To submit a new entry for inclusion in the benchmark you must prepare the following information:

• You should submit separate entries for patient-independent (general) and patient-specific (personalized) models.
• You should submit different entries when training on different datasets. Here an entry is considered as a trained model. This means you should submit different entries if you use different training routines to train on different datasets.
• Your results file should contain annotations from at least one of the four publicly available datasets. The result file should contain the detections for all the subjects in the dataset. If you have tested your model on several datasets you should concatenate the different results files in a single file to upload on this platform.
• It is your responsibility to guarantee independence between your training set and your test set. You should use Leave-One-Out cross-validation, K-fold cross-validation, or train on an independent dataset. We provide guidelines for cross-validation on the description of the framework.
• The results file should follow the data format described in the framework. You can download an example of a result file here (TODO upload file).

All entries are manually reviewed before inclusion in the benchmark. You might be contacted for further information prior to inclusion.

* required fields

Contact Details

People developing model *:
Affiliated institution:
Model name *:
Paper or other resource for more information about the model:
License *:
Link to source code:
Citation details: (how do you want people to reference your work)
Contact email *:

Model Details:

Please provide as much information as possible.

Model Type *:

Patient-independent model (general)
Patient-specific model (personalized)

Signal Processing pipeline:
What is the processing of data from input to output. Include information on pre-processing and post-processing steps.

Model Description:

Model Description *:
What type of model is it? This includes basic model architecture details, such as whether it is a Naive Bayes classifier, a Convolutional Neural Network, etc. Mention important parameters.
Model features *:
What was the window length for classification. For feature based models, which features are used?
Number of channels *:
How many EEG channels does your model use?.

Training procedure:

Training Data *:
What data was used for training the model? For personalized models: how much data was initially used for training. For general models: which datasets were used. Data imbalance:
How did you deal with dataset imbalance? Caveats and Recommendations:
This section should list additional concerns that were not covered in the previous sections. For example, did the results suggest any further testing? Were there any relevant groups that were not represented in the evaluation dataset? Are there additional recommendations for model use?

Results

Upload results file *:
Upload your prediction dataset. The file is a concatenation of predictions on the different datasets. It should follow the file format described in the framework