Narcolepsy is a chronic brain disorder. The mechanism is the impairment of brain controlling of sleep and wakefulness. The cause of this disease is still unclear, but common symptoms include excessive day time sleepiness, cataplexy, hypnogogic hallucination, sleep paralysis, and sleep disturbance. Because these symptoms are easily confused together in many situations, it is difficult for doctors to make the diagnosis. Therefore, medical treatment for patients is always delayed. According to previous research report, narcoleptic patients are often delay diagnosis for 10 to 15 years after the onset of the disease. Clearly, to make the diagnosis of narcolepsy is very difficult. Another cause for the delay is the method for diagnosing narcolepsy, which mainly rely on sleep examination instruments and the testing of hypocretin concentration in the cerebrospinal fluid. However, these tests are difficult to carry out in many areas, and diagnosing narcolepsy is still difficult in many countries. To the patients and their families, developing a fast and accurate method or tool for diagnosing narcolepsy is of the utmost importance.

The purposes of this study are as follows: (1) To collect comprehensive narcolepsy and non-narcolepsy brain imaging data. The difference between the two groups will be analyzed. To find the difference between the Type 1 and Type 2 narcolepsy by brain imaging characteristics. Use these data to find the special parameter by "machine learning" and build a predictive model; (2)To collect comprehensive narcolepsy and non-narcolepsy HLA typing data. Attempt to understand the HLA profile of narcoleptic patients and their parents in Taiwan. To analyze the difference between the two groups of Type 1 and Type 2 narcolepsy. Use these data of HLA typing characteristics to find the special parameter by "machine learning" and to establish a predictive model; and (3) categorize and group narcolepsy clinical data, sleep examination data, and the aforementioned data based on machine learning concept and build a predictive model as the basis for developing a fast and accurate" narcolepsy diagnostic tool or model" in the future. Research method: This is a case control study. There are 400 subjects (age 9 - 45) will be collected. These subjects will be divided into the three following groups: (1) experimental group (narcolepsy Type 1, 200 subjects); (2) experimental group (narcolepsy Type 2, 100 subjects); and (3) control group (age and gender matched non-narcolepsy subjects, 100). The investigators will collect all the clinical data for each subject, including clinical characteristics, sleep examination data, actigraphy, HLA typing, and brain imaging data. Data analysis method: the narcolepsy clinical data, sleep examination data, and the aforementioned data were categorized and grouped through data analysis based on computer machine learning, neural network, and predictive model effectiveness analysis concepts. Then the investigators will built a predictive model based on the results.

Inclusion Criteria:

1. Meet the diagnosis of Type 1 or Type 2 sleep disorder, and comorbid with depression.
2. The age is introduced between 18-60 years old, regardless of gender.
3. Those who agree to participate in the trial and sign the subject's consent form.

Exclusion Criteria:

1. Combined with epilepsy, brain injury or severe organic brain disease or serious heart disease.
2. Combined with serious other mental disorders, such as bipolar disorder, mental retardation or addiction disorders.
3. Not willing to participate in this study or not willing to fill out the consent form.
4. Those who are not suitable to enter the experiment after being evaluated by PI and co PI.

participant and doctor