Depression is a common mental disease characterized by persistent low mood, anhedonia, grief and cognitive impairment that severely affects people’s quality of life. The prevalence of depression is about 2 - 4% worldwide, and 1.7 - 2% in China. According to statistics of the World Health Organization (WHO), more than 350 million people suffered different degrees of depression worldwide. And a related study of a meta-analysis of 50371 patients from 118 studies found that the correct recognition rate for depression was only 47.3%. So with the high incidence and low recognition rate of depression, to explore simple, objective, accurate evaluation methods or biomarkers for depression detection is a major public-health challenge. The contest is dedicated to the advancement of research on the evaluation of depression based on physiological signals.
The competition dataset is a depressive disorder 128-channel resting-state electroencephalogram dataset for analysis of mental disorders. Collected from 128-channel HydroCel Geodesic Sensor Net (Electrical Geodesics Inc., Oregon Eugene, USA) and Net Station acquisition software (version 4.5.4). The location of the 128 electrodes placement (E1 to E128) is shown in Fig. 1. The sampling frequency was 250 Hz. All the raw electrode signals were referenced to the Cz. The impedance of each electrode was checked prior to recording, to ensure good contact, and was kept below 50 kΩ.
128 channel HydroCel Geodesic Sensor Net (HCGSN)
For now, the dataset includes data mainly from clinically depressed patients and matching normal controls. All our patients were carefully diagnosed and selected by professional psychiatrists in hospitals. We will provide the participants with part of the data as a training set which can be downloaded. The participants are required to use the data to classify the characteristics of depression patients and normal people through artificial intelligence algorithms, and can automatically classify the given test set data.
IEEE Healthcom 2020 Commitee
IEEE Healthcom 2020 aims at bringing together interested parties from around the world working in the healthcare field to exchange ideas, discuss innovative and emerging solutions, and develop collaborations.
Ubiquitous Awareness and Intelligent Solutions Lab of Lanzhou University
Founded in January 2009, UAIS Lab relies on the Academy of Information Science and Engineering of Lanzhou University to conduct research in the fields of pervasive computing theory and application, psychophysiology, emotional computing, semantic web and ontology. The laboratory implements the laboratory director responsibility system under the guidance of the academic committee. It has a young research team with multidisciplinary characteristics. The average age is under 35 years old. The researchers have professional backgrounds in computer, medicine and information processing. At present, there are five professors, three associate professors, one lecturer and eight secondary professionals, and a group of doctoral and postgraduate students. In recent years, the laboratory has undertaken research projects such as the National Natural Science Foundation, the National “973” Program, the International Cooperation of the Ministry of Science and Technology, EU International Cooperation, Enterprise Cooperation, and the “985” and “211” Construction of Lanzhou University, and established with domestic and foreign scientific research units. Joint laboratories, strengthen international exchanges, and jointly cultivate high-level scientific and technological talents.
The contest requires 1-3 people to participate in the team, and each player can only join one team. Each team submits a maximum of 5 results per day.
Estimated size: 50-150 teams.
(1)First stage：Sign up
June 1, 2020 - Contest start.
Aug 31, 2020 - The registration is closed, and no individual registration or team changes will be accepted thereafter.
Oct 1-17, 2020 - The result submission deadline. Before this date, contestants may submit the model code and training results to the organizer multiple times, and the final submission will be regarded as the final result. Teams that fail to submit before the deadline will be regarded as abandoned.
Oct 17- 20, 2020 - The organizer uses the test set data to verify the model classification effect of each participating team, and gives the final ranking.
(4)Final stage：Announce results and award bonuses
Oct 20, 2020 - The organizer will announce the results of the competition on the Healthcom 2020 conference official website; during the conference, the prize money will be distributed to the winning teams.
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.
1. Data recording and storage
The data files named with “D” prefix represent data from patients with major depressive disorder (MDD), and the data files named with “N” prefix represent data from normal controls (NC). The format of EEG files are .csv. In validation and testing set, the files are named as their ids. The submission file should be a csv file with two columns: the first column is the file name (without ‘.csv’), the other is the prediction, 0 is normal (NC), and 1 is major depressive disorder(MDD).
The uploaded files named with D, N contain all of the signals (e.g. D1 means MDD patients with id 1, N1 means normal control with id 1.). Each subject has 40 2s segments(e.g. D1-1.csv,D1-2.csv,…,D1-40.csv) . After loading the .csv files, each file is a matrix. The dimension of the matrix is 128 * 500. 128 is the number of electrodes. The first 128 signals are from the electrode E1 to electrode E128. 500 is the sampling point.
2. Training set
14 outpatients diagnosed with MDD, as well as 17 NC are in the training set. The total number of samples is 1240（(14+17) * 40).
3. Validation set
5 outpatients diagnosed with MDD, as well as 6 NC are in the validation set. The total number of samples is 440 ((5 + 6) * 40).
4. Test set
5 outpatients diagnosed with MDD, as well as 6 NC are in the testing set. The total number of samples is 440 ((5 + 6) * 40).
5. Sample Submission
Sample submission file defines the format of the submitted files.
The main purpose of this classification task is to use artificial intelligence algorithms to complete the feature classification of normal people and depression patients based on the resting EEG data provided.
Input: Resting EEG data fragment of subjects(.csv file containing 128*500 matrix).
Output: The results of the classification of the depressive disorder of the subjects to which this data segment belongs.
Submissions will be evaluated based on their F1 score. The F1 score measures accuracy using the precision and recall.Precision is the ratio of true positives (TP) to all predicted positives (TP + FP). Recall is the ratio of true positives to all actual positives (TP + FN). The F1 score is given by:
First prize，1 team，RMB ￥10,000（Before tax）
Second prize，1 team，RMB ￥6,000（Before tax）
Third prize，1 team，RMB ￥3,000（Before tax）
Note: The amount and proportion of the bonus may be adjusted according to the number of participants, and the specific announcement will be subject to the deadline for registration.
The participating teams are welcome to submit a short paper based on your achievements, about 1-2 pages in length, excellent short papers will be included in the papers published in IEEE journals. The short paper will be searched by EI, and the intellectual property rights belong to the participating teams.
Intellectual property and academic integrity
1.The intellectual property rights of the entries belong to the team.
2.Teams and individual contest participants shall consciously abide by the relevant intellectual property laws and regulations, and shall not infringe on the intellectual property rights or other rights and interests of others. The organizers, contractors and co-organizers of this contest shall not be liable for any legal consequences.
3.Teams and individual contest participants should ensure academic integrity. Once academic misconduct such as code plagiarism or technical plagiarism is discovered, they will be disqualified immediately.
Website of Conference: https://healthcom2020.ieee-healthcom.org/
Dataset description paper: http://modma.lzu.edu.cn/data/publications/
QQ group: 1067755371
1.The right to interpret the charter belongs to the organizer. If you have any questions, please send an email to firstname.lastname@example.org or put it in the official QQ group.
2.The schedule and schedule of the competition listed in this charter are planned. If there is any adjustment due to force majeure, a notice will be issued on the official website of the conference and the official QQ group in time.
IEEE Healthcom 2020 Committee
May 28, 2020