NPRP12S-0227-190164

A Novel Contactless ANFIS-EMG Wearable Device for Diabetic Sensorimotor Polyneuropathy (DSPN)

Biomedical implants are used to improve the quality of function of jeopardized structures or structures that have undergone serious bone and joint deterioration in human bodies. These implants are artificial medical devices that are inserted directly into the body of a patient. Implants are used to substitute the diseased or lost biological structure in order to restore function and form to original performance. Therefore, implants assist in enhancing the quality of life and human being's longevity for patients whose structures have been jeopardized [1]. However, implants often face challenges related to biocompatibility, biostability, and suitability of the tribo-bio-mechanical properties. This often leads to loss of function of the implant, the consequences of which are discomfort and pain. Most commonly used implants are hip, knee, spinal substitutions and joint replacements. Joints in human body agonize due to degenerative diseases like arthritis, which results in loss of function or pain [2]. This kind of diseases cause bone mechanical properties degradation owing to lack of normal biological self-healing progression or disproportionate loading. It has been reported that about 20% of people with age higher than forty agonize because of degenerative diseases. In addition, the older population is increasing faster than expected. Solutions for these problems lie in the development of more suitable biomaterials for implants that closely resemble bone dentin, a natural

Diabetes is one of the rapidly increasing diseases in Qatar and around the world that causes one of the highest numbers of mortalities [1]. 19% of Qatar's population between the ages of 20 and 79 suffered from diabetes as of 2018 while this number is estimated to double by 2035 [2]. Globally, this number is predicted to rise to 55% by 2035. Diabetic sensorimotor polyneuropathy (DSPN) is a key consequence of diabetes mellitus, which is also the most common and costly complication involving sensory, motor, and autonomic nerves in both type 1 (T1D) and type 2 (T2D) diabetes and can affect 50–90% of diabetic patients [3, 4]. DSPN can have a negative impact on the gait of patients with diabetes [5] as there is a close connection between DSPN and anomalies in gait with an augmented risk of falls, particularly in older diabetic patients. Besides, DSPN is also an early indicator for non-healing diabetic wounds and diabetic foot ulcers, which account for one of the most common complications of diabetes, leading to increased healthcare cost, decreased quality of life, infections, amputations, and death. Early detection and better classification tools for DSPN can allow correct diagnosis and treatment of painful diabetic neuropathy as well as timely intervention to prevent foot ulceration, amputation and other diabetic complications. 

It has been recently shown using the pioneering imaging technique of corneal confocal microscopy [6] that remarkable abnormalities exist in gait, in not only patients with diabetes [7], but also in patients having reduced glucose tolerance and pre-diabetic symptoms [8]. These abnormalities have been linked to the severity of small fibre neuropathy, suggesting that alterations in small fibers and motor function and gait occur at very early stages of diabetes, and identification of these alterations will allow greater control of risk factors to limit nerve damage and to apply interventions such as vitamin D replacement and exercises to improve gait and limit falls. The wide prevalence of DSPN and its projected increase in the near future has further necessitated a need for the development of better diagnostic tools for early and/or better detection of DSPN. 

In this project, we propose to develop a novel contactless wearable electromyogram (EMG) system with an associated graphical user interface (GUI) for early, reliable, and robust detection of DSPN, stratification of severity, and assessment of the functional consequences (Figure 1). The project will start with the development of a contactless capacitive electrode for the acquisition of clean EMG signals from lower-limb muscles, and an adaptive neuro-fuzzy inference system (ANFIS) model will be trained utilizing the database created in the project. The ANFIS will be implemented in a host system (personal computer) for the diagnosis and stratification of DSPN severity. The system comprises of two subsystems: i) a contactless wearable capacitive sensor-based EMG signal acquisition and conditioning system; and ii) an adaptive digital filter with real-time signal processing and ANFIS classification (host system). The vastus lateralis (VL), Gastrocnemius Lateralis (GL), tibialis anterior (TA), and gastrocnemius medialis (GM) will be acquired by the contactless EMG system during several complete gait cycles when the patients will be instructed to walk at a self-selected rhythm on a 5m walkway of a portable force plate system. The vertical ground reaction force (vGRF) generated throughout several full gait cycles will be captured in the host system in synchrony to the four muscle EMG signals from wearable to identify the onset of muscle activation. The real-time adaptive digital filter will be the front-end of the classifier host subsystem to provide the dynamic filtering solution throughout the EMG signal measurement. The EMG signals’ linear envelopes curve will be produced after the offset removal, full-wave rectification, moving-average low-pass filtering, normalization by the mean of the EMG signal, and segmented to individual muscle activation in reference to vGRF data. The highest peaks and the delay to reach their peaks for the four different muscles, and inter-muscle activation parameter (co-activation index) will be calculated as EMG variables. The clinical gold-standard for DSPN severity detection technique, nerve conduction study (NCS) will be used for labeling the subject groups while creating the database. The severity of DSPN will be classified as: absent, mild, moderate and severe. The ANFIS will be trained using EMG variables (as inputs) to classify the subjects into 4 DSPN classes. The performance evaluation of the contactless ANFIS-EMG system will be carried out by comparing its classification accuracy with the NCS based DSPN stratification. We believe that the proposed system for real time assessment of DSPN will enable quantification of neuromuscular function to enhance the healthcare of diabetic patients in relation to identifying high-risk individuals to improve their risk factors and to target them for exercise to improve strength and gait as well as vitamin D replacement strategies. 

The project has four specific major aims:

Aim 1: To design and implement a contactless EMG system to generate clean cEMG signal. 

Aim 2: To design and implement a real-time and fully-automatic ANFIS classifier implemented in the host system for DSPN severity classification during gait dynamics.

Aim 3: To create a cEMG and NCS database for patients with differing severity of DSPN.

Aim 4: To validate the effectiveness of the Contactless ANFIS-EMG system to stratify DSPN severity with the NCS.

Outcome Summary:

Journal Published: 23     Journal Accepted: 2     Journal Under-review: 3    To be Communicated: 6    Patent Filed: 4

Articles Published (23):

1.  Ng CL, Reaz MB, Crespo ML, Cicuttin A, Shapiai MI, Ali SH, Kamal NB, Chowdhury ME. A Flexible Capacitive Electromyography Biomedical Sensor for Wearable Healthcare Applications. IEEE Transactions on Instrumentation and Measurement. 2023 Jun 6.

2.  Mahmud S, Hossain MS, Chowdhury ME, Reaz MB. MLMRS-Net: Electroencephalography (EEG) motion artifacts removal using a multi-layer multi-resolution spatially pooled 1D signal reconstruction network. Neural Computing and Applications. 2023 Apr;35(11):8371-88. 

3.  Faisal MA, Chowdhury ME, Mahbub ZB, Pedersen S, Ahmed MU, Khandakar A, Alhatou M, Nabil M, Ara I, Bhuiyan EH, Mahmud S. NDDNet: a deep learning model for predicting neurodegenerative diseases from gait pattern. Applied Intelligence. 2023 Mar, 53, pages 20034–20046.

4.  Ng CL, Reaz MB, Ali SH, Crespo ML, Cicuttin A, Chowdhury ME, Kiranyaz S, Kamal NB. Powerline interference suppression of a textile-insulated capacitive biomedical sensor using digital filters. Measurement. 2023 Feb 15;207:112425. 

5.  Shuzan MN, Chowdhury ME, Reaz MB, Khandakar A, Abir FF, Faisal MA, Ali SH, Bakar AA, Chowdhury MH, Mahbub ZB, Uddin MM. Machine learning-based classification of healthy and impaired gaits using 3D-GRF signals. Biomedical Signal Processing and Control. 2023 Mar 1;81:104448.

6.  Hossain MS, Mahmud S, Khandakar A, Al-Emadi N, Chowdhury FA, Mahbub ZB, Reaz MB, Chowdhury ME. MultiResUNet3+: A Full-Scale Connected Multi-Residual UNet Model to Denoise Electrooculogram and Electromyogram Artifacts from Corrupted Electroencephalogram Signals. Bioengineering. 2023 May 10;10(5):579.

7.  Khandakar A, Chowdhury ME, Reaz MB, Kiranyaz S, Hasan A, Rahman T, Ali SH, Bakar AA, Podder KK, Chowdhury MH, Faisal MA. DSPNet: A Self-ONN Model for Robust DSPN Diagnosis From Temperature Maps. IEEE Sensors Journal. 2023 Jan 19;23(5):5370-81.

8.  Khandakar A, Mahmud S, Chowdhury ME, Reaz MB, Kiranyaz S, Mahbub ZB, Md Ali SH, Bakar AA, Ayari MA, Alhatou M, Abdul-Moniem M. Design and implementation of a smart insole system to measure plantar pressure and temperature. Sensors. 2022 Oct 7;22(19):7599.

9.  Mahmud S, Khandakar A, Chowdhury ME, AbdulMoniem M, Reaz MB, Mahbub ZB, Sadasivuni KK, Murugappan M, Alhatou M. Fiber Bragg Gratings based smart insole to measure plantar pressure and temperature. Sensors and Actuators A: Physical. 2023 Feb 1;350:114092.

10. Khandakar A, Chowdhury ME, Reaz MB, Ali SH, Abbas TO, Alam T, Ayari MA, Mahbub ZB, Habib R, Rahman T, Tahir AM. Thermal change index-based diabetic foot thermogram image classification using machine learning techniques. Sensors. 2022 Feb 24;22(5):1793.

11. Hossain MS, Chowdhury ME, Reaz MB, Ali SH, Bakar AA, Kiranyaz S, Khandakar A, Alhatou M, Habib R, Hossain MM. Motion artifacts correction from single-channel EEG and fNIRS signals using novel wavelet packet decomposition in combination with canonical correlation analysis. Sensors. 2022 Apr 21;22(9):3169.

12. Haque F, Reaz MB, Chowdhury ME, Shapiai MI, Malik RA, Alhatou M, Kobashi S, Ara I, Ali SH, Bakar AA, Bhuiyan MA. A machine learning-based severity prediction tool for the Michigan neuropathy screening instrument. Diagnostics. 2023 Jan 11;13(2):264.

13. Haque F, Reaz MB, Chowdhury ME, Kiranyaz S, Ali SH, Alhatou M, Habib R, Bakar AA, Arsad N, Srivastava G. Performance Analysis of Conventional Machine Learning Algorithms for Diabetic Sensorimotor Polyneuropathy Severity Classification Using Nerve Conduction Studies. Computational Intelligence and Neuroscience. 2022 Apr 25;2022.

14. Haque F, Reaz MB, Chowdhury ME, Ezeddin M, Kiranyaz S, Alhatou M, Ali SH, Bakar AA, Srivastava G. Machine learning-based diabetic neuropathy and previous foot ulceration patients detection using electromyography and ground reaction forces during gait. Sensors. 2022 May 5;22(9):3507.

15. Hossain MS, Reaz MB, Chowdhury ME, Ali SH, Bakar AA, Kiranyaz S, Khandakar A, Alhatou M, Habib R. Motion artifacts correction from EEG and fNIRS signals using novel multiresolution analysis. IEEE Access. 2022 Mar 11;10:29760-77.

16. Khandakar A, Chowdhury ME, Reaz MB, Ali SH, Kiranyaz S, Rahman T, Chowdhury MH, Ayari MA, Alfkey R, Bakar AA, Malik RA. A novel machine learning approach for severity classification of diabetic foot complications using thermogram images. Sensors. 2022 Jun 2;22(11):4249. 

17. Md. Ahasan Atick Faisal, Muhammad E. H. Chowdhury, Amith Khandakar, Md Shafayet Hossain, Mohammed Alhatou, Sakib Mahmud, Iffat Ara, Shah Imran Sheikh, Mosabber Uddin Ahmed, “An Investigation to Study the Effects of Tai Chi on Human Gait Dynamics using Classical Machine Learning” in Computers in Biology and Medicine (2022) Jan 6;142:105184. doi: 10.1016/j.compbiomed.2021.105184.

18. Haque, Fahmida, Mamun Bin Ibne Reaz, Muhammad EH Chowdhury, Sawal Hamid Md Ali, Ahmad Ashrif A. Bakar, Tawsifur Rahman, Syoji Kobashi, Chitra A. Dhawale, and Mohammad Arif Sobhan Bhuiyan. "A nomogram-based diabetic sensorimotor polyneuropathy severity prediction using Michigan neuropathy screening instrumentations." Computers in Biology and Medicine 139 (2021): 104954.

19. Khandakar, Amith, Muhammad EH Chowdhury, Mamun Bin Ibne Reaz, Sawal Hamid Md Ali, Md Anwarul Hasan, Serkan Kiranyaz, Tawsifur Rahman, Rashad Alfkey, Ahmad Ashrif A. Bakar, and Rayaz A. Malik. "A machine learning model for early detection of diabetic foot using thermogram images." Computers in Biology and Medicine 137 (2021): 104838.

20. Haque, Fahmida, Mamun Bin Ibne Reaz, Muhammad Enamul Hoque Chowdhury, Geetika Srivastava, Sawal Hamid Md Ali, Ahmad Ashrif A. Bakar, and Mohammad Arif Sobhan Bhuiyan. "Performance Analysis of Conventional Machine Learning Algorithms for Diabetic Sensorimotor Polyneuropathy Severity Classification." Diagnostics 11, no. 5 (2021): 801.

21. Zaman, Kh Shahriya, Mamun Bin Ibne Reaz, Sawal Hamid Md Ali, Ahmad Ashrif A. Bakar, and Muhammad Enamul Hoque Chowdhury. "Custom Hardware Architectures for Deep Learning on Portable Devices: A Review." IEEE Transactions on Neural Networks and Learning Systems (2021).

22. Haque, Fahmida, Mamun BI Reaz, Muhammad EH Chowdhury, Fazida H. Hashim, Norhana Arsad, and Sawal HM Ali. "Diabetic Sensorimotor Polyneuropathy Severity Classification Using Adaptive Neuro Fuzzy Inference System." IEEE Access 9 (2021): 7618-7631.

23. Ng CL, Reaz MB, Crespo ML, Cicuttin A, Chowdhury ME. Characterization of capacitive electromyography biomedical sensor insulated with porous medical bandages. Scientific reports. 2020 Sep 10;10(1):14891.

Articles Accepted (2):

1.  Khandakar A, Faisal MA, Chowdhury ME, Reaz MB, Ali SH, Razak MI, Bakar AA, Mahmud S, Malik RA. Laser Induced Graphene based Smart Insole to Measure Plantar Temperature. IEEE Sensors Journal. 2023 Sep 26.

2.  Md. Ahasan Atick Faisal, Muhammad E. H. Chowdhury, Sakib Mahmud, Amith Khandakar, Mosabber Uddin Ahmed, Abdulrahman Alqahtani, “Robust and Novel Attention Guided MultiResUnet Model for 3D Ground Reaction Force and Moment Prediction from Foot Kinematics” in Neural Computing and Applications. 2023

Articles Under-review (3):

1. Sakib Mahmud, Muhammad E. H. Chowdhury, Serkan Kiranyaz, Nasser Al Emadi, Anas M. Tahir, Md Shafayet Hossain, Amith Khandakar, Somaya Al-Maadeed, “Restoration of Motion-Corrupted EEG Signals using Attention-Guided Operational CycleGAN” is submitted to Engineering Application of Artificial Intelligent (2023). 

2. Md Nazmul Islam Shuzan, Muhammad E. H. Chowdhury, Mamun Bin Ibne Reaz, Moajjem Hossain Chowdhury, Sawal Hamid Md Ali, Zaid Bin Mahbub, Abdulrahman Alqahtani, Amith Khandakar, Ahmad Ashrif A. Bakar, Mohd Ibrahim Bin Shapiai@Abd Razak, Rayaz A Malik, “Self-GRFNet: A Lightweight Self-Organizing Neural Networks for Classifying Healthy and Impaired Gaits” is submitted to IEEE Sensor Journal (2023).

3. Fahmida Haque, Mamun Bin Ibne Reaz, Muhammad E. H. Chowdhury, Serkan Kiranyaz, Mohamed Abdelmoniem, Emadeddin Hussein, Mohammed Shaat, Sawal Hamid Md Ali, Ahmad Ashrif A Bakar, Geetika Srivastava, Mohammad Arif Sobhan Bhuiyan, Mohd Hadri Hafiz Mokhtar, “Machine Learning-based Diabetic Sensorimotor Polyneuropathy Severity Classification using Biomechanical Features during Gait” is submitted in Sensors (2023).

Articles to be Communicated (6):

1. Tawsifur Rahman, Muhammad E. H. Chowdhury, Ioannis N. Petropoulos, Amith Khandakar, Maryam Ferdousi, Uazman Alam, Erik Lovblom, Bruce Perkins, Daniele Pacaud, Rodica Pop-Busui, Roni Shtein, Katie Edwards, Nathan Efron, Rayaz A Malik, “Deep Learning-based Corneal Nerve Analysis Accurately Diagnoses Patients with Diabetic Peripheral Neuropathy” will be submitted to Diabetic Care, 2023. 

2. Mehrin Nawaz, Muhammad E. H. Chowdhury, Ioannis N. Petropoulos, Amith Khandakar, Mamun Bin Ibne Reaz, Anwarul Hasan, Mohamed Alhatou, Rayaz A Malik,  “Deep learning-based model for classifying age-controlled healthy and type 2 diabetes populations using wide-field corneal subbasal nerve plexus mosaics” will be submitted to Scientific Reports, 2023. 

3. Adiba Tabassum Chowdhury, Muhammad E. H. Chowdhury, Rumana Akhter, Zaid Bin Mahbub, Amith Khandakar, Mamun Bin Ibne Reaz, Anwarul Hasan, Mohamed Alhatou, Rayaz A Malik, “A novel machine learning scoring system to stratify the severity of diabetic neuropathy patients using nerve conduction studies” will be submitted to Computer in Biology and Medicine, 2023. 

4. Khandakar, Amith, Muhammad EH Chowdhury, Mamun Bin Ibne Reaz, Sawal Hamid Md Ali, Md Anwarul Hasan, Serkan Kiranyaz, Tawsifur Rahman, Rashad Alfkey, Ahmad Ashrif A. Bakar, and Rayaz A. Malik. "A novel deep learning framework to stratify the severity of diabetic neuropathy patients using thermogram images." in Computers in Biology and Medicine, 2024. 

5. Khandakar, Amith, Muhammad EH Chowdhury, Mamun Bin Ibne Reaz, Md Anwarul Hasan, Serkan Kiranyaz, Mohammed Alhatou and Rayaz A. Malik. “A portable thermal imaging platform to stratify the severity of diabetic neuropathy patients” in Journal of IEEE Sensor, 2024.

6. Nazmul Islam Shuzan, Adiba Tabassum Chowdhury, Muhammad EH Chowdhury, Mamun Bin Ibne Reaz, Rumana Akhter, Zaid Bin Mahbub, Rayaz Malik, “A multimodal framework using surface EMG and vertical ground reaction force to classify the diabetic polyneuropathy patients” in Diabetic Care, 2024.

2.  Amith Khandakar, Muhammad E.H. Chowdhury, Mamun Bin Ibne Reaz, Mamun Bin Ibne Reaz, Sawal Hamid Md Ali, Serkan Kiranyaz, Rashad Alfkey, Ahmad Ashrif A. Bakar, Rayaz A. Malik, Mohamed Arselene Ayari, Moajjem Hossain Chowdhury, Kanchon Kanti Podder “Self-DFUNet -A Novel Machine Learning Model for stratification of Early Diabetic Foot complication using Thermogram Images.”, US Patent Application, 2023. 

4.  Sakib Mahmud, Md Shafayet Hossain, Muhammad E. H. Chowdhury, Mamun Bin Ibne Reaz “A Deep Learning Model to Remove Motion Artifacts from Electroencephalography (EEG) Signals”, US Patent App.  63/367,174.

Research Team (Core) 

Graduate Students (Ph.D.)

Graduate Students (M.S.)

Research Assistants

Photo Gallary