Longitudinal neural and vascular recovery following ultraflexible neural electrode implantation in aged mice

Abstract

Flexible neural electrodes improve the recording longevity and quality of individual neurons by promoting tissue-electrode integration. However, the intracortical implantation of flexible electrodes inevitably induces tissue damage. Understanding the longitudinal neural and vascular recovery following the intracortical implantation is critical for the ever-growing applications of flexible electrodes in both healthy and disordered brains. Aged animals are of particular interest because they play a key role in modeling neurological disorders, but their tissue-electrode interface remains mostly unstudied. Here we integrate in-vivo two-photon imaging and electrophysiological recording to determine the time-dependent neural and vascular dynamics after the implantation of ultraflexible neural electrodes in aged mice. We find heightened angiogenesis and vascular remodeling in the first two weeks after implantation, which coincides with the rapid increase in local field potentials and unit activities detected by electrophysiological recordings. Vascular remodeling in shallow cortical layers preceded that in deeper layers, which often lasted longer than the recovery of neural signals. By six weeks post-implantation vascular abnormalities had subsided, resulting in normal vasculature and microcirculation. Putative cell classification based on firing pattern and waveform shows similar recovery time courses in fast-spiking interneurons and pyramidal neurons. These results elucidate how structural damages and remodeling near implants affecting recording efficacy, and support the application of ultraflexible electrodes in aged animals at minimal perturbations to endogenous neurophysiology.

Keywords: Chronic; Electrophysiology; In vivo two-photon imaging; Tissue-electrode interface; Ultraflexible electrode; Vasculature.

Fig. 1:. Ultraflexible PI NET permits longitudinal 2P imaging of microvasculature and concurrent neural recording.

A: key microfabrication steps of PI-NET. B, C: photos of NETs (4 shank, 32 channel in total) on the fabrication substrate (B) and after released from the substrate and immersed in water (C). D, E: Photo (D) and schematics (E) showing NET co-implanted with a cranial window to accommodate concurrent 2P imaging and electrophysiological recordings longitudinally. Animal was awake and head-fixed for all measurements. F: a representative 3D reconstructed image of vasculature surrounding three NET shanks at 6 weeks after implantation. Golden ribbon: pseudo-colored NETs. G: a representative maximum intensity projection (MIP) in xy plane (at 200 μm deep). H: A zoom-in binarized image of G for quantification of micro-vessel volume fraction. White: binarized result. I: representative laminar cortical recordings from a 32-channel NET array at week six after implantation. Scale bars: 200 μm (B, C, F, G), 100 μm (H), 1 mm (inset in D); 20 ms (I, horizontal) and 500 μV (I, vertical).

Fig. 2:. Vascular remodeling and neural recovery is the most pronounced in the first two weeks of implantation.

A: MIP of xz sections (90 mm thick in Y) of a representative animal showing vascular recovery at the NET interface. Dashed orange lines depict NET. B: Volume fraction of vasculature as a function of distance from the NET showing the spatial extent of injury at multiple time points after implantation. C: Volume fraction of vasculature as a function of time post implantation from within 200 μm of the NET implantation site and more than 300 μm away; values are mean±SD (n = 16). D: representative laminar recordings from one NET shank (8 channels) showing spontaneous LFP and spiking activities. E – G: normalized LFP at 30 – 60 Hz (E), firing rate (FR) and unit per channel (F), and spike amplitude (G) as a function of days post implantation; values are mean ± SE (n = 30). H: scatter plot of normalized LFP and vasculature volume fraction showing a sharp increase in LFP at relative low vascularization. The solid line is guide for the eye. Scale bars: 100 μm (A); 20 ms (D: horizontal), and 200 μV (D, vertical).

Fig. 3:. Microvascular remodeling shows location and depth dependence with little influence on neural electrophysiological activity.

A: vasculature images at multiple cortical depths and time points after implantation from one representative animal. Arrow denotes the implantation site. NETs are pseudo-colored as yellow. Scalebar: 200 μm. B: cortical-depth dependence of the averaged time needed for regions within 200 μm of NETs to reach 80% vessel density relative to regions far away. values are mean±SE (n = 12). C: color-coded vessel recovery time as a function of cortical depth and distance away from NETs (n = 12). D: color-coded vessel recovery time from a subset of animals in C representing faster-than-average recovery time course (n = 6). E – G: Spike rate (E), amplitude (F), and spike yield per channel (G) in the group of faster recovery vasculatures and the rest. No significant difference was detected; values are mean±SE (n = 9).

Fig. 4:. Large-scale recording reveals changes of single-unit activity over time, cortical depth, and cell types.

A: Example recordings of single units from a single 32-channel NET shank. Numbers denote the recording sites along the shank. Multiple waveforms were detected at each site. Scale bars: 200 μV (vertical) and 2 ms (horizontal). B – C: single unit yield per channel, firing rate (B), and amplitudes (C) as a function of days after implantation; values are mean±SE (n = 16). D – F: cortical depth dependence of single unit firing rate (D), amplitude (E), and single-unit yield per channel (F) as a function of days post implantation. G: scatter plot of 4593 single units in 3D space for purtative cell classification (blue, interneurons; red, remaining neurons). H: example histogram distribution of trough-to-peak showing bi-modal distributions. I: Number of putative narrow waveform interneuron and their ratio as a function of days post implantation. Scale bars: 2 ms (A, horizontal), 200 μV (A, vertical).

Fig. 5:. Vessel diameter remained relative stable but vasospasms were detected occasionally.

A: color-coded binarized imaging stack showing the three categories of vessels. Orange: large vessels; green: medium vessels; grey: small vessels. Golden: NETs B: Vessel diameters as a function of time post implantation for the three vessel categories. C: 2P imaging stacks on the implantation day, Day 7, Day 14, and Day 42 post implantation from two animals showing diverse vessel changes. Vessel of interest marked in orange. All images are MIP of xy sections (100 μm thick stack in Z). D: Diameter of the marked vessel in (C) as a function of time post implantation; values are mean±SE (n = 7).

Fig. 6:. Microcirculation in perfused capillaries near the implantation sites remained stable over time.

A: representative 2P imaging marking the position of a capillary (dashed line) for the line scans in B. Scale bar: 200 μm. Green, vasculature; golden Yellow: NET; B: Matrix of line scans showing movement of RBCs as dark stripes, the slope of which gives the blood flow speed. C – E: Box plot of RBC velocity (C), flux (D), and hematocrit (E) values from week 1 to week 6 post implantation. Whisker plots show the 25th percentile and the 75th percentile of the data set. For (C) η_p² = 0.022, *P < 0.05; n = 513 capillaries (6 mice); RBC velocity; (D) η_p² = 0.023, **P < 0.01; n = 515 capillaries (6 mice); RBC flux; (E) n = 466 capillaries (6 mice); hematocrit, by one-way ANOVA.