Interpretable simultaneous localization of MRI
corpus callosum and classification of atypical
Parkinsonian disorders using YOLOv5
1
st
Vamshi Krishna Kancharla
IIIT - Bangalore
Karnataka, India
2
nd
Debanjali Bhattacharya
IIIT - Bangalore
Karnataka, India
3
rd
Neelam Sinha
IIIT - Bangalore
Karnataka, India
4
th
Jitender Saini
Neuroimaging and Interventional Radiology
NIMHANS - Bangalore
Karnataka, India
5
th
Pramod Kumar Pal
Dept. of Neurology
NIMHANS - Bangalore
Karnataka, India
6
th
Sandhya M.
Neuroimaging and Interventional Radiology
NIMHANS - Bangalore
Karnataka, India
Abstract—Structural MRI(S-MRI) is one of the most versatile
imaging modality that revolutionized the anatomical study of
brain in past decades. The corpus callosum (CC) is the principal
white matter fibre tract, enabling all kinds of inter-hemispheric
communication. Thus, subtle changes in CC might be associated
with various neurological disorders. The present work proposes
the potential of YOLOv5-based CC detection framework to
differentiate atypical Parkinsonian disorders (PD) from healthy
controls (HC). With 3 rounds of hold-out validation, mean
classification accuracy of 92% is obtained using the proposed
method on a proprietary dataset consisting of 20 healthy subjects
and 20 cases of APDs, with an improvement of 5% over SOTA
methods (CC morphometry and visual texture analysis) that
used the same dataset. Subsequently, in order to incorporate the
explainability of YOLO predictions, Eigen CAM based heatmap
is generated for identifying the most important sub-region in CC
that leads to the classification. The result of Eigen CAM showed
CC mid-body as the most distinguishable sub-region in classifying
APDs and HC, which is in-line with SOTA methodologies and
the current prevalent understanding in medicine.
Index Terms—Structural MRI, Corpus Callosum, Atypical
Parkinsonism, YOLOv5, Eigen CAM
I. INTRODUCTION
Parkinson’s disease (PD) is a slowly progressing neurode-
generative movement disorder that mainly affects the motor
system. The common motor and non-motor symptoms of PD
include an ongoing loss of motor control (example, shaking,
stiffness, slow movement, postural instability), loss of sense
of smell, cognitive changes and many others. The category
of Parkinsonism includes the classic form of PD and its
other atypical variants, commonly referred “Parkinson’s Plus
Syndromes” or “Atypical Parkinsonism”. These Atypical PDs
(APD) are more severe and generally spread more rapidly;
making it less treatable as compared to classical PD. Multiple
System Atrophy (MSA) and Progressive Supranuclear Palsy
(PSP) are two examples of such APDs, which are included in
this study. Both PSP and MSA highlight the classical features
of PD with additional rare symptoms that distinguish them
from simple PD; although the clinical differentiation of PD,
MSA and PSP at the early stage is extremely difficult [1]–[3].
While there exists several literature that have studied the
whole brain and regional brain atrophy to quantify neu-
rodegeneration, little works have been performed on brain
corpus callosum (CC) especially in case of APDs, despite
its high relevance in neurodegeneration [4]–[9]. The CC is
the principal and the largest WM fibre track, situated at the
center part of CNS. With more than 300 million fibres it
connects two cerebral hemispheres and enabling all kinds of
interhemispheric communication [10], [11]. Therefore, subtle
structural changes in callosal architechture can be associated
with cognitive and behavioral deficits seen in neurological
diseases. The CC structure in brain is shown in Figure 2.
Being easily distinguishable in mid-sagittal MRI plane, CC
in this MRI plane lends itself well for extensive research
[1]–[9]. Hence, in this work instead of analyzing changes in
different sub-cortical brain regions, we focus in studying the
degeneration of this particular CC structure in brain, seen in
Parkinsonian disorders.
The present work is an extension of our previously pub-
lished works that have quantified callosal degeneration in
Parkinsonian disorders using CC morphometry and texture
analysis [1]–[4]. However, one major limitation of our pre-
vious studies was it relied on manual extraction of CC from
MRI images. With advances in deep learning, some recent
studies have used deep neural architecture for automatically
segment CC structures from MRI image volume [12]–[14].
However, no studies have been reported on automatic detection
of CC to classify Parkinsonian disorders from healthy controls
(HC). This motivates us to examine the ability of deep learning
based YOLO (You Only Look Once) framework to localize
CC region in MRI volume while simultaneously classifying
Parkinsonian disorders from controls. However, interpretibility
arXiv:2306.00473v1 [eess.IV] 1 Jun 2023
下载后可阅读完整内容，剩余4页未读，立即下载
陈书予
- 粉丝: 2w+
- 资源: 20
我的内容管理
展开
最新资源
- 解决本地连接丢失无法上网的问题
- BIOS报警声音解析：故障原因与解决方法
- 广义均值移动跟踪算法在视频目标跟踪中的应用研究
- C++Builder快捷键大全：高效编程的秘密武器
- 网页制作入门：常用代码详解
- TX2440A开发板网络远程监控系统移植教程：易搭建与通用解决方案
- WebLogic10虚拟内存配置详解与优化技巧
- C#网络编程深度解析：Socket基础与应用
- 掌握Struts1：Java MVC轻量级框架详解
- 20个必备CSS代码段提升Web开发效率
- CSS样式大全：字体、文本、列表样式详解
- Proteus元件库大全：从基础到高级组件
- 74HC08芯片：高速CMOS四输入与门详细资料
- C#获取当前路径的多种方法详解
- 修复MySQL乱码问题：设置字符集为GB2312
- C语言的诞生与演进：从汇编到系统编程的革命
资源上传下载、课程学习等过程中有任何疑问或建议，欢迎提出宝贵意见哦~我们会及时处理！
点击此处反馈
