CAUEEG: Chung-Ang University Hospital EEG dataset for automatic EEG diagnosis research
This repository is the official page of the CAUEEG dataset presented in "Deep learning-based EEG analysis to classify mild cognitive impairment for early detection of dementia: algorithms and benchmarks" from the CNIR (CAU NeuroImaging Research) team.
Please refer to the academic paper, "Deep learning-based EEG analysis to classify mild cognitive impairment for early detection of dementia: algorithms and benchmarks," for detailed information.
❗ Note: The use of the CAUEEG dataset is allowed for only academic and research purposes 👩🎓👨🏼🎓.
- Researchers wishing to utilize the CAUEEG dataset should turn out this form as approval from the local ethics committee is required.
- After reviewing the response, we will send you a mail with a download link and password.
The file tree for the CAUEEG data set is as follows:
caueeg-dataset
| annotation.json
| annotation.xlsx
| dementia.json
| abnormal.json
|
+---event
| 00001.json
| 00002.json
| ...
| 01388.json
|
\---signal
+---edf
| 00001.edf
| 00002.edf
| ...
| 01388.edf
caueeg-dataset/annotation.json: Diagnostic annotation file for each EDF and event pair.caueeg-dataset/annotation.xlsx: Excel version ofannotation.jsonfor visualization.caueeg-dataset/dementia.json: Annotation file for theCAUEEG-Dementiabenchmark.caueeg-dataset/dementia-no-overlap.json: Annotation file for theCAUEEG-Dementiabenchmark, excluded all patient-overlapping cases with training data from the test and validation sets.caueeg-dataset/abnormal.json: Annotation file for theCAUEEG-Abnormalbenchmark.caueeg-dataset/abnormal-no-overlap.json: Annotation file for theCAUEEG-Abnormalbenchmark, excluded all patient-overlapping cases with training data from the test and validation sets.caueeg-dataset/event/*.json: Event descriptions during EEG recording.caueeg-dataset/signal/edf/*.edf: EEG signals written in European data format (EDF).
The implementation code for using the CAUEEG dataset and benchmarks via the PyTorch library is available in a public repository at caueeg-ceednet.
With caueeg-ceednet, here is an example of loading the CAUEEG dataset:
import pprint
from datasets.caueeg_script import *
config_data, full_eeg_dataset = load_caueeg_full_dataset(dataset_path='local/dataset/caueeg-dataset/',
load_event=False)
pprint.pprint(config_data, width=250)
print('\n', '-' * 100, '\n')
pprint.pprint(full_eeg_dataset[0])
print('\n', '-' * 100, '\n')
pprint.pprint(full_eeg_dataset[3]){'dataset_name': 'CAUEEG dataset',
'signal_header': ['Fp1-AVG', 'F3-AVG', 'C3-AVG', 'P3-AVG', 'O1-AVG', 'Fp2-AVG', 'F4-AVG', 'C4-AVG', 'P4-AVG', 'O2-AVG', 'F7-AVG', 'T3-AVG', 'T5-AVG', 'F8-AVG', 'T4-AVG', 'T6-AVG', 'FZ-AVG', 'CZ-AVG', 'PZ-AVG', 'EKG', 'Photic']}
----------------------------------------------------------------------------------------------------
{'age': 78,
'serial': '00001',
'signal': array([[ 0., -11., -13., ..., 0., 0., 0.],
[ 29., 33., 34., ..., 0., 0., 0.],
[ -3., -6., -3., ..., 0., 0., 0.],
...,
[ -4., -2., 1., ..., 0., 0., 0.],
[112., 67., 76., ..., 0., 0., 0.],
[ -1., -1., -1., ..., 0., 0., 0.]]),
'symptom': ['mci', 'mci_amnestic', 'mci_amnestic_rf']}
----------------------------------------------------------------------------------------------------
{'age': 78,
'serial': '00004',
'signal': array([[ 30., 34., 36., ..., 0., 0., 0.],
[ 34., 25., 24., ..., 0., 0., 0.],
[ 11., 15., 15., ..., 0., 0., 0.],
...,
[-10., -6., -4., ..., 0., 0., 0.],
[ 23., 32., 30., ..., 0., 0., 0.],
[ -1., 0., -1., ..., 0., 0., 0.]]),
'symptom': ['dementia', 'ad', 'load']}
With load_event option triggered, the dataset also loads event data, as follows:
config_data, full_eeg_dataset = load_caueeg_full_dataset(dataset_path='local/dataset/caueeg-dataset/',
load_event=True)
pprint.pprint(config_data, width=250)
print('\n', '-' * 100, '\n')
pprint.pprint(full_eeg_dataset[0]){'dataset_name': 'CAUEEG dataset',
'signal_header': ['Fp1-AVG', 'F3-AVG', 'C3-AVG', 'P3-AVG', 'O1-AVG', 'Fp2-AVG', 'F4-AVG', 'C4-AVG', 'P4-AVG', 'O2-AVG', 'F7-AVG', 'T3-AVG', 'T5-AVG', 'F8-AVG', 'T4-AVG', 'T6-AVG', 'FZ-AVG', 'CZ-AVG', 'PZ-AVG', 'EKG', 'Photic']}
----------------------------------------------------------------------------------------------------
{'age': 78,
'event': [[0, 'Start Recording'],
[0, 'New Montage - Montage 002'],
[36396, 'Eyes Open'],
[72518, 'Eyes Closed'],
[73862, 'Eyes Open'],
[75248, 'Eyes Closed'],
[76728, 'swallowing'],
[77978, 'Eyes Open'],
[79406, 'Eyes Closed'],
[79996, 'Photic On - 3.0 Hz'],
[80288, 'Eyes Open'],
[81296, 'Eyes Closed'],
[82054, 'Photic Off'],
[84070, 'Photic On - 6.0 Hz'],
[84488, 'Eyes Open'],
[85538, 'Eyes Closed'],
[86086, 'Photic Off'],
[88144, 'Photic On - 9.0 Hz'],
[90160, 'Photic Off'],
[91458, 'Eyes Open'],
[92218, 'Photic On - 12.0 Hz'],
[92762, 'Eyes Closed'],
[94198, 'Photic Off'],
[94742, 'Eyes Open'],
[95708, 'Eyes Closed'],
[96256, 'Photic On - 15.0 Hz'],
[98272, 'Photic Off'],
[100330, 'Photic On - 18.0 Hz'],
[102346, 'Photic Off'],
[102596, 'Eyes Open'],
[103856, 'Eyes Closed'],
[104361, 'Photic On - 21.0 Hz'],
[106420, 'Photic Off'],
[106880, 'Eyes Open'],
[107804, 'Eyes Closed'],
[108435, 'Photic On - 24.0 Hz'],
[110452, 'Photic Off'],
[111080, 'Eyes Open'],
[112004, 'Eyes Closed'],
[112509, 'Photic On - 27.0 Hz'],
[114528, 'Photic Off'],
[114864, 'Eyes Open'],
[116124, 'Eyes Closed'],
[116544, 'Photic On - 30.0 Hz'],
[118602, 'Photic Off'],
[126672, 'artifact'],
[134030, 'Move'],
[135584, 'Eyes Open'],
[136668, 'Eyes Closed'],
[139818, 'Eyes Open'],
[141414, 'Eyes Closed'],
[145000, 'Paused']],
'serial': '00001',
'signal': array([[ 0., -11., -13., ..., 0., 0., 0.],
[ 29., 33., 34., ..., 0., 0., 0.],
[ -3., -6., -3., ..., 0., 0., 0.],
...,
[ -4., -2., 1., ..., 0., 0., 0.],
[112., 67., 76., ..., 0., 0., 0.],
[ -1., -1., -1., ..., 0., 0., 0.]]),
'symptom': ['mci', 'mci_amnestic', 'mci_amnestic_rf']}
Here is an example of loading the CAUEEG-Dementia benchmark:
config_data, train_dataset, val_dataset, test_dataset = load_caueeg_task_datasets(dataset_path='local/dataset/caueeg-dataset/',
task='dementia',
load_event=False)
pprint.pprint(config_data)
print('\n', '-' * 100, '\n')
pprint.pprint(train_dataset[0])
print('\n', '-' * 100, '\n')
pprint.pprint(val_dataset[0])
print('\n', '-' * 100, '\n')
pprint.pprint(test_dataset[0])Similarly, the CAUEEG-Abnormal benchmark is loaded as follows:
config_data, train_dataset, val_dataset, test_dataset = load_caueeg_task_datasets(dataset_path='local/dataset/caueeg-dataset/',
task='abnormal',
load_event=False)
pprint.pprint(config_data)
print('\n', '-' * 100, '\n')
pprint.pprint(train_dataset[0])
print('\n', '-' * 100, '\n')
pprint.pprint(val_dataset[0])
print('\n', '-' * 100, '\n')
pprint.pprint(test_dataset[0])One can replace task='dementia' to task='dementia-no-overlap' or task='abnormal' to task='abnormal-no-overlap' if wanting to use the 'no-overlap' version, respectively.
To specify the split and benchmark, here is the way:
config, test_dataset = load_caueeg_task_split(dataset_path='local/dataset/caueeg-dataset/',
task='abnormal', split='test', load_event=False)
pprint.pprint(config_data)
print('\n', '-' * 100, '\n')
pprint.pprint(test_dataset[0]){'class_label_to_name': ['Normal', 'Abnormal'],
'class_name_to_label': {'Abnormal': 1, 'Normal': 0},
'task_description': 'Classification of [Normal] and [Abnormal] symptoms',
'task_name': 'CAUEEG-Abnormal benchmark'}
----------------------------------------------------------------------------------------------------
{'age': 57,
'class_label': 0,
'class_name': 'Normal',
'serial': '00560',
'signal': array([[ 7., 41., 49., ..., 0., 0., 0.],
[ 23., 26., 27., ..., 0., 0., 0.],
[-16., -20., -18., ..., 0., 0., 0.],
...,
[ 32., 31., 31., ..., 0., 0., 0.],
[177., 228., 142., ..., 0., 0., 0.],
[ 0., -1., 0., ..., 0., 0., 0.]]),
'symptom': ['normal', 'cb_normal']}
For more advanced usages of the CAUEEG dataset and two corresponding benchmarks, refer to this notebook.
If you found this dataset helpful, please cite the paper below.
@article{kim2023deep,
title={Deep learning-based EEG analysis to classify normal, mild cognitive impairment, and dementia: algorithms and dataset},
author={Kim, Min-jae and Youn, Young Chul and Paik, Joonki},
journal={NeuroImage},
pages={120054},
year={2023},
publisher={Elsevier}
}