Description of the problem

Hi there. I have some EEG data recorded using the Nihon Kohden (NK) system, this data includes .EEG files which store the raw EEG recording, but also the extension files: .21E, .PNT, .LOG ( https://mne.tools/0.21/auto_tutorials/io/plot_20_reading_eeg_data.html#nihon-kohden-eeg-data-eeg-21e-pnt-log)

The .LOG extension file store the annotated event time stamps. I need precise annotation times but I'm only getting timestamps rounded to the nearest full second. Has anybody dealt with this file format or have suggestions on how to add more precise annotation support

Steps to reproduce

import mne

f = mne.io.read_raw_nihon(file_path)
f.annotations

Link to data

No response

Expected results

I tried to use the EDF browser Nihon Kohden to EDF+ tool. The output had more precise annotations - including milliseconds. Will also try to validate this on BESA

Actual results

mne.io.read_raw_nihon.annotations returns the 'onset' timestamp rounded to the nearest second

Additional information

I've played around with the _read_nihon_annotations method in
https://github.com/mne-tools/mne-python/blob/main/mne/io/nihon/nihon.py without any luck. Unable to decode to LOG files into a format other than the datetime '%H%M%S' format provided