1
|
Matsubara K, Ohgami Y, Okamura K, Aoto S, Fukami M, Shimada Y. Machine learning trial to detect sex differences in simple sticker arts of 1606 preschool children. Minerva Pediatr (Torino) 2024; 76:343-349. [PMID: 38842380 DOI: 10.23736/s2724-5276.21.06067-5] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [MESH Headings] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Indexed: 06/07/2024]
Abstract
BACKGROUND Previous studies suggested that drawings made by preschool boys and girls show distinguishable differences. However, children's drawings on their own are too complexly determined and inherently ambiguous to be a reliable indicator. In the present study, we attempted to develop a machine learning algorithm for classification of sex of the subjects using children's artworks. METHODS We studied three types of simple sticker artworks from 1606 Japanese preschool children aged 51-83 months (803 boys and 803 girls). Those artworks were processed into digitalized data. Simulated data based on the original data were also generated. Logistic regression approach was applied to each dataset to make a classifier, and run on each dataset in a stratified ten-fold cross-validation with hyperparameter tuning. A probability score was calculated in each sample and utilized for sex classification. Prediction performance was evaluated using accuracy, recall, and precision scores, as well as learning curves. RESULTS Two models created from the original and simulated data showed comparably low metrics. The distributions of probability scores in the samples from boys and girls mostly overlapped and were indistinguishable. Learning curves of the models showed an extremely under-fitted pattern. CONCLUSIONS Our machine learning algorithm was unable to distinguish simple sticker arts created by boys and girls. More complex tasks will enable to develop an accurate classifier.
Collapse
Affiliation(s)
- Keiko Matsubara
- Department of Molecular Endocrinology, National Research Institute for Child Health and Development, Tokyo, Japan
| | - Yuko Ohgami
- Department of Child Development and Education, Faculty of Humanities, Wayo Women's University, Chiba, Japan
| | - Koji Okamura
- Department of Systems BioMedicine, National Research Institute for Child Health and Development, Setagaya, Tokyo, Japan
| | - Saki Aoto
- Medical Genome Center, National Research Institute for Child Health and Development, Setagaya, Tokyo, Japan
| | - Maki Fukami
- Department of Molecular Endocrinology, National Research Institute for Child Health and Development, Tokyo, Japan -
| | - Yukiko Shimada
- Department of Child Studies, Faculty of Human Development, Kokugakuin University, Kanagawa, Japan
| |
Collapse
|
2
|
Skaramagkas V, Boura I, Spanaki C, Michou E, Karamanis G, Kefalopoulou Z, Tsiknakis M. Detecting Minor Symptoms of Parkinson's Disease in the Wild Using Bi-LSTM with Attention Mechanism. SENSORS (BASEL, SWITZERLAND) 2023; 23:7850. [PMID: 37765907 PMCID: PMC10535804 DOI: 10.3390/s23187850] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Key Words] [MESH Headings] [Track Full Text] [Subscribe] [Scholar Register] [Received: 08/14/2023] [Revised: 09/02/2023] [Accepted: 09/11/2023] [Indexed: 09/29/2023]
Abstract
Parkinson's disease (PD) is a neurodegenerative disorder characterized by motor and nonmotor impairment with various implications on patients' quality of life. Since currently available therapies are only symptomatic, identifying individuals with prodromal, preclinical, or early-stage PD is crucial, as they would be ideal candidates for future disease-modifying therapies. Our analysis aims to develop a robust model for accurate PD detection using accelerometer data collected from PD and non-PD individuals with mild or no tremor during phone conversations. An open-access dataset comprising accelerometer recordings from 22 PD patients and 11 healthy controls (HCs) was utilized. The data were preprocessed to extract relevant time-, frequency-, and energy-related features, and a bidirectional long short-term memory (Bi-LSTM) model with attention mechanism was employed for classification. The performance of the model was evaluated using fivefold cross-validation, and metrics of accuracy, precision, recall, specificity, and f1-score were computed. The proposed model demonstrated high accuracy (98%), precision (99%), recall (98%), specificity (96%), and f1-score (98%) in accurately distinguishing PD patients from HCs. Our findings indicate that the proposed model outperforms existing approaches and holds promise for detection of PD with subtle symptoms, like tremor, in the wild. Such symptoms can present in the early or even prodromal stage of the disease, and appropriate mobile-based applications may be a practical tool in real-life settings to alert individuals at risk to seek medical assistance or give patients feedback in monitoring their symptoms.
Collapse
Affiliation(s)
- Vasileios Skaramagkas
- Institute of Computer Science, Foundation for Research and Technology Hellas (FORTH), GR-700 13 Heraklion, Greece;
- Department of Electrical and Computer Engineering, Hellenic Mediterranean University, GR-710 04 Heraklion, Greece
| | - Iro Boura
- School of Medicine, University of Crete, GR-710 03 Heraklion, Greece; (I.B.); (C.S.)
- Department of Basic and Clinical Neuroscience, Institute of Psychiatry, Psychology and Neuroscience, King’s College London, London WC2R 2LS, UK
| | - Cleanthi Spanaki
- School of Medicine, University of Crete, GR-710 03 Heraklion, Greece; (I.B.); (C.S.)
- Department of Neurology, University Hospital of Heraklion, GR-715 00 Heraklion, Greece
| | - Emilia Michou
- School of Health Rehabilitation Sciences, Department of Speech and Language Therapy, University of Patras, GR-265 04 Patras, Greece;
| | - Georgios Karamanis
- Department of Neurology, Patras University Hospital, GR-264 04 Patras, Greece; (G.K.); (Z.K.)
| | - Zinovia Kefalopoulou
- Department of Neurology, Patras University Hospital, GR-264 04 Patras, Greece; (G.K.); (Z.K.)
| | - Manolis Tsiknakis
- Institute of Computer Science, Foundation for Research and Technology Hellas (FORTH), GR-700 13 Heraklion, Greece;
- Department of Electrical and Computer Engineering, Hellenic Mediterranean University, GR-710 04 Heraklion, Greece
| |
Collapse
|
3
|
Göker H. Automatic detection of Parkinson's disease from power spectral density of electroencephalography (EEG) signals using deep learning model. Phys Eng Sci Med 2023; 46:1163-1174. [PMID: 37245195 DOI: 10.1007/s13246-023-01284-x] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Key Words] [MESH Headings] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 07/26/2022] [Accepted: 05/18/2023] [Indexed: 05/29/2023]
Abstract
Parkinson's disease (PD) is characterized by slowed movements, speech disorders, an inability to control muscle movements, and tremors in the hands and feet. In the early stages of PD, the changes in these motor signs are very vague, so an objective and accurate diagnosis is difficult. The disease is complex, progressive, and very common. There are more than 10 million people worldwide suffering from PD. In this study, an EEG-based deep learning model was proposed for the automatic detection of PD to support experts. The EEG dataset comprises signals recorded by the University of Iowa from 14 PD patients and 14 healthy controls. First of all, the power spectral density values (PSDs) of the frequencies between 1 and 49 Hz of the EEG signals were calculated separately using periodogram, welch, and multitaper spectral analysis methods. 49 feature vectors were extracted for each of the three different experiments. Then, the performances of support vector machine, random forest, k-nearest neighbor, and bidirectional long-short-term memory (BiLSTM) algorithms were compared using the PSDs feature vectors. After the comparison, the model integrating welch spectral analysis and the BiLSTM algorithm showed the highest performance as a result of the experiments. The deep learning model achieved satisfactory performance with 0.965 specificity, 0.994 sensitivity, 0.964 precision, 0.978 f1-score, 0.958 Matthews correlation coefficient, and 97.92% accuracy. The study is a promising attempt to detect PD from EEG signals and it also provides evidence that deep learning algorithms are more effective than machine learning algorithms for EEG signal analysis.
Collapse
Affiliation(s)
- Hanife Göker
- Health Services Vocational College, Gazi University, 06830, Ankara, Turkey.
| |
Collapse
|
4
|
Cardoso Mendes L, Abreu Rosa de Sá A, Alves Marques I, Morère Y, de Oliveira Andrade A. RehaBEElitation: the architecture and organization of a serious game to evaluate motor signs in Parkinson's disease. PeerJ Comput Sci 2023; 9:e1267. [PMID: 37346638 PMCID: PMC10280492 DOI: 10.7717/peerj-cs.1267] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Key Words] [Track Full Text] [Figures] [Journal Information] [Subscribe] [Scholar Register] [Received: 08/16/2022] [Accepted: 02/08/2023] [Indexed: 06/23/2023]
Abstract
Background The use of serious games (SG) has received increasing attention in health care, and can be applied for both rehabilitation and evaluation of motor signs of several diseases, such as Parkinson's disease (PD). However, the use of these instruments in clinical practice is poorly observed, since there is a scarcity of games that, during their development process, simultaneously address issues of usability and architectural design, contributing to the non-satisfaction of the actual needs of professionals and patients. Thus, this study aimed to present the architecture and usability evaluation at the design stage of a serious game, so-called RehaBEElitation, and assess the accessibility of the game. Methods The game was created by a multidisciplinary team with experience in game development and PD, taking into consideration design guidelines for the development of SG. The user must control the movements of a bee in a 3D environment. The game tasks were designed to mimic the following movements found in the gold-standard method tool-Movement Disorder Society-Unified Parkinson's Disease Rating Scale (MDS-UPDRS)-for the assessment of individuals with PD: hand opening and closing, hand extension and flexion, hand adduction and abduction, finger tapping, and forearm supination and pronation. The user interacts with the game using a wearable interface device that embeds inertial and tactile sensors. The architecture of RehaBEElitation was detailed using the business process model (BPM) notation and the usability of the architecture was evaluated using the Nielsen-Shneiderman heuristics. Game accessibility was evaluated by comparing the overall scores of each phase between 15 healthy participants and 15 PD patients. The PD group interacted with the game in both the ON and OFF states. Results The system was modularized in order to implement parallel, simultaneous and independent programming at different levels, requiring less computational effort and enabling fluidity between the game and the control of the interface elements in real time. The developed architecture allows the inclusion of new elements for patient status monitoring, extending the functionality of the tool without changing its fundamental characteristics. The heuristic evaluation contemplated all the 14 heuristics proposed by Shneiderman, which enabled the implementation of improvements in the game. The evaluation of accessibility revealed no statistically significant differences (p < 0.05) between groups, except for the healthy group and the PD group in the OFF state of medication during Phase 3 of the game. Conclusions The proposed architecture was presented in order to facilitate the reproduction of the system and extend its application to other scenarios. In the same way, the heuristic evaluation performed can serve as a contribution to the advancement of the SG design for PD. The accessibility evaluation revealed that the game is accessible to individuals with PD.
Collapse
Affiliation(s)
- Luanne Cardoso Mendes
- Centre for Innovation and Technology Assessment in Health (NIATS), Faculty of Electrical Engineering, Federal University of Uberlândia, Universidade Federal de Uberlândia, Uberlândia, Minas Gerais, Brazil
- Laboratoire de Conception, d’Optimisation et de Modélisation des Systèmes (LCOMS), Université de Lorraine, Metz, Moselle, France
| | - Angela Abreu Rosa de Sá
- Assistive Technology Laboratory, Faculty of Electrical Engineering (NTA), Universidade Federal de Uberlândia, Uberlândia, Minas Gerais, Brazil
| | - Isabela Alves Marques
- Centre for Innovation and Technology Assessment in Health (NIATS), Faculty of Electrical Engineering, Federal University of Uberlândia, Universidade Federal de Uberlândia, Uberlândia, Minas Gerais, Brazil
- Laboratoire de Conception, d’Optimisation et de Modélisation des Systèmes (LCOMS), Université de Lorraine, Metz, Moselle, France
| | - Yann Morère
- Laboratoire de Conception, d’Optimisation et de Modélisation des Systèmes (LCOMS), Université de Lorraine, Metz, Moselle, France
| | - Adriano de Oliveira Andrade
- Centre for Innovation and Technology Assessment in Health (NIATS), Faculty of Electrical Engineering, Federal University of Uberlândia, Universidade Federal de Uberlândia, Uberlândia, Minas Gerais, Brazil
| |
Collapse
|
5
|
Loh HW, Hong W, Ooi CP, Chakraborty S, Barua PD, Deo RC, Soar J, Palmer EE, Acharya UR. Application of Deep Learning Models for Automated Identification of Parkinson's Disease: A Review (2011-2021). SENSORS 2021; 21:s21217034. [PMID: 34770340 PMCID: PMC8587636 DOI: 10.3390/s21217034] [Citation(s) in RCA: 17] [Impact Index Per Article: 5.7] [Reference Citation Analysis] [Abstract] [Key Words] [MESH Headings] [Track Full Text] [Download PDF] [Figures] [Subscribe] [Scholar Register] [Received: 09/20/2021] [Revised: 10/07/2021] [Accepted: 10/19/2021] [Indexed: 12/18/2022]
Abstract
Parkinson’s disease (PD) is the second most common neurodegenerative disorder affecting over 6 million people globally. Although there are symptomatic treatments that can increase the survivability of the disease, there are no curative treatments. The prevalence of PD and disability-adjusted life years continue to increase steadily, leading to a growing burden on patients, their families, society and the economy. Dopaminergic medications can significantly slow down the progression of PD when applied during the early stages. However, these treatments often become less effective with the disease progression. Early diagnosis of PD is crucial for immediate interventions so that the patients can remain self-sufficient for the longest period of time possible. Unfortunately, diagnoses are often late, due to factors such as a global shortage of neurologists skilled in early PD diagnosis. Computer-aided diagnostic (CAD) tools, based on artificial intelligence methods, that can perform automated diagnosis of PD, are gaining attention from healthcare services. In this review, we have identified 63 studies published between January 2011 and July 2021, that proposed deep learning models for an automated diagnosis of PD, using various types of modalities like brain analysis (SPECT, PET, MRI and EEG), and motion symptoms (gait, handwriting, speech and EMG). From these studies, we identify the best performing deep learning model reported for each modality and highlight the current limitations that are hindering the adoption of such CAD tools in healthcare. Finally, we propose new directions to further the studies on deep learning in the automated detection of PD, in the hopes of improving the utility, applicability and impact of such tools to improve early detection of PD globally.
Collapse
Affiliation(s)
- Hui Wen Loh
- School of Science and Technology, Singapore University of Social Sciences, Singapore 599494, Singapore
| | - Wanrong Hong
- Cogninet Brain Team, Cogninet Australia, Sydney, NSW 2010, Australia
| | - Chui Ping Ooi
- School of Science and Technology, Singapore University of Social Sciences, Singapore 599494, Singapore
| | - Subrata Chakraborty
- Faculty of Engineering and Information Technology, University of Technology Sydney, Sydney, NSW 2007, Australia
| | - Prabal Datta Barua
- Cogninet Brain Team, Cogninet Australia, Sydney, NSW 2010, Australia
- Faculty of Engineering and Information Technology, University of Technology Sydney, Sydney, NSW 2007, Australia
- School of Business (Information Systems), Faculty of Business, Education, Law & Arts, University of Southern Queensland, Toowoomba, QLD 4350, Australia
| | - Ravinesh C Deo
- School of Sciences, University of Southern Queensland, Springfield, QLD 4300, Australia
| | - Jeffrey Soar
- School of Business (Information Systems), Faculty of Business, Education, Law & Arts, University of Southern Queensland, Toowoomba, QLD 4350, Australia
| | - Elizabeth E Palmer
- Centre of Clinical Genetics, Sydney Children's Hospitals Network, Randwick, NSW 2031, Australia
- School of Women's and Children's Health, University of New South Wales, Randwick, NSW 2031, Australia
| | - U Rajendra Acharya
- School of Science and Technology, Singapore University of Social Sciences, Singapore 599494, Singapore
- School of Business (Information Systems), Faculty of Business, Education, Law & Arts, University of Southern Queensland, Toowoomba, QLD 4350, Australia
- School of Engineering, Ngee Ann Polytechnic, Singapore 599489, Singapore
- Department of Bioinformatics and Medical Engineering, Asia University, Taichung 413, Taiwan
- Research Organization for Advanced Science and Technology (IROAST), Kumamoto University, Kumamoto 860-8555, Japan
| |
Collapse
|