Transdiagnostic connectome-based prediction of craving

Garrison, Kathleen A.; Sinha, Rajita; Potenza, Marc N.; Gao, Siyuan; Liang, Qinghao; Lacadie, Cheryl; Scheinost, Dustin

doi:10.1101/2021.05.21.21257620

Cited by 5 publications

(11 citation statements)

References 78 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…We focused on the neutral condition in this study. Functional connectivity under this condition contributed the most to predicting individual differences in craving (28). Furthermore, basal craving was strongly associated with substance use (e.g., days of alcohol use) during relapse (17,35).…”

Section: Participantsmentioning

confidence: 91%

“…Participants in both datasets first completed a baseline condition (no sensory input) followed by a task condition where a neutral/relaxed state was induced via distinct sensory modalities (Supplementary Materials). Participants listened to a personalized audio script in the imagery dataset to induce relaxation (28). For the visual stimuli dataset, individuals viewed a series of neutral/relaxing images (33,34).…”

Section: Participantsmentioning

confidence: 99%

“…Recent studies have identified brain activation (27) and functional connectivity patterns (28,29) distinctively associated with craving. But little is known about how these brain regions dynamically interact across time to contribute to craving.…”

Section: Introductionmentioning

confidence: 99%

“…But little is known about how these brain regions dynamically interact across time to contribute to craving. For instance, a brain-based craving network significantly predicted craving in a transdiagnostic sample, highlighting contributions of the default mode, salience, and subcortical networks (28). However, since the employed approach required calculating functional connectivity over a minutes-long period, how the craving network's temporal fluctuations give rise to this affective state remains elusive.…”

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

Network state dynamics underpin craving in a transdiagnostic population

Ye,

Garrison,

Lacadie

et al. 2023

Preprint

View full text Add to dashboard Cite

Emerging fMRI brain dynamic methods present a unique opportunity to capture how brain region interactions across time give rise to evolving affective and motivational states. As the unfolding experience and regulation of affective states affect psychopathology and well-being, it is important to elucidate their underlying time-varying brain responses. Here, we developed a novel framework to identify network states specific to an affective state of interest and examine how their instantaneous engagement contributed to its experience. This framework investigated network state dynamics underlying craving, a clinically meaningful and changeable state. In a transdiagnostic sample of healthy controls and individuals diagnosed with or at risk for craving-related disorders (N=252), we utilized connectome-based predictive modeling (CPM) to identify craving-predictive edges. An edge-centric timeseries approach was leveraged to quantify the instantaneous engagement of the craving-positive and craving-negative networks during independent scan runs. Individuals with higher craving persisted longer in a craving-positive network state while dwelling less in a craving-negative network state. We replicated the latter results externally in an independent group of healthy controls and individuals with alcohol use disorder exposed to different stimuli during the scan (N=173). The associations between craving and network state dynamics can still be consistently observed even when craving-predictive edges were instead identified in the replication dataset. These robust findings suggest that variations in craving-specific network state recruitment underpin individual differences in craving. Our framework additionally presents a new avenue to explore how the moment-to-moment engagement of behaviorally meaningful network states supports our changing affective experiences.

show abstract

Section: Participantsmentioning

confidence: 91%

Section: Participantsmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Network state dynamics underpin craving in a transdiagnostic population

Ye,

Garrison,

Lacadie

et al. 2023

Preprint

View full text Add to dashboard Cite

show abstract

“…Additionally, although IGD was not included in this study, a craving network identified across conditions suggests that craving may involve difficulties redirecting attention from internal self-related processes as linked to the DMN (Garrison et al. 2021 ). Given the relevance of craving to IGD, additional research should explore this possible role for the DMN in IGD.…”

Section: Independent Component Analysis In Internet Gaming Disordermentioning

confidence: 99%

Independent component analysis for internet gaming disorder

Mestre‐Bach

Granero

Fernández‐Aranda

et al. 2023

Dialogues in Clinical Neuroscience

View full text Add to dashboard Cite

Introduction: There is a growing interest in the study of the neurobiological correlates of internet gaming disorder (IGD), and new techniques are beginning to be implemented for this purpose, such as independent component analysis (ICA). Aims: The present narrative review aimed to explore the studies that had used ICA for the study of the different brain networks possibly associated with IGD. Methods: We specifically focussed on three of the main networks: default-mode network, executive-control and salience networks. Results: Most studies have identified alterations in these three brain networks in individuals with IGD, which may be involved in the development and maintenance of this disorder. Conclusion: More studies are needed to deepen an understanding of the specific role of each in the symptomatology and treatment of IGD.

show abstract

The effects of data leakage on connectome-based machine learning models

Rosenblatt¹,

Tejavibulya²,

Jiang³

et al. 2023

Preprint

Self Cite

View full text Add to dashboard Cite

Predictive modeling has now become a central technique in neuroimaging to identify complex brain-behavior relationships and test their generalizability to unseen data. However, data leakage, which unintentionally breaches the separation between data used to train and test the model, undermines the validity of predictive models. Although previous literature suggests that leakage is generally pervasive in machine learning, few studies have empirically evaluated the effects of leakage in neuroimaging data. Here, using over 500 different pipelines spanning four large neuroimaging datasets and three phenotypes, we evaluated six forms of leakage fitting into three broad categories: feature selection, covariate correction, and lack of independence between subjects. As expected, leakage via feature selection and repeated subjects drastically inflated prediction performance. Notably, other forms of leakage had only minor effects (e.g., leaky site correction) or even decreased prediction performance (e.g., leaky covariate regression). In some cases, leakage affected not only prediction performance, but also model coefficients, and thus neurobiological interpretations. Overall, our results illustrate the variable effects of leakage on prediction pipelines and underscore the importance of avoiding data leakage to improve the validity and reproducibility of predictive modeling.

show abstract

Transdiagnostic connectome-based prediction of craving

Cited by 5 publications

References 78 publications

Network state dynamics underpin craving in a transdiagnostic population

Network state dynamics underpin craving in a transdiagnostic population

Independent component analysis for internet gaming disorder

The effects of data leakage on connectome-based machine learning models

Contact Info

Product

Resources

About