The Spatial Distributions of Neuroimaging-guided Personalized Targets for TMS in Depression

116 

Abstract Submission 

Gai Kong¹, Lijiang Wei², Sirui Wang¹, Jijun Wang¹, Chaozhe Zhu², Yingying Tang¹

¹Shanghai Mental Health Center, Shanghai Jiaotong University School of Medicine, Shanghai, China, ²Beijing Normal University, State Key Laboratory of Cognitive Neuroscience and Learning, Beijing, China

Gai Kong  
Shanghai Mental Health Center, Shanghai Jiaotong University School of Medicine
Shanghai, China

Lijiang Wei  
Beijing Normal University, State Key Laboratory of Cognitive Neuroscience and Learning
Beijing, China

Sirui Wang  
Shanghai Mental Health Center, Shanghai Jiaotong University School of Medicine
Shanghai, China

Jijun Wang  
Shanghai Mental Health Center, Shanghai Jiaotong University School of Medicine
Shanghai, China

Chaozhe Zhu  
Beijing Normal University, State Key Laboratory of Cognitive Neuroscience and Learning
Beijing, China

Yingying Tang  
Shanghai Mental Health Center, Shanghai Jiaotong University School of Medicine
Shanghai, China

Repetitive transcranial magnetic stimulation (rTMS) over the left dorsolateral prefrontal cortex (DLPFC) has been an effective treatment for major depressive disorder (MDD) (1). Recent studies have further developed a precision rTMS strategy with personalized targets guided by personalized functional connectivity (FC) between the DLPFC and the subgenual cingulate cortex (sgACC) to improve the antidepressant effects in MDD (2). As depression is a disorder with distinct symptom domains, whether the spatial distribution of personalized DLPFC targets are related to these symptom domains remains uncertain. We thus explore whether the personalized DLPFC targets can be clustered into distinct subtypes and exam their association with clinical characteristics.

We recruited a total of 133 MDD patients from two centers (Shanghai Mental Health Center (SMHC) and Suzhou Guangji Hospital (SZGJ) with a diagnosis of MDD using the Diagnostic and Statistical Manual of Mental Disorders (DSM-4) criteria. All MDD patients completed the 17-item Hamilton Depression Scale (HAMD-17), Hamilton Anxiety Scale (HAMA), and Montgomery-Asberg Depression Rating Scale (MADRS) assessments. Resting-state functional MRI (fMRI) data were acquired and preprocessed. After the quality control of fMRI images, 120 MDD were included in the following analysis. We calculated FC between each voxel within the left DLPFC and a sgACC-based seed, as described in detail in our previous study (3), and determined a personalized DLPFC target with a maximum negative DLPFC-sgACC FC. Then, the k-means clustering method was applied to cluster MDD subgroups based on the spatial distributions of their personalized targets. Between-subgroup comparisons were performed for the demographic and clinical characteristics.

We obtained two distinct clusters of personalized targets in MDD patients, as shown in Fig 1. One subgroup had personalized targets over the anterior part of the left DLPFC (in red in Fig.1A, termed as the anterior subgroup), and the other subgroup had personalized targets over the posterior part (in blue in Fig.1A, termed as the posterior subgroup). The anterior subgroup, constituting the majority (73.3%), were closer to the group-average DLPFC target with a median distance of 12.5mm. In comparison, those in the posterior subgroup (26.7%) had a median distance of 50.5mm. There were no significant differences in gender, education, or illness duration between these two subgroups (p>0.05). However, we observed significant between-subgroup differences in various anxiety indicators. The posterior subgroup showed higher HAMD-17 total scores (p=0.034) and HAMA total scores (p=0.008) than the anterior subgroup. Factor scores of anxiety/somatization (p=0.001) in HAMD-17 and factor scores of somatic (p=0.030) and psychic anxiety (p=0.020) in HAMA were higher in the posterior subgroup than the anterior subgroup, whereas all other factors did not differ between two subgroups.

In the present study, we identified two MDD subgroups with distinct spatial distributions of their personalized DLPFC targets based on the DLPFC-sgACC FC. Most MDD patients had anterior DLPFC targets close to the group-average DLPFC target, consistent with the clinical rTMS efficacy in MDD. More importantly, we identified another posterior MDD subgroup, who had more posterior DLPFC targets and presented severe anxiety symptoms. The heterogeneity of personalized target distributions within the DLPFC highlights the importance of individualized rTMS intervention in MDD.

TMS ¹

Psychiatric (eg. Depression, Anxiety, Schizophrenia) ²

Connectivity (eg. functional, effective, structural)

fMRI Connectivity and Network Modeling

BOLD fMRI

FUNCTIONAL MRI

Transcranial Magnetic Stimulation (TMS)

Other - Major depressive disorder; Functional connectivity, Personalized target; Group-average target; Subtypes