Joseph Costello awarded Rackham Predoc to support research on brain-machine interfaces

Costello is working to restore mobility to individuals by developing improved brain-machine interfaces.
Joseph Costello

Joseph Costello, PhD student in ECE, was awarded a Rackham Predoctoral Fellowship for his research that focuses on high performance, power efficient brain-machine interfaces.

Brain-machine interfaces (BMIs) are a promising solution for restoring mobility and communication to people who suffer from sensorimotor impairments, including spinal cord injury, limb amputations, stroke, and neurodegenerative diseases. 

“Tasks we take for granted like drinking from a cup, interacting with a phone, or communicating words can be challenging or impossible for individuals suffering the loss of limb mobility or speech,” explained Costello. “Brain-machine interfaces (BMIs) offer a promising solution for restoring mobility and communication.”

BMIs used in the cerebral cortex consist of an electrode array implanted in the brain, a decoding algorithm to directly translate neural signals into the user’s intended actions, and an output device.  

Brain-machine interfaces (BMIs) offer a promising solution for restoring mobility and communication.

Joseph Costello

While BMIs have successfully controlled computer cursors and robotic arms, they are still limited in the accuracy of the decoding algorithm and consume too much power for an implantable system. Costello wants to fix these limitations.

His research will focus on: (1) developing accurate decoding algorithms that work well in real-time, (2) optimizing the power consumption of the neural signal processing pipeline, and (3) applying these findings for prosthetic control with human study participants.

A recent project involves working to improve prosthetic hand function for people who have had upper limb amputation or paralysis. He evaluated a type of artificial neural network called a recurrent neural network (RNN) against other neural network algorithms in real-time for continuous decoding of finger movements. And he found that RNN decoders could be used to control simple movements as accurately as a non-disabled hand. 

“He is going to be one of the rock stars of the next generation of neural prosthetics,” said Cynthia Chestek, Associate Professor of Biomedical Engineering, Robotics and ECE, and Director of the Cortical Neuroprosthetics Lab.

Costello has 13 publications to his name, and another in preparation. He is first author for two of the publications, one of which was selected as a spotlight paper at the Conference on Neural Information Processing Systems (NeurIPS 2023).

He received his B.S.E. in electrical engineering at the University of Michigan, and has already acquired extensive experience in the intersection of engineering and neuroscience through work at Boston Scientific Neuromodulation and several neuroscience labs at Michigan. As an undergraduate student, Costello participated in the Clear Lung Project with the group M-HEAL, where he led the engineering sub-team to design, build, and test a device to diagnose pneumonia in underdeveloped countries. He also developed an app to help a Parkinson’s patient with advanced gait abnormalities through an electronic foot vibrator and connected phone application. This technology earned him a Cognizant Making the Future Scholarship 2016. As a graduate student, he received an NSF Graduate Research Fellowship. 

Costello has mentored several undergraduate and graduate students in Chestek’s lab. In addition, he is a volunteer with Science Education & Engagement for Kids (SEEK) at underserved elementary schools, and is also a national and world championship qualifier on the Michigan Triathlon Team. He is co-advised by Prof. Cynthia Chestek and David Blaauw, the Kensall D. Wise Collegiate Professor of Electrical Engineering and Computer Science.