Perturbation in silico Indicates the Crucial Role of NAc in Abstinence from Meth Addiction
Poster No:
458
Submission Type:
Abstract Submission
Authors:
Jiaqi Zhang1, Yaoyao Du2, Jun Liu2, Tianzi Jiang1
Institutions:
1Brainnetome Center, Institute of Automation, Chinese Academy of Sciences, Beijing, Beijing, 2Department of Radiology, Second Xiangya Hospital of Central South University, Changsha, Hunan, China, Changsha, Hunan
First Author:
Jiaqi Zhang
Brainnetome Center, Institute of Automation, Chinese Academy of Sciences
Beijing, Beijing
Brainnetome Center, Institute of Automation, Chinese Academy of Sciences
Beijing, Beijing
Co-Author(s):
Yaoyao Du
Department of Radiology, Second Xiangya Hospital of Central South University, Changsha, Hunan, China
Changsha, Hunan
Department of Radiology, Second Xiangya Hospital of Central South University, Changsha, Hunan, China
Changsha, Hunan
Jun Liu
Department of Radiology, Second Xiangya Hospital of Central South University, Changsha, Hunan, China
Changsha, Hunan
Department of Radiology, Second Xiangya Hospital of Central South University, Changsha, Hunan, China
Changsha, Hunan
Tianzi Jiang
Brainnetome Center, Institute of Automation, Chinese Academy of Sciences
Beijing, Beijing
Brainnetome Center, Institute of Automation, Chinese Academy of Sciences
Beijing, Beijing
Introduction:
Methamphetamine (MA) addiction imposes a significant public health burden and various intervention methods have limited therapeutic efficacy. To explore an innovative and effective intervention strategy, we carried out a one-year longitudinal study for 62 MA users. MRI scans and MA craving questionnaires (MCQs) were collected before (MA1) and after long-term abstinence (MA2). We established the whole-brain computational models to investigate crucial regions associated with abstinence from methamphetamine addiction. To the best of our knowledge, the application of this approach to investigate MA addiction is currently lacking in the existing studies. Our work may serve as a reference for the practical application of this approach in future addiction therapeutics.
Methods:
Probabilistic Metastable Substates (PMS)
Brain activity is not static, so to capture its metastable nature, we employed the PMS method, which is consistent with previous literature. By leveraging the PMS, we were able to identify and characterize recurrent substates, enabling us to quantitatively assess the brain's state for subsequent model fitting and in silico perturbations. First, we obtained phase coherence matrices (dFC). Then, we extracted the leading eigenvectors to capture the temporal evolution of dFC. Lastly, we applied the k-means clustering algorithm to identify the PMS space. In our subsequent analysis, we selected a cluster solution with k = 8, as it is the smallest number of clusters that could effectively distinguish the three distinct groups of subjects.
Whole-Brain Computational Model
The brain can be conceptualized as a complex network, with nodes representing specific brain regions and edges representing the structural connectivity obtained via diffusion MRI. Here, each of the N = 246 brain regions was modeled as the Hopf model, which is a canonical model for characterizing features of brain dynamics and studying perturbation dynamics. To simulate the dynamics of the entire brain, we coupled each Hopf model with the empirical structure connectivity and adjusted the strength of these connections using a global coupling parameter. This coupling parameter played a crucial role in scaling the constraints imposed by the structural connectivity, allowing for a more accurate representation of the brain's dynamics.
Before conducting a model fitting to determine the optimal model, it is essential to establish a measurement that can quantitatively assess the distance between simulated data and empirical data. In alignment with prior literature, we adopted the symmetrized Kullback-Leibler (KL) distance as our distance metric.
Brain activity is not static, so to capture its metastable nature, we employed the PMS method, which is consistent with previous literature. By leveraging the PMS, we were able to identify and characterize recurrent substates, enabling us to quantitatively assess the brain's state for subsequent model fitting and in silico perturbations. First, we obtained phase coherence matrices (dFC). Then, we extracted the leading eigenvectors to capture the temporal evolution of dFC. Lastly, we applied the k-means clustering algorithm to identify the PMS space. In our subsequent analysis, we selected a cluster solution with k = 8, as it is the smallest number of clusters that could effectively distinguish the three distinct groups of subjects.
Whole-Brain Computational Model
The brain can be conceptualized as a complex network, with nodes representing specific brain regions and edges representing the structural connectivity obtained via diffusion MRI. Here, each of the N = 246 brain regions was modeled as the Hopf model, which is a canonical model for characterizing features of brain dynamics and studying perturbation dynamics. To simulate the dynamics of the entire brain, we coupled each Hopf model with the empirical structure connectivity and adjusted the strength of these connections using a global coupling parameter. This coupling parameter played a crucial role in scaling the constraints imposed by the structural connectivity, allowing for a more accurate representation of the brain's dynamics.
Before conducting a model fitting to determine the optimal model, it is essential to establish a measurement that can quantitatively assess the distance between simulated data and empirical data. In alignment with prior literature, we adopted the symmetrized Kullback-Leibler (KL) distance as our distance metric.
Results:
1. We quantitatively characterized the brain states and identified the significant brain states related to craving scores, indicating distinct spatiotemporal functional patterns among the three cohorts, namely the MA1, MA2, and HC.
2. We constructed whole-brain computational models and identified the nucleus accumbens (NAc) as a potential intervention target through in silico perturbations for each region across the whole brain, which is aligned with previous clinical experiments. Our observations of the underlying functional alterations also indicated reduced craving scores and improved cognitive functions.
3. We observed that the impact of perturbing other brain regions across the whole brain was related to their connectivity with NAc in the context of MA1, suggesting the ability of NAc to identify other potential intervention targets. Notably, this relationship was absent within the context of MA2, suggesting a potential decrease in the abstinence influence of the nucleus accumbens, or possibly that NAc has recovered during long-term abstinence.
2. We constructed whole-brain computational models and identified the nucleus accumbens (NAc) as a potential intervention target through in silico perturbations for each region across the whole brain, which is aligned with previous clinical experiments. Our observations of the underlying functional alterations also indicated reduced craving scores and improved cognitive functions.
3. We observed that the impact of perturbing other brain regions across the whole brain was related to their connectivity with NAc in the context of MA1, suggesting the ability of NAc to identify other potential intervention targets. Notably, this relationship was absent within the context of MA2, suggesting a potential decrease in the abstinence influence of the nucleus accumbens, or possibly that NAc has recovered during long-term abstinence.
Conclusions:
In summary, our findings not only offer a new perspective highlighting the central role of NAc in abstinence from methamphetamine addiction but also offer potential avenues for advanced translational interventions in addiction therapy.
Disorders of the Nervous System:
Psychiatric (eg. Depression, Anxiety, Schizophrenia) 1
Modeling and Analysis Methods:
Classification and Predictive Modeling 2
fMRI Connectivity and Network Modeling
Novel Imaging Acquisition Methods:
BOLD fMRI
Diffusion MRI
Keywords:
Addictions
Atlasing
Data analysis
FUNCTIONAL MRI
Modeling
MRI
Psychiatric Disorders
Therapy
Other - whole-brain computational model
1|2Indicates the priority used for review
·Overall analysis pipeline
Provide references using author date format
Deco, G., et al. (2017). "The dynamics of resting fluctuations in the brain: metastability and its dynamical cortical core." Scientific reports 7(1): 3095.
Deco, G., et al. (2019). "Awakening: Predicting external stimulation to force transitions between different brain states." Proceedings of the National Academy of Sciences 116(36): 18088-18097.
Du, Y., et al. (2022). "Changes in ALFF and ReHo values in methamphetamine abstinent individuals based on the Harvard‐Oxford atlas: A longitudinal resting‐state fMRI study." Addiction Biology 27(1): e13080.
Fan, L., et al. (2016). "The human brainnetome atlas: a new brain atlas based on connectional architecture." Cerebral cortex 26(8): 3508-3526.
Fox, M. D., et al. (2014). "Resting-state networks link invasive and noninvasive brain stimulation across diverse psychiatric and neurological diseases." Proceedings of the National Academy of Sciences 111(41): E4367-E4375.