Network localization of gray matter atrophy in addiction
Presented During:
Wednesday, June 26, 2024: 11:30 AM - 12:45 PM
COEX
Room:
Grand Ballroom 101-102
Poster No:
646
Submission Type:
Abstract Submission
Authors:
Min Wang1, Zhoukang Wu1, Liangjiecheng Huang1, Xiaochu Zhang1, Xiaosong He1
Institutions:
1University of Science and Technology of China, Hefei, Anhui
First Author:
Co-Author(s):
Introduction:
Imaging meta-analyses of addiction have summarized symptom-brain region correspondence patterns, revealing diagnosis-specific and transdiagnostic effects [1]. However, group average differences are not representative of individual cases [2]. There are both clinical and neuroanatomical variabilities at the single-subject level in addiction, complicating it challenging to develop neuroimaging biomarkers to track disease severity, progression, and treatment response. Here we demonstrate through normative model [3] and lesion network mapping [4] techniques that regional gray matter atrophy patterns across patients with addiction are highly heterogeneous, yet these deviations can be embedded in common functional circuits and networks.
Methods:
T1-weighted data from 333 patients with substance use disorder and 957 healthy controls (HC) were collected from 3 public databases [5]. Voxel-based morphometry analyses via CAT12 were conducted to segment these images into gray matter, white matter, and cerebrospinal fluid. We used age, gender, intracranial volume, and site effect as covariates to build a voxel-wise normative model for the gray matter volume (GMV) of HC. Normative model can be used to define a normative range of variation against which new individuals are compared. Applying the model to each patient's data, we can generate a personalized whole-brain voxel-wise GMV deviation map. We identified all voxels with GMV deviation score < -2 as regions with extreme atrophy for each patient, corresponding to a 2 standard deviations below the mean GMV of HC, controlling for covariates. We overlaid binarized atrophy maps from all patients to identify regions consistently showing atrophy in the highest number of patients.
Next, an "atrophy functional connectivity (FC) network" was derived for each patient, defined as brain network functionally connected to aforementioned atrophic regions. Using the binarized atrophy regions as seed, FC with the rest of the brain was calculated using the GSP1000 resting functional dataset [6]. For the 1000 FC maps generated for each atrophy seed for each patient, a one-sample t test was used to infer regions that were significantly connected to the seed atrophy region via FC and to obtain an FWE-corrected atrophy FC network map. Finally, we calculated the overlap ratio of these atrophic FC networks.
Next, an "atrophy functional connectivity (FC) network" was derived for each patient, defined as brain network functionally connected to aforementioned atrophic regions. Using the binarized atrophy regions as seed, FC with the rest of the brain was calculated using the GSP1000 resting functional dataset [6]. For the 1000 FC maps generated for each atrophy seed for each patient, a one-sample t test was used to infer regions that were significantly connected to the seed atrophy region via FC and to obtain an FWE-corrected atrophy FC network map. Finally, we calculated the overlap ratio of these atrophic FC networks.
Results:
We first verified that the age distributions and model fitting parameters in Fig1.A supported our hypothesis that the HC's demographic information can predict GMV well. We then overlapped the binary atrophy masks across all patients and within each subgroup of substance addiction. Surprisingly, regardless of drug type, only 10%-20% of patients showed consistent atrophy at same locations, suggest that at individual level, voxels exceeding -2 deviations distributed heterogeneously across each patient cohort (Fig1.B). In contrast, traditional group-level statistical inferences on the whole-brain deviation maps can still identify prominent gray matter atrophy pattens both across and within each subgroup (Fig1.C).
By examining the overlay map of atrophy FC network we found that, although the locations of extreme atrophy were heterogeneous among patients, in fact, the atrophic structures of more than 60% of patients are functionally embedded into a homogeneous network, involving medial prefrontal cortex, middle/posterior cingulate cortex, and occipital lobe (Fig2.A). Importantly, this atrophic FC network was stable across all three drug subtypes (Fig2.B).
By examining the overlay map of atrophy FC network we found that, although the locations of extreme atrophy were heterogeneous among patients, in fact, the atrophic structures of more than 60% of patients are functionally embedded into a homogeneous network, involving medial prefrontal cortex, middle/posterior cingulate cortex, and occipital lobe (Fig2.A). Importantly, this atrophic FC network was stable across all three drug subtypes (Fig2.B).
Conclusions:
Although high heterogeneity of GMV deviations is a general characteristic of addiction, these deviations are often coupled to common functional circuits and networks, offering a putative neural substrate for phenotypic similarities among individuals assigned the same diagnosis. Specifically, the common involvement of default mode network across subtype may shed light on the understanding of the progress of structural atrophy in patients with addition.
Disorders of the Nervous System:
Psychiatric (eg. Depression, Anxiety, Schizophrenia) 1
Modeling and Analysis Methods:
Connectivity (eg. functional, effective, structural)
fMRI Connectivity and Network Modeling 2
Keywords:
Addictions
FUNCTIONAL MRI
Morphometrics
STRUCTURAL MRI
1|2Indicates the priority used for review
Provide references using author date format
2. Wolfers, T. et al. Mapping the heterogeneous phenotype of schizophrenia and bipolar disorder using normative models. JAMA Psychiat 75, 1146–1155 (2018).
3. Rutherford, S. et al. The normative modeling framework for computational psychiatry. Nat. Protoc. 17, 1711-1734 (2022).
4. Aaron, MT. et al. Network localization of clinical, cognitive, and neuropsychiatric symptoms in Alzheimer’s disease. Brain. 143(4), 1249-1260 (2020).
5. Wei, D. et al. Structural and functional brain scans from the cross-sectional Southwest University adult lifespan dataset. Sci Data 5, 180134 (2018).
6. Holmes AJ. et al. Brain Genomics Superstruct Project initial data release with structural, functional, and behavioral measures. Sci Data 2: 1–16 (2015).