Unraveling the Mysteries of the Female Brain: Sex Patterns in ADHD
揭开女性大脑的神秘面纱: 多动症的性别模式
加入WiDS数据挑战,提升您的数据科学技能,同时解决一个有趣且重要的社会问题!
您将使用fMRI数据,任务是建立一个模型,以预测一个人的性别和是否有ADHD(注意力缺陷多动障碍)。
WiDS数据挑战全球赛是与安·S·鲍威尔女性大脑健康倡议(WBHI)合作开发的,参与单位包括康奈尔大学和加州大学圣巴巴拉分校。
数据集和支持由健康大脑网络(HBN)提供。
在发展中出现的神经精神疾病,如焦虑、抑郁、自闭症和注意力缺陷多动障碍(ADHD),在男性和女性的症状表现和程度上有所不同。
ADHD在大约11%的青少年中发生,其中约14%的男孩和8%的女孩被诊断为ADHD。
有证据表明,ADHD的女孩常常未被诊断,因为她们更多表现出注意力不集中等症状,这些症状更难被发现。
未被诊断的ADHD女孩会继续遭受影响她们心理健康和功能能力的症状。
在今年的WiDS数据挑战中,参与者需要建立一个模型,预测个体的性别和ADHD诊断。
儿童和青少年的功能性脑成像数据
社会人口统计、情感和育儿信息
挑战问题和任务:“与ADHD相关的脑活动模式是什么?它们在男性和女性之间是否不同,如果是的话,如何不同?”
任务是创建一个多结果模型,以预测两个独立的目标变量:
1) ADHD(1=是,0=否) 2)女性(1=是,0=否)
Data cleaning and preprocessing for model development and analysis
Understanding correlations
Regression model selection (statistical, machine learning)
Regression model fitting/testing (cross validation, avoiding over/underfitting, stratification of data during training)
Explaining factors that drive the performance of the model
Multi-outcome prediction
F1分数是精确率和召回率的调和平均值,因此它可以在一个指标中同时反映精确率和召回率。F分数的最高值为1.0,表示完美的精确率和召回率,而最低值为0,表示精确率和召回率均为零。
https://en.wikipedia.org/wiki/F-score
由于本次挑战的主题是揭示性别不平等,并且ADHD的诊断对女性来说更难以预测,因此在本次比赛中,我们对女性ADHD病例(ADHD_Outcome=1,Sex_F=1)赋予2倍的权重。 在我们计算F1分数时,加权F1分数会在每一列上进行计算,然后将这两个分数平均,以得出最终的Kaggle排行榜分数。
## 提交格式
对于测试数据集中的每个参与者(即每一行),提交文件应包含三列:participant_id
、ADHD_Outcome
和Sex_F
。 participant_id
应为整数,ADHD_Outcome
和Sex_F
均应为实数。每一行的这三个值应以逗号分隔。
文件应包含一个标题,并具有以下格式:
在比赛期间,排行榜是根据大约51%的测试数据计算的。 比赛结束后,最终排名将基于其余49%的数据进行计算。因此,最终排行榜排名可能与比赛期间的不同。
The Role of AI in Advancing Women’s Brain Health Research (27 min)
Sex-specific differences in the healthy and disordered brain (16 min)
More information on the dataset and how it was processed
Functional Connectivity Matrices Processed Through RBC (7 min)
Follow along dataset videos with this summary script
Global Datathon Workshop #1: Introduction to the Challenge and Dataset (17 min)
Global Datathon Workshop #2: Dataset Preprocessing and Preparation with corresponding Jupyter notebook (57 min)
Global Datathon Workshop #3: Building & Evaluating a Machine Learning Model with corresponding Jupyter notebook (55 min)
Statistical Approaches on Vectorized Connectomes for Brain-Behavior Mapping (36 min)
Geometric Approaches for Processing Brain Connectomes (19 min)
Full data dictionaries for Color Vision Test, Strength and Difficulties Questionnaire, and Alabama Parenting Questionnaire
1st place
$3,000 Kaggle cash prize for the team
WiDS Worldwide event ticket for each team member
WiDS Datathon award certificate for each team member
2nd place
$3,000 Kaggle cash prize for the team
WiDS Worldwide event ticket for each team member
WiDS Datathon award certificate for each team member
3rd place
$3,000 Kaggle cash prize for the team
WiDS Worldwide event ticket for each team member
WiDS Datathon award certificate for each team member
4th place
$3,000 Kaggle cash prize for the team
WiDS Worldwide event ticket for each team member
WiDS Datathon award certificate for each team member
5th place
$3,000 Kaggle cash prize for the team
WiDS Worldwide event ticket for each team member
WiDS Datathon award certificate for each team member
January 7, 2025 - Start Date. Register here with WiDS Worldwide to participate in the datathon.
April 16, 2025 - Leaderboard Prize form opens. Fill out a brief form for students, high schoolers & first-timer participants to be eligible for prizes.
April 27, 2025 - Entry and Team Merger Deadline. This is the last day participants may join or merge teams.
April 30, 2025 - Final Submission Deadline. Prize form also closes at this time
May 7, 2025 - Winners announced. Top performers may also be recognized throughout the competition.
All deadlines are at 11:59 PM UTC on the corresponding day unless otherwise noted. The competition organizers reserve the right to update the contest timeline if they deem it necessary.