Injecting electronics with a syringe. Reproduced from Liu. et al., Nature Publishing Group.

Neural activity is often recorded with rigid electrode arrays that are implanted via surgical openings in the skull. The inflexible nature of such electrodes can require compromises in their positioning and may lead to shifts in their location during long-term recordings. To overcome these issues, ideal recording electrodes would be flexible and integrate into the tissue. Charles Lieber and his team at Harvard University have developed flexible meshes that can be injected into biomaterials and tissues and that contain electronics suitable for recording neural activity or other applications.

According to Lieber, the idea of flexible, injectable electronics has evolved over several years and builds on tissue engineering work in his lab. He had become increasingly disillusioned with chip-based work and was looking for less invasive ways to interface with tissues. This required “reducing overall feature sizes, increasing flexibility and making the connectivity look more and more like something that is natural,” says Lieber.

In addition to the flexibility of the material, the ability to simply inject the electronics further reduces the invasiveness of the tool. However, to allow the meshes to be inserted into syringes, the researchers had to make some simple adjustments to the mesh design. “You don't want this to crumple up like a ball of aluminum foil or it will get jammed in the needle. You want it to scroll up like a piece of paper ... inside the needle,” says Lieber. They achieved this by giving the mesh subunits a parallelogram shape, which reduced stiffness in the direction that the mesh rolls up while maintaining rigidity in the other direction.

Once the mesh is injected, its behavior depends on the surrounding tissue or material. Within soft, flexible surroundings, the mesh can expand and unfold. However, when injected into the dense tissue of a mouse brain, it can expand only a little and thus does not cut into and damage the tissue. More importantly, the mesh has mechanical properties and feature sizes similar to those of neural tissue, yielding almost no immunoreactivity in contrast to other electrodes.

The researchers demonstrated the suitability of the injectable electronics by using them to record local field potentials in the mouse hippocampus. Standard analysis methods even allowed them to identify single-unit action potentials in their recordings. A benefit of the injectable electronics is that the researchers can position them with a precision of 10 micrometers, and they know exactly where the recording electrodes are located on the mesh. Lieber thinks that these meshes can even be placed more precisely than rigid probes. In addition, the injected meshes enable chronic recordings, as they are incorporated into the tissue and their position remains stable. “We are able to record really nice single-unit data over literally 4 months,” mentions Lieber.

Finally, the injectable electronics enable experiments that were not previously possible. In collaboration with Josh Sanes at Harvard University, Lieber has started to perform electrophysiological recordings in the mouse retina, which was previously not accessible in vivo. Because the injectable electronics are essentially transparent, they do not interfere with vision. Lieber even dreams of recording from multiple sites simultaneously, such as the retina and the visual cortex, which should be possible as the electronics are so easy to inject.