Minji Lee
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Minji Lee has a total of 31 co-authors. Here are the top co-authors by number of shared publications:
- Jung Kyung Hong
Department of Psychiatry, Seoul National University Bundang Hospital, Seongnam, Republic of Korea
5 shared publications - Young-In Yoon
Department of Psychiatry, Seoul National University Bundang Hospital, Seongnam, Republic of Korea
5 shared publications - Dongheon Lee
Asleep Inc., Seoul, Republic of Korea
3 shared publications - Clete A. Kushida
Division of Sleep Medicine, Department of Psychiatry and Behavioral Sciences, Stanford University School of Medicine, Stanford, CA, United States
2 shared publications - Daewoo Kim
Asleep Co., Ltd., Seoul, Republic of Korea
2 shared publications - Hai Hong Tran
Asleep Inc., Seoul, Korea
2 shared publications
Latest Publications
- 2 October 2026
- Published by Springer Nature in Sleep and Breathing
- Vol. 30 (5) , 278
Associations between weather variables and sleep architecture: A large-scale analysis using consumer sleep data
- 2 October 2026
- Published by Springer Nature in Sleep and Breathing
- Vol. 30 (5) , 278
Associations between weather variables and sleep architecture: A large-scale analysis using consumer sleep data
- Open Access
- 25 October 2025
- Published by Korean Neuropsychiatric Association in Psychiatry Investigation
- Vol. 22 (10) , 1153-1163
Exploring the Relationship Between Depression and Attention-Deficit/Hyperactivity Disorder Symptoms in Children Using Brain Activity Monitoring
Objective Although depression is a common comorbidity in children with attention-deficit/hyperactivity disorder (ADHD), its neurophysiological relationship according to each symptom has rarely been explored. This study aimed to inform clinical practice by exploring the neurophysiological underpinnings of depression comorbidity in ADHD.Methods We conducted a cross-sectional study of 87 children with ADHD (68 males). Resting quantitative electroencephalography (qEEG) recordings were collected with eyes closed. We used various questionnaires to evaluate ADHD symptoms severity, depression, and anxiety. Pearson correlation coefficients were used to investigate the relationship between the z-score relative spectral power of qEEG and each psychological symptom. Data were analyzed using IBM SPSS 27.0.Results The study’s findings indicated that theta activity at the frontal, central, and parietal locations had a negative correlation with the severity of ADHD symptoms in children diagnosed with ADHD. In contrast, alpha activity in these same regions demonstrated a positive correlation with ADHD symptom severity. Additionally, delta activity in the regions was negatively correlated with depression severity.Conclusion These findings suggest that alpha and theta activity might serve as a reliable neurophysiological marker of ADHD symptom severity, while delta activity might function as a reliable biological marker of depression severity in children with ADHD. However, further research is needed to generalize the results of this study.
- 25 October 2025
- Published by Korean Neuropsychiatric Association in Psychiatry Investigation
- Vol. 22 (10) , 1153-1163
Open AccessExploring the Relationship Between Depression and Attention-Deficit/Hyperactivity Disorder Symptoms in Children Using Brain Activity Monitoring
Objective Although depression is a common comorbidity in children with attention-deficit/hyperactivity disorder (ADHD), its neurophysiological relationship according to each symptom has rarely been explored. This study aimed to inform clinical practice by exploring the neurophysiological underpinnings of depression comorbidity in ADHD.Methods We conducted a cross-sectional study of 87 children with ADHD (68 males). Resting quantitative electroencephalography (qEEG) recordings were collected with eyes closed. We used various questionnaires to evaluate ADHD symptoms severity, depression, and anxiety. Pearson correlation coefficients were used to investigate the relationship between the z-score relative spectral power of qEEG and each psychological symptom. Data were analyzed using IBM SPSS 27.0.Results The study’s findings indicated that theta activity at the frontal, central, and parietal locations had a negative correlation with the severity of ADHD symptoms in children diagnosed with ADHD. In contrast, alpha activity in these same regions demonstrated a positive correlation with ADHD symptom severity. Additionally, delta activity in the regions was negatively correlated with depression severity.Conclusion These findings suggest that alpha and theta activity might serve as a reliable neurophysiological marker of ADHD symptom severity, while delta activity might function as a reliable biological marker of depression severity in children with ADHD. However, further research is needed to generalize the results of this study.
- 11 April 2025
- Published by Elsevier in Journal of Psychiatric Research
- Vol. 186, 129-136
Transcranial alternating current stimulation in subjects with insomnia symptoms: A randomized, double-blind and controlled study
- 11 April 2025
- Published by Elsevier in Journal of Psychiatric Research
- Vol. 186, 129-136
Transcranial alternating current stimulation in subjects with insomnia symptoms: A randomized, double-blind and controlled study
- Open Access
- 1 January 2024
- Published by Wiley in Acta Neurologica Scandinavica
- Vol. 2024 (1)
Impact of Depression on Cognitive Function and Phenoconversion to Neurodegenerative Diseases in Patients With Isolated REM Sleep Behavior Disorder
Background: This study was aimed at analyzing cognitive function and quantitative electroencephalogram (qEEG) in patients with isolated REM sleep behavior disorder (iRBD) based on the presence of depression and at evaluating the impact of depression on phenoconversion to neurodegenerative diseases. Methods: Individuals diagnosed with iRBD via polysomnography were included. Based on the presence of depression, patients were categorized into two groups. Neuropsychological tests and qEEG were conducted following the diagnosis of iRBD, and outcomes were compared between the two groups. Patients were regularly followed to monitor their phenoconversion status. Cox regression analysis was performed to assess the hazard ratio associated with depression. Results: Ninety iRBD patients (70% males) were included, with a median age of 66.3 years. Depression was identified in 26 (28.9%) of these patients. The depressed group showed significantly poorer performance only in color reading subtest of Stroop ( p = 0.029) compared to the nondepressed group, showing reduced processing speed. In qEEG, relative gamma power ( p = 0.034) and high gamma power ( p = 0.020) in the parietal region were significantly higher in the depressed group than in the nondepressed group. Depression was associated with a hazard ratio of 3.32 for the risk of phenoconversion to neurodegenerative diseases in iRBD patients ( p = 0.011). Conclusion: Depressive symptoms in iRBD patients should be closely monitored as they could aggravate cognitive dysfunction and increase the risk of phenoconversion to neurodegenerative diseases.
- 1 January 2024
- Published by Wiley in Acta Neurologica Scandinavica
- Vol. 2024 (1)
Open AccessImpact of Depression on Cognitive Function and Phenoconversion to Neurodegenerative Diseases in Patients With Isolated REM Sleep Behavior Disorder
Background: This study was aimed at analyzing cognitive function and quantitative electroencephalogram (qEEG) in patients with isolated REM sleep behavior disorder (iRBD) based on the presence of depression and at evaluating the impact of depression on phenoconversion to neurodegenerative diseases. Methods: Individuals diagnosed with iRBD via polysomnography were included. Based on the presence of depression, patients were categorized into two groups. Neuropsychological tests and qEEG were conducted following the diagnosis of iRBD, and outcomes were compared between the two groups. Patients were regularly followed to monitor their phenoconversion status. Cox regression analysis was performed to assess the hazard ratio associated with depression. Results: Ninety iRBD patients (70% males) were included, with a median age of 66.3 years. Depression was identified in 26 (28.9%) of these patients. The depressed group showed significantly poorer performance only in color reading subtest of Stroop ( p = 0.029) compared to the nondepressed group, showing reduced processing speed. In qEEG, relative gamma power ( p = 0.034) and high gamma power ( p = 0.020) in the parietal region were significantly higher in the depressed group than in the nondepressed group. Depression was associated with a hazard ratio of 3.32 for the risk of phenoconversion to neurodegenerative diseases in iRBD patients ( p = 0.011). Conclusion: Depressive symptoms in iRBD patients should be closely monitored as they could aggravate cognitive dysfunction and increase the risk of phenoconversion to neurodegenerative diseases.
- Open Access
- 1 June 2023
- Published by JMIR Publications Inc. in Journal of Medical Internet Research
- Vol. 25 (1) , e46216
Prediction of Sleep Stages Via Deep Learning Using Smartphone Audio Recordings in Home Environments: Model Development and Validation
Background: The growing public interest and awareness regarding the significance of sleep is driving the demand for sleep monitoring at home. In addition to various commercially available wearable and nearable devices, sound-based sleep staging via deep learning is emerging as a decent alternative for their convenience and potential accuracy. However, sound-based sleep staging has only been studied using in-laboratory sound data. In real-world sleep environments (homes), there is abundant background noise, in contrast to quiet, controlled environments such as laboratories. The use of sound-based sleep staging at homes has not been investigated while it is essential for practical use on a daily basis. Challenges are the lack of and the expected huge expense of acquiring a sufficient size of home data annotated with sleep stages to train a large-scale neural network. Objective: This study aims to develop and validate a deep learning method to perform sound-based sleep staging using audio recordings achieved from various uncontrolled home environments. Methods: To overcome the limitation of lacking home data with known sleep stages, we adopted advanced training techniques and combined home data with hospital data. The training of the model consisted of 3 components: (1) the original supervised learning using 812 pairs of hospital polysomnography (PSG) and audio recordings, and the 2 newly adopted components; (2) transfer learning from hospital to home sounds by adding 829 smartphone audio recordings at home; and (3) consistency training using augmented hospital sound data. Augmented data were created by adding 8255 home noise data to hospital audio recordings. Besides, an independent test set was built by collecting 45 pairs of overnight PSG and smartphone audio recording at homes to examine the performance of the trained model. Results: The accuracy of the model was 76.2% (63.4% for wake, 64.9% for rapid-eye movement [REM], and 83.6% for non-REM) for our test set. The macro F1-score and mean per-class sensitivity were 0.714 and 0.706, respectively. The performance was robust across demographic groups such as age, gender, BMI, or sleep apnea severity (accuracy 73.4%-79.4%). In the ablation study, we evaluated the contribution of each component. While the supervised learning alone achieved accuracy of 69.2% on home sound data, adding consistency training to the supervised learning helped increase the accuracy to a larger degree (+4.3%) than adding transfer learning (+0.1%). The best performance was shown when both transfer learning and consistency training were adopted (+7.0%). Conclusions: This study shows that sound-based sleep staging is feasible for home use. By adopting 2 advanced techniques (transfer learning and consistency training) the deep learning model robustly predicts sleep stages using sounds recorded at various uncontrolled home environments, without using any special equipment but smartphones only.
- 1 June 2023
- Published by JMIR Publications Inc. in Journal of Medical Internet Research
- Vol. 25 (1) , e46216
Open AccessPrediction of Sleep Stages Via Deep Learning Using Smartphone Audio Recordings in Home Environments: Model Development and Validation
Background: The growing public interest and awareness regarding the significance of sleep is driving the demand for sleep monitoring at home. In addition to various commercially available wearable and nearable devices, sound-based sleep staging via deep learning is emerging as a decent alternative for their convenience and potential accuracy. However, sound-based sleep staging has only been studied using in-laboratory sound data. In real-world sleep environments (homes), there is abundant background noise, in contrast to quiet, controlled environments such as laboratories. The use of sound-based sleep staging at homes has not been investigated while it is essential for practical use on a daily basis. Challenges are the lack of and the expected huge expense of acquiring a sufficient size of home data annotated with sleep stages to train a large-scale neural network. Objective: This study aims to develop and validate a deep learning method to perform sound-based sleep staging using audio recordings achieved from various uncontrolled home environments. Methods: To overcome the limitation of lacking home data with known sleep stages, we adopted advanced training techniques and combined home data with hospital data. The training of the model consisted of 3 components: (1) the original supervised learning using 812 pairs of hospital polysomnography (PSG) and audio recordings, and the 2 newly adopted components; (2) transfer learning from hospital to home sounds by adding 829 smartphone audio recordings at home; and (3) consistency training using augmented hospital sound data. Augmented data were created by adding 8255 home noise data to hospital audio recordings. Besides, an independent test set was built by collecting 45 pairs of overnight PSG and smartphone audio recording at homes to examine the performance of the trained model. Results: The accuracy of the model was 76.2% (63.4% for wake, 64.9% for rapid-eye movement [REM], and 83.6% for non-REM) for our test set. The macro F1-score and mean per-class sensitivity were 0.714 and 0.706, respectively. The performance was robust across demographic groups such as age, gender, BMI, or sleep apnea severity (accuracy 73.4%-79.4%). In the ablation study, we evaluated the contribution of each component. While the supervised learning alone achieved accuracy of 69.2% on home sound data, adding consistency training to the supervised learning helped increase the accuracy to a larger degree (+4.3%) than adding transfer learning (+0.1%). The best performance was shown when both transfer learning and consistency training were adopted (+7.0%). Conclusions: This study shows that sound-based sleep staging is feasible for home use. By adopting 2 advanced techniques (transfer learning and consistency training) the deep learning model robustly predicts sleep stages using sounds recorded at various uncontrolled home environments, without using any special equipment but smartphones only.
- Open Access
- 2 February 2023
- Published by JMIR Publications Inc. in JMIR Preprints
Prediction of Sleep Stages Via Deep Learning Using Smartphone Audio Recordings in Home Environments: Model Development and Validation (Preprint)
BACKGROUND The growing public interest and awareness regarding the significance of sleep is driving the demand for sleep monitoring at home. In addition to various commercially available wearable and nearable devices, sound-based sleep staging via deep learning is emerging as a decent alternative for their convenience and potential accuracy. However, sound-based sleep staging has only been studied using in-laboratory sound data. In real-world sleep environments (homes), there is abundant background noise, in contrast to quiet, controlled environments such as laboratories. The use of sound-based sleep staging at homes has not been investigated while it is essential for practical use on a daily basis. Challenges are the lack of and the expected huge expense of acquiring a sufficient size of home data annotated with sleep stages to train a large-scale neural network. OBJECTIVE This study aims to develop and validate a deep learning method to perform sound-based sleep staging using audio recordings achieved from various uncontrolled home environments. METHODS To overcome the limitation of lacking home data with known sleep stages, we adopted advanced training techniques and combined home data with hospital data. The training of the model consisted of 3 components: (1) the original supervised learning using 812 pairs of hospital polysomnography (PSG) and audio recordings, and the 2 newly adopted components; (2) transfer learning from hospital to home sounds by adding 829 smartphone audio recordings at home; and (3) consistency training using augmented hospital sound data. Augmented data were created by adding 8255 home noise data to hospital audio recordings. Besides, an independent test set was built by collecting 45 pairs of overnight PSG and smartphone audio recording at homes to examine the performance of the trained model. RESULTS The accuracy of the model was 76.2% (63.4% for wake, 64.9% for rapid-eye movement [REM], and 83.6% for non-REM) for our test set. The macro F1-score and mean per-class sensitivity were 0.714 and 0.706, respectively. The performance was robust across demographic groups such as age, gender, BMI, or sleep apnea severity (accuracy 73.4%-79.4%). In the ablation study, we evaluated the contribution of each component. While the supervised learning alone achieved accuracy of 69.2% on home sound data, adding consistency training to the supervised learning helped increase the accuracy to a larger degree (+4.3%) than adding transfer learning (+0.1%). The best performance was shown when both transfer learning and consistency training were adopted (+7.0%). CONCLUSIONS This study shows that sound-based sleep staging is feasible for home use. By adopting 2 advanced techniques (transfer learning and consistency training) the deep learning model robustly predicts sleep stages using sounds recorded at various uncontrolled home environments, without using any special equipment but smartphones only. CLINICALTRIAL
- 2 February 2023
- Published by JMIR Publications Inc. in JMIR Preprints
Open AccessPrediction of Sleep Stages Via Deep Learning Using Smartphone Audio Recordings in Home Environments: Model Development and Validation (Preprint)
BACKGROUND The growing public interest and awareness regarding the significance of sleep is driving the demand for sleep monitoring at home. In addition to various commercially available wearable and nearable devices, sound-based sleep staging via deep learning is emerging as a decent alternative for their convenience and potential accuracy. However, sound-based sleep staging has only been studied using in-laboratory sound data. In real-world sleep environments (homes), there is abundant background noise, in contrast to quiet, controlled environments such as laboratories. The use of sound-based sleep staging at homes has not been investigated while it is essential for practical use on a daily basis. Challenges are the lack of and the expected huge expense of acquiring a sufficient size of home data annotated with sleep stages to train a large-scale neural network. OBJECTIVE This study aims to develop and validate a deep learning method to perform sound-based sleep staging using audio recordings achieved from various uncontrolled home environments. METHODS To overcome the limitation of lacking home data with known sleep stages, we adopted advanced training techniques and combined home data with hospital data. The training of the model consisted of 3 components: (1) the original supervised learning using 812 pairs of hospital polysomnography (PSG) and audio recordings, and the 2 newly adopted components; (2) transfer learning from hospital to home sounds by adding 829 smartphone audio recordings at home; and (3) consistency training using augmented hospital sound data. Augmented data were created by adding 8255 home noise data to hospital audio recordings. Besides, an independent test set was built by collecting 45 pairs of overnight PSG and smartphone audio recording at homes to examine the performance of the trained model. RESULTS The accuracy of the model was 76.2% (63.4% for wake, 64.9% for rapid-eye movement [REM], and 83.6% for non-REM) for our test set. The macro F1-score and mean per-class sensitivity were 0.714 and 0.706, respectively. The performance was robust across demographic groups such as age, gender, BMI, or sleep apnea severity (accuracy 73.4%-79.4%). In the ablation study, we evaluated the contribution of each component. While the supervised learning alone achieved accuracy of 69.2% on home sound data, adding consistency training to the supervised learning helped increase the accuracy to a larger degree (+4.3%) than adding transfer learning (+0.1%). The best performance was shown when both transfer learning and consistency training were adopted (+7.0%). CONCLUSIONS This study shows that sound-based sleep staging is feasible for home use. By adopting 2 advanced techniques (transfer learning and consistency training) the deep learning model robustly predicts sleep stages using sounds recorded at various uncontrolled home environments, without using any special equipment but smartphones only. CLINICALTRIAL
- Open Access
- 31 May 2021
- Published by Korean College of Neuropsychopharmacology in Clinical Psychopharmacology and Neuroscience
- Vol. 19 (2) , 323-333
Relationship Between the Loudness Dependence of the Auditory Evoked Potential and the Severity of Suicidal Ideation in Patients with Major Depressive Disorder
Mingyu Hwang, Yeon Jung Lee, Minji Lee, Byungjoo Kang, Yun Sung Lee, Jaeuk Hwang, Sung-il Woo, Sang-Woo Hahn. Clin Psychopharmacol Neurosci 2021;19:323-33. https://doi.org/10.9758/cpn.2021.19.2.323
- 31 May 2021
- Published by Korean College of Neuropsychopharmacology in Clinical Psychopharmacology and Neuroscience
- Vol. 19 (2) , 323-333
Open AccessRelationship Between the Loudness Dependence of the Auditory Evoked Potential and the Severity of Suicidal Ideation in Patients with Major Depressive Disorder
Mingyu Hwang, Yeon Jung Lee, Minji Lee, Byungjoo Kang, Yun Sung Lee, Jaeuk Hwang, Sung-il Woo, Sang-Woo Hahn. Clin Psychopharmacol Neurosci 2021;19:323-33. https://doi.org/10.9758/cpn.2021.19.2.323