Accurate control is one of the biggest challenges in prosthetic hands. A prosthetic must move only when the user intends it to, as unintended movements can cause frustration, reduce trust, and create safety risks.

To address this issue, I explored the main control technologies used in prosthetics: EMG (muscle signals) and EEG (brain signals). Electromyography (EMG) measures the electrical activity produced when muscles contract. When your brain sends a signal to move your hand, tiny electrical impulses travel through motor neurons to the muscle fibers. Electrodes placed on the skin detect these electrical changes. The strength and pattern of the signal can then be analyzed and translated into commands, such as opening or closing a prosthetic hand. Electroencephalography (EEG) measures electrical activity directly from the brain. Neurons in the brain communicate using small electrical impulses, and when large groups of neurons fire together, they create detectable voltage patterns. Electrodes placed on the scalp record these signals. Specific brainwave patterns can be associated with different intentions, such as imagining a movement, and these patterns can be processed to control external devices.

Here shows the pros and cons comparing the two systems which affects how well a prosthetic performs. EMG advantages then disadvantages:

• Fast response
• Strong, clear signals
• Widely used and reliable
• Can activate accidentally
• Needs usable muscles
• Muscle fatigue can reduce accuracy

EEG advantages then disadvantages:

• Works without muscles
• Captures user intent directly
• Allows advanced control options
• Slower response
• Noisy and less precise
• Requires careful calibration

Based on this comparison, I began to question: why rely on only one signal source if both provide valuable but different information? Through further research, I found that combining EMG and EEG into a hybrid control system could significantly improve prosthetic performance. In this model, the system would collect input from both brain signals (EEG) and muscle signals (EMG) simultaneously. An intelligent processing system;such as an AI classifier could analyze patterns from both data streams in real time. Within a second(more or less as technology improves), it could determine the user’s intended action and decide how to execute it. For example, in routine movements, EMG could handle fast execution because of its strong and immediate signals. In safety-critical situations, such as gripping a fragile object or operating near hazards, the system could require confirmation from both EEG (intention) and EMG (execution) before activating movement. This dual-layer verification would reduce accidental activation and improve safety.

Over time, as technology advances, the system could incorporate adaptive learning. It could adjust to muscle fatigue, changes in signal patterns, or individual user differences. Additional factors such as movement history, environmental context, or pressure sensors in the prosthetic hand could also be integrated to further refine decision-making. By dynamically deciding when to rely on EMG, EEG, or both together, a hybrid system could provide faster response, higher accuracy, improved safety, and greater user confidence—making prosthetic control more natural and reliable.

Based on my data and simulation, EEG and EMG each have important limitations when used alone. EMG provides fast and precise muscle detection but can be unreliable due to noise, fatigue, or involuntary activation.

EEG detects user intent before movement occurs but is more susceptible to signal noise and misinterpretation.

When combined:

• EEG helps confirm intentional movement
• EMG verifies physical execution
• False activations are reduced
• Overall accuracy, safety, and reliability improve

By combining EEG and EMG, the strengths of one system compensate for the weaknesses of the other, filling gaps in intent detection, reliability, and control accuracy.

This hybrid approach creates a more accurate and dependable prosthetic control system by requiring both intent (EEG) and action (EMG) before movement occurs. The combined signals improve safety, reduce false activations, and increase user trust. In the future, advanced machine learning could further personalize control, adapt to muscle fatigue, and make prosthetic devices feel more natural, intuitive, and reliable for users.

Articles & Web Sources Oitzman, M. (2025, September 9). How BrainCo robotic hands are changing lives. The Robot Report. Retrieved from https://www.therobotreport.com/how-brainco-robotic-hands-changing-lives/ BrainCo. (2026). BrainCo Bionic Hand: Reclaim independence & embrace life. Retrieved from https://www.brainco.cn/en-US/news/sk8qg3fv1txyetl6jk97ijwt BrainCo. (2025). Rediscovering the everyday: How the BrainCo Bionic Hand is restoring daily life. Retrieved from https://www.brainco.cn/en-US/news/sizwr1fahhj4se7t5vnuggtc Shen, K., Zhang, Z., & Guo, S. (2026). Comparative study on different control methods of limb prostheses. Academic Journal of Science and Technology. Retrieved from https://drpress.org/ojs/index.php/ajst/article/view/32273 Kumar, P. S. (2025). BIONIX: A wireless, low-cost prosthetic arm with dual-signal EEG and EMG control. arXiv. https://doi.org/10.48550/arXiv.2512.16929 Comparative control of rehabilitation wheelchair using periocular electromyography and electroencephalography. (2024). Biomedical Signal Processing and Control, 90, Article 105854. https://doi.org/10.1016/j.bspc.2023.105854 Other Academic References Hasan, M. K., Wahid, S. R., Rahman, F., Maliha, S. K. (2022). Grasp-and-lift detection from EEG signals using convolutional neural network. arXiv. https://doi.org/10.48550/arXiv.2202.06128 Jafarzadeh, M., Hussey, D. C., & Tadesse, Y. (2019). Deep learning approach to control of prosthetic hands with electromyography signals. arXiv. https://doi.org/10.48550/arXiv.1909.09910 News / Feature Story Choi, B. (2022, August 13). Benjamin Choi: 17‑year‑old invents pathbreaking prosthetic arm. TIMnovate. Retrieved from https://timnovate.com/2022/08/13/benjamin-choi-17-year-old-invents-pathbreaking-prosthetic-arm/ Along with these, it includes Youtube videos and all the people on Stack Overflow who helped answer some questions about coding the simulations. ChatGPT also assisted in explaining scientific concepts, clarifying questions, and helping me better understand topics related to my science fair project.

I would like to sincerely thank my mom for her constant support and encouragement throughout this project. I am especially grateful to Mr. Degelder for all his guidance and help along the way. I also appreciate all my teachers for their support, as well as my friends and everyone else who encouraged me during this journey.