Systematic organization of cortical thickness co-alterations in substance use disorders.

540 

Abstract Submission 

Sofie Valk¹, Foivos Georgiadis², Meike Hettwer³, Sophia Thomopoulos⁴, Paul Thompson⁵, Scott Mackey⁶, Patricia Conrod⁷, Hugh Garavan⁸, Clara Moreau⁹, Boris Bernhardt¹⁰, Matthias Kirschner¹¹, ENIGMA Addiction Working Group¹²

¹Max Planck Institute for Human Cognitive and Brain Sciences, Leipzig, Germany, ²University of Zurich, Zurich, Switzerland, ³Forschungszentrum Jülich, Jülich, North Rhine-Westphalia, ⁴USC, Marina del Rey, CA, ⁵USC, Marina Del Rey, CA, ⁶The University of Vermont, Burlington, VT, ⁷University of Montreal, Montreal, Quebec, ⁸University of Vermont, Burlington, VT, ⁹University of Southern California, Los Angeles, CA, ¹⁰Montreal Neurological Institute and Hospital, Montreal, Quebec, ¹¹Geneva University Hospitals (HUG), Thonex, Geneva, ¹²UCLA, Burlington, Vermont

Sofie Valk  
Max Planck Institute for Human Cognitive and Brain Sciences
Leipzig, Germany

Foivos Georgiadis  
University of Zurich
Zurich, Switzerland

Meike Hettwer  
Forschungszentrum Jülich
Jülich, North Rhine-Westphalia

Sophia Thomopoulos  
USC
Marina del Rey, CA

Paul Thompson  
USC
Marina Del Rey, CA

Scott Mackey  
The University of Vermont
Burlington, VT

Patricia Conrod  
University of Montreal
Montreal, Quebec

Hugh Garavan  
University of Vermont
Burlington, VT

Clara Moreau, PhD  
University of Southern California
Los Angeles, CA

Boris Bernhardt  
Montreal Neurological Institute and Hospital
Montreal, Quebec

Matthias Kirschner  
Geneva University Hospitals (HUG)
Thonex, Geneva

ENIGMA Addiction Working Group  
UCLA
Burlington, Vermont

Substance use disorders are highly comorbid with other neuropsychiatric disorders and share with them widespread structural brain alterations (Eaton, Rodriguez-Seijas, Carragher, & Krueger, 2015; Mackey et al., 2019; Reich-Erkelenz, Schmitt, & Falkai, 2015). Previous work has shown that structural alterations across neuropsychiatric conditions are organized in a systematic fashion, linked to the intrinsic organization of the human connectome (Hettwer et al., 2022). However, to what extent similar coordinated co-alteration effects extend to substance use disorders remains to be established. Here, we investigated substance use co-alteration networks (pathological structural covariance) to elucidate concordant macroscale principles of illness and substance-use effects across the cortex.

We derived maps of case-control differences in cortical thickness from 2,847 patients with six substance use disorders (SUDs: alcohol, amphetamines, cocaine, opioids, cannabis, and nicotine; Fig. 1A) and 1,951 non-affected individuals from the Enhancing NeuroImaging Genetics through Meta-Analysis (ENIGMA) Addiction Working Group (Thompson et al., 2020; Mackey, 2020). We investigated substance use co-alteration networks via inter-regional SUD association with cortical thickness (Fig. 1B). Hub regions are regions that show high co-alteration profiles across disorders and were computed as the sum of top 20% covariance strengths in each region. Next, we identified potential substance use co-alteration epicenters, i.e., regions whose normative connectivity profiles (based on data from the Human Connectome Project, (Van Essen et al., 2013)) are correlated with substance use co-alteration hubs. Last, we compared the effects of substance use disorder-related co-alterations and neuropsychiatric disease-related co-alterations based on prior work (Hettwer et al., 2022). To do so, we probed regional embedding and inter-relational organizational axes (Vos de Wael et al., 2020).

We identified organizational principles of shared cortical thickness alterations across SUDs (SUDcov). Co-alteration network hubs followed normative functional (r=0.527, pspin<0.05) and structural (r=0.314, pspin<0.05) connectivity patterns (Fig. 1C). Functional epicenters included temporal regions, precuneus, right frontal cortex and hippocampus and amygdala, and structural epicenters were found in posterior/mid cingulate, preSMA, right sensory-motor regions, and the thalamus (Fig. 1C). We observed a high overlap between SUDcov and structural covariance patterns of cortical thickness, except in inferior temporal areas and mOFC (Fig. 2A and B). Similar associations were observed when comparing SUDcov with neuropsychiatric cross-disorder impact (Fig. 2C). Studying low-dimensional organizations of SUDcov, the primary gradient of substance use disorder differentiated OFC, inferior temporal regions, and anterior cingulate from the rest of the cortex and mirrored the second covariance and neuropsychiatric co-alteration gradients (pspin<0.05). The second SUDcov gradient differentiated parietal and OFC areas from the rest of the cortex (Fig 2DE). Overall, SUD co-alterations varied stronger along G2cov (median r=0.25) relative to G1cov (median r=0.15), and the reverse was observed in neuropsychiatric co-alteration gradients (G1 median r=0.33, G2 median r=0.15). Hierarchical clustering of SUD and neuropsychiatric co-alteration patterns further underlined this differentiation (Fig. 2F).

The shared association of SUDs with cortical thickness covaries in a network-like fashion. These patterns partly mirror those of normative structural covariance, underscoring the differentiation between inferior/paralimbic and superior regions of the cerebral cortex by SUDs as probed by a gradient framework (Valk et al., 2020). We observed differentiation between SUDs versus neuropsychiatric co-alteration networks, possibly related to a differentiation between co-morbidity and neurodevelopment (Hettwer et al., 2022).

Psychiatric (eg. Depression, Anxiety, Schizophrenia) ¹

Cortical Anatomy and Brain Mapping ²

Addictions

Computational Neuroscience

Psychiatric Disorders

Other - Disease co-alterations