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
2.4.1. ABIDE Data Preprocessing
1. 省流版
1.1. 心得
(1)摘要没说方法,只说了应用上
(2)intro也没咋写方法,说贡献是提出了一个模型hhhh,真会说话
(3)公式写得太冗杂了,没必要重复很多次,下标也太多了
(4)我怎么没理解它FC是怎么来的?似乎没提到时间序列和皮尔逊什么的,通用吗?随意吗?是我没看到吗?
(5)分别用5 fold和10 fold分别凑一张对比表.......我 啊这
(6)医学部分也没好好说,感觉能二区大概是实验做的多吧(虽然也没有多大用)
1.2. 论文总结图
2. 论文逐段精读
2.1. Abstract
①Facing the heterogeneity of ASD, traditional functional connectivity (FC) methods can not accurately classify ASD and HC
②They proposed a functional graph discriminative network (FGDN)
2.2. Introduction
①Introducing ASD and the meaning of pre-diagnosis for ASD patients
②Noting how fMRI applies to ASD diagnosis
③Listing some machine learning methods on ASD diagnosis(是比较早期的,所以倒不至于和related work重复)
④Further introducing and explain the graph representation methods
⑤Their contributions: a) proposed FGDN, b) high classification performance, c) researched biomarkers
2.3. Materials and Methods
2.3.1. Graph Construction
①Graph construction:
where and denote node set and edge set in graph , denotes the adjacency matrix(可能需要特别注意一下,因为其他的一般都用). And the adjacency matrix constructed by Gaussian distance function and k-nearest neighbor:
②FC matrix construction:
where denotes the class (0 means ASD and HC is 1)
③Two templates:
with
2.3.2. Proposed Graph Model
①用了谱图卷积和一阶近似...什么的。但是感觉也是用的别人的方法,又再次介绍了一下。我就不赘述了
②The overall framework:
③Atlas: AAL 116, HO 118 and MODL 128
④Graph kernel: 64
⑤Order of Chebyshev: 3
⑥Activation function: Sigmoid
⑦Loss function: cross entropy
2.4. Results
2.4.1. ABIDE Data Preprocessing
①Sample: 866 with 402 ASD and 464 HC
②FC: Ledoit–Wolf regularized shrinkage estimator(对什么数据使用啊????)
③ in NN: 20
2.4.2. Experiments
①Validation: 5 or 10-fold cross validation
②Optimizer: Adam
③Learning rate: 0.0001
④Weight decay: 0.0005
⑤Batch size: 16
⑥Dropout rate: 0.1
⑦5 fold comparison:
⑧10-fold comparison:
2.5. Discussion
①Loss and training curve(感觉也没什么用):
②Training ratio(一般来说增加训练样本都更好啦...):
③Leave-one-site-out performance:
④5 most discriminative brain regions by t-test:
2.6. Conclusion
......
3. Reference
Li, J. et al. (2021) 'Identification of Autism Spectrum Disorder With Functional Graph Discriminative Network', Frontiers in Neuroscience. doi: 10.3389/fnins.2021.729937