Skip to content

CHB-MIT preprocessing: Potential labeling issue for seizure windows #29

Description

@XDUzhangzy

I am not working on epilepsy research, so I may not fully understand how seizure annotation is typically handled.
Please correct me if I am mistaken.
I was reviewing the labeling logic and wanted to confirm if this might be a potential issue.

The current code labels a window as seizure only if it contains the seizure start or end point:

Image

What about a window that falls completely inside a seizure interval (containing neither the start nor the end point)? For example:

Seizure interval: [1000, 5000]

Window: [2000, 4560]

Intuitively, I would expect this window to be labeled as seizure as well, since it contains seizure activity. However, under the current logic, it seems it would be labeled as 0.

Is this behavior by design? Or should the logic check for overlap between the window and the seizure interval instead?

Thank you for your time and for sharing the code!

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions