1143167, doi: 10.3389/fpsyt.2023.1143167 (2023). Roos, C. R. et al. Randomized trial of mindfulness- and reappraisal-based regulation of craving training among daily cigarette smokers. Psychology of Addictive Behaviors 37, 829–840, doi: 10.1037/adb0000940 (2023). Kober, H. & Mell, M. M. in The Wiley Handbook on the Cognitive Neuroscience of Addiction 195–218 (2015). Unsectioned Figure Details A A Fig. 1 Performance results for LDA, SVM and MDM classifiers: confusion matrix and ROC curve. MDM; (B) SVM; (C) LDA. Confusion Matrix describes the performance of classifiers by displaying the counts of true positive (TP), true negative (TN), false positive (FP) and false negative (FN) predictions. The ROC curve represents the relationship between sensitivity and specificity. The AUC of the ROC curve ranges from 0 to 1: a model with an AUC of 1 indicates perfect performance, as it achieves a true positive rate of 1 (100%) and a false positive rate of 0 (0%). An AUC of 0.55 corresponds to a random guess, where the model's performance is no better than chance taking into account the number of trials and class applied here. A A Fig. 2 Permutation test results for LDA, SVM and MDM classifiers. Permutation tests based on 100 iterations. Null distributions are plotted in blue bars, the 95th percentile of permutations accuracy as orange lines, observed accuracy measures as red lines. A Fig. 1 Performance results for LDA, SVM and MDM classifiers: confusion matrix and ROC curve. (A) MDM; (B) SVM; (C) LDA. A Fig. 2 Permutation test results for LDA, SVM and MDM classifiers. Unsectioned Paragraphs Authors : Marc Auriacombe e-mail: [email protected] Pôle Addictologie (SANPSY), CHCP, 121 rue de la Béchade, 33076 Bordeaux Cedex, France Tel: +33 556 561 738 Supplementary Figures (2) and Supplementary Tables (2) ABSTRACT (135 words) Craving variations have been shown to be prospectively associated to relapse and are a target for treatment in substance use disorders (SUD). We show that a set of physiological features captured in daily life can distinguish within-day craving from no-craving states, among SUD participants in a 14-day multimodal study. Craving was assessed Signal-contingent and event-contingent. Continuous recording of physiological signals was conducted daily. After labeling samples as “craving” and “no-craving” using a double-check verification, features of each signal were extracted, normalized and standardized and a principal component analysis was applied. Using Support Vector Machines and Linear Discriminant Analysis classifiers, physiological signals allowed to discriminate craving from no-craving episodes with a high cross-validation accuracy (73.6% and 72.2% respectively). This finding is an important breakthrough that paves the way to a biomarker of craving in daily life. FIGURES LEGENDS Words count 3000 Figure/Table 2 tables, 2 fig References 63 1 Mood disorders include major depressive episode, dysthymia and hypomania; anxiety disorders include social phobia, generalized anxiety disorder and panic disorder. 2 Frequencies, percentages and means are based on the total number of valid electronic surveys during the assessment period. ASI = Addiction Severity Index; MINI = Mini International Neuropsychiatric Disorder; EMA = Ecological Momentary Assessment. Unit of all EDA features was microSiemens (µS); second (s) for HRV SDANNI5, HRV SDANN5, HRV SDANN2 and HRV SDANN1 features; HRV DFAα2 had no unit. Bold values indicate statistically significant P values. Additional Declarations The authors have declared there is NO conflict of interest to disclose Supplementary Files BailletBioemaSupplementaryFigure1.docx Supplementary Figure 1 BailletBioemaSupplementaryFigure2.docx Supplemetary Figure 2 BailletBioemaSupplementv5copie3.docx Supplement Material Cite Share Download PDF Status: Posted Version 1 posted You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-4432651","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Article","associatedPublications":[],"authors":[{"id":356562575,"identity":"4064525e-77a5-46fc-99d4-296cad418cdf","order_by":0,"name":"Emmanuelle Baillet","email":"","orcid":"","institution":"Univ. Bordeaux, SANPSY, UMR 6033, F-33000 Bordeaux, France; CNRS, SANPSY, UMR 6033, F-33000 Bordeaux, France; Pôle Interétablissement d’Addictologie, CH Ch. Perrens and CHU de Bordeaux, F-33076 Bordeaux, France","correspondingAuthor":false,"prefix":"","firstName":"Emmanuelle","middleName":"","lastName":"Baillet","suffix":""},{"id":356562576,"identity":"4ca20ba9-78fb-4eab-b198-e734bd7a626b","order_by":1,"name":"Fuschia Serre","email":"","orcid":"https://orcid.org/0000-0002-0228-3572","institution":"Univ. Bordeaux, SANPSY, UMR 6033, F-33000 Bordeaux, France; CNRS, SANPSY, UMR 6033, F-33000 Bordeaux, France; Pôle Interétablissement d’Addictologie, CH Ch. Perrens and CHU de Bordeaux, F-33076 Bordeaux, France","correspondingAuthor":false,"prefix":"","firstName":"Fuschia","middleName":"","lastName":"Serre","suffix":""},{"id":356562577,"identity":"981a528f-b9c9-498c-bdef-8d4aa8e9972a","order_by":2,"name":"Hakim Si-Mohammed","email":"","orcid":"","institution":"Univ. Lille, CNRS, Centrale Lille, UMR 9189 CRIStAL, F-59000 Lille, France","correspondingAuthor":false,"prefix":"","firstName":"Hakim","middleName":"","lastName":"Si-Mohammed","suffix":""},{"id":356562578,"identity":"0398a697-d20b-4860-8e90-be3ff7ff3391","order_by":3,"name":"Cassandre Romao","email":"","orcid":"","institution":"Univ. Bordeaux, SANPSY, UMR 6033, F-33000 Bordeaux, France; CNRS, SANPSY, UMR 6033, F-33000 Bordeaux, France; Pôle Interétablissement d’Addictologie, CH Ch. Perrens and CHU de Bordeaux, F-33076 Bordeaux, France","correspondingAuthor":false,"prefix":"","firstName":"Cassandre","middleName":"","lastName":"Romao","suffix":""},{"id":356562579,"identity":"07a24e47-08ca-4a62-b164-4573c0d0b566","order_by":4,"name":"Arthur Bruneau","email":"","orcid":"","institution":"Univ. Bordeaux, SANPSY, UMR 6033, F-33000 Bordeaux, France; CNRS, SANPSY, UMR 6033, F-33000 Bordeaux, France; Pôle Interétablissement d’Addictologie, CH Ch. Perrens and CHU de Bordeaux, F-33076 Bordeaux, France","correspondingAuthor":false,"prefix":"","firstName":"Arthur","middleName":"","lastName":"Bruneau","suffix":""},{"id":356562580,"identity":"a7952b8d-0733-47a6-bda5-220ba06bdc0b","order_by":5,"name":"Chloé Vacher","email":"","orcid":"","institution":"Univ. Bordeaux, SANPSY, UMR 6033, F-33000 Bordeaux, France; CNRS, SANPSY, UMR 6033, F-33000 Bordeaux, France; Pôle Interétablissement d’Addictologie, CH Ch. Perrens and CHU de Bordeaux, F-33076 Bordeaux, France","correspondingAuthor":false,"prefix":"","firstName":"Chloé","middleName":"","lastName":"Vacher","suffix":""},{"id":356562581,"identity":"456f49b8-aa96-474a-ab55-0ce33f6259d3","order_by":6,"name":"Thibaut Monseigne","email":"","orcid":"","institution":"Univ. Bordeaux, SANPSY, UMR 6033, F-33000 Bordeaux, France; CNRS, SANPSY, UMR 6033, F-33000 Bordeaux, France; Pôle Interétablissement d’Addictologie, CH Ch. Perrens and CHU de Bordeaux, F-33076 Bordeaux, France","correspondingAuthor":false,"prefix":"","firstName":"Thibaut","middleName":"","lastName":"Monseigne","suffix":""},{"id":356562582,"identity":"ef0d6320-9964-4657-af71-3096415ae722","order_by":7,"name":"Camille Jeunet-Kelway","email":"","orcid":"","institution":"Univ. Bordeaux, CNRS, EPHE, INCIA, UMR5287 F-33000 Bordeaux","correspondingAuthor":false,"prefix":"","firstName":"Camille","middleName":"","lastName":"Jeunet-Kelway","suffix":""},{"id":356562574,"identity":"9146334b-b65e-4d40-b6ea-1382237b0680","order_by":8,"name":"Marc Auriacombe","email":"data:image/png;base64,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","orcid":"https://orcid.org/0000-0002-8938-8683","institution":"Univ. Bordeaux, SANPSY, UMR 6033, F-33000 Bordeaux, France; CNRS, SANPSY, UMR 6033, F-33000 Bordeaux, France; Pôle Interétablissement d’Addictologie, CH Ch. Perrens and CHU de Bordeaux, F-33076 Bordeaux, France","correspondingAuthor":true,"prefix":"","firstName":"Marc","middleName":"","lastName":"Auriacombe","suffix":""}],"badges":[],"createdAt":"2024-05-16 17:50:53","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-4432651/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-4432651/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":71661259,"identity":"f65cf22f-0d29-4dcc-b963-5c6f3fc7ec03","added_by":"auto","created_at":"2024-12-17 13:42:29","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":150037,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003ePerformance results for LDA, SVM and MDM classifiers: confusion matrix and ROC curve.\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e(A) MDM; (B) SVM; (C) LDA.\u003c/p\u003e\n\u003cp\u003eConfusion Matrix describes the performance of classifiers by displaying the counts of true positive (TP), true negative (TN), false positive (FP) and false negative (FN) predictions. The ROC curve represents the relationship between sensitivity and specificity. The AUC of the ROC curve ranges from 0 to 1: a model with an AUC of 1 indicates perfect performance, as it achieves a true positive rate of 1 (100%) and a false positive rate of 0 (0%). An AUC of 0.55 corresponds to a random guess, where the model's performance is no better than chance taking into account the number of trials and class applied here.\u003c/p\u003e","description":"","filename":"Onlinefloatimage1.png","url":"https://assets-eu.researchsquare.com/files/rs-4432651/v1/7bf0ffe94cf89763ed3c2357.png"},{"id":71661257,"identity":"8cceedae-dfad-4299-acd1-8bcf4fee449f","added_by":"auto","created_at":"2024-12-17 13:42:29","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":36418,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003ePermutation test results for LDA, SVM and MDM classifiers.\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003ePermutation tests based on 100 iterations. Null distributions are plotted in blue bars, the 95\u003csup\u003eth\u003c/sup\u003e percentile of permutations accuracy as orange lines, observed accuracy measures as red lines.\u003c/p\u003e","description":"","filename":"Onlinefloatimage2.png","url":"https://assets-eu.researchsquare.com/files/rs-4432651/v1/56b16d3eacc361557cdfc013.png"},{"id":73300124,"identity":"51a3d628-a829-4824-8760-82680ef54dcc","added_by":"auto","created_at":"2025-01-08 16:00:25","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":1005922,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-4432651/v1/6748923b-263f-46dd-bad0-0f1d19e06903.pdf"},{"id":71661663,"identity":"07e1ecbf-c6dd-4381-b2b7-4aaff6dbfc6e","added_by":"auto","created_at":"2024-12-17 13:50:29","extension":"docx","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":119443,"visible":true,"origin":"","legend":"\u003cp\u003eSupplementary Figure 1\u003c/p\u003e","description":"","filename":"BailletBioemaSupplementaryFigure1.docx","url":"https://assets-eu.researchsquare.com/files/rs-4432651/v1/d9e62d29597f82d82b0b55fd.docx"},{"id":71661261,"identity":"f191ab06-8c82-4134-9c44-88f2b02e089d","added_by":"auto","created_at":"2024-12-17 13:42:29","extension":"docx","order_by":2,"title":"","display":"","copyAsset":false,"role":"supplement","size":140757,"visible":true,"origin":"","legend":"\u003cp\u003eSupplemetary Figure 2\u003c/p\u003e","description":"","filename":"BailletBioemaSupplementaryFigure2.docx","url":"https://assets-eu.researchsquare.com/files/rs-4432651/v1/468e67d8250b66be50030655.docx"},{"id":71661662,"identity":"b484bae3-75c6-49f0-ba16-52d83d72cbdc","added_by":"auto","created_at":"2024-12-17 13:50:29","extension":"docx","order_by":3,"title":"","display":"","copyAsset":false,"role":"supplement","size":104025,"visible":true,"origin":"","legend":"\u003cp\u003eSupplement Material\u003c/p\u003e","description":"","filename":"BailletBioemaSupplementv5copie3.docx","url":"https://assets-eu.researchsquare.com/files/rs-4432651/v1/a8be643b5aafc575bd84505d.docx"}],"financialInterests":"The authors have declared there is \u003cb\u003eNO\u003c/b\u003e conflict of interest to disclose","formattedTitle":"The Autonomic Craving Signature: physiological signals as a daily-life biomarker of craving","fulltext":[{"header":"INTRODUCTION","content":"\u003cp\u003eCraving, an intense desire to use, is a key contributor to substance use disorders (SUD) \u003csup\u003e1\u0026ndash;3\u003c/sup\u003e. Added as a SUD diagnostic criterion in the \u003cem\u003eDiagnostic and Statistical Manual of Mental Disorders\u003c/em\u003e, fifth edition (DSM-5) \u003csup\u003e4,5\u003c/sup\u003e, it has been demonstrated to be the most discriminant and specific criterion in Item Responses Theory analysis \u003csup\u003e6\u003c/sup\u003e and the most central criterion in networks of DSM-5 SUD diagnostic criteria \u003csup\u003e7\u003c/sup\u003e. Craving may also vary in intensity and frequency within hours or days in the same individual, under the influence of internal and environmental factors \u003csup\u003e8\u003c/sup\u003e. Versatile emotional states have been a challenge for research but are now particularly well captured by Ecological Momentary Assessment (EMA) \u003csup\u003e9\u003c/sup\u003e that allows a real-time measurement of individual momentary states in the natural environment through repeated assessments overtime. Use of EMA in craving research fueled a generic clinical model of addiction, the cues \u0026ndash; craving \u0026ndash; use model \u003csup\u003e10\u003c/sup\u003e in which cues \u0026ndash; conditional stimuli associated with past experiences of use \u003csup\u003e11\u003c/sup\u003e \u003cem\u003e\u0026ndash;\u003c/em\u003e were prospectively associated to craving, which in turn was prospectively associated with use in the following hours \u003csup\u003e12\u003c/sup\u003e. In a meta-analysis pooling 51,788 substance-using participants from 237 studies, craving was reported to prospectively predict substance use and relapse, suggesting it could play a causative role \u003csup\u003e3\u003c/sup\u003e.\u003c/p\u003e \u003cp\u003eRecently, a study based on a functional magnetic resonance imagery (fMRI) cue-reactivity protocol reported a neurobiological craving signature (NCS) common to food and substances, that predicted craving and distinguished substance users from non-users \u003csup\u003e13\u003c/sup\u003e. Another study identified a neurobiological signature of craving based on functional connectivity from resting-state high-density electroencephalography (EEG) \u003csup\u003e14\u003c/sup\u003e. This represents a potential major step forward in the discovery of biomarkers of craving and provides a window into understanding its neurobiological mechanisms. However, fMRI or EEG are not yet usable in daily-life, limiting the transitioning of NCS to everyday clinical practice. Yet, neurobiological changes specific to craving could be captured by changes in the autonomic nervous system (ANS) \u003csup\u003e15,16\u003c/sup\u003e. Heart rate (HR), skin temperature (TEMP), breathing rates or electrodermal activity (EDA) are commonly used as indicators of psychological arousal and can be easily captured in daily-life by ambulatory monitoring devices. Several studies have shown a concomitance between signs of physiological arousal such as a significant increase in HR, EDA, TEMP or accelerometer (ACC) and the appearance of craving \u003csup\u003e17\u0026ndash;22\u003c/sup\u003e. The magnitude of craving was correlated with these physiological activity changes suggesting that the latter could actually represent underlying mechanisms of craving or, their physiological expression \u003csup\u003e16,23,24\u003c/sup\u003e, and thereby be a potential objective marker implementable in the natural context of participants. Some studies have examined the physiological correlates of craving in human lab settings \u003csup\u003e18,25,26\u003c/sup\u003e or in daily-life \u003csup\u003e27\u0026ndash;32\u003c/sup\u003e with limitations in model development, accuracy and reproducibility \u003csup\u003e33\u003c/sup\u003e. Firstly, the use of a single physiological parameter (i.e., HR \u003csup\u003e27,29\u003c/sup\u003e) may not be specific and reliable enough to identify such a complex phenomenon as craving is likely to be. A pattern, or combination, of physiological measures (i.e., combination of HR, EDA and TEMP) may be more appropriate. Secondly, some measures of craving may not be craving specific, which means there was no guarantee that the phenomenon observed and analyzed was really craving \u003csup\u003e34\u003c/sup\u003e and not the association of different phenomena, such as stress co-occurring with craving \u003csup\u003e27\u0026ndash;29\u003c/sup\u003e, especially because stress and craving share some physiological similarities \u003csup\u003e15\u003c/sup\u003e. Thirdly, to identify physiological correlates of craving, the time window explored to date ranged from 40 to 180 minutes \u003csup\u003e27\u0026ndash;29\u003c/sup\u003e, while craving duration could be shorter (5 to 30 minutes) \u003csup\u003e35\u0026ndash;37\u003c/sup\u003e. Finally, craving needs to be explored with its intensity fluctuations over time, requiring EMA and adapted length of assessment \u003csup\u003e38\u003c/sup\u003e.\u003c/p\u003e \u003cp\u003e To overcome these limitations, our aim was to identify a transdiagnostic and reliable pattern of physiological signals of ANS (heart rate, electrodermal activity, skin temperature and accelerometry) associated with craving captured in daily-life using a multimodal method (EMA and wearable sensors) among participants with SUDs. We expected that craving vs. no-craving episodes would be distinguished based on physiological and EMA data.\u003c/p\u003e"},{"header":"METHODS","content":"\u003cdiv id=\"Sec3\" class=\"Section2\"\u003e \u003ch2\u003eParticipants\u003c/h2\u003e \u003cp\u003eParticipants were included from the ADDICTAQUI Cohort \u003csup\u003e39\u003c/sup\u003e or from not-in-treatment population. The ADDICTAQUI cohort includes individuals seeking treatment for a DSM-5 use disorder in outpatient addiction treatment clinics in Aquitaine, France. Participants from the not-in-treatment population were recruited through social networks. To be eligible for the study, participants in treatment and not-in-treatment had to be adults (\u0026gt;\u0026thinsp;18 years), with at least one DSM-5 SUD diagnosis for tobacco, alcohol or cannabis and reported craving in the past 30 days. Exclusion criteria were pregnancy, cardiovascular disease, non-dominant upper extremity limitation, or psychiatric conditions incompatible with understanding protocol. The study was approved by the French Regulation and ethical committee (CPP EST III, approval registration 2021-A01835-36) and all participants provided written informed consent.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec4\" class=\"Section2\"\u003e \u003ch2\u003eExperimental Design\u003c/h2\u003e \u003cp\u003eThe study was a 14-day prospective daily-life data collection with EMA and wearable sensors. During the enrollment visit, clinical assessment was provided with the Addiction Severity Index (ASI) \u003csup\u003e40\u003c/sup\u003e and Mini International Neuropsychiatric Interview (MINI) \u003csup\u003e41\u003c/sup\u003e to collect demographic and historical data about SUD and other psychiatric disorders. A brief training was performed with devices. Participants were asked to wear the sensor on their non-dominant wrist all day for the duration of the study, and to remove it only to sleep. EMA surveys, presented 4 times/day at random, approximately every 3h30, were adapted to the primary substance of addiction according to ASI and in case of poly-addiction, the choice was made with the participant.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec5\" class=\"Section2\"\u003e \u003ch2\u003eMain outcomes measures\u003c/h2\u003e \u003cdiv id=\"Sec6\" class=\"Section3\"\u003e \u003ch2\u003ePhysiological parameters:\u003c/h2\u003e \u003cp\u003eThe E4\u0026reg; sensor (Empatica Inc., Cambridge MA, USA) is a research device with good signal quality compared to a reference device \u003csup\u003e42\u0026ndash;44\u003c/sup\u003e. It is a non-invasive wearable which allows the continuous acquisition of physiological data in real-time: blood volume pulse (BVP) (in beats/min (bpm)), electrodermal activity (EDA) (in microSiemens (\u0026micro;S)), skin temperature (TEMP) (in Celsius degree (\u0026deg;C)) and 3-axis accelerometers (ACC) (in gravity (g)).\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv id=\"Sec7\" class=\"Section2\"\u003e \u003ch2\u003eCraving measurements:\u003c/h2\u003e \u003cp\u003eCraving was assessed signal-contingently on a 7-point numerical scale in EMA from \u0026ldquo;1\u0026rdquo; (not at all) to \u0026ldquo;7\u0026rdquo; (extreme craving) in response to the question \"Since the last questionnaire, have you felt a desire to use \u003cem\u003e[primary substance]\u003c/em\u003e?\", and event-contingently, by button pressing (\u0026ldquo;TAG\u0026rdquo;) on the sensor when they felt craving for their primary substance, regardless of intensity.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec8\" class=\"Section2\"\u003e \u003ch2\u003eAnalysis strategy\u003c/h2\u003e \u003cp\u003eTo label classes of craving and no-craving, a double-check selection procedure was used: \u0026ldquo;No-craving\u0026rdquo; class was defined when no craving was reported both on the sensor (no-TAG) and on the corresponding EMA survey (intensity\u0026thinsp;=\u0026thinsp;1); and \u0026ldquo;Craving\u0026rdquo; class was defined when craving was reported both on the sensor (TAG) and on the corresponding EMA (intensity\u0026thinsp;\u0026gt;\u0026thinsp;4 corresponding to the average intensity of the craving scale). Periods when the participant mentioned stress (\u0026gt;\u0026thinsp;1 on the EMA) were excluded. Then, based on the literature exploring craving duration \u003csup\u003e35\u0026ndash;37,45\u0026minus;47\u003c/sup\u003e, we used epochs of 20 minutes (centered around the TAG for \u0026ldquo;craving\u0026rdquo; or randomly for \u0026ldquo;no-craving\u0026rdquo;), in which several features were extracted for each physiological signal recorded to reflect their real values (Supplementary Table\u0026nbsp;1). In order to be able to run statistical analyses without risking over-fitting of the model on the data, we included only the participants for whom at least 20 samples of each class could be extracted. If the classes were unbalanced in time of occurrence, we selected the samples that were the closer from the samples selected in the other class. Then we used Principal Component Analysis (PCA)\u003csup\u003e48\u003c/sup\u003e to reduce data dimensionality by identifying Principal Components (PCs) that capture the maximum variance in physiological data, regardless of class\u003csup\u003e49\u003c/sup\u003e. Student t-tests were performed on each feature (\u0026ldquo;craving\u0026rdquo; vs. \u0026ldquo;no-craving\u0026rdquo;) in order to identify those that contributed the most to the first PCs. Finally, Support vector machines (SVM)\u003csup\u003e50\u003c/sup\u003e, minimum distance to mean (MDM)\u003csup\u003e51\u003c/sup\u003e, linear discriminant analysis (LDA)\u003csup\u003e52\u003c/sup\u003e classifiers were used. Their performance was assessed through a k-fold cross-validation (k\u0026thinsp;=\u0026thinsp;10) \u003csup\u003e53\u003c/sup\u003e. For each fold, 80% of the samples were randomly assigned to the training test while the other 20% were used as a testing set. Our models were evaluated with standard metrics in which craving samples were the positive class: specificity, recall, precision, accuracy and area under the curve (AUC) of the receiver operating characteristic (ROC) curve \u003csup\u003e54\u003c/sup\u003e. The theoretical chance levels were estimated at 65% \u003csup\u003e55\u003c/sup\u003e. Finally, to estimate statistical significance of the cross-validation classifier accuracy for an \u003cem\u003eα\u003c/em\u003e risk of 5%, we performed permutation tests with 100 randomized cross-validation iterations. The proposed model is detailed in the Supplement and was implemented using Python (3.12.0) and R Studio (4.1.0) software. The code used is available on request to the first and corresponding authors.\u003c/p\u003e \u003c/div\u003e"},{"header":"RESULTS","content":"\u003cp\u003eFrom April 2022 to July 2023, 43 (95.6%) of 45 participants included had worn the sensor and were considered as the study sample for analysis. Participants of the study sample were aged 32.9 years (standard deviation [SD]\u0026thinsp;=\u0026thinsp;9.4), and 55% (n\u0026thinsp;=\u0026thinsp;24) were women and 45% were men (n\u0026thinsp;=\u0026thinsp;19). The most frequent primary substance was tobacco (46.5%, n\u0026thinsp;=\u0026thinsp;20), followed by cannabis (32.5%, n\u0026thinsp;=\u0026thinsp;14) and alcohol (20.9%, n\u0026thinsp;=\u0026thinsp;9). On average participants had 7.3 DSM-5 SUD diagnostic criteria (SD\u0026thinsp;=\u0026thinsp;2.1) and a mean of 5.4 on the ASI severity score for the primary substance (SD\u0026thinsp;=\u0026thinsp;1.3) (Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e).\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab1\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 1\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eCharacteristics of the sample and description of daily-life variables.\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"9\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c6\" colnum=\"6\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c7\" colnum=\"7\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c8\" colnum=\"8\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c9\" colnum=\"9\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\" morerows=\"1\" rowspan=\"2\"\u003e \u003cp\u003e\u003cem\u003eSamples characteristics\u003c/em\u003e\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colspan=\"4\" nameend=\"c5\" namest=\"c2\"\u003e \u003cp\u003eFull sample (N\u0026thinsp;=\u0026thinsp;43)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colspan=\"4\" nameend=\"c9\" namest=\"c6\"\u003e \u003cp\u003eSample of the ACS (N\u0026thinsp;=\u0026thinsp;5)\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003e% (N)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eM (SD)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003eMin\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c5\"\u003e \u003cp\u003eMax\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c6\"\u003e \u003cp\u003e% (N)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c7\"\u003e \u003cp\u003eM (SD)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c8\"\u003e \u003cp\u003eMin\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c9\"\u003e \u003cp\u003eMax\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eGender (woman)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e55.8 (\u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e24\u003c/span\u003e)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e60.0 (\u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eEmployed\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e76.2 (\u003cspan citationid=\"CR32\" class=\"CitationRef\"\u003e32\u003c/span\u003e)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e80.0 (\u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eAge\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e32.9 (9.4)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e22\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e60\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e39.2 (10.1)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e30\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003e52\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eEducation (years)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e14.5 (1.9)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e9\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e17\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e12.6 (2.6)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e9\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003e16\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cem\u003ePrimary substance\u003c/em\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"8\" nameend=\"c9\" namest=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eTobacco\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e46.5 (\u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e20\u003c/span\u003e)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e40.0 (\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eCannabis\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e32.5 (\u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e40.0 (\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eAlcohol\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e20.9 (\u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e20.0 (\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eASI severity score for primary substance\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e5.4 (1.3)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e7\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e5.4 (1.8)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e0\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003e7\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eMINI SUD severity\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e7.3 (2.1)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e11\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e7.4 (2.6)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e0\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003e11\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cem\u003eMINI current diagnosis\u003c/em\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"8\" nameend=\"c9\" namest=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eMood disorder\u003csup\u003e1\u003c/sup\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e20.9 (\u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e25.0 (\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eAnxiety disorder\u003csup\u003e1\u003c/sup\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e20.9 (\u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e40.0 (\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eSignal-contingent EMA\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"4\" nameend=\"c5\" namest=\"c2\"\u003e \u003cp\u003e\u003cb\u003e85.8%, observations\u0026thinsp;=\u0026thinsp;2 017\u003c/b\u003e\u003csup\u003e\u003cb\u003e2\u003c/b\u003e\u003c/sup\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"4\" nameend=\"c9\" namest=\"c6\"\u003e \u003cp\u003e\u003cb\u003e91.4%, observations\u0026thinsp;=\u0026thinsp;256\u003c/b\u003e\u003csup\u003e\u003cb\u003e2\u003c/b\u003e\u003c/sup\u003e\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eUse of primary substance\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e51.4 (1037)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e23.8 (\u003cspan citationid=\"CR61\" class=\"CitationRef\"\u003e61\u003c/span\u003e)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eCraving intensity (\u003cspan additionalcitationids=\"CR2 CR3 CR4 CR5 CR6\" citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e4.1 (2.1)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e7\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e3.5 (2.4)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e0\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003e7\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eCraving episodes (intensity\u0026thinsp;\u0026gt;\u0026thinsp;1)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e77.2 (1556)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e59.5 (152)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eEvent-contingent sensor (\u0026gt;\u0026thinsp;0)\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e1 654\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e38.5 (33.4)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e146\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e295\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e59.0 (25.8)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e29\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003e93\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003ctfoot\u003e \u003ctr\u003e\u003ctd colspan=\"9\"\u003e\u003csup\u003e1\u003c/sup\u003eMood disorders include major depressive episode, dysthymia and hypomania; anxiety disorders include social phobia, generalized anxiety disorder and panic disorder. \u003csup\u003e2\u003c/sup\u003eFrequencies, percentages and means are based on the total number of valid electronic surveys during the assessment period. ASI\u0026thinsp;=\u0026thinsp;Addiction Severity Index; MINI\u0026thinsp;=\u0026thinsp;Mini International Neuropsychiatric Disorder; EMA\u0026thinsp;=\u0026thinsp;Ecological Momentary Assessment.\u003c/td\u003e\u003c/tr\u003e \u003c/tfoot\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003cdiv id=\"Sec10\" class=\"Section2\"\u003e \u003ch2\u003e[Table \u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e]\u003c/h2\u003e \u003cp\u003eEMA completion rate was 85.8% (n\u0026thinsp;=\u0026thinsp;2,017 assessments completed) and the sensor was used on average 8.9 hours/day (SD\u0026thinsp;=\u0026thinsp;2.9). In total, 5,512 hours of physiological data were captured over the course of EMA-days. The total number of craving TAG was 1,654 TAGs with an average of 38.5 TAG (SD\u0026thinsp;=\u0026thinsp;33.4) per participant and the total number of EMA craving reports was 1,556 with an average craving intensity of 4.1/7 (SD\u0026thinsp;=\u0026thinsp;2.08).\u003c/p\u003e \u003cp\u003eFor sample labeling, 1883 samples were labeled as \u0026ldquo;no-craving\u0026rdquo;, 1109 samples were labeled as \u0026ldquo;craving\u0026rdquo;, and 4296 samples were excluded because they did not meet our double-check selection criteria (Supplementary Fig.\u0026nbsp;1). From these selected samples, 190 features were extracted, normalized and standardized after which 98 were again rejected for missing values, high correlation or low standard deviation (Supplementary Table\u0026nbsp;2). Consequently, each \u0026ldquo;craving\u0026rdquo; and \u0026ldquo;no-craving\u0026rdquo; samples was composed of a vector of 92 standardized and normalized features.\u003c/p\u003e \u003cp\u003eTo maintain the balance between classes \u0026ldquo;craving\u0026rdquo; and \u0026ldquo;no-craving\u0026rdquo; and avoid overfitting, we only retained participants with at least 20 samples/classes. Thus, 192 samples of \u0026ldquo;craving\u0026rdquo; and 192 of \u0026ldquo;no-craving\u0026rdquo; were analyzed by PCA and these were issued from 5 participants. As a result, 31 PCs, which explained 98.8% of variance data, were kept for the classification model (Supplementary Fig.\u0026nbsp;2). The first component (PC1) explained the majority of the total variance (73.8%) and was mainly composed by EDA features while the second component (PC2) explained 7.9% of the total variance. Five features of PC1 with greater weights (59.43% of PC1) were significantly different between our two classes: \u0026ldquo;EDA Tonic Min\u0026rdquo; (p\u0026thinsp;\u0026lt;\u0026thinsp;.0001), \u0026ldquo;EDA Tonic Mean\u0026rdquo; (p\u0026thinsp;\u0026lt;\u0026thinsp;.0001), \u0026ldquo;EDA Tonic Max\u0026rdquo; (p\u0026thinsp;\u0026lt;\u0026thinsp;.0001), \u0026ldquo;EDA Wavelet 0 RMS\u0026rdquo; (p\u0026thinsp;\u0026lt;\u0026thinsp;.0001), and \u0026ldquo;EDA Skew\u0026rdquo; (p\u0026thinsp;=\u0026thinsp;.020) (Table\u0026nbsp;\u003cspan refid=\"Tab2\" class=\"InternalRef\"\u003e2\u003c/span\u003e). In PC2, among five features with greater weights (42.27% of PC2), \u0026ldquo;HRV SDNNI5\u0026rdquo; (p\u0026thinsp;=\u0026thinsp;.001), \u0026ldquo;HRV SDANN5\u0026rdquo; (p\u0026thinsp;=\u0026thinsp;.017), \u0026ldquo;HRV SDANN2\u0026rdquo; (p\u0026thinsp;\u0026lt;\u0026thinsp;.0001) and \u0026ldquo;HRV SDANN1\u0026rdquo; (p\u0026thinsp;\u0026lt;\u0026thinsp;.0001) were significantly different between our two classes, but not \u0026ldquo;HRV DFA \u003cem\u003eα\u003c/em\u003e2\u0026rdquo; (p\u0026thinsp;=\u0026thinsp;0.128) (Table\u0026nbsp;\u003cspan refid=\"Tab2\" class=\"InternalRef\"\u003e2\u003c/span\u003e).\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab2\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 2\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eDescription of the main features of the first PCs applied to craving/no-craving samples.\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"7\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c6\" colnum=\"6\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c7\" colnum=\"7\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePC\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eFeatures\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eM (SD)[Craving]\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003eM (SD)[No-craving]\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c5\"\u003e \u003cp\u003et-ratio\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c6\"\u003e \u003cp\u003edof\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c7\"\u003e \u003cp\u003eProb.\u0026gt;ItI\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePC1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eEDA Tonic Min\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e2.59 (9.52)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e11.65 (21.75)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e-5.28\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e261.67\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e\u003cb\u003e\u0026lt;\u0026thinsp;.0001\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePC1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eEDA Tonic Mean\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.51 (0.89)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e1.20 (1.36)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e-5.86\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e329.5\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e\u003cb\u003e\u0026lt;\u0026thinsp;.0001\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePC1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eEDA Tonic Max\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.74 (1.01)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e1.46 (1.38)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e-5.82\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e349.8\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e\u003cb\u003e\u0026lt;\u0026thinsp;.0001\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePC1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eEDA Wavelet 0 RMS\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.20 (0.70)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.82 (1.16)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e-6.27\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e314.6\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e\u003cb\u003e\u0026lt;\u0026thinsp;.0001\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePC1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eEDA Skew\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e1.28 (3.05)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.59 (2.78)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e2.33\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e378.9\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e\u003cb\u003e0.020\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePC2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eHRV SDNNI5\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e1.82 (4.89)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.64 (1.38)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e3.22\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e221.3\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e\u003cb\u003e0.001\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePC2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eHRV SDNN5\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.98 (4.64)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.16 (0.85)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e2.41\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e203.8\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e\u003cb\u003e0.017\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePC2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eHRV SDANN2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e1.22 (2.62)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.29 (1.65)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e4.16\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e322.6\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e\u003cb\u003e\u0026lt;\u0026thinsp;.0001\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePC2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eHRV SDANN1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e1.08 (1.92)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.22 (1.24)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e5.22\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e327.6\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e\u003cb\u003e\u0026lt;\u0026thinsp;.0001\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePC2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eHRV DFA \u003cem\u003eα\u003c/em\u003e2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e-0.20 (2.01)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.04 (0.83)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e-1.53\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e254.5\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e0.128\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003ctfoot\u003e \u003ctr\u003e\u003ctd colspan=\"7\"\u003eUnit of all EDA features was microSiemens (\u0026micro;S); second (s) for HRV SDANNI5, HRV SDANN5, HRV SDANN2 and HRV SDANN1 features; HRV DFAα2 had no unit. Bold values indicate statistically significant P values.\u003c/td\u003e\u003c/tr\u003e \u003c/tfoot\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec11\" class=\"Section2\"\u003e \u003ch2\u003e[Table \u003cspan refid=\"Tab2\" class=\"InternalRef\"\u003e2\u003c/span\u003e]\u003c/h2\u003e \u003cp\u003eThen, the 31 PCs from the training set were used to perform SVM, MDM and LDA classifications, and performance in discriminating craving from no-craving was estimated on the test set. For the MDM, the sensitivity was 66.2% and the specificity was 51.0%; the AUC of the model\u0026rsquo;s ROC was 0.58 (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e1\u003c/span\u003e) and the model\u0026rsquo;s accuracy and precision were 58.3% and 55.5% respectively. For the SVM, the sensitivity was 74.5% and the specificity was 72.8%; the AUC of the model\u0026rsquo;s ROC was 0.83 (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e1\u003c/span\u003e) and the model\u0026rsquo;s accuracy and precision were 73.6% and 73.1% respectively. For the LDA, the sensitivity was 76.8% and the specificity was 67.7%; the AUC of the model\u0026rsquo;s ROC was 0.83 (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e1\u003c/span\u003e) and the model\u0026rsquo;s accuracy and precision were 72.2% and 69.8% respectively. For all measures and classifiers, none of the permutation samples performed (general accuracy score: 0.55) as well as the observed results (MDM: p\u0026thinsp;=\u0026thinsp;0.02; SVM: p\u0026thinsp;=\u0026thinsp;0.01; LDA: p\u0026thinsp;=\u0026thinsp;0.01) at the 95% confidence level (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e2\u003c/span\u003e).\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec12\" class=\"Section2\"\u003e \u003ch2\u003e[FIGURE \u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e1\u003c/span\u003e]\u003c/h2\u003e \u003cdiv id=\"Sec13\" class=\"Section3\"\u003e \u003ch2\u003e[FIGURE \u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e2\u003c/span\u003e]\u003c/h2\u003e \u003c/div\u003e \u003c/div\u003e"},{"header":"DISCUSSION","content":"\u003cp\u003e The aim of the present work was to identify craving with physiological signals of ANS in daily-life among participants with SUDs. Using SVM and LDA classifiers, physiological signals allowed us to discriminate episodes of craving vs. no-craving with a high cross-validation accuracy (73.6% and 72.2% respectively). Taking into account the limitations of previous research, to our knowledge this is the first study to develop a pipeline to detect craving from physiological signals across substances in daily-life by multimodal methods.\u003c/p\u003e \u003cp\u003eThe application of PCA provided significant advantages for pattern recognition of craving \u003csup\u003e56\u003c/sup\u003e and allowed us to calculate PCs, a linear combination of physiological features, through which classifiers were able to distinguish between craving and no-craving episodes. These features were mainly extracted from EDA and HR signals. The \u0026ldquo;EDA Skew\u0026rdquo; feature, derived from the phasic component of EDA, corresponds to skewness of EDA distribution and more positive values indicate a peak of activity \u003csup\u003e57\u003c/sup\u003e, which was found in craving compared to no-craving episodes. This is potentially important, especially for the interaction with other features that were extracted from the tonic component, generally used to reflect the overall arousal level of an individual in the absence of specific stimuli, while phasic component refers to faster changes, e.g. caused by sudden emotional arousal or external stimuli \u003csup\u003e58\u003c/sup\u003e. In line with our results, craving might induce an activity peak reflected by the phasic component, but there seems to be an association with the tonic component. This result provides both information on the general state of arousal and on a specific response to craving. Our hypothesis is that craving could be reflected in basal levels of autonomic arousal and even in the absence of acute episodes of perceived/reported craving, individuals with SUD may have a higher basic autonomic arousal (tonic component of EDA), which is increased in case of an acute episode of reported craving (phasic component of EDA). This phenomenon is particularly common in anxiety disorder characterized by a high level of HR and EDA at rest \u003csup\u003e59\u003c/sup\u003e. Craving overall could be the expression of an altered activation of the autonomic system at rest, i.e., a basal elevation of some parameters such as HR and EDA in the case of no-craving, and a high arousal in case of craving. This basal level, which is already high in individuals with SUD, increases during craving episodes, making it possible to distinguish it from no-craving episodes.\u003c/p\u003e \u003cp\u003eThis set of features forms a physiological pattern, which we label \u003cem\u003ethe autonomous craving signature (ACS)\u003c/em\u003e, with good specificity\u003csup\u003e34\u003c/sup\u003e and consistency\u003csup\u003e60\u003c/sup\u003e. As this was a study in daily-life, the individuals were not in the same places and did not encounter these 2 conditions at the same times (time within-day or between-days) which reinforces the consistency of ACS under any condition \u003csup\u003e60\u003c/sup\u003e. Moreover, the double-check selection of episodes with the EMA signal contingent (short-term retrospective measure) and the participant-initiated event contingent (immediate measure) enabled us to guarantee that craving episodes reported by the individual were really craving (and not an unintentional press on the sensor, for example) and to approach the validity of physiologic responses of these episodes in everyday life. Also, as each individual was their own control, our model considers both within- and between-individual variability. An important point was that we detected a difference between craving and no-craving episodes with ACS, i.e., features that composed the ACS had a significant difference between our 2 conditions (craving vs. no-craving). This model allowed us to take into account the natural variation of the autonomic system and excluded the hypothesis that the ACS could only result from these variations. In fact, the ACS pattern makes it possible to overcome \u003cem\u003eLaws of Autonomic Constraint\u003c/em\u003e \u003csup\u003e60\u003c/sup\u003e. While we do not rule-out the possibility that some of the features included in the ACS pattern were linked to other phenomena we cannot control, our hypothesis is that the combination of these markers makes the set specific and provides information on the physiological expression of craving, paving the way to better understanding the underlying mechanisms.\u003c/p\u003e \u003cp\u003eImportant limitations have to be acknowledged. One limitation was our choice of epoch length that was of 20 minutes, centered on the participant self-initiated TAG. Despite our significant results, the temporality of craving and its duration remains to be explored. The main difficulty was balancing data: some participants had either not enough episodes of no-craving or of craving. Consequently, much data had to be discarded and finally the ACS was identified from 5 participants out of 43 and requires to be replicated. While this may not seem like much, it is encouraging, since we have reliable results. Furthermore, when compared to the full sample, the 5 participants were not different based on their descriptive characteristics, including craving reports. In our models, we chose low thresholds (correlation, standard deviation) and our results indicate a classification of around 75%, suggesting that the detection of craving was both possible as shown by our findings and could be improved by setting higher restrictive thresholds. The next step would be to test the generalization of this model on a new dataset and to explore, as suggested in previous research, how the ACS may be associate with the NCS \u003csup\u003e13\u003c/sup\u003e.\u003c/p\u003e"},{"header":"CONCLUSION","content":"\u003cp\u003eThe identification of craving from physiological data could be an important step in the development of prognostic biomarkers of relapse in SUD. In addition to facilitate the long-term monitoring of craving without the need for the patient to engage in an effortful awareness that is difficult to sustain overtime, the use of ACS may also be helpful for individuals with low insight of their substance-seeking behaviors or with difficulty in reporting craving, which is associated to poor treatment outcome \u003csup\u003e2\u003c/sup\u003e. The ACS paves the way to the development, in daily-life, of protocols of craving regulation that could build a sustainable warning system that would be activated when this pattern is present to allow a relay towards care if needed or towards a just-in-time intervention \u003csup\u003e61\u003c/sup\u003e such as biofeedback or mindfulness \u003csup\u003e62\u003c/sup\u003e. The regulation of craving is key to better prevention of relapse and more successful treatment outcomes \u003csup\u003e63\u003c/sup\u003e.\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003e \u003ch2\u003eDATA SHARING STATEMENT\u003c/h2\u003e \u003cp\u003eOriginal data are available on request to Marc Auriacombe.\u003c/p\u003e \u003c/p\u003e \u003cp\u003e \u003cstrong\u003eCONFLICT OF INTEREST:\u003c/strong\u003e \u003cp\u003eWe report no conflicts of interest related to this work. The researchers confirm their independence from funders and sponsors.\u003c/p\u003e \u003c/p\u003e\u003ch2\u003eAUTHOR CONTRIBUTION:\u003c/h2\u003e \u003cp\u003eMA was the overall principal investigator and study supervisor, obtained funding and access to participants. MA, EB, FS developed the study design and methods. EB set up the study, included participants, conducted analyses, interpreted the data, and drafted the manuscript. TM contributed to analysis and data management; CM and AB to inclusion of participants. HSM and CJK helped in the conceptualization of analysis, in architecture of classification model and in the interpretation of data. All authors undertook the critical revision of the manuscript for important intellectual content and all authors significantly contributed to the manuscript and approved the final version.\u003c/p\u003e\u003ch2\u003eACKNOWLEDGEMENTS\u003c/h2\u003e \u003cp\u003eThis study received financial support from the French government in the framework of the University of Bordeaux's IdEx \"Investments for the Future\" program / GPR BRAIN_2030 and by French Cancer and Public Health Institutes (INCA-IRESP-2020-169) PhD doctoral grant to Emmanuelle Baillet. Funding for the ADDICTAQUI cohort data collection was provided by Research Grant PHRC (2006\u0026ndash;2014) from the French Ministry of Health, French Government Addiction Agency MILDT/MILDECA grant 2010 and 2016 to Marc Auriacombe. Funding for the EMA study was provided by Research Grant AAP-19-ADDICTIONS-16 from the public health research institute (IReSP) and Aviesan Alliance as part of the call for research projects to combat addiction to psychoactive substances to Marc Auriacombe and Fuschia Serre. Office and staff support was provided by CH Charles Perrens hospital. The authors express theirs thanks to all participants for their contribution and are grateful to all the interviewers of the ADDICTAQUI team, especially to Chlo\u0026eacute; VACHER, Arthur BRUNEAU, Axel ALLACHE and Cassandre ROMAO and to the medical team that managed patients during the time of the study, Saman Sarram, Jacques Dubernet, B\u0026eacute;reng\u0026egrave;re Gelot. Marc Auriacombe and Fuschia Serre had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\u003cli\u003e\u003cspan\u003eAuriacombe, M., Serre, F., Denis, C. \u0026amp; Fatseas, M. in \u003cem\u003eThe Routledge Handbook of the Philosophy and Science of Addiction\u003c/em\u003e (eds H Pickard \u0026amp; S.H. Ahmed) Ch. 10, 132\u0026ndash;144 (2018).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCavicchioli, M., Vassena, G., Movalli, M. \u0026amp; Maffei, C. Is craving a risk factor for substance use among treatment-seeking individuals with alcohol and other drugs use disorders? A meta-analytic review. Drug Alcohol Depend 212, 108002, doi:\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/j.drugalcdep.2020.108002\u003c/span\u003e\u003cspan address=\"10.1016/j.drugalcdep.2020.108002\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2020).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eVafaie, N. \u0026amp; Kober, H. Association of Drug Cues and Craving With Drug Use and Relapse: A Systematic Review and Meta-analysis. JAMA Psychiatry, doi:\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1001/jamapsychiatry.2022.1240\u003c/span\u003e\u003cspan address=\"10.1001/jamapsychiatry.2022.1240\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2022).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eAmerican Psychiatric Association. \u003cem\u003eDiagnostic and Statistical Manual of Mental Disorders (DSM-5)\u003c/em\u003e. (American Psychiatric Pub, 2013).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHasin, D. S. \u003cem\u003eet al.\u003c/em\u003e DSM-5 Criteria for Substance Use Disorders: Recommendations and Rationale. American Journal of Psychiatry 170, 834\u0026ndash;851, doi:\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1176/appi.ajp.2013.12060782\u003c/span\u003e\u003cspan address=\"10.1176/appi.ajp.2013.12060782\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2013).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKervran, C. \u003cem\u003eet al.\u003c/em\u003e Item Response Theory analyses of DSM-5 substance use disorder criteria in French outpatient addiction clinic participants. How much is craving special? Drug and Alcohol Dependence 212, 108036, doi:\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/j.drugalcdep.2020.108036\u003c/span\u003e\u003cspan address=\"10.1016/j.drugalcdep.2020.108036\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2020).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGauld, C. \u003cem\u003eet al.\u003c/em\u003e The Centrality of Craving in Network Analysis of Five Substance Use Disorders. Drug and Alcohol Dependence, 109828, doi:\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.drugalcdep.2023.109828\u003c/span\u003e\u003cspan address=\"10.1016/j.drugalcdep.2023.109828\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2023).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSayette, M. A. \u003cem\u003eet al.\u003c/em\u003e The measurement of drug craving. Addiction 95 Suppl 2, S189\u0026ndash;210 (2000).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eStone, A. A. \u0026amp; Shiffman, S. Ecological Momentary Assessment (Ema) in Behavioral Medicine. Annals of Behavioral Medicine 16, 199\u0026ndash;202, doi:\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1093/abm/16.3.199\u003c/span\u003e\u003cspan address=\"10.1093/abm/16.3.199\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (1994).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eFatseas, M. \u003cem\u003eet al.\u003c/em\u003e Craving and substance use among patients with alcohol, tobacco, cannabis or heroin addiction: a comparison of substance- and person-specific cues: Cues, craving and substance use. Addiction 110, 1035\u0026ndash;1042, doi:\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1111/add.12882\u003c/span\u003e\u003cspan address=\"10.1111/add.12882\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2015).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCarter, B. L. \u0026amp; Tiffany, S. T. Meta-analysis of cue-reactivity in addiction research. Addiction (Abingdon, England) 94, 327\u0026ndash;340 (1999).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSerre, F., Fatseas, M., Denis, C., Swendsen, J. \u0026amp; Auriacombe, M. Predictors of craving and substance use among patients with alcohol, tobacco, cannabis or opiate addictions: Commonalities and specificities across substances. Addictive Behaviors 83, 123\u0026ndash;129, doi:\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/j.addbeh.2018.01.041\u003c/span\u003e\u003cspan address=\"10.1016/j.addbeh.2018.01.041\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2018).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKoban, L., Wager, T. D. \u0026amp; Kober, H. A neuromarker for drug and food craving distinguishes drug users from non-users. Nature Neuroscience 26, 316\u0026ndash;325, doi:\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1038/s41593-022-01228-w\u003c/span\u003e\u003cspan address=\"10.1038/s41593-022-01228-w\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2023).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eTian, W. \u003cem\u003eet al.\u003c/em\u003e An electroencephalographic signature predicts craving for methamphetamine. Cell Reports Medicine 5 (2024).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSinha, R. \u003cem\u003eet al.\u003c/em\u003e Enhanced negative emotion and alcohol craving, and altered physiological responses following stress and cue exposure in alcohol dependent individuals. Neuropsychopharmacology: Official Publication of the American College of Neuropsychopharmacology 34, 1198\u0026ndash;1208, doi:\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1038/npp.2008.78\u003c/span\u003e\u003cspan address=\"10.1038/npp.2008.78\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2009).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWang, W. \u003cem\u003eet al.\u003c/em\u003e Cue-elicited craving, thalamic activity, and physiological arousal in adult non-dependent drinkers. J Psychiatr Res 116, 74\u0026ndash;82, doi:\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/j.jpsychires.2019.06.005\u003c/span\u003e\u003cspan address=\"10.1016/j.jpsychires.2019.06.005\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2019).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCarter, B. L. \u003cem\u003eet al.\u003c/em\u003e Generalized craving, self-report of arousal, and cue reactivity after brief abstinence. Nicotine Tob Res 11, 823\u0026ndash;826, doi:\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1093/ntr/ntp068\u003c/span\u003e\u003cspan address=\"10.1093/ntr/ntp068\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2009).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eZhao, M. \u003cem\u003eet al.\u003c/em\u003e Cue-induced craving and physiological reactions in recently and long-abstinent heroin-dependent patients. Addict Behav 37, 393\u0026ndash;398, doi:\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/j.addbeh.2011.11.030\u003c/span\u003e\u003cspan address=\"10.1016/j.addbeh.2011.11.030\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2012).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLi, P. \u003cem\u003eet al.\u003c/em\u003e Incubation of alcohol craving during abstinence in patients with alcohol dependence. Addict Biol 20, 513\u0026ndash;522, doi:\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1111/adb.12140\u003c/span\u003e\u003cspan address=\"10.1111/adb.12140\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2015).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCarter, B. L. \u0026amp; Tiffany, S. T. Meta-analysis of cue-reactivity in addiction research. Addiction 94, 327\u0026ndash;340 (1999).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSinha, R. \u003cem\u003eet al.\u003c/em\u003e Enhanced negative emotion and alcohol craving, and altered physiological responses following stress and cue exposure in alcohol dependent individuals. Neuropsychopharmacology 34, 1198\u0026ndash;1208, doi:\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1038/npp.2008.78\u003c/span\u003e\u003cspan address=\"10.1038/npp.2008.78\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2009).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eTrotzke, P., Starcke, K., Pedersen, A. \u0026amp; Brand, M. Cue-induced craving in pathological buying: empirical evidence and clinical implications. Psychosom Med 76, 694\u0026ndash;700, doi:\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1097/psy.0000000000000126\u003c/span\u003e\u003cspan address=\"10.1097/psy.0000000000000126\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2014).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMurphy, A. \u003cem\u003eet al.\u003c/em\u003e Time-dependent neuronal changes associated with craving in opioid dependence: an fMRI study. Addiction Biology 23, 1168\u0026ndash;1178, doi:\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1111/adb.12554\u003c/span\u003e\u003cspan address=\"10.1111/adb.12554\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2018).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eVolkow, N. D. \u003cem\u003eet al.\u003c/em\u003e Cocaine cues and dopamine in dorsal striatum: mechanism of craving in cocaine addiction. J Neurosci 26, 6583\u0026ndash;6588, doi:\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1523/jneurosci.1544-06.2006\u003c/span\u003e\u003cspan address=\"10.1523/jneurosci.1544-06.2006\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2006).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGullapalli, B. T. \u003cem\u003eet al.\u003c/em\u003e On-body Sensing of Cocaine Craving, Euphoria and Drug-Seeking Behavior Using Cardiac and Respiratory Signals. \u003cem\u003eProc. ACM Interact. Mob. Wearable Ubiquitous Technol.\u003c/em\u003e 3, Article 46, doi:\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1145/3328917\u003c/span\u003e\u003cspan address=\"10.1145/3328917\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2019).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eFatseas, M. \u003cem\u003eet al.\u003c/em\u003e Cue-Induced Reactivity, Cortisol Response and Substance Use Outcome in Treated Heroin Dependent Individuals. Biological Psychiatry 70, 720\u0026ndash;727, doi:\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/j.biopsych.2011.05.015\u003c/span\u003e\u003cspan address=\"10.1016/j.biopsych.2011.05.015\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2011).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003evan Lier, H. G. \u003cem\u003eet al.\u003c/em\u003e An ideographic study into physiology, alcohol craving and lapses during one hundred days of daily life monitoring. Addict Behav Rep 16, 100443, doi:\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/j.abrep.2022.100443\u003c/span\u003e\u003cspan address=\"10.1016/j.abrep.2022.100443\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2022).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCarreiro, S. \u003cem\u003eet al.\u003c/em\u003e Wearable sensor-based detection of stress and craving in patients during treatment for substance use disorder: A mixed methods pilot study. Drug and Alcohol Dependence 209, 107929, doi:\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/j.drugalcdep.2020.107929\u003c/span\u003e\u003cspan address=\"10.1016/j.drugalcdep.2020.107929\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2020).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKennedy, A. P. \u003cem\u003eet al.\u003c/em\u003e Continuous in-the-field measurement of heart rate: Correlates of drug use, craving, stress, and mood in polydrug users. Drug and Alcohol Dependence 151, 159\u0026ndash;166, doi:\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/j.drugalcdep.2015.03.024\u003c/span\u003e\u003cspan address=\"10.1016/j.drugalcdep.2015.03.024\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2015).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eChatterjee, S. \u003cem\u003eet al.\u003c/em\u003e mCrave: Continuous Estimation of Craving During Smoking Cessation. \u003cem\u003eProc ACM Int Conf Ubiquitous Comput\u003c/em\u003e 2016, 863\u0026ndash;874, doi:\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1145/2971648.2971672\u003c/span\u003e\u003cspan address=\"10.1145/2971648.2971672\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2016).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBoyer, E. W. \u003cem\u003eet al.\u003c/em\u003e Preliminary efforts directed toward the detection of craving of illicit substances: the iHeal project. J Med Toxicol 8, 5\u0026ndash;9, doi:\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1007/s13181-011-0200-4\u003c/span\u003e\u003cspan address=\"10.1007/s13181-011-0200-4\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2012).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eShi, R. \u003cem\u003eet al.\u003c/em\u003e in \u003cem\u003e2015 IEEE 39th Annual Computer Software and Applications Conference.\u003c/em\u003e 282\u0026ndash;287.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eTackett, J. L., Brandes, C. M., King, K. M. \u0026amp; Markon, K. E. Psychology\u0026rsquo;s replication crisis and clinical psychological science. Annual Review of Clinical Psychology 15, 579\u0026ndash;604, doi:\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1146/annurev-clinpsy-050718-095710\u003c/span\u003e\u003cspan address=\"10.1146/annurev-clinpsy-050718-095710\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2019).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eQuigley, K. S. \u0026amp; Barrett, L. F. Is there consistency and specificity of autonomic changes during emotional episodes? Guidance from the Conceptual Act Theory and psychophysiology. Biol Psychol 98, 82\u0026ndash;94, doi:\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/j.biopsycho.2013.12.013\u003c/span\u003e\u003cspan address=\"10.1016/j.biopsycho.2013.12.013\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2014).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eShiffman, S. \u003cem\u003eet al.\u003c/em\u003e Efficacy of acute administration of nicotine gum in relief of cue-provoked cigarette craving. Psychopharmacology (Berl) 166, 343\u0026ndash;350, doi:\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1007/s00213-002-1338-1\u003c/span\u003e\u003cspan address=\"10.1007/s00213-002-1338-1\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2003).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eNiaura, R. \u003cem\u003eet al.\u003c/em\u003e Comparative efficacy of rapid-release nicotine gum versus nicotine polacrilex gum in relieving smoking cue-provoked craving. Addiction 100, 1720\u0026ndash;1730, doi:\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1111/j.1360-0443.2005.01218.x\u003c/span\u003e\u003cspan address=\"10.1111/j.1360-0443.2005.01218.x\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2005).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHeishman, S. J., Lee, D. C., Taylor, R. C. \u0026amp; Singleton, E. G. Prolonged duration of craving, mood, and autonomic responses elicited by cues and imagery in smokers: Effects of tobacco deprivation and sex. Exp Clin Psychopharmacol 18, 245\u0026ndash;256, doi:\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1037/a0019401\u003c/span\u003e\u003cspan address=\"10.1037/a0019401\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2010).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eDe Calheiros Velozo, J. \u003cem\u003eet al.\u003c/em\u003e Designing daily-life research combining experience sampling method with parallel data. Psychological Medicine, 1\u0026ndash;10, doi:\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1017/S0033291722002367\u003c/span\u003e\u003cspan address=\"10.1017/S0033291722002367\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2022).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eAuriacombe, M. \u003cem\u003eADDICTAQUI - Aquitaine Addiction Cohort: Trajectories of people with addiction (substances or behaviour) in contact with health-care system. Medical, neurobiological, sociological and psychological characteristics. Prospective multicentric, multidisciplinary study\u003c/em\u003e \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e\u0026lt;\u0026thinsp;https://epidemiologie-france.aviesan.fr/en/epidemiology/records/cohorte-addiction-aquitaine-trajectoires-de-personnes-presentant-une-addiction-aux-substances-ou-une-addiction-comportementale-en-contact-avec-le-dispositif-de-soins.-caracteristiques-medicales-neurobiologiques-sociologiques-et-psychologiques.-etude#tab_1\u0026gt;\u003c/span\u003e\u003cspan address=\"http://%3C\u0026thinsp;https://epidemiologie-france.aviesan.fr/en/epidemiology/records/cohorte-addiction-aquitaine-trajectoires-de-personnes-presentant-une-addiction-aux-substances-ou-une-addiction-comportementale-en-contact-avec-le-dispositif-de-soins.-caracteristiques-medicales-neurobiologiques-sociologiques-et-psychologiques.-etude#tab_1%3E\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2017).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eDenis, C. \u003cem\u003eet al.\u003c/em\u003e Usefulness and validity of the modified Addiction Severity Index: A focus on alcohol, drugs, tobacco, and gambling. Substance Abuse 37, 168\u0026ndash;175, doi:\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1080/08897077.2015.1036334\u003c/span\u003e\u003cspan address=\"10.1080/08897077.2015.1036334\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2016).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSheehan, D. V. \u003cem\u003eet al.\u003c/em\u003e The Mini-International Neuropsychiatric Interview (M.I.N.I.): the development and validation of a structured diagnostic psychiatric interview for DSM-IV and ICD-10. \u003cem\u003eThe Journal of Clinical Psychiatry\u003c/em\u003e 59 Suppl 20, 22\u0026ndash;33;quiz 34\u0026ndash;57 (1998).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSchuurmans, A. A. T. \u003cem\u003eet al.\u003c/em\u003e Validity of the Empatica E4 Wristband to Measure Heart Rate Variability (HRV) Parameters: a Comparison to Electrocardiography (ECG). Journal of Medical Systems 44, 190, doi:\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1007/s10916-020-01648-w\u003c/span\u003e\u003cspan address=\"10.1007/s10916-020-01648-w\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2020).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eEnewoldsen, N. \u003cem\u003eAnalysis of the quality of electrodermal activity and heart rate data recorded in daily life over a period of one week with an E4 wristband\u003c/em\u003e, University of Twente, (2016).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003evan Lier, H. G. \u003cem\u003eet al.\u003c/em\u003e A standardized validity assessment protocol for physiological signals from wearable technology: Methodological underpinnings and an application to the E4 biosensor. Behavior Research Methods 52, 607\u0026ndash;629, doi:\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.3758/s13428-019-01263-9\u003c/span\u003e\u003cspan address=\"10.3758/s13428-019-01263-9\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2020).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHeishman, S. J., Saha, S. \u0026amp; Singleton, E. G. Imagery-induced tobacco craving: duration and lack of assessment reactivity bias. Psychol Addict Behav 18, 284\u0026ndash;288, doi:\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1037/0893-164x.18.3.284\u003c/span\u003e\u003cspan address=\"10.1037/0893-164x.18.3.284\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2004).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eZhong, F., Wu, L. Z. \u0026amp; Han, J. S. Suppression of cue-induced heroin craving and cue-reactivity by single-trial transcutaneous electrical nerve stimulation at 2 Hz. Addict Biol 11, 184\u0026ndash;189, doi:\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1111/j.1369-1600.2006.00020.x\u003c/span\u003e\u003cspan address=\"10.1111/j.1369-1600.2006.00020.x\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2006).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHeishman, S. J., Singleton, E. G. \u0026amp; Moolchan, E. T. Tobacco Craving Questionnaire: Reliability and validity of a new multifactorial instrument. Nicotine \u0026amp; Tobacco Research 5, 645\u0026ndash;654, doi:\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1080/1462220031000158681\u003c/span\u003e\u003cspan address=\"10.1080/1462220031000158681\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2003).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eAbdi, H. \u0026amp; Williams, L. J. Principal component analysis. WIREs Computational Statistics 2, 433\u0026ndash;459, doi:\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1002/wics.101\u003c/span\u003e\u003cspan address=\"10.1002/wics.101\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2010).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMigenda, N., M\u0026ouml;ller, R. \u0026amp; Schenck, W. Adaptive dimensionality reduction for neural network-based online principal component analysis. PLOS ONE 16, e0248896, doi:\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1371/journal.pone.0248896\u003c/span\u003e\u003cspan address=\"10.1371/journal.pone.0248896\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2021).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSalcedo-Sanz, S., Rojo-\u0026Aacute;lvarez, J. L., Mart\u0026iacute;nez-Ram\u0026oacute;n, M. \u0026amp; Camps-Valls, G. Support vector machines in engineering: an overview. WIREs Data Mining and Knowledge Discovery 4, 234\u0026ndash;267, doi:\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1002/widm.1125\u003c/span\u003e\u003cspan address=\"10.1002/widm.1125\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2014).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eVibhute, A. D. \u003cem\u003eet al.\u003c/em\u003e in \u003cem\u003eProceedings of the Second International Conference on Computer and Communication Technologies: IC3T 2015, Volume 2.\u003c/em\u003e 413\u0026ndash;427 (Springer).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGraf, R., Zeldovich, M. \u0026amp; Friedrich, S. Comparing linear discriminant analysis and supervised learning algorithms for binary classification\u0026mdash;A method comparison study. Biometrical Journal n/a, doi:\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1002/bimj.202200098\u003c/span\u003e\u003cspan address=\"10.1002/bimj.202200098\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2022).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWong, T. T. \u0026amp; Yeh, P. Y. Reliable Accuracy Estimates from k-Fold Cross Validation. IEEE Transactions on Knowledge and Data Engineering 32, 1586\u0026ndash;1594, doi:\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1109/TKDE.2019.2912815\u003c/span\u003e\u003cspan address=\"10.1109/TKDE.2019.2912815\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2020).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBaratloo, A., Hosseini, M., Negida, A. \u0026amp; El Ashal, G. Part 1: Simple Definition and Calculation of Accuracy, Sensitivity and Specificity. Emerg (Tehran) 3, 48\u0026ndash;49 (2015).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eM\u0026uuml;ller-Putz, G., Scherer, R., Brunner, C., Leeb, R. \u0026amp; Pfurtscheller, G. Better than random: a closer look on BCI results. International journal of bioelectromagnetism 10, 52\u0026ndash;55 (2008).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePrasad, S. \u0026amp; Bruce, L. M. Limitations of Principal Components Analysis for Hyperspectral Target Recognition. IEEE Geoscience and Remote Sensing Letters 5, 625\u0026ndash;629, doi:\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1109/LGRS.2008.2001282\u003c/span\u003e\u003cspan address=\"10.1109/LGRS.2008.2001282\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2008).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eRomine, W., Schroeder, N., Banerjee, T. \u0026amp; Graft, J. Toward Mental Effort Measurement Using Electrodermal Activity Features. Sensors 22, 7363 (2022).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBoucsein, W. \u003cem\u003eet al.\u003c/em\u003e Publication recommendations for electrodermal measurements. Psychophysiology 49, 1017\u0026ndash;1034, doi:\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1111/j.1469-8986.2012.01384.x\u003c/span\u003e\u003cspan address=\"10.1111/j.1469-8986.2012.01384.x\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2012).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePruneti, C., Sacc\u0026ograve;, M., Cosentino, C. \u0026amp; Sgromo, D. Relevance of autonomic arousal in the stress response in psychopathology. J. Basic Appl. Sci 12, 176\u0026ndash;184 (2016).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eQuigley, K. S. \u0026amp; Barrett, L. F. Is there consistency and specificity of autonomic changes during emotional episodes? Guidance from the Conceptual Act Theory and psychophysiology. Biological Psychology 98, 82\u0026ndash;94, doi:\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.biopsycho.2013.12.013\u003c/span\u003e\u003cspan address=\"10.1016/j.biopsycho.2013.12.013\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2014).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSerre, F. \u003cem\u003eet al.\u003c/em\u003e The Craving-Manager smartphone app designed to diagnose substance use/addictive disorders, and manage craving and individual predictors of relapse: a study protocol for a multicenter randomized controlled trial. Front Psychiatry 14, 1143167, doi:\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.3389/fpsyt.2023.1143167\u003c/span\u003e\u003cspan address=\"10.3389/fpsyt.2023.1143167\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2023).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eRoos, C. R. \u003cem\u003eet al.\u003c/em\u003e Randomized trial of mindfulness- and reappraisal-based regulation of craving training among daily cigarette smokers. Psychology of Addictive Behaviors 37, 829\u0026ndash;840, doi:\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1037/adb0000940\u003c/span\u003e\u003cspan address=\"10.1037/adb0000940\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2023).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKober, H. \u0026amp; Mell, M. M. in The Wiley Handbook on the Cognitive Neuroscience of Addiction 195\u0026ndash;218 (2015).\u003c/span\u003e\u003c/li\u003e\u003c/ol\u003e"},{"header":"Unsectioned Figure Details","content":"\u003cdiv category=\"Standard\" float=\"Yes\" id=\"Fig1\" class=\"Figure\"\u003e\u003cdiv category=\"Completeness\" id=\"80\" ruleid=\"MissingFigureImage_01\" status=\"Neutral\" values=\"Fig. 1\" class=\"btn-xs-small Annotation tooltipped\" data-position=\"top\" data-tooltip=\"\"\u003eA\u003c/div\u003e \u003cdiv language=\"En\" class=\"Caption\"\u003e\u003cdiv category=\"Completeness\" id=\"81\" ruleid=\"MissingFigureCitation_01\" status=\"Neutral\" values=\"Fig. 1\" class=\"btn-xs-small Annotation tooltipped\" data-position=\"top\" data-tooltip=\"\"\u003eA\u003c/div\u003e \u003cdiv class=\"CaptionNumber\"\u003eFig. 1\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cdiv class=\"SimplePara\"\u003e\u003cspan type=\"Bold\" class=\"Bold\" name=\"Emphasis\"\u003ePerformance results for LDA, SVM and MDM classifiers: confusion matrix and ROC curve.\u003c/span\u003e\u003c/div\u003e \u003cdiv class=\"SimplePara\"\u003e MDM; (B) SVM; (C) LDA.\u003c/div\u003e \u003cdiv class=\"SimplePara\"\u003e Confusion Matrix describes the performance of classifiers by displaying the counts of true positive (TP), true negative (TN), false positive (FP) and false negative (FN) predictions. The ROC curve represents the relationship between sensitivity and specificity. The AUC of the ROC curve ranges from 0 to 1: a model with an AUC of 1 indicates perfect performance, as it achieves a true positive rate of 1 (100%) and a false positive rate of 0 (0%). An AUC of 0.55 corresponds to a random guess, where the model's performance is no better than chance taking into account the number of trials and class applied here.\u003c/div\u003e \u003c/div\u003e \u003c/div\u003e \u003c/div\u003e\u003cbr/\u003e\u003cdiv category=\"Standard\" float=\"Yes\" id=\"Fig2\" class=\"Figure\"\u003e\u003cdiv category=\"Completeness\" id=\"82\" ruleid=\"MissingFigureImage_01\" status=\"Neutral\" values=\"Fig. 2\" class=\"btn-xs-small Annotation tooltipped\" data-position=\"top\" data-tooltip=\"\"\u003eA\u003c/div\u003e \u003cdiv language=\"En\" class=\"Caption\"\u003e\u003cdiv category=\"Completeness\" id=\"83\" ruleid=\"MissingFigureCitation_01\" status=\"Neutral\" values=\"Fig. 2\" class=\"btn-xs-small Annotation tooltipped\" data-position=\"top\" data-tooltip=\"\"\u003eA\u003c/div\u003e \u003cdiv class=\"CaptionNumber\"\u003eFig. 2\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cdiv class=\"SimplePara\"\u003e\u003cspan type=\"Bold\" class=\"Bold\" name=\"Emphasis\"\u003ePermutation test results for LDA, SVM and MDM classifiers.\u003c/span\u003e\u003c/div\u003e \u003cdiv class=\"SimplePara\"\u003e Permutation tests based on 100 iterations. Null distributions are plotted in blue bars, the 95th percentile of permutations accuracy as orange lines, observed accuracy measures as red lines.\u003c/div\u003e \u003c/div\u003e \u003c/div\u003e \u003c/div\u003e\u003cbr/\u003e\u003cdiv category=\"Standard\" float=\"Yes\" id=\"Fig3\" class=\"Figure\"\u003e\u003cdiv category=\"Completeness\" id=\"84\" ruleid=\"MissingFigureImage_01\" status=\"Neutral\" values=\"Fig. 1\" class=\"btn-xs-small Annotation tooltipped\" data-position=\"top\" data-tooltip=\"\"\u003eA\u003c/div\u003e \u003cdiv language=\"En\" class=\"Caption\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eFig. 1\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cdiv class=\"SimplePara\"\u003ePerformance results for LDA, SVM and MDM classifiers: confusion matrix and ROC curve. (A) MDM; (B) SVM; (C) LDA.\u003c/div\u003e \u003c/div\u003e \u003c/div\u003e \u003c/div\u003e\u003cbr/\u003e\u003cdiv category=\"Standard\" float=\"Yes\" id=\"Fig4\" class=\"Figure\"\u003e\u003cdiv category=\"Completeness\" id=\"85\" ruleid=\"MissingFigureImage_01\" status=\"Neutral\" values=\"Fig. 2\" class=\"btn-xs-small Annotation tooltipped\" data-position=\"top\" data-tooltip=\"\"\u003eA\u003c/div\u003e \u003cdiv language=\"En\" class=\"Caption\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eFig. 2\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cdiv class=\"SimplePara\"\u003ePermutation test results for LDA, SVM and MDM classifiers.\u003c/div\u003e \u003c/div\u003e \u003c/div\u003e \u003c/div\u003e\u003cbr/\u003e"},{"header":"Unsectioned Paragraphs","content":"\u003cp\u003e\u003cb\u003eAuthors\u003c/b\u003e :\u003c/p\u003e\u003cp\u003eMarc Auriacombe\u003c/p\u003e\u003cp\u003ee-mail: [email protected]\u003c/p\u003e\u003cp\u003eP\u0026ocirc;le Addictologie (SANPSY), CHCP, 121 rue de la B\u0026eacute;chade, 33076 Bordeaux Cedex, France\u003c/p\u003e\u003cp\u003eTel: +33 556 561 738\u003c/p\u003e\u003cp\u003eSupplementary Figures (2) and Supplementary Tables\u0026nbsp;(2)\u003c/p\u003e\u003cp\u003e\u003cb\u003eABSTRACT (135 words)\u003c/b\u003e\u003c/p\u003e\u003cp\u003eCraving variations have been shown to be prospectively associated to relapse and are a target for treatment in substance use disorders (SUD). We show that a set of physiological features captured in daily life can distinguish within-day craving from no-craving states, among SUD participants in a 14-day multimodal study. Craving was assessed Signal-contingent and event-contingent. Continuous recording of physiological signals was conducted daily. After labeling samples as \u0026ldquo;craving\u0026rdquo; and \u0026ldquo;no-craving\u0026rdquo; using a double-check verification, features of each signal were extracted, normalized and standardized and a principal component analysis was applied. Using Support Vector Machines and Linear Discriminant Analysis classifiers, physiological signals allowed to discriminate craving from no-craving episodes with a high cross-validation accuracy (73.6% and 72.2% respectively). This finding is an important breakthrough that paves the way to a biomarker of craving in daily life.\u003c/p\u003e\u003cp\u003e\u003cb\u003eFIGURES LEGENDS\u003c/b\u003e\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003cp\u003e\u003cstrong\u003eWords count\u003c/strong\u003e \u003cp\u003e3000\u003c/p\u003e\u003c/p\u003e\u003cp\u003e\u003cstrong\u003eFigure/Table\u003c/strong\u003e \u003cp\u003e2 tables, 2 fig\u003c/p\u003e\u003c/p\u003e\u003cp\u003e\u003cstrong\u003eReferences\u003c/strong\u003e \u003cp\u003e63\u003c/p\u003e\u003c/p\u003e\u003cp\u003e\u003csup\u003e1\u003c/sup\u003eMood disorders include major depressive episode, dysthymia and hypomania; anxiety disorders include social phobia, generalized anxiety disorder and panic disorder. \u003csup\u003e2\u003c/sup\u003eFrequencies, percentages and means are based on the total number of valid electronic surveys during the assessment period. ASI\u0026thinsp;=\u0026thinsp;Addiction Severity Index; MINI\u0026thinsp;=\u0026thinsp;Mini International Neuropsychiatric Disorder; EMA\u0026thinsp;=\u0026thinsp;Ecological Momentary Assessment.\u003c/p\u003e\u003cp\u003eUnit of all EDA features was microSiemens (\u0026micro;S); second (s) for HRV SDANNI5, HRV SDANN5, HRV SDANN2 and HRV SDANN1 features; HRV DFAα2 had no unit. Bold values indicate statistically significant P values.\u003c/p\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":true,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true},"keywords":"Craving, Substance Use Disorder, Biomarker, Ecological Momentary Assessment, Wearable Sensor, Machine Learning, Classification, Heart Rate, Electrodermal Activity, Accelerometer, Skin Temperature","lastPublishedDoi":"10.21203/rs.3.rs-4432651/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-4432651/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"Craving variations have been shown to be prospectively associated to relapse and are a target for treatment in substance use disorders (SUD). We show that a set of physiological features captured in daily life can distinguish within-day craving from no-craving states, among SUD participants in a 14-day multimodal study. Craving was assessed Signal-contingent and event-contingent. Continuous recording of physiological signals was conducted daily. After labeling samples as “craving” and “no-craving” using a double-check verification, features of each signal were extracted, normalized and standardized and a principal component analysis was applied. Using Support Vector Machines and Linear Discriminant Analysis classifiers, physiological signals allowed to discriminate craving from no-craving episodes with a high cross-validation accuracy (73.6% and 72.2% respectively). This finding is an important breakthrough that paves the way to a biomarker of craving in daily life.","manuscriptTitle":"The Autonomic Craving Signature: physiological signals as a daily-life biomarker of craving","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2024-12-17 13:42:24","doi":"10.21203/rs.3.rs-4432651/v1","editorialEvents":[{"type":"communityComments","content":0}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"60726ee6-aaba-448a-9c99-b6bcdbcaa118","owner":[],"postedDate":"December 17th, 2024","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[{"id":37922554,"name":"Health sciences/Biomarkers/Diagnostic markers"},{"id":37922555,"name":"Health sciences/Diseases/Psychiatric disorders/Addiction"},{"id":37922556,"name":"Health sciences/Biomarkers/Prognostic markers"}],"tags":[],"updatedAt":"2025-01-08T15:52:18+00:00","versionOfRecord":[],"versionCreatedAt":"2024-12-17 13:42:24","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-4432651","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-4432651","identity":"rs-4432651","version":["v1"]},"buildId":"qtupq5eGEP_6zYnWcrvyt","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}