One option is to convert the EDF recording into ordinary CSV files first and then import those files into Mathematica.

I recently created an open-source tool called edf2csv: https://edf2csv.vercel.app/

The basic command is: bash npx edf2csv recording.edf

It creates an output folder containing:

signals.csv
channels.csv
annotations.csv
metadata.json

You can then import the signal data into Mathematica using something like:

data = Import["recording_csv/signals.csv", "Dataset"]

The tool works locally, supports EDF, EDF+, BDF, and BDF+, and preserves the original units and sampling rates. It also separates channels with different sampling rates instead of inventing interpolated samples.

For disclosure, I am the developer of the tool. I built it partly because importing and correctly interpreting EDF recordings often requires substantially more work than importing an ordinary table. Would love to hear your thoughts using this tool!