CMU faculty work on biosensors

For faculty from Carnegie Mellon University and the University of Pittsburgh, a project that started as creating a sensor for measuring bone stress has turned into a project for revolutionizing how humans and technology interact. Also known as Brain Computer Interface (BCI), biosensors are implantable micromechanical devices that allow human neural activity to work in conjunction with technology.
One common use would be allowing a person to control a prosthetic limb by just thinking. These devices would allow a handicapped or medically ill person to have a better quality of life.

Another use of the BCI interface is for epilepsy patients, who suffer from random seizures. Xin Li, a research scientist in the ECE department who is working on signal processing, said, “If we come up with a system which can continuously measure the signals from the brain, and then do some signal processing, [we can] try to predict if the seizure will happen or not. If [the seizure] happens, we can control a device, for example, a drug delivery device to control the seizures.” This project is a collaborative effort combining the electrical and computer engineering, robotics, and medical engineering departments at Carnegie Mellon, and the medical center at Pitt under one roof: the Center for Implantable Medical Microsystems (CIMM).

CIMM is part of the Institute for Complex Engineered Systems (ICES), which was started in spring of this year and is currently headed by ECE professor Gary Fedder.

The project is still in the developing phase and research is ongoing.

The biosensors that are still being developed would consist of a 32-input array of electrodes that connect to nerves in a part of the brain called the dura. These electrodes would pick up electrical signals or voltage off of neurons and transmit them through complex circuitry to a signal-processing device.

From there, they would be wirelessly transmitted to a computer, where the characteristics of the signals, such as amplitude or frequency, would be decoded and translated into algorithms.

The computer would apply the algorithms and the device would behave accordingly.

Ken Mai, an assistant professor in the ECE department, is a principal investigator working on Brain Implantable Computing Platforms, the part of the sensor that gets implanted into the brain. “This is in contrast to existing implants that are primarily sensors and actuators that rely on external computational resources. By implanting the computation, we contend that we can significantly increase the capability of the implants and the sophistication of the applications,” Mai said.

One problem that the team is facing is the question of how to pack the circuitry and a strong enough power source into a tiny package.
Mai proposed a solution to this problem saying, “[A solution is to generate] 3-D stacking of chips for smaller form factors.” Another problem is that if the strong power source is compacted into a tiny area, it would lead to high heat levels which could cause tissue damage in the brain. To counteract this, the team has suggested the use of sub-threshold circuits and the implementation of a wireless power delivery method.

Even if a solution to all these problems is found, there is always the problem of having the body reject the sensor as a response from the immune system.

Even though the project is less than a year old, Fedder and Mai both believe that the project can be used in human trials within four to five years.

Fedder said, “The technology exists today to create what we’re talking about, but there hasn’t been that much research. It is certainly possible within five years. It’s a real collaborative effort, with multidisciplinary fields that requires the effort of many different people.”
Mai said, “I would expect that implantable devices will have widespread use in medicine leading many people to have implants at least on a temporary basis.”