Advanced Deep Learning Spectroscopy of Scalogram Infused CNN Classifiers for Robust Identification of Post‐Hypoxic Epileptiform EEG Spikes

Abbasi, Hamid; Gunn, Alistair J.; Unsworth, Charles P.

doi:10.1002/aisy.202000198

Cited by 14 publications

(18 citation statements)

References 32 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Meanwhile, the most commonly adopted feature that is widely implemented in EEG signal characterization includes timedomain (TD), time-frequency domain (TFD), frequency domain (FD), fourier transform (FT), discrete wavelet transform (DWT), and continuous wavelet transform (CWT)-based features [30]. Abbasi models to detect HI brain injury and got satisfactory results [31]. Logesparan et al [32] used various statistical feature extraction methods on EEG datasets but concentrated on only two features, "relative power" and "line length," which produced better performance in seizure detection.…”

Section: Feature Extraction and Selection Techniques Applied Inmentioning

confidence: 99%

EEG-Based Epileptic Seizure Detection via Machine/Deep Learning Approaches: A Systematic Review

Ahmad

Wang

Zhu

et al. 2022

Computational Intelligence and Neuroscience

View full text Add to dashboard Cite

Epileptic seizure is one of the most chronic neurological diseases that instantaneously disrupts the lifestyle of affected individuals. Toward developing novel and efficient technology for epileptic seizure management, recent diagnostic approaches have focused on developing machine/deep learning model (ML/DL)-based electroencephalogram (EEG) methods. Importantly, EEG’s noninvasiveness and ability to offer repeated patterns of epileptic-related electrophysiological information have motivated the development of varied ML/DL algorithms for epileptic seizure diagnosis in the recent years. However, EEG’s low amplitude and nonstationary characteristics make it difficult for existing ML/DL models to achieve a consistent and satisfactory diagnosis outcome, especially in clinical settings, where environmental factors could hardly be avoided. Though several recent works have explored the use of EEG-based ML/DL methods and statistical feature for seizure diagnosis, it is unclear what the advantages and limitations of these works are, which might preclude the advancement of research and development in the field of epileptic seizure diagnosis and appropriate criteria for selecting ML/DL models and statistical feature extraction methods for EEG-based epileptic seizure diagnosis. Therefore, this paper attempts to bridge this research gap by conducting an extensive systematic review on the recent developments of EEG-based ML/DL technologies for epileptic seizure diagnosis. In the review, current development in seizure diagnosis, various statistical feature extraction methods, ML/DL models, their performances, limitations, and core challenges as applied in EEG-based epileptic seizure diagnosis were meticulously reviewed and compared. In addition, proper criteria for selecting appropriate and efficient feature extraction techniques and ML/DL models for epileptic seizure diagnosis were also discussed. Findings from this study will aid researchers in deciding the most efficient ML/DL models with optimal feature extraction methods to improve the performance of EEG-based epileptic seizure detection.

show abstract

Section: Feature Extraction and Selection Techniques Applied Inmentioning

confidence: 99%

EEG-Based Epileptic Seizure Detection via Machine/Deep Learning Approaches: A Systematic Review

Ahmad

Wang

Zhu

et al. 2022

Computational Intelligence and Neuroscience

View full text Add to dashboard Cite

show abstract

“…We have previously shown that micro-scale sharp-wave EEG patterns with amplitudes between 20-80 µV and a duration between 70 and 250 ms (frequency range: 4 to 14.3 Hz, in the θ (4-8 Hz), α (8-12 Hz), and lower-beta β band (i.e., 12-14.3 Hz)) were superimposed on a suppressed EEG background, within the latent phase of fetal sheep data (103 ± 1 days, human brain maturation equivalent ~28-30 weeks) which are a reliable marker for hypoxic-ischemic encephalopathy (HIE) (see Figure 1c,d) [6,15]. We have shown that EEG sharp-waves can help to predict the latent phase of injury after an HI insult, where a larger number of sharp-waves within the first 30 min post-HI is associated with greater subcortical neuronal survival in the caudate nucleus (r = 0.80).…”

Section: Micro-scale Sharp-wave Eeg Waveforms-an Hi Biomarkermentioning

confidence: 99%

“…Our team has previously shown that in preterm fetal sheep these EEG waveforms emerge in the form of micro-scale sharp-waves and gamma spike transients, and that they are significantly correlated with the subcortical brain damage post HI injury [6,15]. Our team has successfully developed and validated automated advanced signal processing technology, based on deep-learning, for the identification, quantification, and localization of these patterns in preterm fetal sheep data (accuracy > 99%) [15,16]. We have previously shown that micro-scale sharp-waves, in particular, contain timing information related to the evolution of injury in the latent phase after HI.…”

Section: Introductionmentioning

confidence: 99%

“…Hence, there is a significant need for developing automated signal processing techniques to assist with reliable assessments of EEG in suspected cases of brain injury [4,19,[25][26][27][28][29]. One approach is to focus on automatic identification and quantification of micro-scale EEG patterns as well as seizures in order to determine the timing and effectively target treatment for infants who might benefit [9,15,17,25,26,30,31].…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Early Signatures of Brain Injury in the Preterm Neonatal EEG

Abbasi,

Battin,

Butler

et al. 2023

Signals

Self Cite

View full text Add to dashboard Cite

Reliable prognostic biomarkers are needed to support the early diagnosis of brain injury in extremely preterm infants, and to develop effective neuroprotective protocols that are tailored to the progressing phases of injury. Experimental and clinical research shows that severity of neuronal damage is correlated with changes in the electroencephalogram (EEG) after hypoxic-ischemia (HI). We have previously reported that micro-scale sharp-wave EEG waveforms have prognostic utility within the early hours of post-HI recordings in preterm fetal sheep, before injury develops. This article aims to investigate whether these subtle EEG patterns are translational in the early hours of life in clinical recordings from extremely preterm newborns. This work evaluates the existence and morphological similarity of the sharp-waves automatically identified throughout the entire duration of EEG data from a cohort of fetal sheep 6 h after HI (n = 7, at 103 ± 1 day gestation) and in recordings commencing before 6 h of life in extremely preterm neonates (n = 7, 27 ± 2.0 weeks gestation). We report that micro-scale EEG waveforms with similar morphology and characteristics (r = 0.94) to those seen in fetal sheep after HI are also present after birth in recordings started before 6 h of life in extremely preterm neonates. This work further indicates that the post-HI sharp-waves show rapid morphological evolution, influenced by age and/or severity of neuronal loss, and thus that automated algorithms should be validated against such signal variations. Finally, this article discusses the need for more focused research on the early assessment of EEG changes in preterm infants to help determine the timing of brain injury to identify biomarkers that could assist in targeting novel therapies for particular phases of injury.

show abstract

“…In addition, and crucially, the non-linear transformations implemented by the multiple layers of the deep neural networks can extract low-level features that make preprocessing unnecessary. This makes this approach a very suitable tool for classification and regression tasks applied to EEG signals [1,[38][39][40][41].…”

Section: Introductionmentioning

confidence: 99%

A 1D CNN for high accuracy classification and transfer learning in motor imagery EEG-based brain-computer interface

2021

View full text Add to dashboard Cite

Objective: Brain-computer interface (BCI) aims to establish communication paths between the brain processes and external devices. Different methods have been used to extract human intentions from electroencephalography (EEG) recordings. Those based on motor imagery (MI) seem to have a great potential for future applications. These approaches rely on the extraction of EEG distinctive patterns during imagined movements. Techniques able to extract patterns from raw signals represent an important target for BCI as they do not need labor-intensive data pre-processing. Approach: We propose a new approach based on a 10-layer one-dimensional convolution neural network (1D-CNN) to classify five brain states (four MI classes plus a ‘baseline’ class) using a data augmentation algorithm and a limited number of EEG channels. In addition, we present a transfer learning method used to extract critical features from the EEG group dataset and then to customize the model to the single individual by training its outer layers with only 12-minute individual-related data. Main results: The model tested with the ‘EEG Motor Movement/Imagery Dataset’ outperforms the current state-of-the-art models by achieving a 99.38% accuracy at the group level. In addition, the transfer learning approach we present achieves an average accuracy of 99.46%. Significance: The proposed methods could foster future BCI applications relying on few-channel portable recording devices and individual-based training.

show abstract

Advanced Deep Learning Spectroscopy of Scalogram Infused CNN Classifiers for Robust Identification of Post‐Hypoxic Epileptiform EEG Spikes

Cited by 14 publications

References 32 publications

EEG-Based Epileptic Seizure Detection via Machine/Deep Learning Approaches: A Systematic Review

EEG-Based Epileptic Seizure Detection via Machine/Deep Learning Approaches: A Systematic Review

Early Signatures of Brain Injury in the Preterm Neonatal EEG

A 1D CNN for high accuracy classification and transfer learning in motor imagery EEG-based brain-computer interface

Contact Info

Product

Resources

About