Home / Current Issue / Paper 1716837
Nightmare Disorder Assistant Tool
Subject area: Science,Engineering and Technology · Area of research: Deep Learning
DOI: https://doi.org/10.64388/IREV9I10-1716837
Abstract
Nightmare Disorder is a sleep condition characterized by recurrent distressing dreams that lead to abrupt awakenings, emotional distress, and impaired sleep quality. It is strongly associated with psychiatric conditions such as post-traumatic stress disorder (PTSD), anxiety, and depression. Traditional diagnostic approaches rely heavily on subjective measures including patient self-reports, sleep diaries, and clinician interpretation, which often lack consistency and scalability. This paper proposes an automated EEG-based framework for detecting nightmare episodes using deep learning techniques. The system leverages spectrogram-based feature extraction to transform EEG signals into time-frequency representations, enabling effective use of Convolutional Neural Networks (CNNs) for pattern recognition. The model is trained on REM sleep EEG data, including perturbed samples designed to simulate abnormal neural activity associated with nightmares. It also incorporates a severity analysis module that evaluates the intensity of detected episodes based on spectral deviations. A web-based dashboard built using React.js and Next.js provides an interactive interface for users to upload EEG segments and visualize results. FastAPI is used as the backend framework to integrate the trained model and handle inference requests efficiently. Experimental results demonstrate classification accuracy ranging from 89% to 94.6%, indicating strong performance in distinguishing between normal and abnormal REM patterns. The system provides an objective, scalable, and interpretable solution for nightmare detection, contributing toward AI-driven advancements in sleep disorder diagnostics.
Keywords
EEG, Sleep Disorder, REM Sleep, Nightmare Disorder, Deep Learning.
How to cite this paper
@article{1716837,
author = {Amoolya S, Nischitha B S},
title = {Nightmare Disorder Assistant Tool},
journal = {Iconic Research And Engineering Journals},
year = {2026},
volume = {9},
number = {10},
pages = {3437-3444},
issn = {2456-8880},
url = {https://www.irejournals.com/formatedpaper/1716837.pdf},
abstract = {Nightmare Disorder is a sleep condition characterized by recurrent distressing dreams that lead to abrupt awakenings, emotional distress, and impaired sleep quality. It is strongly associated with psychiatric conditions such as post-traumatic stress disorder (PTSD), anxiety, and depression. Traditional diagnostic approaches rely heavily on subjective measures including patient self-reports, sleep diaries, and clinician interpretation, which often lack consistency and scalability. This paper proposes an automated EEG-based framework for detecting nightmare episodes using deep learning techniques. The system leverages spectrogram-based feature extraction to transform EEG signals into time-frequency representations, enabling effective use of Convolutional Neural Networks (CNNs) for pattern recognition. The model is trained on REM sleep EEG data, including perturbed samples designed to simulate abnormal neural activity associated with nightmares. It also incorporates a severity analysis module that evaluates the intensity of detected episodes based on spectral deviations. A web-based dashboard built using React.js and Next.js provides an interactive interface for users to upload EEG segments and visualize results. FastAPI is used as the backend framework to integrate the trained model and handle inference requests efficiently. Experimental results demonstrate classification accuracy ranging from 89% to 94.6%, indicating strong performance in distinguishing between normal and abnormal REM patterns. The system provides an objective, scalable, and interpretable solution for nightmare detection, contributing toward AI-driven advancements in sleep disorder diagnostics.},
keywords = {EEG, Sleep Disorder, REM Sleep, Nightmare Disorder, Deep Learning.},
month = {April},
doi = {https://doi.org/10.64388/IREV9I10-1716837}
}