Main
The extensive lines of bidirectional communication between the brain and visceral organs facilitate integration of internally arising interoceptive cues that are critical for survival. The gut–brain communication exemplifies one such important pathway wherein humoral and neural signals emerging from the abdominal viscera relay metabolic information to the brain for maintaining energy balance. Besides the well-known homeostatic functions, recent evidence suggests that the consciously imperceptible gut-to-brain signals can also modulate higher-level cognitive processes such as motivation, affect, learning and memory1,2,3,4,5. These findings create opportunities for co-opting such brain–organ neural circuits to develop minimally invasive autonomic neuromodulation therapies for otherwise intractable metabolic and neurological disorders such as treatment-resistant depression, obesity or diabetes6,7. However, identification of mechanisms underlying brain–viscera communication that influence neurocognitive states has remained challenging, which is at least in part due to the dearth of implantable bio-integrated multifunctional devices that allow safe, long-term deployment in anatomically and physiologically disparate organs of behaving animals. Traditionally, the fabrication of bio-integrated devices has relied on the use of resource-intensive lithographic techniques adapted from the semiconductor industry that require specialized cleanroom environments8,9,10,11,12,13,14,15. The thin film processing nature of lithography necessitates independent fabrication of individual modalities of the device stack, followed by careful manual assembly, thereby making this approach unsuitable for rapid customization16,17,18,19,20,21,22. Consequently, there remains a need for monolithic and scalable fabrication approaches that do not compromise the design flexibility, multimodality, functional sophistication and long-term biocompatibility of bioelectronic interfaces. Here we introduce a strategy that bridges this technological gap and demonstrate its potential in experiments spanning neural circuits in the brain and the gut.
We develop multifunctional bioelectronic interfaces based on polymer fibers embedded with solid-state microelectronic components using thermal drawing23 (Fig. 1a). We leverage the top-down nature of thermal drawing to produce, in a single step, several tens of meters of microscale fibers (~1,000 rodent-scale probes) that can host: (1) surface localized microscale light emitting diodes (µLEDs) for optogenetics; (2) microscale thermal sensors for precision thermometry; (3) microelectrodes for electrophysiology; and (4) microfluidic channels for drug and gene delivery (Fig. 1b). We demonstrate that mechanical properties of such fibers can be engineered to produce architectures compatible with implantation in the deep-brain and the tortuous, mobile gastrointestinal tract. Furthermore, we also develop a modular wireless control circuit, NeuroStack, to interface with the microelectronic fibers (Fig. 1c)24 that permits real-time, programmable light delivery across multiple independent channels and wireless data transfer for recording of local tissue temperature in untethered behaving mice.
Fig. 1: Schematic illustration of microelectronics-integrated multifunctional fibers that enable wireless modulation of brain and gut neural circuits.
a, High-throughput, monolithic fabrication of multifunctional polymer fibers using thermal drawing yields several tens of meters of continuous fibers (~1,000 rodent-scale probes) with tunable mechanics and solid-state microelectronic components. b, Such fibers can host multiple independently addressable µLEDs for optogenetics, microelectrodes for extracellular electrophysiology, microfluidics for gene/chemical payload delivery and thermal sensors for tissue thermometry—all in a miniature footprint. c, NeuroStack, a custom designed modular wireless control circuit, enables real-time programmable optical stimulation and data transfer for recording of tissue temperature. d, The multifunctional microelectronic fibers together with NeuroStack allow for wireless modulation of neural circuits in the deep-brain and in the small intestine of awake behaving mice.
We demonstrate that microelectronic fibers can be chronically implanted into the brain and the intestine of mice (Fig. 1d). The stiff, yet flexible, fibers designed for the brain can accurately target deep-brain nuclei such as the ventral tegmental area (VTA), where we deliver a viral vector carrying channelrhodopsin-2 (ChR2) to dopaminergic (DA) neurons through the integrated microfluidic channel. The colocalized electrodes and μLEDs on the same fiber permit longitudinal recording of spontaneous and optically evoked neural activity following delivery of genetic payload, while thermal sensors enable concomitant deep-brain thermometry. On the other hand, the soft and compliant gut fibers capable of delivering light and nutrients in the intestinal lumen allow direct modulation of gastrointestinal neural circuitry, including epithelial sensory cells in the proximal and distal small intestine and the upper-gut innervating vagal afferents. Moreover, multisite devices permit simultaneous implantation of multifunctional fibers in the gut and the brain, allowing us to probe central neural representation of postingestive nutrient sensing. By coupling the VTA implanted fibers to NeuroStack, we show that wireless programmable optogenetic stimulation of DA neurons elicits a reward behavior. Similarly, the soft gut fibers enable wireless intraluminal gut optogenetics targeting the sparsely distributed enteroendocrine/neuropod cells in the duodenum and ileum that modulate feeding behavior. Finally, we uncover that optogenetic stimulation of vagal afferents from the gut lumen produces a rewarding phenotype, thereby demonstrating direct modulation of central nervous system function from the intestine in behaving mice. We anticipate that these illustrative applications will foreshadow widespread use of wireless multifunctional microelectronic fibers to study brain–viscera and multi-organ neural communication pathways.
Results
Multifunctional microelectronic fibers for the brain
To produce multifunctional microelectronic fibers for interrogation of brain neural circuits, we designed a multilayer polycarbonate (PC) preform (glass transition temperature Tg = 150 °C, Young’s modulus E = 1.8–3.2 GPa) (Fig. 2a) that was thermally drawn into a functional fiber (Fig. 2b, Supplementary Fig. 1a–d and Supplementary Video 1), while simultaneously feeding spools of interconnect (silver-copper, Ag-Cu, 40-µm diameter) and recording electrode microwires (tungsten, 25-µm diameter) (Supplementary Note 1). The overall cross-sectional geometry of the preform (Fig. 2c) was conserved during the draw (Fig. 2d), yielding ~50 m of functional fiber (Fig. 2e) with dimensions of 370.7 ± 2.8 µm × 190.4 ± 3.4 µm (mean ± s.d., n = 5 sections; Supplementary Fig. 1e). The fibers were assembled into implantable probes (Supplementary Fig. 2a–f and Fig. 2f) by mounting blue (peak emission wavelength λ = 470 nm) and green (λ = 527 nm) µLEDs (InxGa1−xN, 270 × 210 × 50 µm3) along the fiber surface followed by deposition of a thin layer of parylene-C as a biofluid barrier coating. Optical micrographs of independently addressable µLEDs and microfluidic infusion in the final fiber device are shown in Fig. 2g–i. While the above approach shows integration of microelectronic components within the fibers postdraw, we also demonstrate embedding of semiconductor devices inside multifunctional fibers during fiber drawing, indicating further scalability of this platform (Supplementary Note 2 and Supplementary Fig. 3a–k)25.
Fig. 2: Fabrication and characterization of multifunctional brain fibers.
a, Preform layout and assembly of a multifunctional brain fiber that comprises interconnect channels in the central PC layer and a precursor to microfluidic channel and recording electrodes in the bottom PC layer. b, Schematic of the thermal drawing process with simultaneous feeding of interconnect microwires (40 µm Ag-Cu) and recording electrodes (25 µm tungsten). c, Photograph of the preform highlighting channels for interconnects, electrodes and microfluidics. d, An optical micrograph of a fiber cross-section showing conserved features. e, Several meters of as-drawn fiber depicting the scalability of fabrication (left) and a close-up view of an acutely bent flexible fiber (right). f, A fully assembled multifunctional fiber device with I/O pins for microelectrodes and µLEDs, access tubing for microfluidic channel and a ground wire. g,h, Independently addressable blue (g) and green (h) μLEDs at the distal end of the device. i, Simultaneous microfluidic fluid delivery and blue µLED under operation at the device tip. j, Electrochemical impedance spectrum of the tungsten microelectrodes in 1 × PBS (n = 3 independent samples). Inset shows variation of electrode impedance with bending deformation. k, Optical intensity output (blue trace) and efficiency (red trace) of μLED (λ = 470 nm) integrated within a fiber with varying input electrical power (n = 6 independent samples), where arrows indicate the y axes for respective plots. l, Steady-state calibration curve of the in-fiber thermal sensor at 32–42 °C (n = 4 independent samples). m, Variation of local temperature as recorded by the thermal sensor during operation of an adjacently placed blue μLED on the same fiber at different stimulation frequencies (35.2 mW mm−2, 10-ms pulse). n, Characterization of the fiber microfluidic channel showing output speed (red trace) and return rate (blue trace) at varying input injection speeds (n = 4 independent samples), where arrows indicate the y axes for respective plots. o, Mechanical finite element model compares the displacement of the probe tip for steel, silica, bare PC fiber and microelectronic PC fiber at displacements of the brain tissue between 10 and 100 µm. All shaded areas and error bars represent s.d.; data are presented as mean ± s.d. µFluidic, microfluidic.
Source data
Characterization of brain fibers
Incorporation of tungsten microwires in fibers afforded low-impedance microelectrodes (|Z| = 46.3 ± 6 kΩ at 1 kHz) for electrophysiology without compromising device flexibility (Fig. 2j and inset). The electrode impedance exhibited a negligible increase upon immersion in phosphate-buffered saline (PBS) over 7 weeks, and no leakage current was observed through the polymer cladding (Supplementary Fig. 4a,b). The light intensity from the integrated blue µLEDs was tunable over a range between 0.6 mW mm−2 and 70 mW mm−2 (Fig. 2k), which is sufficient for optogenetic modulation of behavior mediated by microbial rhodopsins such as ChR2 (refs. 26,27). The robust bonding of µLEDs was corroborated by the stable light output even at large bending angles up to 90°, while long-term immersion tests in PBS demonstrated functionality for at least 7 weeks (Supplementary Fig. 4c,d). We applied finite element modeling (FEM) to investigate how illumination intensity and volume varied with distance from the μLED at different input intensities (Supplementary Fig. 5a–d). We find that even a moderate intensity of 30 mW mm−2 covers a tissue volume of ~0.75 mm3 near the fiber surface, sufficient for optogenetic modulation of most brain nuclei in mice.
We leveraged the temperature-dependent current-voltage (I–V) characteristics of the diodes (InxGa1−xN µLED, λ = 470 nm) to record heat dissipation in the tissue during operation of the neighboring µLED28. A linear dependence of diode current on temperature defined the sensor calibration curve (Fig. 2l). Consistent with the thermal FEM (Supplementary Fig. 6a–d), the sensor detected a negligible temperature rise of 0.085 °C from a colocated µLED (λ = 470 nm, ~30 mW mm−2) operating at 40 Hz, which is well below the ~2 °C heating that occurs during clinical electrical deep-brain stimulation (Fig. 2m and Supplementary Fig. 7)29. To assess the microfluidic functionality, we measured the return rate of fluid infusion through the fiber (Fig. 2n) which was found to be in the range of 80–100% at physiologically relevant infusion speeds of 20–100 nl s−1 for intracranial injections. Since packaging of multiple functions in a neural probe can be at odds with achieving flexible device mechanics, we measured the bending stiffness of the fibers in the single-cantilever mode to mimic their anchoring to the skull. The fiber stiffness ranged between 25 and 33 N m−1, which is substantially lower than that of silica (132 N m−1, 0.4-mm diameter, 1-cm length) and stainless-steel (792 N m−1, 0.4-mm diameter, 1-cm length) probes of similar dimensions (Supplementary Fig. 8a). This was further corroborated through the mechanical FEM, which estimated a relative micromotion between the fiber tip and brain tissue that was ~2–4 orders of magnitude lower than silica and steel implants of similar dimensions (Fig. 2o and Supplementary Fig. 8b–f)30.
Microelectronic fibers for the gut
Unlike the brain, the gastrointestinal tract precludes implantation of rigid devices owing to a tortuous anatomy of the lumen which is encased in a delicate tissue, through which ingested food and fluids must pass31. Hence, we created multifunctional microelectronic fibers that are 10–15 times more compliant than the brain interfaces described above. These fibers permitted site-specific delivery of light and nutrients in the intestinal lumen of behaving mice. As their cladding (Fig. 3a), these fibers leveraged thermoplastic triblock elastomer poly(styrene-b-ethylene-co-butylene-b-styrene) (SEBS) (Tg = 140 °C, E = 3–5 MPa)32 which is ~103 times softer than the PC cladding used in microelectronic brain fibers. A flexible conductive polyethylene composite was employed for integration of metallic interconnects within the elastomer cladding, and a thin layer of PC was applied to maintain structural integrity at the drawing conditions. The preform was drawn (Fig. 3b) into ~50 m of continuous microscale fiber (535 × 315 µm2) with a largely conserved cross-sectional geometry (Fig. 3c–e), while simultaneously incorporating interconnect microwires through convergence drawing. The fully assembled gut fiber devices (Fig. 3f) were ~8.5 cm long and hosted six µLEDs at the distal end that could be operated as two independently addressable sets of blue and green µLEDs (three each, Fig. 3f inset) and a microfluidic outlet situated ~0.8–1 mm posterior to the first µLED pair.
Fig. 3: Fabrication and characterization of soft, multifunctional gut fiber.
a, Layout of the multilayered gut fiber preform, where SEBS layers for convergence and microfluidic channel are molded from an inverse aluminum mold followed by preform assembly. b, The resultant preform is drawn into several meters of soft fiber using convergence-based thermal drawing. c, Digital image showing the cross-section of an assembled preform. d, Cross-sectional micrograph of the gut fiber highlighting conserved features. e, Several meters of as-drawn fiber wrapped around a spooler demonstrating scalable fabrication. f, Digital image of a multifunctional gut fiber device. Inset shows three green (left) and blue μLEDs (middle) on the fiber and dual optofluidic modality (right), where infusion of a PBS bolus is accompanied by operation of a blue μLED. g, Cyclic buckling of the gut fiber over 104 cycles at 1-mm and 5-mm displacements and corresponding normalized light output from the fiber μLED (n = 3 independent samples). h, Light output from fiber μLEDs subjected to deformations at radii of curvature of 90° (n = 3 independent samples) and 180° angles (n = 3 independent samples). i, Bending stiffness of the fiber with SEBS cladding (n = 3 independent samples) and PC cladding (n = 3 independent samples) with identical cross-sections in comparison with a 400-µm silica waveguide (dashed). j, Mechanical FEM depicting stress distribution profiles in rigid silica (top), stiff PC (middle) and soft SEBS (bottom) fibers. k, Intraluminal implantation of a rigid silica fiber results in rupture of intestinal tissue (left), while soft gut fiber permits intestinal implantation that can negotiate the luminal curvature without tissue damage. l, FEM simulated elastic strain in Ag-Cu interconnects at varying radii of curvature. m, Corresponding spatial strain distribution in a gut fiber bent at radii of curvature of 5.6 cm (top) and 0.7 cm (bottom). n, Microfluidic return rate through the gut fiber at injection speeds between 1 and 5 µl s−1 relevant to intraluminal nutrient infusion in a straight and bent geometry (n = 3 independent samples). All shaded areas and error bars represent s.d.; data are presented as mean ± s.d. CPE, carbon doped polyethylene; r, radius.
Source data
Characterization of gut fibers
In the gut fibers, the surface-mounted µLEDs enable a laterally directed illumination profile which allows spatial targeting of epithelial cells and vagal afferents from within the lumen. This illumination profile is in contrast to an anatomically mismatched dorsal–ventrally oriented light cone of a typical silica fiber (Supplementary Fig. 9a). The cumulative light output from the three axially distributed µLEDs remained stable in PBS over at least 4 weeks (Supplementary Fig. 9b,c), while the optical output on the outer surface of the intestinal wall was only modestly attenuated by the presence of the intestinal tissue (Supplementary Fig. 9d). Using FEM, the optical penetration depth in the gut wall was found to be in the range 0.15–1 mm, and the illumination volume was estimated to be between 0.9 and 8.8 mm3 for input intensities ranging from 20 to 100 mW mm−2 (Supplementary Fig. 10a–f). This is sufficient to broadly illuminate the subepithelial mucosa layer 50–100 µm beneath the mucosal membrane that receives dense vagal innervation33. The corresponding temperature change in the gut wall was found to be negligible for input optical intensities between 20 and 100 mW mm−2, while the µLED separation (~1 cm) was sufficient to prevent any co-operative heat buildup (Supplementary Fig. 11a–e).
As the gut continuously undergoes peristaltic distortion, we performed cyclic buckling tests to assess the mechanical integrity of the fibers at deformations of 1 mm and 5 mm for up to 104 cycles, which had no impact on device performance (Fig. 3g). The fiber robustness was further supported by negligible changes in light output during 90° and 180° bending deformation at radii of 2–10 cm, which corresponds to strains exceeding those experienced by devices during surgical implantation process (Fig. 3h). We hypothesized that the mechanical compliance of the gut fiber would minimize the force exerted on the intestinal lumen, which is critical for chronic bio-integration in behaving animals. To mimic a surgically implanted gut fiber affixed to the abdominal wall at one end, their stiffness was evaluated in a single-cantilever bending mode. The gut fibers exhibited stiffness of 2–5 N m−1 across frequency ranges of heartbeat, respiration, locomotion and peristalsis (Fig. 3i), which is considerably lower than the stiffness of identical fibers composed entirely of stiffer plastic such as PC (70–75 N m−1) and similarly sized commercial silica fibers (400-µm diameter, 132 N m−1). FEM of stress distribution profiles also qualitatively captured these experimentally observed stiffness trends (Fig. 3j). Unsurprisingly, during an intraluminal implantation in a mouse small intestine, the rigid silica fiber punctured the mucosal membrane, and thus was unsuitable for in vivo use, whereas the soft gut fiber readily negotiated the lumen curvature without damaging the epithelial tissue (Fig. 3k). Since the surgical procedure requires bending of the gut fiber at acute radii, we simulated strain distribution in the copper interconnects which have the lowest yield strain among the fiber constituents and confirmed the strain to be below the elastic limit of 0.3% for radii>0.5 cm (Fig. 3l,m). Finally, fluid infusion through the microfluidic channel of gut fiber at injection speeds relevant to intestinal nutrient delivery (1–5 µl s−1)34 yielded a high return rate in the range of 60–90% under both straight and bent conditions (Fig. 3n).
Wireless operation of microelectronic fibers
Incorporation of microelectronics in polymer fibers provides an opportunity for wireless bidirectional operation that can facilitate behavioral assays in untethered subjects. To realize this, we engineered a miniature (15.5 mm), light-weight (1.1 g) modular platform24, NeuroStack, that enabled programmable wireless optical stimulation across two independent channels and data transfer for real-time temperature recording (Fig. 4a). The circuit features a Bluetooth Low Energy (BLE) communication protocol via a 2.4-GHz wireless link and an on-board miniature rechargeable battery for stable operation. This not only allows easy deployment in animal behavior studies with minimal user intervention, but also permits real-time programming of up to four devices across independent channels without any specialized equipment from a computer connected to an nRF52840 Development Kit that acts as a base station. The modular design of the circuit enables straightforward customization of experiments and allows integration of additional functions (Fig. 4b) within the same area footprint which is limited in small rodents. Similarly, the detachable architecture of the module obviates the need for the animals to carry subdermal electronics which are prone to misalignment and malfunction (see Supplementary Note 3 for additional discussion).
Fig. 4: NeuroStack, a custom designed modular wireless circuit for microelectronic fibers, allows programmable light delivery and physiological recording.
a,b, Schematic illustration (left) and digital images (right) highlighting important circuit components of a primary module (a) and intensity module (b). c–f, Images showing independent wireless control of blue (c) and green (d) µLEDs within a brain fiber and blue (e) and green (f) µLEDs in the gut fiber. g, Circuit layout of NeuroStack highlighting the power management block, temperature sensing block and intensity control block. h, Real-time control of optical stimulation frequency between 10 and 40 Hz. i, Real-time control of optical stimulation duty cycle between 20% and 80%. j, Optical pulse shaping through control of pulse rise and fall times between 5 and 15 ms. k, Intensity module permits real-time control of µLED brightness with corresponding photographs at 2.4-, 2.6- and 2.8-V bias. l, Calibration of the fiber thermal sensor using NeuroStack under steady-state conditions of 29–40 °C. m, Wireless recording of temperature transients with the thermal sensor in the microelectronic fiber and its comparison with a commercial thermocouple in wired mode. d, diameter; w, weight; CPU, central processing unit; DAC, digital-to-analog converter; GND, ground; GPIO, general-purpose input/output; LDO, low-dropout regulator; RF, radio frequency; SPI, serial peripheral interface; SWCLK, serial wire clock; SWD, serial-wire debug; SWDIO, serial-wire debug data input/output ; VDD, voltage drain-to-drain; ADC, analog-to-digital converter.
Source data
The primary module of NeuroStack carries an MDBT42V wireless microcontroller (with Nordic nRF52832 chip) and a rechargeable battery for orientation-independent power supply (Fig. 4a, Supplementary Fig. 12a–d and Supplementary Videos 2–4). During operation, the primary module relies on the microcontroller’s general-purpose input/output (GPIO) pins to drive current through the fiber μLEDs (Fig. 4c–f), while the header pins at the top support attachment of an intensity module (Fig. 4b). The intensity module can control the intensity of optical stimulation through a digital-to-analog converter which is programmed from the serial interface and allows transient shaping of optical stimulation pulses, a feature that is important for minimizing coupled electromagnetic artifacts in electrophysiological recordings during optical stimulation. The temperature sensing circuit uses one of the in-fiber μLEDs as a temperature sensor, whereas the other channel is assigned for optical stimulation via the second μLED. A constant voltage below the turn-on voltage is applied to the sensor diode using a low-dropout regulator to measure current variations. The sensed current is then amplified using a single-stage inverting differential amplifier. The internal analog-to-digital converter of the microcontroller then digitizes the amplified analog signal and transmits it wirelessly to the computer. The default sampling rate for temperature recording is set to 200 Hz and the temperature sensing function can also be turned off through the system software to save power when not needed. In addition to the main components, the module also hosts a 32-kHz crystal oscillator and inductors for power management. The device is programmed/debugged with a 6-pin J-link interface and controlled remotely through an intuitive software platform (see the Methods section for software details). The NeuroStack power breakdown (Supplementary Fig. 12e,f) scales with the µLED duty cycle and other functions including temperature sensing and intensity control. Figure 4c–f shows wireless operation of the independently addressable µLEDs within the brain and gut fibers connected to NeuroStack. The complete electrical layout of the primary module, together with temperature sensing and intensity control circuits, appears in Fig. 4g. We characterized the capability of NeuroStack to control the frequency, duty cycle, pulse shape and intensity of optical stimulation in real-time from the user interface. The creation of the stimulation pulse consists of two phases controlled by software timers. One of the timers manages the change in state between pulsing and resting period, thereby controlling frequency of optical stimulation (Fig. 4h), while the other timer sets the duty cycle during the pulsing period (Fig. 4i). The same principle is used for the intensity control to set the maximum intensity and rise/fall times of the pulses (Fig. 4j,k). The wireless temperature recording function was evaluated via steady-state measurements on a hotplate which showed linear dependence of the measured current on the surrounding temperature (Fig. 4l), while the dynamic response upon successive immersions in a temperature-controlled water bath matched closely to that of wired recordings from a commercial thermocouple (Fig. 4m).
Multimodal interrogation of midbrain DA neurons
The combined optical, electrical, fluidic and thermometry functions in the brain fibers enabled multiple experiments in mice for at least 2 months following implantation (Fig. 5a). As a validation study, we first targeted the DA neurons in the VTA, a key node of the reward and motivation pathways35. The fibers were stereotactically implanted in the VTA of DAT::Cre transgenic mice which express Cre-recombinase under the dopamine transporter (DAT) promoter. The integrated microfluidic channel within the fiber permitted delivery of an adeno-associated virus (adeno-associated virus serotype 5 (AAV5)) carrying ChR2 gene in a Cre-dependent construct (Ef1α::DIO-ChR2-mCherry) or a control construct (Ef1α::DIO-mCherry) to the VTA during a single-step surgery (Fig. 5b(i–iii))36, and robust expression of ChR2 was observed in the VTA sagittal sections (Fig. 5c). Emergence of electrophysiological potentials recorded through integrated microelectrodes in response to optical stimulation via fiber µLEDs revealed the time course of Cre-dependent opsin expression in DA neurons (Fig. 5d and Supplementary Fig. 13). As it is known that recording electrodes may exhibit optical stimulation artifacts (for example, Becquerel effect)37, we confirmed the physiological origins of optically evoked signals and devised an artifact mitigation strategy by transient pulse shaping (Supplementary Figs. 14a–i and 15a–h and Supplementary Note 4). Optically evoked multiunit neural activity was reliably recorded for at least 2 months in awake moving mice (Supplementary Fig. 16a–o). We hypothesized that the flexible fibers could enable stable recording of spontaneous single neuron activity over extended periods due to their reduced micromotion relative to brain tissue. After confirming the functional stability of implanted electrodes up to 6 months (Supplementary Fig. 17a,b), we recorded the spontaneous single-unit activity from putative VTA neurons in chronically implanted mice for 4 weeks (Fig. 5e and Supplementary Fig. 18). Additional examples of single neuron electrophysiology at weeks 2 (n = 3 mice) and 4 (n = 3 mice) appear in Supplementary Figs. 19a–i and 20a–i, respectively.
Fig. 5: Multimodal interrogation of deep-brain neural circuit and wireless programmable optogenetics during behavior.
a, Experimental timeline for in vivo validation of various fiber functionalities. b, Gene delivery through the microfluidic channel and fiber implantation in the same surgical procedure (i); photograph of one-step surgery highlights the fluid injection setup and the implanted fiber probe (ii); fully recovered animal ~1 month postsurgery carrying a wireless module (iii). c, Expression of Cre-dependent ChR2-mCherry construct 4 weeks following microfluidic AAV5 delivery into the VTA of DAT::Cre mice; (top) blue, DAPI; (middle) red, mCherry; (bottom) merge. d, Longitudinal electrophysiological recording of optically evoked neural activity in the VTA following AAV5 delivery. e, Spontaneous neural activity recorded from VTA at week 4 postimplantation (top); corresponding average action potential waveforms of two isolated neurons (bottom). f, Wireless intracranial temperature recording in mice (n = 6) exploring an open-field chamber during simultaneous wireless photostimulation (shaded region). g, Fiber thermal sensors detect brain hypothermia in the VTA induced by intraperitoneal injection of ketamine-xylazine mixture at 30-mg kg−1 (n = 3 mice) and 60-mg kg−1 (n = 3 mice) doses. h, Place preference assay during real-time wirelessly programmable optical stimulation in DAT-Cre mice. i,k,m, Percentage preference to the chamber coupled to the rewarding optical stimulation at baseline and on test day for mice transduced with ChR2-mCherry or mCherry in the VTA at three different wireless photostimulation conditions (top). i, On versus off. ChR2-mCherry: P = 4.16 × 10–5, t = –9.03, d.f. = 7; mCherry: P = 0.183, t = –1.50, d.f. = 6. k, Phasic versus tonic. ChR2-mCherry: P = 2.02 × 10–4, t = –9.65, d.f. = 5; mCherry: P = 0.403, t = –0.88, d.f. = 7. m, Blue-light versus green-light stimulation. ChR2-mCherry: P = 3.58 × 10–5, t = –9.24, d.f. = 7; mCherry: P = 0.55; t = –0.64, d.f. = 5 (**P
>>> Read full article>>>
Copyright for syndicated content belongs to the linked Source : Nature.com – https://www.nature.com/articles/s41587-023-01833-5
