NeuroRoots, a bio-inspired, seamless brain machine interface for long-term recording in delicate brain regions

Abstract

Scalable electronic brain implants with long-term stability and low biological perturbation are crucial technologies for high-quality brain-machine interfaces that can seamlessly access delicate and hard-to-reach regions of the brain. Here, we created "NeuroRoots," a biomimetic multi-channel implant with similar dimensions (7 μm wide and 1.5 μm thick), mechanical compliance, and spatial distribution as axons in the brain. Unlike planar shank implants, these devices consist of a number of individual electrode "roots," each tendril independent from the other. A simple microscale delivery approach based on commercially available apparatus minimally perturbs existing neural architectures during surgery. NeuroRoots enables high density single unit recording from the cerebellum in vitro and in vivo. NeuroRoots also reliably recorded action potentials in various brain regions for at least 7 weeks during behavioral experiments in freely-moving rats, without adjustment of electrode position. This minimally invasive axon-like implant design is an important step toward improving the integration and stability of brain-machine interfacing.

FIG. 1.

(a) NeuroRoots overview and assembly. 3D model of the NeuroRoots and the different configurations of the tip. Design 1 and 2 are zoomed-in representations of the tip with the electrodes organized in 150 and 25 μm depth spacing, respectively. A zoomed-in representation of one tetrode-like set of electrodes and the cross-sectional representation are also presented. Parylene C substrate is represented in light gray and platinum in dark brown. (b) Microscope picture image of the NeuroRoots with Design 1 and Design 2. (c) Assembly method using capillary and surface-tension effects to draw the electrodes onto the microwire. Bottom insets show a microscope picture of the electrode leads after delamination from the fabrication substrate and after lamination on a substrate. (d) Microscopy showing the electrodes assembled onto electrosharpened microwires, demonstrating their spacing is set by the initial position in the array.

FIG. 2.

(a) Apparatus for chronic recordings in freely moving rats. A picture of a rat 4 days post-surgery. The implant connector is protected by a cap on top of the scaffold presented on the right. The 3D exploded-view shows the main parts that compose the NeuroRoots platform. Left pictures show the adaptor and the NeuroRoots assembled before insertion and a zoomed-in picture of the guiding system with the NeuroRoots assembled onto the microwire. (b) Implantation strategy of the NeuroRoots into deep-brain regions. The assembled microwire and NeuroRoots are implanted into the desired brain region through a 3D printed scaffold hat (left). Once the NeuroRoots are released, the microwire is removed, leaving only the electrodes implanted into the brain tissue (right). (c) X-ray microtomography scanning showing the electrode distribution after implantation using PEDOT:PSS based devices. (d) Microscope image of a NeuroRoots placed onto a brain slice of cell CA1 of the hippocampus for scale. The electrodes of 10 μm in diameter are similar in size to neuron soma (Cresyl violet staining).

FIG. 3.

(a) Representative recordings of raw traces. (i) Two seconds recording of six consecutive channels. Scale: 200 ms. (ii) Zoomed-in view of the blue box presented in (i) shows characteristic activity while the green zoomed-in green box show the activity of a neighbor electrode. Scale: 10 ms. (iii) Action potential of a hippocampal neuron recorded during the acute experiment. Scale: 0.5 SD and 1 ms. (b) Cluster stability assessment during the chronic experiment. (i) Representative APs of the same neuron and interneuron over the 7 weeks of the experiment. Scale neuron: 15 μV, 1 ms, interneuron 15 μV, and 0.2 ms. (ii) Overlay of averaged APs corresponding to a cluster tracked over 7 weeks. Scale: 4 SD and 0.5 ms. Each cluster corresponds to a different electrode. Each color is representative of recordings from a different week as describes in the colored legend. (c) Behavioral analysis. (i) Animal trajectories in gray on a double Y-Maze (1.4 × 1.2 m²). Spikes from an example cell overlaid in red. (ii) Estimated firing fields for the same cell; warmer color indicates increased firing activity.

FIG. 4.

(a) Microscope image of NeuroRoots implanted into a cerebellar slice. (b) In vitro recordings from a cerebellar slice. Typical activity of the 32 channels during recordings. Each trace is (−270; 200] μV and 1 s. (C) Microscope image of the histological slice with the NeuroRoots footprint measured at 38 μm in diameter. (d) In vivo recordings in the cerebellum. Typical activity from 4 electrodes after off-line processing, scale bar 100 μV, 4 ms. A typical single unit extracted from one of the channels during in vivo recordings is presented on the right. Red and blue colors correspond to two distinct single units and the bold line represents the averaged activity of the spike collected. Scale bar: 200 μV and 0.2 ms.