Visualizing Spatiotemporal Dynamics of Somatosensory Cortex Neurons During Electroacupuncture Using Two-Photon In Vivo Imaging

Summary

This protocol integrates cranial windows in Thy1-GCaMP6f transgenic mice, standardized electroacupuncture, and two-photon imaging to visualize millisecond-scale cortical ensemble dynamics during stimulation, revealing layer-specific plasticity.

Abstract

This protocol establishes a comprehensive experimental framework for investigating cortical circuit dynamics during EA (electroacupuncture) stimulation. The methodology integrates three key components: chronic cranial window implantation in Thy1-GCaMP6f transgenic mice, standardized EA stimulation at the ST36 acupoint using 2/100 Hz biphasic pulses, and high-resolution two-photon calcium imaging. This integrated approach enables real-time visualization of neural ensemble activity in the primary somatosensory (S1) cortex with millisecond temporal precision, capturing previously inaccessible details of network-level responses to neuromodulation. The technique successfully reveals distinct layer-specific activation patterns and longitudinal plasticity changes, providing critical insights into the cortical mechanisms underlying the effects of EA. By overcoming the fundamental spatiotemporal limitations inherent in conventional fMRI and electrophysiological techniques, this high-resolution platform offers unprecedented analytical capabilities for mapping dynamic neural circuits. The protocol's robust design and reproducible outcomes make it particularly valuable for optimizing targeted neuromodulation therapies and advancing our understanding of circuit-level plasticity in response to peripheral stimulation.

Introduction

Electroacupuncture, a modern adaptation of traditional acupuncture incorporating controlled electrical currents applied via needles at specific acupoints, is recognized for modulating neural circuits involved in pain processing, motor control, and autonomic function. Its effects are believed to involve complex signaling cascades initiated by peripheral sensory nerve activation (primarily Aδ and C fibers), leading to neurotransmitter release (e.g., opioids, serotonin, GABA) and subsequent modulation at spinal, brainstem, and cortical levels^1,2. Key structures implicated include somatosensory cortices. Within the S1 cortex, which processes tactile and nociceptive information, EA likely influences both local inhibitory-excitatory balances and long-range connectivity^3,4. However, the precise temporal sequence of activation/inhibition across different neuronal populations (e.g., pyramidal neurons, interneurons) and layers within S1 during and immediately after EA stimulation remains largely unmapped due to the methodological constraints mentioned previously^5,6. Understanding these real-time dynamics is crucial for optimizing EA parameters and predicting therapeutic outcomes.

Traditional methods for investigating the neuro-modulatory effects of EA (electroacupuncture) on cortical circuits, such as behavioral studies and post-mortem analyses, offer valuable insights but are unable to capture real-time neural activity with high spatiotemporal resolution^7,8. To overcome these limitations, in vivo two-photon calcium imaging emerges as a powerful alternative, enabling direct visualization of neural ensemble dynamics during stimulation^8,9. This technique provides unparalleled advantages for chronic longitudinal studies, allowing researchers to observe -- with cellular resolution -- the rapid and distributed activity across cortical layers in awake, behaving animals. Unlike acute electrophysiology, which offers high temporal resolution but limited spatial coverage, two-photon imaging facilitates large-scale monitoring of hundreds to thousands of neurons simultaneously within a defined field of view^9,10. Furthermore, compared to functional magnetic resonance imaging (fMRI), which excels in whole-brain coverage but operates at slower temporal scales and lacks cellular specificity, two-photon calcium imaging delivers fine-grained detection of activity from specific neuronal populations on a timescale relevant to synaptic and circuit-level processes^11,12. Thus, two-photon imaging is uniquely positioned to bridge a critical gap in our understanding of how peripheral EA stimuli are translated into spatiotemporally precise cortical activation patterns, particularly in regions such as the primary somatosensory cortex (S1). As the primary hub for somatosensory processing (including signals from acupoints), and given that EA robustly modulates its activity, S1 represents a strategic gateway for studying cortical circuit dynamics induced by electroacupuncture.

Two-photon laser scanning microscopy (2PLSM) coupled with genetically encoded calcium indicators (GECIs), such as GCaMP6, represents a revolutionary tool for neuroscience. It enables chronic, high-resolution imaging of neural activity in vivo with cellular and subcellular specificity¹³. GECIs fluoresce in proportion to intracellular calcium transients, serving as a reliable proxy for neuronal firing¹⁴. Crucially, 2PLSM's use of near-infrared excitation light provides superior depth penetration (hundreds of microns) compared to conventional microscopy, allowing visualization of activity in superficial and even deeper cortical layers (e.g., L2/3, L4, L5). Its inherent optical sectioning minimizes out-of-focus fluorescence and photodamage, making it ideal for longitudinal studies over weeks or months^15,16,17. This technology represents a revolutionary advance in EA research, with the potential to fundamentally transform our understanding of it, offering the potential to track the millisecond-scale dynamics of hundreds to thousands of individual neurons simultaneously within a defined cortical volume during stimulation.

Protocol

All experiments utilized 8-12-week-old transgenic Thy1-GCaMP6f mice¹⁸. All experimental procedures and animal care protocols were conducted in accordance with the National Institutes of Health (NIH) Guide for the Care and Use of Laboratory Animals and were approved by the Huazhong University of Science and Technology Animal Care and Use Committee (IACUC; Protocol Approval Number: 3603). Every effort was made to minimize the number of animals used and their suffering. A schematic of the experimental setup is shown in Figure 1.

1. Chronic cranial window implantation

1. Anesthesia induction and stereotaxic stabilization
   1. Initiate anesthesia by placing the 8-12-week-old Thy1-GCaMP6f mouse into an induction chamber supplied with 4-5% isoflurane (vol/vol) in 100% medical-grade oxygen at a flow rate of 1 L/min.
   2. Upon loss of righting reflex (typically within 2-3 min), transfer the animal to a thermostatically controlled heating pad (maintained at 37 °C) integrated into a digital stereotaxic frame.
   3. Secure the mouse using a nose cone delivering 1.5-2.0% isoflurane for surgical maintenance. Apply ophthalmic ointment (e.g., Puralube Vet Ointment) to both eyes to prevent corneal desiccation.
   4. Insert non-traumatic ear bars carefully into the external auditory meatus, ensuring symmetrical positioning without excessive pressure that could cause tissue damage or tympanic membrane rupture¹⁹.
   5. Secure the maxillary palate using the nose clamp, avoiding airway obstruction.
   6. Confirm deep anesthesia by absence of pedal withdrawal reflex to firm toe pinch and stable, slow respiratory rate (80-120 breaths/min). Continuously monitor core body temperature via a rectal probe and adjust the heating pad accordingly.
   7. Administer subcutaneous (s.c.) pre-emptive analgesia (e.g., Carprofen, 5 mg/kg) and fluid support (sterile saline, 0.5-1 mL s.c.). Verify stable physiological parameters (SpO₂ > 95%, respiration regular) throughout setup.
2. Surgical exposure of the skull
   1. Shave the dorsal scalp meticulously using electric clippers, followed by application of a chemical depilatory cream for complete hair removal. Rinse thoroughly with sterile saline.
   2. Perform sequential antisepsis: Scrub the exposed skin three times, alternating between povidone-iodine (10%) and 70% isopropyl alcohol swabs, finishing with a chlorhexidine gluconate (2%) solution.
   3. Using sterile micro-dissecting scissors and fine forceps (Dumont #5), make a midline sagittal incision (~15-20 mm in length) through the skin and underlying connective tissue, starting ~2 mm posterior to the eyes and extending to the interparietal-occipital junction.
   4. Use micro-hemostatic forceps or bipolar cautery (low setting) to control capillary bleeding along the incision margins. Gently retract the skin flaps laterally using curved micro-retractors or 5-0 silk sutures anchored to the surrounding drape.
   5. Expose the entire dorsal calvarium, including frontal, parietal, and interparietal bones. Using a micro-curette (#00), meticulously scrape away the periosteum - a fibrous vascular membrane adherent to the bone surface - under high magnification (surgical microscope).
   6. Apply gentle, tangential strokes to avoid gouging the skull or damaging sutures (bregma, lambda, sagittal, coronal). Achieve a smooth, avascular bone surface essential for precise craniotomy and subsequent adhesive bonding.
   7. Irrigate the field frequently with sterile, warm (37 °C) 0.9% saline to remove bone dust and maintain tissue hydration.
3. Stereotaxic targeting of the S1 hindlimb region
   1. Using a high-precision stereotaxic manipulator arm, position the tip of a fine surgical marker pen or micro-drill bit directly over the anatomical reference point bregma (the intersection of the coronal and sagittal sutures). Confirm coordinates under microscope visualization (10x).
   2. Adjust the manipulator to move the marker precisely anteroposterior (AP) -1.5 mm (posterior to bregma) and mediolateral (ML) +2.0 mm (right hemisphere) relative to the midline (sagittal suture).
      NOTE: This location corresponds to the primary somatosensory cortex (S1) hindlimb representation area (S1HL).
   3. Lightly etch a small crosshair (<0.5 mm) onto the exposed parietal bone surface at this coordinate, ensuring visibility without penetrating the bone.
   4. Verify the absence of major surface vasculature traversing the marked site; if significant vessels are present, adjust ML coordinates minimally (±0.1-0.2 mm) to avoid them, as dural vessels can bleed profusely if breached and impede optical clarity.
   5. Document the final adjusted coordinates.
4. Precise craniotomy execution
   1. Center a sterile, disposable 3-mm diameter biopsy punch precisely over the marked S1HL coordinate. Apply firm, steady downward pressure while rotating the punch gently to score the outer cortical bone layer completely around the circumference. Remove the punch.
   2. Using a high-speed micro-surgical drill (e.g., with 0.5-0.8 mm carbide burr) attached to a stereotaxic arm under constant 0.9% saline irrigation (cooling and debris removal), carefully thin the bone within the scored circle.
   3. Drill in short bursts using a light, tangential milling motion, progressing systematically from the periphery towards the center. Monitor bone translucency closely under magnification.
   4. Cease drilling immediately upon observing a subtle "blushing" appearance of the underlying dura mater, indicating minimal bone thickness (<50 µm). Switch to fine micro-forceps (Dumont #5/45) and microscissors to lift and peel away the remaining thin bone fragments meticulously, working radially outward from the center.
   5. Preserve the dura mater intact - ensure it appears translucent and glistening without tears or significant bleeding. Control any minor dural bleeding immediately by applying gentle pressure with a sterile compressed foam pledget soaked in 0.9% saline or thrombin.
   6. Irrigate the exposed dura copiously with warm, sterile artificial cerebrospinal fluid (aCSF) or 0.9% saline to remove all bone debris. Avoid desiccation^17,20.
5. Chronic cranial window implantation and head-bar fixation
   1. Prepare a sterile 3-mm diameter, #1 thickness (approx. 150 µm) circular glass coverslip. Optionally coat its inner surface with a thin layer of silicone elastomer for optical clarity and reduced adhesion.
   2. Position the coverslip precisely over the craniotomy using fine forceps. Prepare a fast-curing, light-cured dental cement (e.g., Metabond or Charisma) according to manufacturer instructions.
   3. Working rapidly (<90 s working time), apply a thin bead of cement around the perimeter of the coverslip, directly onto the dry, clean skull surface. Ensure the cement seals the coverslip edge completely but does not wick underneath onto the dura.
   4. Immediately cure the cement using a dental curing light (blue light, 450 nm) for 20-30 s per quadrant. Reinforce the seal with multiple layers of a slower-curing, 5 mL, self-cure acrylic dental cement (e.g., Jet Repair Acrylic or Paladur), building a robust, low-profile (<2 mm height) cement cap that extends several millimeters beyond the craniotomy edge onto the surrounding skull.
   5. Embed a lightweight, custom-made stainless steel or titanium head-bar (e.g., 15-20 mm long with central aperture) into the cement cap during its initial curing phase.
   6. Ensure the head-bar is positioned parallel to the sagittal suture and centered over the window, providing rigid, balanced fixation points for future head restraint under the microscope without obstructing optical access.
   7. Allow the cement to cure fully (≥15 min). Verify complete hemostasis and window clarity²¹.
6. Postoperative recovery and monitoring
   1. Following cement curing, carefully release the mouse from the ear bars and nose clamp. Discontinue isoflurane and administer 100% oxygen until spontaneous respiration resumes fully.
   2. Administer buprenorphine HCl (0.05-0.1 mg/kg s.c.) immediately post-op for analgesia. Transfer the animal to a pre-warmed (30-32 °C), oxygen-supplemented recovery chamber lined with clean paper bedding.
   3. Monitor continuously until sternal recumbency and full ambulation are regained (typically 30-60 min). Administer a second dose of carprofen (5 mg/kg s.c.) 24 h post-op. Provide warmed lactated Ringer's solution (1 mL s.c.) for hydration support if needed.
   4. House the mouse singly in a standard cage enriched with nesting material to prevent cage-mates from damaging the head implant. Maintain environmental conditions (22-24 °C, 40-60% humidity, 12:12 h light/dark cycle) with ad libitum access to food and hydrogel.
   5. Monitor the mouse at least twice daily for 7 days post-surgery for signs of pain/distress (hunched posture, piloerection, reduced activity, weight loss >15%), infection (purulent discharge, swelling, redness), neurological deficit, or head-bar loosening.
   6. Allow a minimum recovery period of ≥7 days before commencing any experimental procedures (imaging or EA) to ensure resolution of acute surgical inflammation, restoration of normal behavior and glymphatic clearance, and stabilization of baseline cortical activity. Weigh the animal daily to track recovery²².

2. Electroacupuncture stimulation

1. Precise anatomical localization of ST36 acupoint
   1. Target the ST36 ("Zusanli") acupoint based on established anatomical landmarks in the mouse hindlimb. Prior to needle insertion, palpate the proximal anterior tibia to identify the tibial tuberosity - the prominent bony prominence located just distal to the knee joint where the patellar ligament inserts.
   2. Using fine calipers or a sterile marker, measure 2 mm lateral (towards the fibula) and 2 mm caudal (posterior) relative to the center of the tibial tuberosity. This location corresponds to the tibialis anterior muscle belly, superficial to the deep peroneal nerve branch and proximal to the bifurcation of the anterior tibial artery.
   3. Confirm the absence of major superficial blood vessels at the site via transillumination if necessary.
      NOTE: Accuracy is critical, as slight deviations (>0.5 mm) can significantly alter the neural structures stimulated and subsequent cortical responses in S1. Landmark identification should be performed consistently by trained personnel to minimize inter-operator variability^23,24.
      NOTE: Methods for ensuring needle stability are: (i) Following needle insertion and fixation, an adaptation period is allowed for the mouse to acclimate and reduce initial struggling. (ii) The vibration frequency generated by the electroacupuncture stimulator (e.g., 2 Hz) is inherently low in amplitude. Furthermore, the fine needles used (e.g., Φ0.16 mm) possess sufficient rigidity to effectively resist this minimal vibration. (iii) Throughout the experiment, the operator continuously monitors the needle tip position. Trial data will be excluded if any significant displacement caused by animal movement is observed.
2. Sterile needle insertion technique
   1. Following skin disinfection (alternating 70% ethanol and chlorhexidine scrub), insert a sterile, single-use stainless steel acupuncture needle (nominal diameter: 0.18 mm, equivalent to 36 gauge; length: 13 mm) perpendicularly into the skin at the marked ST36 coordinate.
   2. Advance the needle steadily through the skin, subcutaneous tissue, and into the underlying tibialis anterior muscle using a smooth rotary motion to minimize tissue trauma.
      NOTE: The target insertion depth is 3 mm from the skin surface. This depth ensures optimal engagement of type II/III muscle afferents (Aδ fibers) and potentially somatic small-diameter fibers without risking contact with the underlying tibia or deep neurovascular bundles.
   3. Secure the needle hub firmly using a small sterile adhesive pad or custom clip to prevent dislodgement during subsequent electrical stimulation and potential subtle limb movements.
   4. Verify stable needle placement by observing a slight, transient local muscle fasciculation upon insertion - a physiological indicator of proper intramuscular positioning.
   5. Maintain strict aseptic technique throughout to prevent infection, particularly in chronic stimulation paradigms²³.
3. Electroacupuncture stimulation parameterization and delivery
   1. Connect the insulated handle of the inserted needle to the anode (+) output channel of a programmable electroacupuncture stimulator (e.g., HANS Acupoint Nerve Stimulator, model LH202).
   2. Position a reference cathode (-) electrode, typically a small stainless steel plate or clip electrode coated with conductive gel, on the ipsilateral hind paw (e.g., plantar surface near heel).
   3. Ensure secure contact and low impedance (<5 kΩ) at both sites. Program the stimulator to deliver biphasic square-wave pulses (balanced positive/negative phases) to minimize electrode polarization and tissue damage.
   4. Set the stimulation paradigm to dense-disperse mode: alternating bursts of 2 Hz (low frequency) for 3 s followed by 100 Hz (high frequency) for 3 s.
      NOTE: This combined frequency pattern (often denoted as 2/100 Hz) is empirically demonstrated to enhance endogenous opioid (e.g., β-endorphin, enkephalin) and monoamine (e.g., serotonin, norepinephrine) release in the CNS more effectively than single frequencies alone.
   5. Adjust the current intensity incrementally starting at 0.1 mA until a mild, rhythmic paw twitch (flexion of digits or ankle) is visually observed - typically reached between 1.0 mA in adult mice under light anesthesia. This twitch threshold indicates sufficient recruitment of motor efferents and associated large-diameter afferents (Aβ fibers), confirming physiological activation without causing distress or vigorous struggling. Maintain this calibrated intensity throughout the 20 min stimulation session^25,26.
4. Critical procedural controls and physiological monitoring
   1. Stimulation artifact mitigation
      1. Physically separate EA cables from imaging equipment/headstage.
      2. Implement robust grounding: Connect the stimulator's ground terminal to a dedicated earth ground and ensure the microscope frame and animal heating pad share a common ground point.
      3. Temporally synchronize stimulation pulses with the microscope's blanking signal if available, or employ post-hoc signal processing to remove residual electrical artifacts from calcium traces.
   2. Physiological stability
      1. Continuously monitor core body temperature (maintained at 37.0 ± 0.5 °C via feedback-controlled heating pad), respiration rate (80-120 breaths/min), and SpO₂ (>95%) throughout the 20-min EA session.
      2. Maintain light isoflurane anesthesia (1.0-1.5%) for immobility during combined imaging/stimulation. Continuously observe the induced paw twitch; if it diminishes or ceases, check needle contact and impedance, and adjust intensity slightly (≤0.1 mA increments) to maintain consistent, mild motor activation.
      3. Discontinue stimulation immediately if signs of distress (e.g., tachypnea, vigorous struggling) occur.
   3. Reproducibility
      1. Use identical needle type, insertion depth, stimulator model, and calibration method (paw twitch threshold) across all animals and sessions. Document the final stimulation intensity used for each subject.
      2. Allow a minimum 5-min post-stimulation recovery period before terminating anesthesia or performing additional procedures^27,28.

3. Two-photon imaging during EA

1. Animal preparation, head fixation, and physiological maintenance
   1. Following ≥7 days of post-surgical recovery, head-fix the Thy1-GCaMP6f mouse using the implanted head-bar within a custom-designed, vibration-damped fixation apparatus mounted on a motorized XY stage.
   2. Position the animal comfortably on a low-friction, air-lubricated cylindrical treadmill, allowing unrestrained hindlimb movement. Maintain light surgical anesthesia via nosecone delivery of 0.5-0.8% isoflurane in 100% O₂. This reduced isoflurane level minimizes suppression of neural activity while ensuring immobility sufficient for high-resolution imaging.
   3. Continuously monitor and record core body temperature (maintained at 37.0 ± 0.2 °C via feedback-controlled heating pad), respiration rate (100-140 breaths/min), and peripheral oxygen saturation (SpO₂ > 95%).
   4. Apply ophthalmic gel to prevent corneal drying during prolonged imaging. Ensure the hindlimb ipsilateral to the S1HL cranial window is accessible for EA needle insertion at ST36.
   5. Allow 15-20 min for physiological stabilization after head fixation before initiating imaging²⁹.
2. Optical targeting and optimization for layer II/III imaging
   1. Precisely determine the focal plane to acquire high-quality neuronal signals. Follow this verification procedure.
      1. Reduce the laser power to a low level (typically 10-20% of the imaging power, e.g., 5-10 mW) to avoid unnecessary photobleaching and photodamage during plane location.
      2. Using the Z-axis drive of the two-photon microscope, move the objective slowly in micron steps. Begin at the cranial window surface and advance deeper into the brain tissue.
      3. In the live preview mode, identify the focal plane using these visual indicators as described in steps 3.2.1.4 and 3.2.1.5.
      4. For the correct focal plane on a neuronal soma layer: Confirm that the image shows clear, bright, punctate structures (neuronal somata) appearing round or oval with good background contrast. Ensure cell boundaries are sharp and exhibit a distinct "salt-and-pepper" appearance.
      5. For an incorrect focal plane: If the image appears blurry or homogeneous without well-defined punctate structures, recognize that the focus is in a region devoid of somata. If large, dark, unstructured vascular shadows are visible, identify this as the blood vessel layer.
   2. Verification of the appropriate fluorescence baseline
      1. Confirm a healthy and stable fluorescence baseline before applying any experimental stimulus to ensure valid interpretation of subsequent calcium signals (ΔF/F). Perform steps 3.2.2.2-3.2.2.4.
      2. Adjust the laser power to the formal imaging level (e.g., 30-50 mW).
      3. Acquire images for a few seconds in a fast, low-resolution preview mode.
      4. Observe the real-time displayed mean fluorescence intensity (or directly monitor the image brightness). Verify that the signal remains stable and shows no signs of rapid photobleaching before proceeding to experimental data acquisition.
   3. Center the imaging field of view within the cranial window.
      1. Apply ultrasonic gel or distilled water between the objective and coverslip for immersion coupling. Using infrared Dodt gradient contrast or epifluorescence illumination, identify surface vasculature and align the field of view (FOV) to avoid large surface vessels (>20 µm diameter).
      2. Switch to two-photon excitation mode using a tunable femtosecond-pulsed Ti:Sapphire laser. Set the excitation wavelength to 920 nm - optimal for exciting GCaMP6f (peak ~920 nm) while minimizing green autofluorescence and photodamage compared to shorter wavelengths.
      3. Gradually increase laser power (typically 20-50 mW measured at the sample) while focusing through the transparent dura to the cortical surface (pia mater). Adjust the Z-position of the objective using a piezoelectric nano-focus device to reach a depth of 200 ± 10 µm below the pial surface, corresponding to cortical Layer II/III.
      4. Confirm depth via characteristic neuronal density and morphology: a dense population of small-to-medium pyramidal neuron somata (10-15 µm diameter) with visible apical dendrites.
      5. Optimize laser power and photomultiplier tube (PMT) gain settings to achieve strong baseline GCaMP6f fluorescence (F₀) without saturation (pixel intensity ~50-70% of detector maximum) and minimal background^30,31.
         NOTE: During imaging, the average power of the 920 nm excitation light was precisely maintained at 30-50 mW at the objective. This power level was validated through pilot experiments on animal tissue subjected to identical surgical procedures, ensuring effective signal acquisition for scan durations of up to 25 min while minimizing photodamage and fluorescence photobleaching.
3. High-speed calcium imaging acquisition protocol
   1. Configure the resonant scanner for high-speed frame acquisition:
      Frame size: 512 × 512 pixels (typical FOV: ~450 × 450 µm, pixel size ~0.88 µm)
      Frame rate: 7 frames per second (fps) achieved via bidirectional scanning (~14 kHz line rate). This rate adequately samples GCaMP6f decay kinetics (τ~500 ms) while balancing signal-to-noise ratio (SNR) and minimizing photobleaching.
      ​Pixel dwell time: ~0.8 µs (optimized for SNR at 7 Hz).
   2. Initiate continuous time-lapse acquisition using acquisition software (e.g., ScanImage, Prairie View). Record a 5-min baseline period prior to EA stimulation onset to establish pre-stimulus neural activity dynamics (F₀).
   3. At precisely t = 5 min, commence EA stimulation at ST36 while imaging continues uninterrupted. Acquire data continuously throughout the 20-min EA stimulation period. Total scan duration = 25 min (1500 s, ~10,500 frames).
   4. Save data as a multi-TIFF stack or proprietary binary format with embedded timestamps synchronized to EA trigger pulses via a TTL input channel.
   5. Monitor laser power stability and focus drift (<2 µm) in real-time; pause briefly for refocusing only if significant drift occurs (e.g., >5 µm), noting the timepoint^32,33.
4. Rigorous motion correction and spatial registration
   1. Post-acquisition, process the raw imaging stack using NoRMCorre (Non-Rigid Motion Correction) in MATLAB to eliminate spatial artifacts from residual micromovements (e.g., breathing, pulse).
   2. First, split the stack into overlapping spatial patches (e.g., 64 × 64 pixels, 50% overlap). For each patch, compute a reference template by averaging high-SNR frames from the baseline period.
   3. Using Fourier-based cross-correlation, calculate X-Y translational shifts for every frame relative to the template. Apply these shifts via subpixel interpolation.
   4. Next, address non-rigid deformations: model local distortions within patches using piecewise affine transformations (translation + shearing). Iteratively update the template to improve registration accuracy. Key parameters:
      max_shift: 20 pixels (limit for translation)
      max_dev: 3 pixels (max non-rigid deviation)
      ​bin_width: 50 frames (temporal binning for template updates)
   5. Visualize correction efficacy by inspecting the frame-to-frame correlation plot (should stabilize near 1.0 post-correction) and generating a mean image projection difference map (pre- vs. post-correction). Save the motion-corrected stack for downstream analysis. This step is critical for accurate pixel-wise analysis of calcium transients, especially during EA-induced subtle limb movements.

4. Spatiotemporal dynamics analysis

The motion-corrected imaging stack undergoes automated neuronal segmentation and signal extraction using the Suite2p pipeline (v0.12.3) in Python. Key processing stages include:

1. Spatial footprint identification
   1. Apply constrained non-negative matrix factorization (CNMF) to decompose the stack into spatial components.
   2. Initialize regions of interest (ROIs) using a correlation-based approach: merge pixels with spatial cross-correlation > 0.7 and local SNR > 4 into candidate ROIs. Distinguish somatic ROIs from dendrites/axons by size (8-15 µm diameter) and circularity (compactness threshold >0.6).
2. Neuropil contamination correction
   1. For each ROI, define a concentric neuropil annulus (inner radius: 1.5× ROI radius, outer radius: 2.5× ROI radius).
   2. Calculate neuropil coefficient (r_neuropil) via least-squares regression:
      F_corrected(t) = F_raw(t) - 0.7 × F_neuropil(t)
      (Coefficient empirically validated for Thy1-GCaMP6f mice).
3. Signal deconvolution
   1. Infer spiking activity using OASIS (ℓ₁-regularized deconvolution) with AR(2) noise model to estimate spike probability S(t) from ΔF/F traces:
      ΔF/F(t) = k ⋅ S(t) + b + ε
      where 'k' is the calcium transient kernel and ε is Gaussian noise.
4. Quality control
   1. Exclude ROIs with SNR <3, spatial overlap >30%, or neuropil correlation >0.3. Visual verification is performed via Suite2p GUI³⁴.
      Specific operations and interface elements are provided in Supplementary File 1. The motion-corrected TIFF file generated above was directly fed into the Suite2p pipeline. While Suite2p's integrated environment includes its own rigid motion correction, this redundant internal step was disabled since a more precise non-rigid correction was already applied. Automatically extracted neuronal ROIs can be found in Supplementary File 2.

Results

Structural characterization of S1 cortex and baseline neural activity
We began by establishing a high-fidelity structural and functional baseline for our imaging experiments. Tissue morphology analysis revealed discernible structural differences between non-capplied (Figure 2A) and coverslip-applied (Figure 2B) cranial window preparations. While the application of a coverslip introduced minor compression artifacts (arrowheads,

Discussion

To bridge the critical gap in observing electroacupuncture's (EA) cortical effects in real-time, we present a comprehensive integrated protocol combining chronic cranial window implantation for optical access, standardized EA stimulation, and longitudinal two-photon calcium imaging targeted at the primary somatosensory (S1) cortex.

This integrated protocol offers several distinct advantages over traditional methods for studying EA's cortical effects. First, it provides unprecedented spatiotemp...

Materials

List of materials used in this article

Name	Company	Catalog Number	Comments
Carprofen	MCE	N/A	5 mg/kg for analgesia
Digital stereotaxic frame	RWD	Kopf Model 1900
Electroacupuncture stimulator	HANS LH202 (2/100 Hz biphasic pulses)	N/A	HANS LH202 (2/100 Hz biphasic pulses)
Isoflurane	RWD	R510-22-10	Medical grade, 100% oxygen carrier
Micro-drill system	RWD		0.5-mm carbide burr, saline-cooled
Povidone-iodine	MCE	HY-B2234	10% solution
ScanImage	ScanImage	https://scanimage.org/
Sterile saline	Beyuntian	ST341-500ml	0.9% NaCl
Surgical microscope	Zeiss	https://www.zeiss.com/meditec/en/products/surgical-microscopes.html	10-40× magnification
Thy1-GCaMP6f transgenic mice	Cygen	5523978	8-12 weeks old, male/female
Two-photon microscope	Nikon A1MP	Nikon A1MP	Ti:Sapphire laser (920 nm excitation)