Article Not peer-reviewed version

A Machine Learning Approach to

Preictal Phase Detection in EEG

Samayan Bhattacharya * , Alexis Bennett , Kseniia Kriukova , Celina Alba , Dominique Duncan

Posted Date: 26 February 2024

doi: 10.20944/preprints202402.1441.v1

Keywords: epilepsy; seizure; machine learning; kurtosis; spectral entropy

Preprints.org is a free multidiscipline platform providing preprint service that

is dedicated to making early versions of research outputs permanently
available and citable. Preprints posted at Preprints.org appear in Web of
Science, Crossref, Google Scholar, Scilit, Europe PMC.

Copyright: This is an open access article distributed under the Creative Commons
Attribution License which permits unrestricted use, distribution, and reproduction in any
medium, provided the original work is properly cited.
Preprints.org (www.preprints.org) | NOT PEER-REVIEWED | Posted: 26 February 2024 doi:10.20944/preprints202402.1441.v1

Disclaimer/Publisher’s Note: The statements, opinions, and data contained in all publications are solely those of the individual author(s) and
contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting
from any ideas, methods, instructions, or products referred to in the content.

Article
A Machine Learning Approach to Preictal Phase
Detection in EEG
Samayan Bhattacharya * , Alexis Bennett , Kseniia Kriukova , Celina Alba and Dominique
Duncan
Laboratory of Neuro Imaging, USC Stevens Neuroimaging and Informatics Institute, Keck School of Medicine of
USC, University of Southern California, 2025 Zonal Avenue Los Angeles, CA, USA; [email protected] (A.B.);
[email protected] (K.K.); [email protected] (C.A.); [email protected] (D.D.)
* Correspondence: [email protected]; Tel.: +91-9073056877 (F.L.)

Abstract: Epilepsy is a common neurological condition, typically diagnosed using

Electroencephalogram (EEG). Large scale EEG datasets have recently been made publicly
available, allowing the use of advanced Machine Learning (ML) algorithms to analyze EEG patterns
associated with epilepsy. While most existing studies focus on identifying seizures in the EEG, few
have tried to identify preictal EEG segments. Identifying preictal EEG segments are not only useful
in developing early warning systems but also helps inform the course of treatment for the patient. In
this study, we propose to represent EEG segments as images, instead of time-series data, and identify
preictal EEG segments using a preexisting ML algorithm (YOLOv8) designed for image processing.
Multiplexed images (containing the original EEG signal represented on a 2D grid, kurtosis, and
spectral entropy) achieve the best accuracy of 95.75% on the dataset while images just containing the
EEG signal result in an accuracy of 91.25%. Using only spectrograms, generated from the original
EEG signal, results in an accuracy of 90.15%.

Keywords: epilepsy; seizure; machine learning; kurtosis; spectral entropy

1. Introduction
Epilepsy is a common neurological condition, characterized by spontaneous seizures, and affecting
about 1% of the global population [1]. This results in a direct medical cost of $24.4 billion in the U.S.
annually [2]. Despite the best efforts by healthcare providers towards early diagnosis and treatment of
the symptoms of epilepsy, around 50% of patients continue to experience seizures, leading to a lower
quality of life [3]. Electroencephalography (EEG) is commonly used for diagnosing seizures in humans
and for studying seizures in animal models. The low cost and easy collection of EEG allows its use in
wearable devices for continuous monitoring. This allows the prediction of seizures, hence reducing
its adverse effects [4]. Better seizure prediction techniques increase the effectiveness of intervention
devices that prevent the occurrence of a seizure by electrical stimulation or by administering acute
medication. Electrical brain activity observed through EEG can vary depending on the subject’s
condition and state, and is typically categorized into four main distinct phases. These include the
interictal phase, which consists of normal brain activity and is also present in healthy individuals;
the preictal phase, which precedes a seizure; the ictal phase, which starts with the seizure onset and
concludes with the seizure’s end; and the postictal phase, which follows the ictal phase and lasts until
the next interictal phase begins.
Animal models are often used for studying seizures and developing better models for seizure
prediction in a research setting. This is because it is legal, ethical, and more affordable to induce
neurological conditions in other animals and obtain continuous recordings spanning several hours.
While traditionally, EEG is analyzed by human experts, it is infeasible and increases inter-observer
variability for multi-hour recordings [5–7].
Most studies focus on the automatic detection of seizures, however, few try to identify patterns
in an EEG preceding a seizure. This is probably because there is no clear definition of how long

© 2024 by the author(s). Distributed under a Creative Commons CC BY license.

Preprints.org (www.preprints.org) | NOT PEER-REVIEWED | Posted: 26 February 2024 doi:10.20944/preprints202402.1441.v1

2 of 9

before the occurrence of a seizure the preictal phase starts and there are several interpretations of this
interval [8–10]. The application of mathematical theories of non-linear dynamics initially allowed
the beginning of preictal interval to be identified as the point beyond which brain activity develops
deterministically, leading up to a seizure [11,12]. In other words, it is a point of no return, which once
passed, a seizure will occur. However, the interval between that point and the occurrence of a seizure
varies across patients [11,13].
Among automatic preictal phase-detection methods, statistical approaches compare the
distribution of preictal and interictal EEG segments while some algorithmic approaches perform
a grid search on preictal segments [14]. Machine learning (ML) algorithms have widely been used for
EEG analysis collected both from humans, mostly in clinical settings, and other animals, mostly in
research settings [15–17]. Advanced ML algorithms have proven effective in detecting seizure in both
humans, usually scalp EEG [18–22] and rodents [23].
Very little work has been done on the detection of preictal phase in EEG. Among relevant
literature, [24] describes various feature extraction techniques before classifying ictal, preictal, and
interictal phases of EEG using simple ML algorithms like a Support Vector Machine (SVM). Features
like sharpness of the waves, expressed as kurtosis and spectral entropy, expressing the amount of
information contained in the waves. On the other hand, [25] uses a simple Convolutional Neural
Network (CNN) to perform the same classification using frequency spectrogram. [26] was the first to
used unsupervised learning to identify preictal phase in EEG. They used clustering to group together
interictal patterns. Later, [27–29] to study a relationship between the patterns observed in the preictal
phase and the type of seizure.
CNNs have been widely applied to image classification due to their ability to localize important
features from data. Though not easily accessible, such features can still be investigated by methods
such as Grad-CAM [30]. Their widespread application allows easy access to large models, pre-trained
on large datasets. Transfer learning allows these models to perform better on relatively smaller datasets,
which is very relevant for medical data analysis. YOLOv8 is the latest version of the You Only Look
Once (YOLO) architecture [31]. It uses 50 layers of CNNs with cross-stage partial connections for
better information flow. In this work, we propose to classify raw EEG data into preictal and interictal
segments using a deep convolutional neural network (DCNN). While DCNNs have previously been
applied to EEG data, to the best of our knowledge, our work is the first to use a DCNN to analyze raw
EEG data. This allows explanation techniques to be applied to identify regions of interest in the raw
EEG data.

2. Materials and Methods

2.1. Dataset
We used an EEG dataset, publicly available on Kaggle [32]. It was a part of a competition
organized by the National Institutes of Health (NINDS), the Epilepsy Foundation, and the American
Epilepsy Society. Data was collected, using implantable leads, from 3 dogs with naturally occurring
epilepsy. The voltages were recorded from 16 channels, at a sampling rate of 400Hz, and referenced
to the group average. Recordings spanned from several months, up to a year. Preictal intervals were
considered to be one hour long, ending 5 minutes before the beginning of a seizure. This was done
to avoid contamination from any seizure activity before the start of the seizure, annotated by an
epileptologist. Interictal intervals, of one hour duration, were considered as far apart, from seizures, as
possible. Overall, there were 2744 interictal segments and 211 preictal segments.

2.2. Preprocessing
EEG is downsampled from 400Hz to 100Hz. Since the number of interictal segments is an order of
magnitude greater than the number of preictal segments, we consider the 4 segments produced from
each segment, during downsampling, for preictal segments but consider only the first one out of the 4
Preprints.org (www.preprints.org) | NOT PEER-REVIEWED | Posted: 26 February 2024 doi:10.20944/preprints202402.1441.v1

3 of 9

for interictal segments. This results in 844 preictal segments and 2744 interictal segments, out of which
1000 are considered for a better class balance. Raw EEG is expressed as time series data while CNNs
accept image data. A common approach to representing EEG as image data is in the form of frequency
spectrograms. However, such a representation does not allow the mapping of explanations, generated
by methods such as Grad-CAM, back to the raw EEG data. Hence, instead of representing the raw
EEG in an alternate form, we represent the voltage values as pixels in an image, in a row-major way, as
shown in Figure 1. The channels of a segment are concatenated before laying the voltage values on a
2D grid.
Most pre-trained CNNs use RGB images with 3 channels. The raw EEG laid out on a 2D grid is
one channel, so we include kurtosis and spectral entropy as the two other channels. Both are calculated
with a sliding window of size 100, without overlap. The consecutive values are padded by zeroes.
This is then laid out on a 2D grid, as described above. This results in a sparse matrix, with 99 zeroes
between 2 values as shown in Figure 2. Due to the pooling layers in the CNN, the non-zero values,
get focused on in the deeper layers because any trainable weight, multiplied by the zeroes, results in
zeroes. Since the number of voltage values is not a perfect square, we append zeroes to the end, to
obtain a square of dimensions 980x980. Thus, the resultant images have a dimension of 980x980x3.

2.3. Supervised Learning

Supervised learning algorithms depend on large annotated datasets to train well. However,
given the cost of collecting and annotating medical data, it is often infeasible to train large supervised
learning algorithms, from scratch, for medical applications. Thankfully, due to the popularity of
DCNN models, several pre-trained DCNNs are publicly available. These were trained on real-life RGB
images, from large-scale datasets like Imagenet [33] and COCO datasets [34]. For this work, we use a
pre-trained YOLOv8, from Ultralytics [35]. Due to memory restrictions, we use the medium version of
the YOLOv8 model, with 50 CNN layers and 78.9 million trainable parameters. It was pre-trained on
the COCO dataset [34]. The model is trained for 100 epochs with the default hyperparameters. 80%
of the data is used for training and 20% is used for testing. It is ensured that the copies of preictal
segments, obtained during downsampling do not become part of two different sets. Weights are saved
every 5 epochs and the weights corresponding to the best performance are used for the next step.
Due to memory restrictions during training, we added a padding of 10 on each side of a 980x980x3
image for all 3 channels, resulting in an image of dimensions 1000x1000x3. This was divided into 4
images of dimensions 250x250x3. The 4 images were provided as a batch to the model, with each
image having the label of its parent image.
For comparison, the same model was trained and tested on a dataset containing only the raw EEG
laid on a 2D grid (without the kurtosis and spectral entropy channels). To have 3 channels in these
images, the single channel was copied twice and added to the image. This resulted in a 980x980x3
image, with each 980x980 slice being copies of one another. These were used for training in the same
way as the images with raw eeg, kurtosis and spectral entropy (described in the previous paragraph).
For comparison with the traditional approach, frequency spectrograms were generated for the EEG
segments and used to train and test the model as well.
We also test the performance of the model by training it on data from one animal and testing it on
20% of the data from another animal.

2.4. Generating Explanation

Once the model is trained, we use Grad-CAM [30] to generate heatmaps. The heatmaps indicate
how important each region in the image is to the model. Grad-CAM heatmaps are widely used to
explain the performance of DCNNs on real-life RGB images, for example, if an image is classified as a
cat, the heatmap, desirably, indicates the region of the image containing a cat is the most important.
For our application, the heatmap tells us which regions of the image made the model "think" that it is a
Preprints.org (www.preprints.org) | NOT PEER-REVIEWED | Posted: 26 February 2024 doi:10.20944/preprints202402.1441.v1

4 of 9

preictal segment. The most important regions in the raw EEG are then highlighted to observe relevant
patterns in the preictal EEG. This is summarized in Figure 3.

3. Results
As shown in Table 1, the model performs best when trained on the images containing the raw
EEG, kurtosis, and spectral entropy, with an accuracy of 95.75%, precision of 0.93, recall of 0.98, and
F1 score of 0.95. The performances are close when the model is trained on only EEG and frequency
spectrogram generated from EEG data, with a better performance for the EEG data.
As shown in Table 2, the model performs better, when it is trained and tested on data from the
same animal, than when it is trained on data from one animal and tested on data from another animal.

Table 1. Performance metrics for the model trained and tested on different types of images.
Data Accuracy (%) Precision Recall F1 score
EEG+Kurtosis+Spectral Entropy 95.75 0.932 0.976 0.953
EEG 91.25 0.875 0.933 0.903
Frequency Spectrogram 90.15 0.842 0.942 0.889

Figure 1. A representation of the voltage values in the time series EEG data (represented as natural
numbers) being laid out on a 2D grid.
Preprints.org (www.preprints.org) | NOT PEER-REVIEWED | Posted: 26 February 2024 doi:10.20944/preprints202402.1441.v1

5 of 9

Figure 2. A representation of the voltage values in the 2D image being used to generate 2D images
with Kurtosis and Spectral Entropy values, using a window size of 100.

Table 2. Accuracy (%) of the model trained on EEG+Kurtosis+Spectral Entropy images of the animal in
the row name and tested on images of the animal in the column name.
Dog 1 Dog 2 Dog 3
Dog 1 93.97 89.41 90.77
Dog 2 96.12 97.21 95.54
Dog 3 94.15 91.49 95.32

4. Discussion
In this study, we demonstrated that a pre-existing powerful DCNN can be repurposed to analyze
EEG data in its original form. We observe that including additional features, namely kurtosis and
spectral entropy improves the performance of the model. Kurtosis, broadly speaking, represents
the sharpness of the waves. Seizures are known to contain more sharp spikes, suggesting their
possible relevance for preictal phase identification as well. Spectral entropy represents the amount of
information contained in a wave. It is known to be lower during a seizure. Hence, we hypothesized
that this might be an important feature for identifying the preictal phase as well. Though we considered
these two features, there are a lot of other features that remain unexplored. The same idea also applies
to the DCNN model used. We used YOLOv8 because it is the state of the art, at the moment, for real-life
object detection. However, it is not specialized for EEG analysis. A comprehensive study comparing
the performance of various models trained on different features would be crucial to determine the best
combination for preictal phase detection.
The use of raw EEG to train the model allows us to generate explanations, in the form of
Grad-CAM heatmap, on the raw EEG. This allows us to localize patterns in the EEG that the model
considers relevant to detecting the preictal phase. Such patterns might help experts look for consistency
in the highlighted patterns and potentially discover biomarkers of specific conditions.
We observe that the model performs better when it is trained and tested on data obtained from
the same animal. This might be explained by subtle differences in brain activity among individual
animals with the same underlying condition. There are also minute differences in the position of the
electrodes for different animals. The model seems to be able to pick up on these subtle differences. The
wide disparity in performance, when the model is trained on data from different animals, can also be
attributed to the difference in the number of segments from each animal. The original dataset had 42,
72, and 97 preictal and 500, 1440, and 804 interictal segments from dogs 1, 2, and 3 respectively. We
obtained 4 segments from each preictal segment by downsampling, as described in the preprocessing
Preprints.org (www.preprints.org) | NOT PEER-REVIEWED | Posted: 26 February 2024 doi:10.20944/preprints202402.1441.v1

6 of 9

section above. We also considered a proportional number of interictal segments such that the total
number of interictal segments was equal to 1000. Thus, our final dataset contained 168, 288, and 388
preictal and 182, 525, and 293 interictal segments from dogs 1, 2, and 3 respectively.

Figure 3. A representation of the time series representation of raw EEG being converted to a 2D image
with Raw EEG + Kurtosis + Spectral Entropy. A grad-CAM heatmap is then generated from the trained
model and superimposed on the 2D image. The priority values, from the heatmap, are then mapped
back to the time series representation of the raw EEG to observe important patterns in the preictal
phase. In the last figure, the complete signal is shown on the bottom, the corresponding heatmap is
shown in the middle and the parts of the signal with the highest importance are shown on top.

5. Conclusions
In this work we demonstrated that (1) A pre-existing DCNN model i.e. YOLOv8 can be used for
EEG analysis, (2) The model performs better when features such as Kurtosis and Spectral entropy are
provided along with raw EEG, (3) The model performs better when it is trained and tested on data
from the same animal, and (4) Grad-CAM heatmaps can be mapped to the raw EEG signal to localize
regions of interest and potentially recognize important patterns. A comprehensive study comparing
more features used to train more neural networks might help obtain better results.

Author Contributions: Conceptualization, S.B.; Investigation, S.B., A.B., K.K., C.A., and D.D.; Methodology,
S.B., A.B., K.K., C.A., and D.D.; Project administration, D.D.; Supervision, D.D.; Validation, S.B., A.B., K.K., C.A.,
and D.D.; Visualization, S.B., A.B., K.K., C.A., and D.D.; Writing—original draft, S.B., A.B., K.K., C.A., and D.D.;
Writing—review and editing, S.B., A.B., K.K., C.A., and D.D. All authors have read and agreed to the published
version of the manuscript.
Preprints.org (www.preprints.org) | NOT PEER-REVIEWED | Posted: 26 February 2024 doi:10.20944/preprints202402.1441.v1

7 of 9

Funding: This study was conducted with the support of the National Institutes of Health (NIH) under award
number R01NS111744.
Institutional Review Board Statement: Not applicable
Informed Consent Statement: Not applicable
Data Availability Statement: All data used in this study is publicly available at https://www.kaggle.com/
competitions/seizure-prediction.
Conflicts of Interest: The authors declare no conflicts of interest. The funders had no role in the design of the
study; in the collection, analyses, or interpretation of data; in the writing of the manuscript; or in the decision to
publish the results.

Abbreviations
The following abbreviations are used in this manuscript:

ML Machine Learning
CNN Convolutional Neural Network
DCNN Deep Convolutional Neural Network
YOLO You Only Look Once
EEG Electroencephalogram
SVM Support Vector Machine
NINDS National Institutes of Health

References
1. Kobau, R.; Zahran, H.; Thurman, D.J.; Zack, M.M.; Henry, T.R.; Schachter, S.C.; Price, P.H. Epilepsy
surveillance among adults–19 states, behavioral risk factor surveillance system, 2005 2008.
2. Begley, C.E.; Famulari, M.; Annegers, J.F.; Lairson, D.R.; Reynolds, T.F.; Coan, S.; Dubinsky, S.; Newmark,
M.E.; Leibson, C.; So, E.; et al. The cost of epilepsy in the United States: an estimate from population-based
clinical and survey data. Epilepsia 2000, 41, 342–351.
3. Loring, D.W.; Meador, K.J.; Lee, G.P. Determinants of quality of life in epilepsy. Epilepsy & Behavior 2004,
5, 976–980.
4. Traynelis, S.F.; Dlugos, D.; Henshall, D.; Mefford, H.C.; Rogawski, M.A.; Staley, K.J.; Dacks, P.A.; Whittemore,
V.; Poduri, A.; of Neurological Disorders, N.I.; et al. Epilepsy benchmarks area III: improved treatment
options for controlling seizures and epilepsy-related conditions without side effects. Epilepsy currents 2020,
20, 23S–30S.
5. Abend, N.S.; Gutierrez-Colina, A.; Zhao, H.; Guo, R.; Marsh, E.; Clancy, R.R.; Dlugos, D.J. Interobserver
reproducibility of electroencephalogram interpretation in critically ill children. Journal of clinical
neurophysiology: official publication of the American Electroencephalographic Society 2011, 28, 15.
6. Gerber, P.A.; Chapman, K.E.; Chung, S.S.; Drees, C.; Maganti, R.K.; Ng, Y.t.; Treiman, D.M.; Little, A.S.;
Kerrigan, J.F. Interobserver agreement in the interpretation of EEG patterns in critically ill adults. Journal of
Clinical Neurophysiology 2008, 25, 241–249.
7. Williams, G.W.; Lüders, H.O.; Brickner, A.; Goormastic, M.; Klass, D.W. Interobserver variability in EEG
interpretation. Neurology 1985, 35, 1714–1714.
8. Mormann, F.; Kreuz, T.; Rieke, C.; Andrzejak, R.G.; Kraskov, A.; David, P.; Elger, C.E.; Lehnertz, K. On the
predictability of epileptic seizures. Clinical neurophysiology 2005, 116, 569–587.
9. Valderrama, M.; Alvarado, C.; Nikolopoulos, S.; Martinerie, J.; Adam, C.; Navarro, V.; Le Van Quyen, M.
Identifying an increased risk of epileptic seizures using a multi-feature EEG–ECG classification. Biomedical
Signal Processing and Control 2012, 7, 237–244.
10. Teixeira, C.A.; Direito, B.; Bandarabadi, M.; Le Van Quyen, M.; Valderrama, M.; Schelter, B.; Schulze-Bonhage,
A.; Navarro, V.; Sales, F.; Dourado, A. Epileptic seizure predictors based on computational intelligence
techniques: A comparative study with 278 patients. Computer methods and programs in biomedicine 2014,
114, 324–336.
11. Blauwblomme, T.; Jiruska, P.; Huberfeld, G. Mechanisms of ictogenesis. In International review of neurobiology;
Elsevier, 2014; Vol. 114, pp. 155–185.
Preprints.org (www.preprints.org) | NOT PEER-REVIEWED | Posted: 26 February 2024 doi:10.20944/preprints202402.1441.v1

8 of 9

12. Iasemidis, L.D. Epileptic seizure prediction and control. IEEE Transactions on Biomedical Engineering 2003,
50, 549–558.
13. Freestone, D.R.; Karoly, P.J.; Peterson, A.D.; Kuhlmann, L.; Lai, A.; Goodarzy, F.; Cook, M.J. Seizure
prediction: science fiction or soon to become reality? Current neurology and neuroscience reports 2015, 15, 1–9.
14. Mormann, F.; Andrzejak, R.G.; Elger, C.E.; Lehnertz, K. Seizure prediction: the long and winding road. Brain
2007, 130, 314–333.
15. Gabor, A. Seizure detection using a self-organizing neural network: validation and comparison with other
detection strategies. Electroencephalography and clinical Neurophysiology 1998, 107, 27–32.
16. Gotman, J. Automatic recognition of epileptic seizures in the EEG. Electroencephalography and clinical
Neurophysiology 1982, 54, 530–540.
17. Ein Shoka, A.A.; Dessouky, M.M.; El-Sayed, A.; Hemdan, E.E.D. EEG seizure detection: concepts, techniques,
challenges, and future trends. Multimedia Tools and Applications 2023, pp. 1–31.
18. Farooq, M.S.; Zulfiqar, A.; Riaz, S. A review on Epileptic Seizure Detection using Machine Learning. arXiv
preprint arXiv:2210.06292 2022.
19. Abdelhameed, A.; Bayoumi, M. A deep learning approach for automatic seizure detection in children with
epilepsy. Frontiers in Computational Neuroscience 2021, 15, 650050.
20. Siddiqui, M.K.; Morales-Menendez, R.; Huang, X.; Hussain, N. A review of epileptic seizure detection using
machine learning classifiers. Brain informatics 2020, 7, 1–18.
21. Raghu, S.; Sriraam, N.; Temel, Y.; Rao, S.V.; Kubben, P.L. EEG based multi-class seizure type classification
using convolutional neural network and transfer learning. Neural Networks 2020, 124, 202–212.
22. Gao, Y.; Gao, B.; Chen, Q.; Liu, J.; Zhang, Y. Deep convolutional neural network-based epileptic
electroencephalogram (EEG) signal classification. Frontiers in neurology 2020, 11, 375.
23. Bergstrom, R.A.; Choi, J.H.; Manduca, A.; Shin, H.S.; Worrell, G.A.; Howe, C.L. Automated identification of
multiple seizure-related and interictal epileptiform event types in the EEG of mice. Scientific reports 2013,
3, 1483.
24. Usman, S.M.; Usman, M.; Fong, S.; et al. Epileptic seizures prediction using machine learning methods.
Computational and mathematical methods in medicine 2017, 2017.
25. Zhou, M.; Tian, C.; Cao, R.; Wang, B.; Niu, Y.; Hu, T.; Guo, H.; Xiang, J. Epileptic seizure detection based on
EEG signals and CNN. Frontiers in neuroinformatics 2018, 12, 95.
26. Le Van Quyen, M.; Soss, J.; Navarro, V.; Robertson, R.; Chavez, M.; Baulac, M.; Martinerie, J. Preictal state
identification by synchronization changes in long-term intracranial EEG recordings. Clinical Neurophysiology
2005, 116, 559–568.
27. Li, F.; Liang, Y.; Zhang, L.; Yi, C.; Liao, Y.; Jiang, Y.; Si, Y.; Zhang, Y.; Yao, D.; Yu, L.; et al. Transition of brain
networks from an interictal to a preictal state preceding a seizure revealed by scalp EEG network analysis.
Cognitive Neurodynamics 2019, 13, 175–181.
28. Quercia, A.; Frick, T.; Egli, F.E.; Pullen, N.; Dupanloup, I.; Tang, J.; Asif, U.; Harrer, S.; Brunschwiler, T.
Preictal onset detection through unsupervised clustering for epileptic seizure prediction. In Proceedings of
the 2021 IEEE International Conference on Digital Health (ICDH). IEEE, 2021, pp. 142–147.
29. Leal, A.; Pinto, M.F.; Lopes, F.; Bianchi, A.M.; Henriques, J.; Ruano, M.G.; de Carvalho, P.; Dourado, A.;
Teixeira, C.A. Heart rate variability analysis for the identification of the preictal interval in patients with
drug-resistant epilepsy. Scientific reports 2021, 11, 5987.
30. Selvaraju, R.R.; Das, A.; Vedantam, R.; Cogswell, M.; Parikh, D.; Batra, D. Grad-CAM: Why did you say
that? arXiv preprint arXiv:1611.07450 2016.
31. Redmon, J.; Divvala, S.; Girshick, R.; Farhadi, A. You only look once: Unified, real-time object detection. In
Proceedings of the Proceedings of the IEEE conference on computer vision and pattern recognition, 2016, pp.
779–788.
32. American Epilepsy Society Seizure Prediction Challenge — kaggle.com. https://www.kaggle.com/
competitions/seizure-prediction. [Accessed 23-02-2024].
33. Deng, J.; Dong, W.; Socher, R.; Li, L.J.; Li, K.; Fei-Fei, L. Imagenet: A large-scale hierarchical image database.
In Proceedings of the 2009 IEEE conference on computer vision and pattern recognition. Ieee, 2009, pp.
248–255.
Preprints.org (www.preprints.org) | NOT PEER-REVIEWED | Posted: 26 February 2024 doi:10.20944/preprints202402.1441.v1

9 of 9

34. Lin, T.Y.; Maire, M.; Belongie, S.; Hays, J.; Perona, P.; Ramanan, D.; Dollár, P.; Zitnick, C.L. Microsoft coco:
Common objects in context. In Proceedings of the Computer Vision–ECCV 2014: 13th European Conference,
Zurich, Switzerland, September 6-12, 2014, Proceedings, Part V 13. Springer, 2014, pp. 740–755.
35. Ultralytics. Home — docs.ultralytics.com. https://docs.ultralytics.com/. [Accessed 23-02-2024].

Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those
of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s)
disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or
products referred to in the content.