Written by Shi Weiyang and Jiang Tianzai

Editor-in-charge - Wang Sizhen

Editor — Binwei Yang

Schizophrenia is a severe psychiatric disorder of unknown etiology and high disability, with a worldwide prevalence of about 1% [1].

On August 25, 2022, a team of researchers from the Brain Network Group Research Center of the Institute of Automation of the Chinese Academy of Sciences and six hospitals in the United Nations published a paper entitled "Two subtypes of schizophrenia identified by an individual-level atypical pattern of tensor-based morphometric" at Cerebral Cortex measurement" article, through machine learning, identified two potential imaging subtypes

With the continuous development of imaging technology, magnetic resonance imaging is widely used in the study

Figure 1 Framework for the identification and analysis of schizophrenia subtypes for heterogeneous problems

(Source: Shi W et al.

Based on the multicenter schizophrenia dataset collected in the early stage of the research group, the researchers used normative modeling[6] to extract individualized patient TBM abnormality patterns, and used sparse clustering to identify two potential subtypes

Fig.

(Source: Shi W et al.

Through statistical analysis, the researchers found that subtype 2 had more severe negative symptoms
than subtype 1.

However, the current definition of negative symptoms remains controversial
.

Thus, in this study, the researchers introduced two more widely adopted tissue models of negative symptoms: a four-factor model of Chen et al.
[7] and a five-factor model defined by Lindenmayer et al.
[8].


In both tissue patterns, the negative symptom score for subtype 2 remains significantly higher than for subtype 1
.

To establish a more granular mapping relationship between imaging and clinical symptoms, the researchers performed an image-clinical association analysis using partial least squares correlation and identified a significantly correlated hidden component (Figure 3C
).

On the one hand, the Cain component reveals the imaging sources of clinical differences between subtypes (Figure 3A), and on the other hand, it also provides a new perspective on the division of negative symptoms of schizophrenia: the first five symptoms most relevant to this component all belong to the consensus term for negative symptoms in the current field (Figure 3B).

Figure 3 Imaging sources using imaging-clinical association analysis to reveal differences in clinical symptoms between subtypes

(Source: Shi W et al.
, Cereb Cortex, 2022)

Based on the published allen human brain atlas (AHBA) data[9], the researchers conducted an image-transcriptome association analysis using partial least squares regression, and attempted to explore the possible heterogeneous biological risk factors in schizophrenia suggested by the two types of imaging anatomical subtypes by looking for a set of gene spatial expression patterns associated with the spatial patterns of imaging differences between the two subtypes
。
The results suggest that some widely reported biological processes of schizophrenia risk, such as chemical synaptic transmission and DNA repair, may be heterogeneous; In addition, oxytocin has been found to be associated with negative symptoms of schizophrenia [10], echoing
the "abnormal patterns of the oxytocin signaling pathway may be heterogeneous in schizophrenia" demonstrated in keeg enrichment analysis.

This also suggests that neuroimaging is an effective tool for linking genes and clinical phenotypes
.

Figure 4 Potential heterogeneous biological factors between subtypes were investigated using image-transcriptome association analysis

(Source: Shi W et al.
, Cereb Cortex, 2022)

In summary, focusing on the heterogeneous dilemma of schizophrenia, the study proposes a potential clinical stratification protocol that uses abnormal TBM patterns at the individual level to identify two potential subtypes of schizophrenia that differ significantly in negative symptoms from an anatomically visual perspective
.

Through image-clinical association analysis, the researchers established a mapping relationship between the anatomical differences and clinical symptoms of the images between subtypes, thereby locating the imaging sources
of the clinical heterogeneity exhibited by the two subtypes.

In addition, using image-transcriptome association analysis, the researchers further explored the potential heterogeneous biological factors in schizophrenia suggested by the two subtypes, providing prospective clues
for related studies.

Of course, whether the two imaging subtypes of schizophrenia identified by the institute can provide effective information for the precise treatment of schizophrenia remains to be further explored
.

In addition, a single modality and a single feature can only capture some of the heterogeneity of schizophrenia, and how to use multimodality and multi-feature to further refine the results of this subtype is also a problem
worth exploring in depth.

Overall, the study provides an attractive solution to alleviate the heterogeneity of schizophrenia, and the subtype classification framework is expected to be extended to other heterogeneous psychiatric disorder subtype classification studies
.

Original link: https://academic.
oup.
com/cercor/advance-article-abstract/doi/10.
1093/cercor/bhac301/6675350?redirectedFrom=fulltext

Jiang Tianzai, a researcher at the Institute of Automation of the Chinese Academy of Sciences, is the corresponding author of the paper, and Weiyang is the first author
of the paper when he is a doctoral student.

This study was co-funded by the Science and Technology Service Network Program of the Chinese Academy of Sciences (KFJ-STS-ZDTP-078), the National Natural Science Foundation of China (31300934, 31620103905), the Frontiers of Science Program of the Chinese Academy of Sciences (QYZDJ-SSW-SMC019), the National Key Research and Development Program (2017YFA0105203), and the Open Research Fund of the State Key Laboratory of Cognitive Neuroscience and Learning (CNLYB2004).

。

[1] The NAR-He Cheng/Su Zhida team found that topoisomerase IIA can regulate adult neurogenesis in the subependymal region

[2] The Sci Adv-Liao Wenbo team has made important progress in the adaptive evolution of amphibian brain volume

[3] J Neuroinflammation—From Changchun/Jian Zhang's team found that targeting proteoglycan receptors after hemorrhagic stroke protects white matter integrity and promotes the recovery of neurological function

[4] Front Aging Neurosci—Zeng Yanbing's team established a predictive model and revealed the effects of behavioral changes on cognitive impairment in the elderly

[5] The Sci Adv-Zhao Cunyou/Chen Rongqing team revealed the mechanism of microRNA inducing social and memory abnormalities in mice: miR-501-3p expression defects enhance glutamate delivery

[6] Sci Adv-Zhang Yi's research group found important neurons that regulate drug addiction behavior

[7] J Infect-Yifei Wang's team revealed that mamdc2, a highly expressed gene in Alzheimer's disease microglia, positively regulates the innate antiviral response of neurovirus infection

[8] Sci Adv—Xia Kun/Shen Yiping/Guo Hui revealed the relationship between key regulatory genes of stress particles and neurodevelopmental disorders

[9] Cell Prolif-Lai Liangxue/Zhang Kun/Zou Qingjian team worked together to successfully build a safe and efficient technology system for directional induction of motor neurons in vivo

[10] Nat Commun– Peng Yueqing's team discovered a new brain region that controls non-REM sleep

【1】Special Training on Biomedical Statistics for Clinical Prediction of R Language (October 15-16, Institute of Genetics and Developmental Biology, Chinese Academy of Sciences, Beijing)

【1】2022 World Artificial Intelligence Conference | Brain-Machine Intelligent Fusion - Connecting the Brain to the Future (Shanghai World Expo Center, September 2)

References (swipe up and down to read)

1.
O'Malley, A.
James, Richard Gabriel Frank, and S‐LT Normand.
"Estimating cost‐offsets of new medications: Use of new antipsychotics and mental health costs for schizophrenia.
" Statistics in medicine 30.
16 (2011): 1971-1988.

2.
Carpenter Jr, William T.
, and Brian Kirkpatrick.
"The heterogeneity of the long-term course of schizophrenia.
" Schizophrenia bulletin 14.
4 (1988): 645-652.

3.
Okhuijsen-Pfeifer, C.
, et al.
"Demographic and clinical features as predictors of clozapine response in patients with schizophrenia spectrum disorders: a systematic review and meta-analysis.
" Neuroscience & Biobehavioral Reviews 111 (2020): 246-252.

4.
Brugger, Stefan P.
, and Oliver D.
Howes.
"Heterogeneity and homogeneity of regional brain structure in schizophrenia: a meta-analysis.
" JAMA psychiatry 74.
11 (2017): 1104-1111.

5.
Alnæs, Dag, et al.
"Brain heterogeneity in schizophrenia and its association with polygenic risk.
" JAMA psychiatry 76.
7 (2019): 739-748.

6.
Marquand, Andre F.
, et al.
"Conceptualizing mental disorders as deviations from normative functioning.
" Molecular psychiatry 24.
10 (2019): 1415-1424.

7.
Chen, Ji, et al.
"Neurobiological divergence of the positive and negative schizophrenia subtypes identified on a new factor structure of psychopathology using non-negative factorization: an international machine learning study.
" Biological psychiatry 87.
3 (2020): 282-293.

8.
Lindenmayer, Jean-Pierre, Ruth Bernstein-Hyman, and Sandra Grochowski.
"Five-factor model of schizophrenia: initial validation.
" Journal of Nervous and Mental Disease (1994).

9.
Hawrylycz, Michael J.
, et al.
"An anatomically comprehensive atlas of the adult human brain transcriptome.
" Nature 489.
7416 (2012): 391-399.

10.
Marder, Stephen R.
, and Silvana Galderisi.
"The current conceptualization of negative symptoms in schizophrenia.
" World Psychiatry 16.
1 (2017): 14-24.