Flexible neural interfaces for brain implants—the pursuit of thinness and high density

Neural interfaces that directly measure brain activity are increasingly employed to elucidate large-scale brain networks and treat intractable neurological disorders. Considering the softness of brain tissue, current efforts to study chronic disorders aim to minimize invasiveness. We discuss recent progress on flexible neural interfaces with high durability under bending and stretching achieved by using organic materials. Multichannel microelectrodes are usually fabricated on thin polymer substrates as sheets and needles to reach superficial and deep brain structures, respectively. An interesting recent trend is the integration of high-density microelectrodes to measure detailed brain functions. The use of numerous measurement points (the current highest values achieved are 62 500 electrodes cm–2 and 3072 channels) can increase the accuracy of brain state estimation. However, further improvement should be devised for integration in plane considering the density of 250 000 neurons cm–2 in approximate intervals of 20 μm. Meanwhile, the ultimate goal of improving flexibility in neural interfaces is long-term implantation. Widely used approaches for thinning polymers (∼1 μm) and reducing the rigidity of neural interfaces compromise robustness due to high gas permeability and water uptake. We quantitatively analyze the technical proficiency of flexible neural interfaces in vivo regarding microelectrode integration and robustness. The solution contact impedance, which is a crucial factor in microelectrode miniaturization, is exhaustively surveyed and compared across PEDOT:PSS, Au, Pt, Pt black, IrOx, gels, and other components that should be designed within the permissible source impedance for the measurement device to ensure high-accuracy and low-noise measurements of brain activity in the order of microvolts. Furthermore, we detail a multifunctional neural interface with stretchability, optical transparency, easy intraoperative handling, and flexible transistor implementation for building an active electrode array, providing a new approach for flexible interfaces in neuroscience and neuroengineering.