Machine learning in biosignals processing for mental health: A narrative review

Sajno, Elena; Bartolotta, Sabrina; Tuena, Cosimo; Cipresso, Pietro; Pedroli, Elisa; Riva, Giuseppe

doi:10.3389/fpsyg.2022.1066317

Cited by 14 publications

(7 citation statements)

References 144 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In recent years machine-learning models were build and were shown to be able to determine emotional states by analyzing EEG data [ 52 ]. By using machine-learning based analysis of EEG data, it might even be possible to objectively differentiate between typical and atypical functioning of FER and to examine a possible dependency on epileptiform discharges [ 53 ]. Bartolini et al applied a synchronous registration of functional MRI and EEG data (fMRI-EEG) to investigate hemodynamic response to intermittent photic stimulation (IPS) in individuals with JME compared to healthy controls.…”

Section: Discussionmentioning

confidence: 99%

Facial Emotion Recognition in Patients with Juvenile Myoclonic Epilepsy

Dunkel,

Strzelczyk,

Schubert-Bast

et al. 2023

JCM

View full text Add to dashboard Cite

Previous studies have found facial emotion recognition (FER) impairments in individuals with epilepsy. While such deficits have been extensively explored in individuals with focal temporal lobe epilepsy, studies on individuals with generalized epilepsies are rare. However, studying FER specifically in individuals with juvenile myoclonic epilepsy (JME) is particularly interesting since they frequently suffer from social and neuropsychological difficulties in addition to epilepsy-specific symptoms. Furthermore, recent brain imaging studies have shown subtle microstructural alterations in individuals with JME. FER is considered a fundamental social skill that relies on a distributed neural network, which could be disturbed by network dysfunction in individuals with JME. This cross-sectional study aimed to examine FER and social adjustment in individuals with JME. It included 27 patients with JME and 27 healthy controls. All subjects underwent an Ekman-60 Faces Task to examine FER and neuropsychological tests to assess social adjustment as well as executive functions, intelligence, depression, and personality traits. Individuals with JME performed worse in global FER and fear and surprise recognition than healthy controls. However, probably due to the small sample size, no significant difference was found between the two groups. A potential FER impairment needs to be confirmed in further studies with larger sample size. If so, patients with JME could benefit from addressing possible deficits in FER and social difficulties when treated. By developing therapeutic strategies to improve FER, patients could be specifically supported with the aim of improving social outcomes and quality of life.

show abstract

Section: Discussionmentioning

confidence: 99%

Facial Emotion Recognition in Patients with Juvenile Myoclonic Epilepsy

Dunkel,

Strzelczyk,

Schubert-Bast

et al. 2023

JCM

View full text Add to dashboard Cite

show abstract

“…Model validation plays a crucial role in biosignal processing, allowing for the assessment of a machine learning model's accuracy and generalization ability using independent data not used during training [100]. In the context of biosignal processing, model validation holds significance due to several reasons.…”

Section: B Unsupervised Learning Methods For Biosignal Processingmentioning

confidence: 99%

Machine Learning Approaches in Bioengineering for Biosignal Processing

Biçer,

shayea

2023

Preprint

View full text Add to dashboard Cite

This survey paper offers a comprehensive review of the recent advances and applications of Machine Learning (ML) approaches in the interdisciplinary field of bioengineering, specifically in the realm of biosignal processing. Biosignals, including electroencephalograms (EEG), electrocardiograms (ECG), and electromyograms (EMG), are inherently complex, presenting significant challenges such as noise, artifacts, variability, and nonlinearity in their processing. However, ML has shown promise in overcoming these hurdles, enabling the extraction of useful features and insights from these signals. The paper outlines how ML is leveraged for processing, analyzing, classifying, and interpreting biosignals for various applications, such as diagnosis, monitoring, rehabilitation, and brain-computer interfaces. Additionally, it discusses the ongoing challenges and potential future directions of ML applications in this field. Through this review, we aim to highlight the critical role of ML in enabling adaptive, personalized, and intelligent systems that interact with biosignals in real-time, with potential implications for improving patient outcomes in various medical conditions.

show abstract

“…Inspired by the brain asymmetry for processing emotion, previous researchers have attempted to improve the accuracy of machine learning models to predict emotional states (Huang et al, 2012 ; Ahmed and Loo, 2014 ; Li et al, 2022 ; Sajno et al, 2022 ). To that aim, researchers have proposed features that capture the brain's asymmetrical behavior in emotion processing.…”

Section: Introductionmentioning

confidence: 99%

Identifying relevant asymmetry features of EEG for emotion processing

Mouri,

Valderrama,

Camorlinga

2023

Front. Psychol.

View full text Add to dashboard Cite

The left and right hemispheres of the brain process emotion differently. Neuroscientists have proposed two models to explain this difference. The first model states that the right hemisphere is dominant over the left to process all emotions. In contrast, the second model states that the left hemisphere processes positive emotions, whereas the right hemisphere processes negative emotions. Previous studies have used these asymmetry models to enhance the classification of emotions in machine learning models. However, little research has been conducted to explore how machine learning models can help identify associations between hemisphere asymmetries and emotion processing. To address this gap, we conducted two experiments using a subject-independent approach to explore how the asymmetry of the brain hemispheres is involved in processing happiness, sadness, fear, and neutral emotions. We analyzed electroencephalogram (EEG) signals from 15 subjects collected while they watched video clips evoking these four emotions. We derived asymmetry features from the recorded EEG signals by calculating the log ratio between the relative energy of symmetrical left and right nodes. Using the asymmetry features, we trained four binary logistic regressions, one for each emotion, to identify which features were more relevant to the predictions. The average AUC-ROC across the 15 subjects was 56.2, 54.6, 51.6, and 58.4% for neutral, sad, fear, and happy, respectively. We validated these results with an independent dataset, achieving comparable AUC-ROC values. Our results showed that brain lateralization was observed primarily in the alpha frequency bands, whereas for the other frequency bands, both hemispheres were involved in emotion processing. Furthermore, the logistic regression analysis indicated that the gamma and alpha bands were the most relevant for predicting emotional states, particularly for the lateral frontal, parietal, and temporal EEG pairs, such as FT7-FT8, T7-T8, and TP7-TP8. These findings provide valuable insights into which brain areas and frequency bands need to be considered when developing predictive models for emotion recognition.

show abstract

Machine learning in biosignals processing for mental health: A narrative review

Cited by 14 publications

References 144 publications

Facial Emotion Recognition in Patients with Juvenile Myoclonic Epilepsy

Facial Emotion Recognition in Patients with Juvenile Myoclonic Epilepsy

Machine Learning Approaches in Bioengineering for Biosignal Processing

Identifying relevant asymmetry features of EEG for emotion processing

Contact Info

Product

Resources

About