Ketamine treatment for MDD based on digital twin brain model
Poster No:
2626
Submission Type:
Abstract Submission
Authors:
Ruonan Zheng1, Junyang Zhang1, Chen Wang1, Zhichao Wang1, Yang He1, Yanling Zhou2, Yuping Ning2, Zhang Yu3, Tianzi Jiang4
Institutions:
1Zhejiang Lab, Hangzhou, Zhejiang, 2The Affiliated Brain Hospital of Guangzhou Medical University, Guangzhou, Guangdong, 3Zhejiang Lab, Hang Zhou, Zhejiang, 4Brainnetome Center, Institute of Automation, Chinese Academy of Sciences, Beijing, Beijing
First Author:
Co-Author(s):
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:
Ketamine has gained popularity as an antidepressant due to its rapid and effective action. However, the mechanisms by which ketamine improves emotional states remain unclear. By leveraging the unique biological characteristics of ketamine as an N-methyl-D-aspartate receptors (NMDAR) antagonist, we have utilized a digital twin brain model [1] to investigate ketamine's influence on brain circuitry. This approach has enabled us to explore the underlying neural mechanism of ketamine from the perspective of excitatory/inhibitory balance.
Methods:
We recruited 40 patients with major depressive disorder (MDD) and 38 normal controls (NC). After quality control and excluding missing data, this study used 31 MDD patients who underwent ketamine treatment (6 sub-anaesthetic doses of ketamine by intravenous infusion over 12 days), including 19 effective (MDDa) and 12 ineffective (MDDb) patients according to Montgomery-Asberg Depression Rating Scale. All subjects underwent structural (voxel-size=1mm isotropic) and functional MRI (voxel-size=3.4mm, thickness=4mm, TR=2s) data using 3.0 T Philips Achieva scanner. More details regarding the study design and MRI data acquisition have been described in previous studies [2, 3].
We simulated the spontaneous brain activity at the whole-brain level using a network model with each node representing a brain region and the links between nodes representing white matter connections. We used the dynamic mean field (DMF) model [4] to simulate the neural activity in each brain region. Additionally, ketamine binds directly to NMDAR, which prevents binding with upstream excitatory neurotransmitters and releasing excitatory/inhibitory neurotransmitters. Based on which, we modified the model (Fig. 1D).We hypothesized similar neural dynamics within each functional network and used the same set of parameters for each network.
We simulated functional dynamics of each brain network using the digital twin brain model and hemodynamic model (Fig. 1) and evaluated the model by comparing the real and simulated functional connectivity matrix using Pearson correlation coefficient (PCC). We predicted the effectiveness of ketamine treatment by measuring the changes between the simulation models before and after and the real states using a grid search on the parameter space of Jexc/Jinh.
We simulated the spontaneous brain activity at the whole-brain level using a network model with each node representing a brain region and the links between nodes representing white matter connections. We used the dynamic mean field (DMF) model [4] to simulate the neural activity in each brain region. Additionally, ketamine binds directly to NMDAR, which prevents binding with upstream excitatory neurotransmitters and releasing excitatory/inhibitory neurotransmitters. Based on which, we modified the model (Fig. 1D).We hypothesized similar neural dynamics within each functional network and used the same set of parameters for each network.
We simulated functional dynamics of each brain network using the digital twin brain model and hemodynamic model (Fig. 1) and evaluated the model by comparing the real and simulated functional connectivity matrix using Pearson correlation coefficient (PCC). We predicted the effectiveness of ketamine treatment by measuring the changes between the simulation models before and after and the real states using a grid search on the parameter space of Jexc/Jinh.
Results:
Using different functional network as regions of interest, we observed significant differences in the frontoparietal network (FPN) between MDDa and NC (Fig. 2A a, b). The FPN in NC models leaned towards excitation, whereas the FPN in MDDa models exhibited a bias towards inhibition, resembling the real state. Post-treatment modeling of MDDa's FPN revealed a shift closer to state I(Fig. 2d), indicating a reduction in NMDAR synaptic coupling among interneurons. Simulations of Jexc/Jinh pairs for the FPN in MDDa (Fig. 2B) suggested that inhibiting only the interneurons aligns closest to a healthy state.
For MDDb, we found no significant pre-treatment inhibition bias or post-treatment inhibition of interneurons in the FPN model (Fig. 2C).
For MDDb, we found no significant pre-treatment inhibition bias or post-treatment inhibition of interneurons in the FPN model (Fig. 2C).
Conclusions:
In this study, we used the digital twin brain model to investigate the circuit-level mechanism of antidepressant action of ketamine. Our findings suggest that in NC, the FPN is typically excited, but in depression, it tends to be inhibited. Specifically, in MDDa, ketamine appears to inhibit FPN interneurons, eliciting antidepressant responses. Conversely, the results from MDDb provide a contrasting perspective for the ineffective response. Our study provides a computational perspective of depression symptoms and the neural mechanism of ketamine treatment.
Modeling and Analysis Methods:
Connectivity (eg. functional, effective, structural)
fMRI Connectivity and Network Modeling 2
Physiology, Metabolism and Neurotransmission :
Pharmacology and Neurotransmission 1
Keywords:
Computational Neuroscience
DISORDERS
FUNCTIONAL MRI
MRI
Pharmacotherapy
WHITE MATTER IMAGING - DTI, HARDI, DSI, ETC
1|2Indicates the priority used for review
Provide references using author date format
[2] Zhou, Y.-L. et al. (2020). Volumetric changes in subcortical structures following repeated ketamine treatment in patients with major depressive disorder: a longitudinal analysis. Transl Psychiatry 10, 1–9.
[3] Wang, C. et al. (2019). Association between depression subtypes and response to repeated-dose intravenous ketamine. Acta Psychiatrica Scandinavica 140, 446–457.
[4] Deco, G. et al. (2014). How Local Excitation-Inhibition Ratio Impacts the Whole Brain Dynamics. Journal of Neuroscience 34, 7886–7898.