[综述笔记]Functional Connectome–Based Predictive Modeling in Autism

论文网址：Functional Connectome–Based Predictive Modeling in Autism - ScienceDirect

英文是纯手打的！论文原文的summarizing and paraphrasing。可能会出现难以避免的拼写错误和语法错误，若有发现欢迎评论指正！文章偏向于笔记，谨慎食用

目录

1. 省流版

1.1. 心得

1.2. 论文总结图

2. 论文逐段精读

2.1. Abstract

2.2. ?

2.3. Autism-Specific Considerations for Predictive Modeling

2.3.1. Predictive Models Offer Biological Insight and Potential Clinical Utility

2.3.2. Balancing Large Sample Sizes, Concerns About Data Decay, and Site Effects

2.3.3. Confounds, Tolerability of the Scanning Environment, and Consequences for Predictive Modeling

2.3.4. Comorbidities and Phenotypic Overlap

2.3.5. Sex Imbalance in Autism and Sex-Specific Effects in Predictive Modeling

2.4. Case-Control Classification: The Case for Focusing on Diagnosis

2.4.1. Brain-Based Features Implicated in Autism Classification Differ Across the Life Span

2.4.2. Clinical Utility: Toward Early Diagnosis

2.5. Dimensional Prediction: Accounting for Complex Symptomatology

2.5.1. Predicting Autism Symptoms

2.5.2. Predicting Phenotypes Relevant for Adaptive Functioning

2.6. Subtyping: Simplifying Complexity by Finding Commonalities

2.6.1. Initial Efforts at Subtyping Connectomes

2.6.2. Toward Subtyping of Brain-Behavior Predictive Models

2.7. Limitations

2.8. Future Directions and Conclusions

3. Reference List

---

1. 省流版

1.1. 心得

（1）数据衰减愣是英文中文都没读懂，到时候看看参考文献

1.2. 论文总结图

2. 论文逐段精读

2.1. Abstract

        ①ASD is a heterogeneous neurodevelopmental disorder

        ②fMRI helps to explore the activity of brain network

2.2. ?

        ①There is limit in empirically validated treatments

        ②They provides biological insight and clinical utility (A) in this paper:

2.3. Autism-Specific Considerations for Predictive Modeling

2.3.1. Predictive Models Offer Biological Insight and Potential Clinical Utility

        ①They predict the significant ROI through FC and phenotypic data

2.3.2. Balancing Large Sample Sizes, Concerns About Data Decay, and Site Effects

        ①Commonly used ASD datasets: the Autism Brain Imaging Data Exchange (ABIDE) and the European Autism Interventions Multicenter Study for Developing New Medications (EU-AIMS)

        ②There are dataset decay and multi-site effects

2.3.3. Confounds, Tolerability of the Scanning Environment, and Consequences for Predictive Modeling

        ①Motion correction methods: global signal regression

        ②Severe ASD patients are hard to control themselves, which causes low quality (high motion artifacts) images. Researchers may exclude these images manually. ⭐ However, the lack of data on severely ill patients will impair the comprehensiveness and reliability of the dataset

barring  prep.除了；除非  v.禁止，阻止(某人做某事)；阻挡；(用铁条或木条)封，堵；拦住  bar的现在分词

2.3.4. Comorbidities and Phenotypic Overlap

        ①Diseases can be classified into different subtypes due to their multiple comorbidities

2.3.5. Sex Imbalance in Autism and Sex-Specific Effects in Predictive Modeling

        ①Sex imbalance charactertized in brain images and behaviour will bring the difficulty of diagnosis

        ②Thus, there should be similar number of different sexes in images, or use IQ to predict

2.4. Case-Control Classification: The Case for Focusing on Diagnosis

2.4.1. Brain-Based Features Implicated in Autism Classification Differ Across the Life Span

        ①It is more distinguishable between age in 5 to 15, rather than 15-30. The functional features may seperate during year pasts.

2.4.2. Clinical Utility: Toward Early Diagnosis

        ①Intervention time and corresponding effect:

        ②ASD classification on infants:

where CSBS denotes Communication and Symbolic Behavior Scales, MSEL denotes Mullen Scales of Early Learning, and RBS-R denotes Repetitive Behaviors Scale-Revised

        ③Fetal imaging may give support to further diagnosis

efficacious  adj.有效的；灵验的；奏效的    fetal  adj.胎儿的；胎儿般的

gestation  n.妊娠(期)；怀孕(期)；(想法、计划等的)构思，酝酿，孕育

2.5. Dimensional Prediction: Accounting for Complex Symptomatology

        ①⭐"Symptoms in a number of psychiatric conditions, including autism, exist on a continuum, and the line between what constitutes adaptive versus divergent behavior is often unclear. "

        ②FC for dimensional study:

continuum n.(相邻两者相似但起首与末尾截然不同的)连续体

2.5.1. Predicting Autism Symptoms

        ①Dimensional prediction of ASD:

interrogate  vt.讯问；审问；(在计算机或其他机器上)查询，询问；盘问

2.5.2. Predicting Phenotypes Relevant for Adaptive Functioning

        ①Modelling the brain is feasible

2.6. Subtyping: Simplifying Complexity by Finding Commonalities

2.6.1. Initial Efforts at Subtyping Connectomes

        ①Subtyping study based FC:

        ②Difference between subtypes in FC:

hallmark  n.特点；特征；金银纯度印记(打在金、银、铂制品上，表示纯度、生产日期及产地)  vt.给…打金银纯度印记

2.6.2. Toward Subtyping of Brain-Behavior Predictive Models

        ①Modelling different subtypes is feasible

triage  n.治疗类选法；伤员鉴别分类；患者鉴别分类；分诊

2.7. Limitations

        ①The FC of ASD and TD are quite similar when scanning time is limit (less than 5 mins). It can be more reliable when increases the scanning time to 15 mins

        ②External biological factors may also affect patients afterward. Therefore, phenotype data may not have strong persuasiveness

2.8. Future Directions and Conclusions

        ①They request early diagnosis, multi-modality, dense scanning

3. Reference List

Horien, C. et al. (2022) 'Functional Connectome–Based Predictive Modeling in Autism', Biological Psychiatry, 92(8): 626-642. doi: Redirecting