China Science Daily: Is there any depression, take two steps

By yqqlm yqqlm

Visit: Alibaba Cloud 6.18 Main Venue | Tmall 6.18 Super Red Packet Main Venue Jingdong 6.18 “Jingxiang Red Packet” Maximum denomination of 18618 yuan

China Science Daily: Is there any depression, take two steps

Schematic diagram of the experiment scene provided by the interviewee

China Science Daily: Is there any depression, take two steps(1)

Depression can be reflected in the gait Zhang Shuanghu draws

Therefore, the research can quickly and effectively identify depression The method has practical significance.

Recently, Zhu Tingshao’s research group at the Key Laboratory of Behavioral Sciences of the Institute of Psychology of the Chinese Academy of Sciences discovered through gait behavior data analysis that depression can be reflected in gait, and different types of gait characteristics have an effect on depression. The contribution of recognition is different, and automatic recognition of depression can be effectively realized based on machine learning technology. Related research results were published online in “Frontiers in Psychiatry”.

With the rapid development of economy and society, people’s pace of life is accelerating and work pressure is significantly increasing, and the number of people with psychological and behavioral problems and mental disorders in my country is also increasing. Statistics show that the lifetime prevalence of depression in China is 6.9%. There are currently 95 million people suffering from depression, and the proportion of depression among students has increased.

In September last year, the National Health Commission issued the “Working Plan for Exploring Special Services for the Prevention and Treatment of Depression”, requiring high schools and institutions of higher learning to incorporate depression screening into student health examinations, and respond to students with abnormal results. Give focused attention. At the same time, we will increase interventions for depression in key populations such as pregnancy and childbirth and the elderly.

Incorporating depression screening into health examinations requires intensified screening and assessment of mental health status, so that groups that are prone to depression can be identified in a timely and early manner, and intervention and treatment can be carried out as soon as possible. But the problem is that in the early stages of depression, many people may not be aware of it at all, let alone intervention.

“There are strict standards for clinical diagnosis of depression.” Li Lingjiang, chairman of the Psychiatric Branch of the Chinese Medical Association and chief physician of the Second Xiangya Hospital of Central South University, told the Chinese Journal of Science, “We mainly use the world The diagnostic criteria for depression (ICD system) developed by the Health Organization and the standards of the United States “Diagnosis and Statistics Manual of Mental Disorders”. The diagnosis of depression requires a specialist to make comprehensive judgments on the patient’s state and behavior based on relevant standards and combined with face-to-face consultation. There are no biological indicators for the diagnosis of mental illness.”

Tong Liang, the attending physician at the Yunnan Provincial Psychiatric Hospital, also said, “Depression cannot be diagnosed only by instruments, but also needs to be combined with symptomology and other judgments, and then combined Some related scales are checked.”

However, there are currently less than 40,000 psychiatrists in China. This has caused a dilemma in the diagnosis and treatment of depression: on the one hand, there is a large number of patients, on the other hand, there is a serious shortage of professional doctors.

“The diagnosis of depression is mainly based on doctor’s diagnosis, but the number of specialists is insufficient.” Zhu Tingshao told the China Science News, “Plus, the level of health doctors who conduct primary screenings varies. It is inevitable that there will be some missed diagnosis and misdiagnosis. We hope to use some ecological behavior (such as gait, posture, speech) analysis to identify depression, and provide some auxiliary information in addition to the doctor’s routine diagnosis to help the doctor make the diagnosis.”

“Existing studies have shown that the brain neural network involved in individual posture control is closely related to depression.” Zhu Tingshao said, “Posture symptoms have been proven to be the basic manifestations of depression. It is related to healthy individuals. In contrast, depressed patients have reduced vertical head movement, smaller limb movements, and lower gait speed during walking.”

When designing the experiment, the researchers unexpectedly discovered that the Microsoft Kinect smart somatosensory device can With a sampling rate of 30 Hz, the three-dimensional coordinate changes of 25 body joints of the human body are captured. Moreover, the Kinect smart somatosensory device has the advantages of non-intrusive, low-cost, and easy-to-use. It can easily collect the gait behavior data of the subjects and identify the depression state. “Its effectiveness in motion capture and motion monitoring has been verified. “.

The study recruited a total of 126 depression patients and 121 healthy individuals. The case group is a depression patient in a mental health medical institution in a certain city. The diagnosis result was completed by a psychiatrist based on the “Manual of Diagnosis and Statistics of Mental Disorders”. The control group is a healthy population recruited from society. All subjects walked back and forth naturally on a 6-meter-long and 1-meter-wide sidewalk for two minutes, and a Kinect intelligent somatosensory device placed at one end recorded gait data.

The researcher preprocessed the collected data. First, the data was segmented, and the data of 2 gait cycles during the process of each subject’s walking facing the Kinect intelligent somatosensory device were intercepted, in order to eliminate a large number of Repetitive data causes low computational efficiency and data redundancy; then a low-pass filter is used to denoise the data of 25 body joints. After the data preprocessing was completed, the researchers extracted 10 kinematics features, 300 time domain features, and 825 frequency domain features, and finally used logistic regression analysis to explore the contribution of different types of gait features to depression recognition, and used machine learning technology Train a depression recognition model.

The results of multiple logistic regression analysis show that kinematics, time domain and frequency domain features can explain the variability of the dependent variable (depression) by 12.55%, 58.36% and 60.71%, respectively. At the same time, the depression recognition model based on gait features constructed by machine learning technology is effective.

Researchers believe that, compared with traditional psychometric methods, this method of depression recognition based on gait data has the characteristics of non-intrusive, retrospective, and automated, so this method is combined with traditional measurement methods It can effectively improve the application range and measurement efficiency of psychological measurement.

“This kind of objective data is more convenient for users.” Zhu Tingshao said, “Because there is no need for huge equipment and no complicated operations, you can get the result by just a few steps. ”

Zhu Tingshao believes that, in principle, during large-scale screening, for example, a certain space is set aside in a hospital and a 3D camera (with three-dimensional recognition function) is installed to participate in the screening. After walking for a minute or two from a short stretch of road, you can use your gait to identify whether you are prone to depression. This method can also be used in schools, factories, nursing homes and even members of the family’s depression tendency screening and early warning.

“From the results of the experiment, it can reach a correlation of above average. In other words, its screening results have a certain reference value.” Zhu Tingshao said, “This kind of walking posture and body posture recognition The advantage of can detect early depression. The earlier depression is detected, the better it is for treatment and rehabilitation.”

It is reported that the hardware device of the auxiliary diagnosis system only has a camera with depth information. This kind of camera is easy to buy on the market, and the price is not expensive. At the same time, in conjunction with the analysis program, the analysis program can intuitively read the results from the computer, or it can be made into a mobile app.

“Currently, some brands of mobile phones also have 3D cameras. As long as the mobile phone has sufficient computing power, self-examination can be performed on home mobile phones.” Zhu Tingshao also emphasized, “This method can be used as an auxiliary diagnosis. Convenient and efficient initial screening or self-examination, but it cannot replace a doctor’s diagnosis. Once you find that there is a possibility of depression, you must go to a specialist for treatment.”

(Originally published in “Chinese Journal of Science” 2021-06 -11 3rd Edition Medicine and Health)