Related Experiment Video
Updated: Nov 5, 2025
12:48
Investigating Social Cognition in Infants and Adults Using Dense Array Electroencephalography dEEG
Published on: June 27, 2011
18.0K
Deep learning applied to electroencephalogram data in mental disorders: A systematic review.
Mateo de Bardeci1, Cheng Teng Ip2, Sebastian Olbrich1
1Department for Psychiatry, Psychotherapy and Psychosomatics, Psychiatric University Hospital Zurich (PUK), Switzerland; University Hospital Zurich, Switzerland; University Zurich, Switzerland.
Biological Psychology
|May 15, 2021
Summary
Deep learning (DL) shows promise for analyzing electroencephalogram (EEG) data in mental disorder research. However, studies need better clinical data and rigorous methods for reliable diagnostic and predictive insights.
Area of Science:
- Neuroscience
- Computer Science
- Psychiatry
Background:
- Deep learning (DL) techniques are increasingly applied in medical research.
- Electroencephalogram (EEG) data offers valuable insights into brain activity for mental disorder studies.
Purpose of the Study:
- To systematically review the application of DL techniques to EEG data for diagnosing and predicting mental disorders.
- To assess the quality of existing EEG-DL studies in clinical, data processing, and DL domains.
Main Methods:
- Searched for EEG studies on psychiatric diseases (ICD-10/DSM-V) using Convolutional Neural Networks (CNNs) or Long Short-Term Memory (LSTMs) networks.
- Examined study quality across clinical features, EEG data processing, and DL methodology.
Main Results:
- Most studies provided sufficient details on EEG acquisition and pre-processing.
- Many studies lacked systematic clinical feature characterization and employed flawed model selection or testing procedures.
Conclusions:
- Future research using DL for psychiatric disorders must enhance clinical data quality.
- Adopting state-of-the-art model selection and testing is crucial for higher research standards and clinical significance.
