Page 126 - 2024-Vol20-Issue2

122 | Assim & Mahmood

TABLE V.
DESCRIPTION FOR POPULAR DATASETS

Fig. 3. DNN models percentage used in paper. Ref Description Website
[56] Bonn University:
TABLE IV. contains five www.upf.edu/web
POPULAR EEG DATASETS [55] classes and 11500 /ntsa/downloads/
[40] EEG samples,
Ref Dataset Participant Channels and at 173.61 Hz, it is physionet.org
29 sampled and recorded. /content/chbmit
[49] Siena Scalp 14 21 A, B, C, D, and E /1.0.0/
EEG Database represent the five
21 categories that www.response
Neurologists at 19 make up the dataset. studie.nl/
19 Using the common
Nilratan Sircar 57 10-20 electrode
23-26 placement approach,
[50] Medical College 150 100 each category
has 100 recorded
and Hospital EEG data.
Each EEG signal is
(NRSMCH) 4097 bytes long.
CHB–MIT
Seoul National contains 24 people
with epileptic seizures,
[51] University 25 ranging in age from
Hospital (SNUH) 3 to 22.
A total of 23 channels
Scalp EEG database are used to monitor
each patient
Ramaiah Medical for 46 hours.
18 women and 5 men.
[52] College and 115 256 Hz is the sample
rate employed.
Hospital (RMCH) REPO2MSE cohort:
Consists of 568
NICU of Helsinki epileptic patients
multichannel scalp
[53] University Hospital 79 EEG recordings
with annotations
, Finland about the start of
seizures provided
Neurology by a knowledgeable
epileptologist.
[54] and sleep center 10 EEG data were
captured at
New Delhi EEG 256 Hz, 512 Hz,
or 1024 Hz.
Children’s

[55] Hospital 22
Boston–MIT

database

[56] Bonn EEG 5
time-series database