# How to use the CHB-MIT dataset

**URL:** <https://mne.discourse.group/t/how-to-use-the-chb-mit-dataset/8937>\
**Category:** Support & Discussions\
**Tags:** preprocessing, eeg\
**Created:** [June 14, 2024, 1:16pm UTC](https://mne.discourse.group/t/how-to-use-the-chb-mit-dataset/8937 "2024-06-14T13:16:32Z")\
**Posts on this page:** 4\
**Page:** 1

<div class="post-metadata">

**Author:** ![AaryaAshok77](https://yyz2.discourse-cdn.com/free1/user_avatar/mne.discourse.group/aaryaashok77/32/3723_2.png) [@AaryaAshok77](https://mne.discourse.group/u/AaryaAshok77)\
**Post date:** [June 14, 2024, 1:16pm UTC](https://mne.discourse.group/t/how-to-use-the-chb-mit-dataset/8937/1 "2024-06-14T13:16:32Z")

</div>

I am trying to use the CHB-MIT dataset for some Deep learning project but the data set consists of different patients who were being monitored using equipment with varying number of channels, how do I go about fixing it?

Version: 1.7  
Operating system: MacOS

---

<div class="post-metadata">

**Author:** ![AaryaAshok77](https://yyz2.discourse-cdn.com/free1/user_avatar/mne.discourse.group/aaryaashok77/32/3723_2.png) [@AaryaAshok77](https://mne.discourse.group/u/AaryaAshok77)\
**Post date:** [June 16, 2024, 11:18am UTC](https://mne.discourse.group/t/how-to-use-the-chb-mit-dataset/8937/2 "2024-06-16T11:18:31Z")

</div>

@cbrnr I noticed that you have helped on other doubts related to CHB-MIT please advice.

---

<div class="post-metadata">

**Author:** ![cbrnr](https://yyz2.discourse-cdn.com/free1/user_avatar/mne.discourse.group/cbrnr/32/1409_2.png) [@cbrnr](https://mne.discourse.group/u/cbrnr)\
**Post date:** [June 17, 2024, 2:38pm UTC](https://mne.discourse.group/t/how-to-use-the-chb-mit-dataset/8937/3 "2024-06-17T14:38:15Z")

</div>

I’m not sure I understand your question. If the data sets contain different channels, and you want a consistent subset of identical channels across all patients, you could select channels that are present for all data sets.

---

<div class="post-metadata">

**Author:** ![AaryaAshok77](https://yyz2.discourse-cdn.com/free1/user_avatar/mne.discourse.group/aaryaashok77/32/3723_2.png) [@AaryaAshok77](https://mne.discourse.group/u/AaryaAshok77)\
**Post date:** [June 17, 2024, 3:51pm UTC](https://mne.discourse.group/t/how-to-use-the-chb-mit-dataset/8937/4 "2024-06-17T15:51:00Z")

</div>

The data set consists of several files few files having 22 channels and some range all the way up to 40. Currently I am doing what you suggested and taking only the 22 common channels and making the model. But I wanted to know if there is any alternative.