In vivo Positron Emission Tomography to Reveal Activity Patterns Induced by Deep Brain Stimulation in Rats

Summary

We describe a preclinical experimental method to evaluate metabolic neuromodulation induced by acute deep brain stimulation with in vivo FDG-PET. This manuscript includes all experimental steps, from stereotaxic surgery to the application of the stimulation treatment and the acquisition, processing, and analysis of PET images.

Abstract

Deep brain stimulation (DBS) is an invasive neurosurgical technique based on the application of electrical pulses to brain structures involved in the patient's pathophysiology. Despite the long history of DBS, its mechanism of action and appropriate protocols remain unclear, highlighting the need for research aiming to solve these enigmas. In this sense, evaluating the in vivo effects of DBS using functional imaging techniques represents a powerful strategy to determine the impact of stimulation on brain dynamics. Here, an experimental protocol for preclinical models (Wistar rats), combined with a longitudinal study [¹⁸F]-fluorodeoxyclucose positron emission tomography (FDG-PET), to assess the acute consequences of DBS on brain metabolism is described. First, animals underwent stereotactic surgery for bilateral implantation of electrodes into the prefrontal cortex. A post-surgical computerized tomography (CT) scan of each animal was acquired to verify electrode placement. After one week of recovery, a first static FDG-PET of each operated animal without stimulation (D1) was acquired, and two days later (D2), a second FDG-PET was acquired while animals were stimulated. For that, the electrodes were connected to an isolated stimulator after administering FDG to the animals. Thus, animals were stimulated during the FDG uptake period (45 min), recording the acute effects of DBS on brain metabolism. Given the exploratory nature of this study, FDG-PET images were analyzed by a voxel-wise approach based on a paired T-test between D1 and D2 studies. Overall, the combination of DBS and imaging studies allows describing the neuromodulation consequences on neural networks, ultimately helping to unravel the conundrums surrounding DBS.

Introduction

The term neurostimulation encompasses a number of different techniques aimed at stimulating the nervous system with a therapeutic objective¹. Among them, deep brain stimulation (DBS) stands out as one of the most widespread neurostimulation strategies in clinical practice. DBS consists of the stimulation of deep brain nuclei with electrical pulses delivered by a neurostimulator, implanted directly into the patient's body, through electrodes placed into the brain target to be modulated by stereotactic surgery. The number of articles evaluating the feasibility of DBS application in different neurological and psychiatric disorders is continuously growing², although only some of them have been approved by the Food and Drug Association (FDA) (i.e., essential tremor, Parkinson's disease, dystonia, obsessive-compulsive disorder, and medically refractory epilepsy)³. Furthermore, a large number of brain targets and stimulation protocols are under research for DBS treatment of many more pathologies than officially approved, but none of them are considered definitive. These inconsistencies in DBS research and clinical procedures may in part be due to a lack of full understanding of its mechanism of action⁴. Therefore, huge efforts are being made to decipher the in vivo effects of DBS on brain dynamics, as every advance, however small, will help refine DBS protocols for greater therapeutic success.

In this context, molecular imaging techniques open a direct window to observe in vivo neuromodulatory effects of DBS. These approaches provide the opportunity not only to determine the impact of DBS while it is being applied but also to unravel the nature of its consequences, prevent undesired side effects and clinical improvement, and even adapt stimulation parameters to the patient's needs⁵. Among these methods, positron emission tomography (PET) using 2-deoxy-2-[¹⁸F]fluoro-D-glucose (FDG) is of particular interest because it provides specific and real-time information on the activation state of different brain regions⁶. Specifically, FDG-PET imaging provides an indirect evaluation of neural activation based on the physiological principle of metabolic coupling between neurons and glial cells⁶. In this sense, several clinical studies have reported DBS-modulated brain activity patterns using FDG-PET (see³ for review). Nevertheless, clinical studies easily incur several drawbacks when focusing on patients, such as heterogeneity or recruitment difficulties, which strongly limit their research potential⁶. This context leads researchers to use animal models of human conditions to evaluate biomedical approaches before their clinical translation or, if already applied in clinical practice, to explain the physiological origin of therapeutic benefits or side effects. Thus, despite the large distances between human pathology and the modeled condition in laboratory animals, these preclinical approaches are essential for a safe and effective transition into clinical practice.

This manuscript describes an experimental DBS protocol for murine models, combined with a longitudinal FDG-PET study, in order to assess the acute consequences of DBS on brain metabolism. The outcomes obtained with this protocol may help to unravel the intricate modulatory patterns induced on brain activity by DBS. Therefore, a suitable experimental strategy to examine in vivo the consequences of stimulation is provided, allowing clinicians to anticipate therapeutic effects under specific circumstances and then adapt stimulation parameters to the patient's needs.

Protocol

Experimental animal procedures were conducted according to the European Communities Council Directive 2010/63/EU, and approved by the Ethics Committee for Animal Experimentation of the Hospital Gregorio Marañón. A graphical summary of the experimental protocol is shown in Figure 1A.

1. Brain target localization by in vivo neuroimaging

1. Animal preparation
   NOTE: Male Wistar rats of ~300 g were used.
   1. Place the animal into an anesthesia induction box and seal the top.
   2. Turn on the sevoflurane vaporizer (5% for induction in 100% O₂). When the rat is anesthetized, switch the gas flow to the nosecone. Confirm the state of anesthesia by pinching the rat paw.
   3. Lay the animal supine on the CT bed, maintaining sevoflurane anesthesia (3% for maintenance in 100% O₂).
2. CT imaging
   NOTE: Selection of amperage, voltage, number of projections, number of shots, and voxel resolution depends on the CT scanner. Here, the following parameters: 340 mA, 40 KV, 360 projections, 8 shots, and 200 µm resolution were used^7,8,9.
   1. Secure the facemask or nose cone to the rat.
   2. Secure the rat body at the head, shoulders, hips, and tail with silk tape to provide enough restraint without damage.
   3. Monitor the rat continuously.
   4. Locate the head in the center of the field of view of the CT scanner.
   5. Proceed to acquire the CT image using acquisition parameters according to the specifications of the scanner.
   6. After 10 min, when the in vivo CT scan has been completed, stop the sevoflurane flow and place the rat into the MRI scanner.
3. MR imaging
   NOTE: Scan acquisition specifications vary among scanners, including different software systems and more importantly, the specific research question. Here, a 7-Tesla scanner was used. A T2-weighted spin-echo sequence ^7,8,9 with TE = 33 ms, TR = 3732 ms, and a slice thickness of 0.8 mm (34 slices), matrix size of 256 x 256 pixels with a FOV of 3.5 x 3.5 cm² was used.
   1. Lay the animal supine on the MRI bed, maintaining sevoflurane anesthesia (3% for maintenance in 100% O₂).
   2. Secure the head to a stereotactic frame placed on the scanner bed to avoid head movements during MRI acquisition. Also, secure the rest of the rat body with silk tape.
   3. Locate the head in the center of the field of view of the MRI scanner.
   4. Once the position is correct, proceed to acquire the MRI image.
   5. When in vivo MRI scanning is complete, stop the sevoflurane flow and place the rat into its cage.
   6. Locate a heating lamp near the cage because rats usually reduce their body temperature during the scan.
   7. Monitor the rat until recovery from anesthesia.
4. Atlas co-localization and target coordinates calculation
   1. Once CT and MRI images are acquired and reconstructed following the scanner's recommendations, co-register the CT and MRI images.
   2. Use an imaging processing software to spatially normalize CT and MRI using an automatic rigid registration algorithm based on mutual information¹⁰.
   3. Localize the Bregma line in the co-registered image, and measure the distance in the anterior/posterior (AP: +3.5 mm), midline/lateral (ML: +0.6 mm), and dorsoventral (DV: -3.4 mm) axis from Bregma to the target (i.e., medial prefrontal cortex, mPFC), according to the Paxinos and Watson rat brain atlas¹¹.
      NOTE: Coordinates from Bregma to the target may differ between rats when weight, size, sex, and breed are different.

2. Stereotaxic surgery

CAUTION: Autoclave all surgical material, implants, and stereotaxic units before use, and disinfect the surgical area to avoid infections and complications which may affect animal welfare. Use sterile surgical gloves and cover the animal with sticky drapes to prevent contamination.

1. Animal preparation and anesthesia
   1. Animals were administered 0.1 mg/kg buprenorphine intraperitoneally the day before the surgery. Place the animal into an anesthesia induction box chamber and seal the top.
   2. Turn on the sevoflurane vaporizer (5% for induction in 100% O₂).
   3. When the rat is recumbent, turn off the sevoflurane vaporizer and remove the rat from the box chamber.
   4. Intraperitoneally administer a mixture of ketamine (100 mg/kg) and xylazine (10 mg/kg) to anesthetize the animal.
   5. Wait until the animal is completely anesthetized. Check the level of anesthesia by pinching the interdigital area.
   6. Shave the area between the ears and the eyes.
2. Placement in the stereotactic frame and craniotomy
   1. Place the animal in the prone position on the stereotactic frame and use the head holding adaptor for rats to maintain the animal in the correct position during the surgery.
   2. Ensure immobility of the head by using the rat ear bars. Be careful with the insertion of the ear bars, as too deep an insertion may damage the eardrum.
   3. Apply ophthalmic lubricating gel to the eyes to prevent dryness during surgery, and cover them with sterile gauze.
   4. Cover the animal using sticky drapes to prevent contamination.
   5. Apply iodopovidone solution to the shaved area and clean it with sterile gauze.
   6. Apply mepivacaine in gel on the shaved area to anesthetize the local area.
   7. Make a longitudinal incision in the skin overlying the skull between the ears, extending 1.5-2 cm from lambda to Bregma (i.e., from the cranial vertex towards the eyes).
   8. Expose the skull with the help of 2 or 3 clamps. Remove the periosteum with a cotton bud and clean the blood with saline solution to expose Bregma and the sagittal sutures. Remove the excess saline solution with gauze.
   9. Scratch the skull surface with a scalpel to improve dental cement adhesion. Clean the area with a cotton bud soaked in hydrogen peroxide.
3. Electrode placement and fixation to the skull
   1. Straighten the electrodes with plastic tweezers to ensure the correct placement during the surgery.
      NOTE: Concentric bipolar platinum-iridium electrodes with the ground are used in this protocol.
   2. Locate one electrode on the holder of the right arm of the stereotactic frame.
      NOTE: It might be necessary to adapt the holder to the electrode to fix it better (see Figure 1B). Make sure that the electrode is parallel to the axis of the holder.
   3. Move the right arm holding the electrode through the stereotaxic frame and place the tip of the electrode exactly over Bregma. Try to bring the electrode tip as close as possible to the skull but without touching it to avoid deformation of the electrode, and note the resulting coordinates for Bregma provided by the stereotaxic frame. Make a mark on the skull indicating the initial position of the electrode with a surgical pen.
   4. Move the holder to the AP and ML coordinates obtained in step 1.4.3 and make a mark on the skull with a surgical pen indicating the position of the electrode target.
   5. Remove the right arm of the stereotactic frame holding the electrode. Be careful not to touch anything with the electrode.
   6. Use a small electric drill to make a hole through the skull (about 1-1.5 mm in diameter) in the target position until the dura is visible. Stop any bleeding using a cotton bud.
   7. Drill 4 holes along the skull to locate 4 screws (preferably stainless steel screws of 2-3 mm length) to increase the surface area of the dental cement and to locate the ground. Attach the 4 screws.
   8. Locate the right arm of the stereotactic frame with the right electrode. Move the arm to the calculated position, which should coincide with the hole. Then, lower the electrode until it touches the dura mater. This position will serve as 0 level in the DV direction.
   9. Insert the tip of the electrode in the DV direction, using the DV position in step 1.4.3. Clean the blood and cerebrospinal fluid around the area of the electrode with a cotton bud.
   10. Attach the ground to one of the screws closest to the electrode.
   11. Apply dental cement around the electrode and screws taking care to shape the dental cement avoiding sharp edges, which could injure the animal. Dental cement is applied in a layer to prevent overheating/thermal injury to the tissue/skull. Thick layers require more time to cure before additional layers are added. Ensure the dental cement is completely hardened before removing the electrode from the holder.
       CAUTION: The preparation of the dental cement produces the emanation of toxic vapors from the mixture, which finishes with the solidification of the cement. Therefore, wear a protection mask effective against chemical gases from this point and until the end of the surgery.
   12. Repeat the same procedure from steps 2.3.2-2.3.11 for the other hemisphere of the brain.
   13. Apply more dental cement to form a cap without covering the electrode. Wait until it hardens.
   14. Use braided natural silk non-absorbable suture 1/0, with a triangle needle, to suture in front and behind the cap. If required, remove the non-absorbable sutures at a particular time according to the body region where they are located. Use an iodopovidone solution to disinfect the surgical area.
   15. Remove the rat from the stereotactic frame.
4. CT imaging for electrode placement confirmation
   1. Perform steps 1.2.4-1.2.5 and see Figure 1C.
   2. Once the in vivo CT scan is complete, place the rat into its cage.
   3. Follow steps 1.3.6. and 1.3.7.
5. Postoperative care
   1. Administer antibiotic (ceftriaxone, 100 mg/kg, subcutaneous) for 5 days and analgesic (buprenorphine, 0.1 mg/kg, intraperitoneal) for 3 days as postoperative care. This antibiotic regimen may be prolonged for 5 days if any signs of infection (redness, swelling, and exudate) are observed around the cap.
   2. Perform a visual inspection of each animal daily, searching for signs of pain or distress, and clean the cap with iodopovidone solution.
   3. Provide intensive care for up to 1 week after surgery.

3. PET/CT imaging acquisition

NOTE: Each animal undergoes two PET/CT studies (i.e., in the absence and during DBS administration) under inhaled anesthesia to assess the acute effects induced by the electrical stimulation. Both scanning sessions follow the same imaging acquisition protocol, being performed 1 week after surgery (D1, without stimulation) and 2 days later (D2, during DBS).

1. Animal preparation and anesthesia
   1. Fast the rat for 8-12 h prior to each PET scan to allow higher brain uptake of FDG, improving the signal-to-noise ratio¹².
   2. Place the animal into an anesthesia induction box and seal the top.
   3. Turn on the sevoflurane vaporizer (5% for induction in 100% O₂).
   4. When the rat is anesthetized, switch the gas flux to the nosecone.
2. FDG injection and uptake period
   CAUTION: FDG is a radiotracer, so consider radioprotection measures to avoid radioactivity exposure. Confirm that the institution has all permission to work with radioactive compounds.
   1. Keep the FDG vial inside a lead-lined cabinet until used to avoid undesirable radioactivity exposure.
   2. Fill a small gauge syringe (~27G) with ~37 MBq of the FDG solution in the less possible volume, as measured in an activimeter.
   3. Place a heating pad under the animal's tail or use infra-red light to dilate the tail veins.
   4. Once the lateral veins are evident in the peak of the tail, clean the area with sanitary alcohol (96%).
   5. Inject the FDG solution through one of the lateral tail veins, approaching the vein with a syringe parallel to its trajectory and with the bevel of the needle facing upwards.
   6. Switch off the anesthesia and place the animal back in its cage to recover completely under a heating lamp.
   7. Allow 45 min of radiotracer uptake before starting the image acquisition session. During this period, keep the animal awake and inside a lead shielded chamber.
   8. In the case of the D2 study, deliver DBS as explained below in section 4 (Electrical stimulation administration) during the FDG uptake period.
3. PET acquisition and imaging reconstruction
   NOTE: PET image acquisition specifications depend on the scanner and the scan time. For this protocol, a static PET image was acquired for 45 min with a small-animal PET/CT scanner, using an energy window of 400-700 keV^7,8,9. Review the specifications of the PET/CT equipment before designing the acquisition protocol.
   1. 45 min after FDG injection, place the animal into an anesthesia induction box and seal the top.
   2. Turn on the sevoflurane vaporizer (5% for induction in 100% O₂).
   3. Transfer the animal to the PET/CT bed and lay it in a supine position, securing the nose to the anesthesia nose cone and maintaining sevoflurane anesthesia (3% for maintenance in 100% O₂). Confirm the state of anesthesia by pinching the rat paw.
   4. Repeat steps 1.2.2 and 1.2.3.
   5. Locate the head in the center of the field of view of the PET scanner.
   6. Acquire the static PET image using acquisition parameters according to the scanner's specifications.
   7. Proceed to reconstruct the image using a 2D-OSEM (ordered subset expectation maximization algorithm) and apply decay and dead time corrections^7,8,9.
   8. When the in vivo PET scan is complete, maintain the flow of sevoflurane to the rat in order to subsequently proceed to the CT acquisition without displacing the animal's head position on the scanner bed.
4. CT acquisition
   1. Without changing the animal's position with respect to the previous PET acquisition, proceed to acquire the CT image.
   2. Repeat steps 1.2.3-1.2.5.
   3. Once the in vivo CT scan is completed, stop the sevoflurane flow and place the rat into its respective cage for recovery.
   4. Follow steps 1.3.6. and 1.3.7.
   5. Maintain the animal into a lead-shielded chamber until complete radioactivity decay.

4. Electrical stimulation administration

NOTE: Electrical stimulation is delivered during the FDG uptake period in the D2 imaging session. For this protocol, the stimulation was delivered with an isolated stimulator, with a high-frequency (130 Hz) electrical stimulation in a constant current mode, 150 µA, and a pulse width of 100 µs^7,13,14.

1. DBS stimulator configuration
   1. Prepare the isolated stimulator and the required wires in a wide and quiet room, with enough space for the animal cages and minimal influence of potentially disturbing stimuli.
   2. Connect the stimulation wires to the swivels to allow animals to freely move within their cages and to the stimulator.
   3. Set the stimulation parameters according to the needs of the study.
   4. Use an oscilloscope to check the current mode, frequency, and pulse width. Confirm the biphasic waveform with a rectangular pulse shape (Figure 1D).
2. DBS delivery
   1. After the D1 imaging session and until the D2 acquisition, subject animals to a daily habituation protocol (45 min/day) to accustom them to the stimulation system and the operator's handling, avoiding undesirable stress responses in D2. Connect the stimulation system to each animal, but without turning on the stimulation.
   2. Once the stimulator has been set up, and the animal has been injected with FDG, connect the swivel to the electrodes and turn on the stimulator.
   3. After 45 min, turn off the stimulator, disconnect the animal from the swivel and quickly transfer it to an anesthesia induction chamber to begin step 3.3.

Figure 1: Experimental design. (A) Summary of the experimental steps followed in this protocol. (B) Representative pictures of a holder adaptation for better fixation of the electrode, with (left) and without (right) an electrode. (C) Fused image of an MRI with a CT of an operated animal, showing the correct electrode placement in the medial prefrontal cortex (mPFC). (D) Screenshot of the oscilloscope screen showing the biphasic stimulation waveform. Please click here to view a larger version of this figure.

5. PET image processing and analysis

NOTE: Follow the same image processing on images from D1 and D2 to obtain comparable data for subsequent voxel-wise statistical analysis.

1. Spatial registration of PET images
   1. Use specialized imaging processing software. The whole registration workflow is illustrated in Figure 2.
   2. Center and crop each PET and CT image to the field of view. Register the PET image to its CT using an automatic rigid registration algorithm based on mutual information¹⁵.
      NOTE: Rigid registration methods are only appropriate whether there are no significant differences in body weight or size between animals. Otherwise, consider using elastic methods.
   3. Register each CT image to a reference CT spatially registered to the Paxinos and Watson rat brain atlas¹¹ as in step 5.1.2. Save the resulting transformation parameters.
   4. Apply the transformation parameters obtained in step 5.1.3. to each registered PET image obtaining the PET image registered to the reference CT image.
   5. Save all the final PET images in Nifti format.
2. Intensity normalization and smoothing of PET images
   NOTE: Intensity normalization and smoothing are performed with different in-house scripts based on publicly available resources.
   1. Smooth the PET images with an isotropic Gaussian kernel of 2 mm of Full-Width Half Maximum (FWHM) to correct possible registration errors.
      NOTE: The size of the smoothing filter will depend on the resolution of the PET acquisition, but it is recommended to use a filter of 2-3 times the voxel size of FWHM.
   2. Normalize the intensity of the PET voxel values using an appropriate reference cluster normalization method¹⁶.
   3. Segment a brain mask from a reference MRI registered to the reference CT image.
   4. Apply the brain mask to each PET image to exclude voxels outside the brain from the voxel-wise analysis.
3. Voxel-wise analysis
   NOTE: The statistical analysis, consisting of a voxel-wise analysis of the PET image data, was performed using specialized imaging analysis software¹⁷.
   1. Compare D1 and D2 PET images using a paired T-test, setting adequate statistical significant thresholds.
   2. Consider as definitive results of the analysis only those clusters larger than 50 adjacent voxels to reduce type I errors.
   3. Represent the results in T-maps overlaid on a T2 MRI, showing the changes in glucose brain metabolism induced by DBS (cold colors for FDG reduction and warm colors for FDG increment).

Figure 2: Micro PET/CT imaging registration workflow. Detailed steps of PET image spatial normalization processing for subsequent voxel-wise analysis with Statistical Parametric Mapping (SPM) software. Please click here to view a larger version of this figure.

Representative Results

The animals were sacrificed using CO₂ at the end of the study or when the animal’s welfare was compromised. An example of a complete PET/CT study from an operated animal is shown in Figure 3. Thus, the electrode inserted into the rat brain can be clearly observed in the CT image shown in Figure 3A. This imaging modality provides good anatomical information and facilitates the registration of FDG-PET images, given that functional modalities tend to be blurrier than structural images (Figures 3A,B). In addition, a merged image of the FDG-PET and CT images of the same animal is shown in Figure 3C.

Figure 3: Micro PET/CT imaging of the rat brain with DBS electrodes implanted in the mPFC. (A) Sagittal section of a CT image. (B) Sagittal section of a FDG-PET image of the same animal as in A. (C) A fused PET/CT image resulted from overlaying A and B images spatially registered to the same stereotaxic space. Please click here to view a larger version of this figure.

The voxel-wise analysis performed with SPM12 software and provided here as an example consisted of a paired T-test between D1 (absence of DBS) and D2 (DBS during FDG uptake) studies, which actually belong to a previously published study⁸. Therefore, Figure 4 shows the brain metabolic differences between both PET sessions as T-maps superimposed on sequential 1 mm thick brain slices from an MRI registered to the reference CT image (CTref). These differences consisted of increases and decreases in FDG uptake showed as warm and cold colors, respectively. Also, a detailed summary of the statistical results obtained from the analysis is shown in Table 1. Here, we indicate the modulated brain region, the brain hemisphere in which the modulation is observed, the T statistic, the size of the cluster in the number of voxels (k), the direction of the modulation (i.e., hypermetabolic or hypometabolic changes), and the p-values obtained at peak and cluster levels. This type of table serves as a detailed description of the modulatory changes observed in the slice overlay figure.

Figure 4: Paired T-test results. T-maps resulting from the voxel-wise analysis overlaid on a T2 MRI registered to the same CTref, showing the metabolic changes induced by an acute DBS protocol (D2 vs. D1). The color bars at the bottom of the image represent T values corresponding to regional increases (warm colors) and decreases (cold colors) of FDG uptake (p < 0.005; k > 50 voxels). Abbrev.: AHiPM/AL - Amygdalohippocampal area posteromedial/anterolateral part, Au - Auditory Cortex, Bstm - Brainstem, Cpu - Caudate-putamen, HTh - Hypothalamus, L - Left hemisphere, PMCo - Posteromedial cortical amygdaloid nucleus, R - Right hemisphere, S1 - Primary somatosensory cortex. This figure has been modified with permission from Casquero-Veiga et al.⁸. Please click here to view a larger version of this figure.

D1 vs D2: Stimulation effect
ROI	Side	T	k	↓/↑	p unc. peak level	FWE	FWE
						peak level	cluster level
Bstm	R & L	18.39	1549	↓	<0.001	0.432	<0.001
AHiPM/AL-PMCo - HTh	L	10.39		↓	<0.001	0.949
CPu	L	37.56	738	↑	<0.001	0.025	<0.001
S1-Au		10.53		↑	<0.001	0.947
CPu-Pir	R	17.74	695	↑	<0.001	0.497	<0.001
S1-Au		10.45		↑	<0.001	0.948

Table 1: Changes in brain metabolism after acute DBS in mPFC. D1 vs D2: Stimulation effect. Structures: AHiPM/AL: Amygdalohippocampal area posteromedial/anterolateral part, Au: Auditory Cortex, Bstm: Brainstem, CPu: Caudate-putamen, HTh: Hypothalamus, Pir: Piriform cortex, PMCo: Posteromedial cortical amygdaloid nucleus, S1: Primary somatosensory cortex. ROI: Region of interest. Side: Right (R) and Left (L). T: t value, k: cluster size. Glucose metabolism: Increase (↑) and Decrease (↓). p: p-value, unc.: uncorrected, FWE: Family wise error correction. This table has been modified with permission from Casquero-Veiga et al.⁸.

Discussion

Given the advances in the understanding of brain function and the neural networks involved in the pathophysiology of neuropsychiatric disorders, more and more research is recognizing the potential of DBS in a wide range of neurologically-based pathologies². However, the mechanism of action of this therapy remains unclear. Several theories have attempted to explain the effects obtained in specific pathological and stimulation circumstances, but the heterogeneity of the proposed studies makes it very difficult to reach definitive conclusions⁴. Therefore, despite great efforts, there is no real consensus, but the number of patients undergoing DBS intervention continues growing¹⁸. Then, understanding the DBS consequences in the brain in vivo will allow to unravel which stimulation parameters and protocols of stimulation are more adequate to the needs of each patient, hence obtaining a better success rate. In this context, non-invasive functional neuroimaging modalities, such as FDG-PET, are essential to shed light on what is really occurring under the direct influence of electrical stimulation in the brain. For instance, in the longitudinal protocol explained here, DBS is delivered during the radiotracer uptake period of the D2 PET image. Thus, comparison of the D2 (DBS-ON) and D1 (DBS-OFF) PET studies enables the visualization of the brain regions that are being modulated by the electrical stimulation in vivo, as the "metabolic trapping" properties of FDG allow recording the cumulative changes that occur directly during the stimulation^13,19.

Altogether, this protocol describes a feasible strategy to evaluate the acute consequences of DBS in the brain in vivo, but the variety of DBS parameter combinations and protocols available is immense (e.g., continuous vs. intermittent treatments²⁰, high vs. low-frequency stimulation²¹), and even the effects of the DBS may differ along with the treatment due to inferring direct changes in the brain network under the stimulation influence²². Furthermore, the number of possibilities becomes even greater considering the increasing number of pathologies for which DBS is recommended²³. Therefore, longitudinal neuroimaging studies aiming to uncover the neural activation patterns that allow predicting the potential response to DBS treatment are of particular clinical relevance^24,25. In this regard, there is a wide number of clinical and preclinical studies which have evaluated the therapeutic effects of different DBS protocols by FDG-PET (see³ for review). Thus, there are several examples in which the studied DBS protocol counteracts the brain metabolic pattern associated with the pathology under treatment, inducing an improvement of the patient symptoms and proving the clinical usefulness of DBS-PET approaches. An example of this is found in the stimulation of the subcallosal cingulate region (SCC) for patients with treatment-resistant depression. SCC is metabolically hyperactive in unmedicated patients with depression²⁶, and this hyperactivation is normalized after remission of depression by pharmacological, psychotherapeutic, or DBS treatment^27,28,29. Of importance, SCC metabolism was higher in those patients who responded to DBS before starting the stimulation in comparison to non-responders. This study showed an 80% accuracy in the prediction of the response to SCC-DBS²⁹, highlighting the importance of imaging biomarkers in selecting potential patients for DBS. Therefore, the explained context reflects a history of clinical success of FDG-PET studies aiming to map the brain metabolic pattern of depression with the therapeutic outcomes obtained with SCC-DBS, which should set the basis for similar approaches focused on other neuropsychiatric disorders and DBS protocols in the future.

In this sense, in order to observe the physiological effects of DBS using FDG-PET, it is particularly relevant to carefully consider the specific timing of the DBS protocol to be scanned. Thus, despite applying the same DBS parameters and the same protocol, the timing for the image acquisition will clearly determine the origin of the observed modulation, which may lead to potential misunderstandings by not considering all the factors involved in the final response obtained⁸. Therefore, while the planning of surgery is determinant in laying the basis for the subsequent therapy, the design of an image acquisition protocol appropriate to the consequences of the stimulation under study is essential to fully understand the molecular mechanism underlying the stimulation treatment applied. Along these lines, several factors can drastically modify the response to a specific DBS protocol (e.g., stimulation parameters, electrode insertion, brain structure targeted, pathology under treatment, duration, and frequency of DBS sessions, etc.)^7,8,30. The phenomena reflected by the data collected in the FDG-PET study will depend on the specific time in the course of therapy at which the images are acquired. Then, all these points open up different research opportunities to explore DBS-induced modulation and contribute to explaining the mechanisms underlying this therapy.

Thus, despite the great differences that separate rodent and human brains, adequate practices should be implemented at all levels, with the aim of developing translational protocols. In this sense, it should not be ignored that DBS requires a highly invasive surgery based on a craniotomy so that electrodes can access deep brain structures³¹. At this point, there are two important sources of infection and inflammatory reaction: on the one hand, the direct exposure of brain tissue during surgery and, on the other hand, the insertion of two exogenous elements into an internal organ, creating an insertional scar by their trajectory towards the stimulation target³². Therefore, sterilization of the surgical equipment, maintaining a clean operating area, and adequate postoperative care based on antibiotic and analgesic treatments³³ are essential to ensure that the subject obtains the greatest benefit from the intervention and in the healthiest conditions. Furthermore, this is of particular relevance in FDG-PET imaging studies, as the occurrence of post-surgical complications can modify the pattern of radiotracer uptake given that inflammatory and infectious processes are clearly seen as hypermetabolic signals³⁴, which may lead to a modified response to treatment or an overestimation of the modulation produced by DBS.

However, this experimental methodology is subjected to some limitations: First, DBS protocols are usually long-term, continuous, and chronic treatments. Here, a neuroimaging protocol is shown to evaluate the acute effects of DBS in real-time. Thus, the timing suggested for neuroimaging studies would not be adequate to obtain information on DBS-induced long-term modulation in near real-time. Nevertheless, it may lay the groundwork for developing different longitudinal approaches to serve as basic knowledge for understanding DBS-derived responses. Secondly, since healthy animals have been used to illustrate this method, the application of the explained techniques to different pathological conditions may require their adaptation to ensure better results and optimal welfare conditions. Finally, voxel-wise analyses require large sample sizes and/or strong correction factors to obtain reliable results, as they are always affected by a problem of multiple statistical comparisons. Nevertheless, the assessment of the consequences of DBS on brain metabolism using FDG-PET with a voxel-wise approach is a great advantage due to the intrinsic exploratory nature of this method, which allows for extensive whole-brain analyses without the need for any prior assumptions.

Despite the explained drawbacks of combining DBS and FDG-PET, these approaches provide a large window of opportunity. Thus, obtaining brain metabolic information non-invasively is a great advantage in the sense that neurophysiological data can be collected from the subject during stimulation and on many different occasions along with the DBS treatment. Moreover, FDG-PET is a neuroimaging technique in the clinical setting, which reinforces the translational approach that motivates this method. Likewise, the use of FDG-PET is a particularly suitable alternative since, unlike other imaging modalities, the signal obtained is not influenced by secondary distortions in the electric or magnetic fields derived from the neurostimulation system, which may impair both image quality and system performance²⁴. On the other hand, research interest in evaluating the modulatory consequences of DBS is not limited to therapeutic benefits. In fact, since DBS is a focal, modulatory and non-permanent neurostimulation therapy, it may also help to unravel the neurofunctional activity pathways evaluated by molecular imaging techniques and in response to electrical stimuli provided by the syste³⁵. This information could be particularly valuable in deciphering unresolved neurophysiological enigmas in healthy and pathological conditions. Finally, the methodology explained in this manuscript provides the ability to observe the effects of DBS-induced neuromodulation in vivo, being a powerful strategy to determine the impact of stimulation during its application. In short, understanding the in vivo effect of DBS will help to understand the desired and undesired effects of this treatment, predict clinical improvement, and ultimately adapt the stimulation protocols to the needs of each patient.

Materials

Name	Company	Catalog Number	Comments
7-Tesla Biospec 70/20 scanner	Bruker, Germany	SN0021	MRI scanner for small animal imaging
Betadine	Meda Pharma S.L., Spain	644625.6	Iodine solution (iodopovidone)
Beurer IL 11	Beurer	SN87318	Infra-red light
Bipolar cable 50 cm w/50 cm mesh covering up to 100 cm	Plastics One, USA	305-305 (CM)
Bipolar cable TT2  50 cm up to 100 cm	Plastics One, USA	305-340/2	Bipolar cable TT2  50 cm up to 100 cm
Buprex	Schering-Plough, S.A	961425	Buprenorphine (analgesic)
Ceftriaxona Reig Jofré 1g IM	Laboratorio Reig Jofré S.A., Spain	624239.1	Ceftriaxone (antibiotic)
Commutator	Plastics One, USA	SL2+2C	4 Channel Commutator for DBS
Concentric bipolar platinum-iridium electrodes	Plastics One, USA	MS303/8-AIU/Spc	Electrodes for DBS
Driller	Bosh	T58704	Driller
FDG	Curium Pharma Spain S.A., Spain	-----	2-[18F]fluoro-2-deoxy-D-glucose (PET radiotracer)
Heating pad	DAGA, Spain	23115	Heating pad
Ketolar	Pfizer S.L., Spain	776211.9	Ketamine (anesthetic drug)
Lipolasic 2 mg/g	Bausch & Lomb S.A, Spain	65277	Ophthalmic lubricating gel
MatLab R2021a	The MathWorks, Inc		Support software for SPM12
MRIcro	McCausland Center for Brain Imaging,  University of South Carolina, USA	v2.1.58-0	Software for imaging preprocessing and analysis
Multimodality Workstation (MMWKS)	BiiG, Spain		Software for imaging processing and analysis
Omicrom VISION VET	RGB Medical Devices, Spain	731100 ReV B	Cardiorrespiratory monitor for small imaging
Prevex Cotton buds	Prevex, Finland	-----	Cotton buds
Sevorane	AbbVie Spain, S.L.U, Spain	673186.4	Sevoflurane (inhalatory anesthesia)
Small screws	Max Witte GmbH	1,2 x 2 DIN 84 A2	Small screws
Standard U-Frame Stereotaxic Instrument for Rat, 18° Ear Bar	Harvard Apparatus, USA	75-1801	Two-arms Stereotactic frame for rat
Statistical Parametric Mapping (SPM12)	The Wellcome Center for Human Neuroimaging, UCL Queen Square Institute of Neurology, UK	SPM12	Software for voxel-wise imaging analysis
STG1004	Multi Channel Systems GmbH, Germany	STG1004	Isolated stimulator
SuperArgus PET/CT scanner	Sedecal, Spain	S0026403	NanoPET/CT scanner for small animal imaging
Suture thread with needle, 1/º	Lorca Marín S.A., Spain	55325	Braided natural silk non-absorbable suture 1/0, with triangle needle
Technovit 4004 (powder and liquid)	Kulzer Technique, Germany	64708471; 64708474	Acrylic dental cement for craniotomy tap
Wistar rats (Rattus norvergicus)	Charles River, Spain	animal facility	Animal model used
Xylagesic	Laboratorios Karizoo, A.A, Spain	572599-4	Xylazine (anesthetic drug)
	Normon S.A., Spain	602910	Mepivacaine in gel for topical use