Put the Metal to the Petal

Modern computing systems may be fantastically powerful, but they pale in comparison to the power — and especially the efficiency — of the human brain. As artificial intelligence and other cutting-edge applications continue to demand a steadily increasing diet of energy that limits their scalability, researchers are being driven to seek out alternative solutions. Without innovation in this area, future progress will ultimately be stymied by an unsustainable consumption of resources.

One way out of this problem may involve developing artificial computing systems that operate more like real brains. But of course there are many challenges standing in the way of such an approach. First and foremost, we do not have a deep understanding of how the brain works, so how can we replicate something that is largely a black box to us?

Of course the answer is that we cannot. We first need to collect the raw data necessary to give us that understanding. An approach presently being used to collect this data makes use of three-dimensional clusters of neurons called neural spheroids. Microelectrode arrays can capture the electrical activity of the neurons in real-time, giving us insights into how they operate under different conditions.

Present microelectrode arrays tend to be planar in shape, however, which means they cannot simultaneously collect data from the surface of the entire neural spheroid. This greatly limits our view of the interactions between neurons that take place as the brain processes information. But with any luck, capturing a more complete picture of brain activity may be easier in the future, thanks to a serendipitous discovery by a research group at the Swiss Federal Institute of Technology Lausanne.

As the team was working to develop soft implants for peripheral nerves, they found that the hydrogels in the implants curled up when exposed to water, rendering them useless. Useless as implants, that is. But that curling comes in very handy when working with neural spheroids. A microelectrode array embedded in the same material will curl up and hug the three-dimensional shape of the tissue without causing damage, allowing it to record electrical activity from the entire surface of the structure simultaneously.

The team calls their device the e-Flower, as it is composed of four flexible, flower-like petals with embedded platinum electrodes. The petals fold up as the hydrogel they are composed of swells with water found in the cell culture medium the neural spheroid is immersed in. That allows the petals to make good contact while being gentle and not requiring any harmful solvents that could damage the tissue or otherwise alter its normal electrical activity.

In the future, the team plans to apply their technology to brain organoids as well, which are composed of multiple cell types and more accurately model real brain activity. In addition to providing us with clues as to how we may improve present technologies, it is also believed that e-Flower will help researchers to better understand the biology of a variety of neurological disorders.