All surgical procedures were approved by the National Animal Studies Committee of France (authorization no. APAFIS20359-2019041816357133v5). We used 27 OF1 (Charles Rivers, Les Oncins, France) adult mice 8–10 weeks old, 30–40 g in weight. We used also 6 mice with Thy1-cyan fluorescent protein (CFP)-23, LysM-green fluorescent protein (GFP), CD11c-yellow fluorescent protein (YFP). Mice were deeply anesthetized with ketamine–xylazine (120 mg/kg, 12 mg/kg, respectively) in saline with (IP) intra-peritoneal injection and supplemented when need it (60 mg/kg, 6 mg/kg). Ocry-gel (Laboratoire TVM, Lempdes, France) was applied to maintain eye moisture and surgical anesthesia was verified by a lack of toe pinch reflex. The hair around the surgical site was shaved, and the skin was cleaned with iodine solution (Vétoquinol). The skin of the skull was incised, and the fascia was gently removed using a scalpel.

For the placement electrode on the cortex, a U-shaped clip were placed on the skull of the left side and glued with primer dental cement (GACD, France) followed by blockout resin photopolymerized by an optic UV lamp (GACD, France). This clip served to fix mice during surgery procedure and image acquisition under the microscope, to a holder. A craniotomy was carried out on the right hemisphere of the head using a high-speed drill with a carbide bur (size = round 1/4; World Precision Instruments, Hitchin, England). The brain was kept free of debris by repeated rinsing with phosphate buffered saline (PBS) during implant and surgery.

Two 50 μm platinum-iridium microelectrodes (Phymep) were placed on the dura matter of the brain and fixed on the skull by primer dental cement followed by the photopolymerized block resin. Prior to the surgery, a glass window of 5 mm diameter and 0.17 mm thickness (Thermo Fisher Scientific, France) was cleaned and dried. Once the electrodes were positioned, the glass window was placed over the skull and electrodes. Gentle pressure was applied to maintain the glass in place while dental cement was applied to the edges of the glass and the surrounding bone and muscle. Finally, an additional layer of dental cement was applied over the primer to reinforce the structure.

For the deep brain ablation, groups of OF1 underwent a surgical procedure as described above. However, instead of placing two wires on the cortex, an implantable electrode was placed into the brain in order to reach the hippocampus. For that, a craniotomy was performed by drilling a hole for the electrode (PT electrode twisted 5 mm, Bilaney Consultants, Germany) (Coordinates: ML: −2.04, AP: +2.07, DV: 1.29 determined by the Paxinos atlas) and another one for the reference in the cerebellum. At the end, electrodes were cemented with Dentalon (PHYMEP, France).

The electroporator device (Electro cell B10-CNRS) is connected through the two wires placed on the brain of the mice. The delivered voltage values depend on the space between the two wires, but it was always equal to 1,750 V/cm. The width of the pulse for both the positive and negative phase was 50 μs the frequency was 10 Hz with 10 repetitions. All TIRE and NTIRE stimulations were performed with the electroporator device as it is possible to configure parameters of stimulation (Supplementary Figure 1).

For the microscopy, animals were injected with 50 μl of Dextran, Rhodamine B, 10,000 MW, Neutral (Thermo Fisher Scientific) (3 mg/ml) in saline was injected in the ophthalmic venous sinus using the retro-orbital method. During imaging, the animals were freely breathing, and the microscope chamber was heated to ∼32°C. We used a tunable femtosecond pulsed laser (Mai-Tai, Spectra Physics, Évry, France) coupled to a Zeiss two-photon microscope (LSM 780) equipped with a Plan-Apochromat 20× water immersion objective lens (NA = 1.0) and five non-descanned detectors (NDD). For each image stack laser, intensity and sometimes the gain, were adjusted according to imaging depth in order to maximize image intensity while minimizing saturation throughout the image stack. We always imaged the region between the two wires using the following strategy. The window was epi-illuminated with LED lights on the side of the objective. To determine the vasoconstriction percentage, blood vessels where divided into two groups: large blood vessel ≥20 μm and small blood vessel <20 μm. Images were analyzed using the ZEN 2.6 software, the diameter of blood vessels before and after stimulation was taken, and the percentage of vasoconstriction was calculated.

For depth NTIRE and TIRE ablation of epileptic foci, mice with implantable hippocampal -CA3- electrode (Twisted Pair, Plastic One^®), after 2 days post-surgery, were connected to a data acquisition system (Intan Technologies, United States) in order to record and stimulate. After finding the after discharge (AD) threshold for each mouse (the minimum current applied to evoke a seizure) the following parameters were applied: 50 Hz, 500 μ s pulse width, train of 500 stimulation pulses, with current set at the AD threshold (see Supplementary Figure 2 for more details). This stimulation was repeated 10 times every 10 min (Lothman et al., 1985; Musto et al., 2009). After that, NTIRE or TIRE ablation was performed with the electroporator device (under anesthesia) and epileptiform events were recorded within the following 12 h.

To detect specific events as spikes, we performed a semi-automatic detection on the signal. The detection part was processed by using Delphos software, a detector of spikes and oscillations used mainly in clinical research (Roehri et al., 2018). Before to analyzed the signal, it has been downsample to 3,750 Hz and filtered in between 20 and 40 Hz. All the spikes detection were reviewed by a single person (EA) via AnyWave software, a visualization software for electrophysiological data (Colombet et al., 2015). To asses the number of epileptic spikes, each mice underwent a 1 h recording before NTIRE treatment and 1 h after the treatment (for non-treatment group, delay in between the two recording session was the same as NTIRE group). Epileptic spikes were detected and an average of the spikes/min was calculated.

Finite Element Method calculations were performed with the COMSOL Multiphysics 5.5 software. The physic used was “Electrical Current.” A custom mouse brain mesh model was designed on Blender software and then imported in COMSOL Multiphysics 5.5. The distance and the voltage applied between the electrodes were taken in account to recreate as closely as possible the experimental conditions of the Cortical NTIRE ablation, 245 V between ≃ 1.4 mm. All plots were made and exported with the built-in tool of COMSOL Multiphysics 5.5 (Supplementary Figure 3).

Conventional histology was performed on twelve section. Four days after the stimulation mice were perfused with saline followed by 4% paraformaldehyde, then brain was carefully removed from the skull and transferred to 4% PFA for post fixation. After 1 day of post-fixation, brain was washed three times in 0.12 PB and then immersed in 20% sucrose in 0.12 PB for 24 h. The day after brain was frozen, with the ventral surface placed on a metal plate, cooled by a surrounding bath of dry ice. Brains were stored at −80°C until cut. The brain tissue was sectioned into coronal section of 30 μm in thickness on Leica CM3500 cryostat (Leica Microsystems Inc., Bannockburn, IL, United States).

For cell counting: all tissue sections lying between the two electrodes were collected. Brain sections to be stained with Nissl Eosin were directly mounted on the microscopic glasses (Polysine™ Microscope Adhesion Slides – Thermo Fisher Scientific) and left for 24 h in the oven.

Then, sections were placed in alcohol 95% two times for 10 min, followed by 5 min in chloroform. Then, they were hydrated by alcohol 95, 80, and 75% for 2 min each. Next, sections were soaked 10 times slowly in distilled water before staining them in the heated (60°C) Nissl’s solution for 1 min. A second wash with water is done for 1 min before immersion of the sections in acetic formalin solution for 6 min. Washing with water for 1 min is done before passing to the dehydration step in alcohol 80, 90, 95, 100, and 100% for 2 min each. Nissl’s sections were used for assessing the cell death.

Date were analyzed using R software. Vasoconstriction, cell loss and spike rate data were separated into groups regarding the ablation protocol (NTIRE, TIRE, and no treatment). After a common normality verification using Shapiro test, groups were compared with T-test or Wilcoxon-test depending on the presence or absence of a normal distribution, respectively.

Figures 2, 3: 9 NTIRE mice were analyzed to assess the percentage of vasoconstriction on NTIRE mice. To determine the vasoconstriction percentage broadly, blood vessels where divided into two groups: large blood vessels ≥20 μm and small blood vessels <20 μm.

Figure 4: 4 OF1 mice were sacrificed, and histological analysis of cell loss was performed for cortex cell-density data, with contralateral cortex used as the control for cell density.

Figure 5: 12 OF1 mice were kindled as described above and six received NTIRE ablations, the six other were not treated to show that there is no decrease of spikes during time. After a common normality verification using Shapiro test, Wilcoxon test were performed.

The experimental setup is introduced in Figures 1A,B, where microelectrodes are mounted on the somatosensory or posterior parietal cortex of mice. When the NTIRE event is visualized in real-time using a 2-photon fluorescence microscope, a brief contraction of vessels can be seen but they remain intact. However, neurons in the region of the NTIRE are destroyed, as depicted in Figure 1C. A thermal ablation, or thermal irreversible electroporation (TIRE) also uses a high electric field, but typically the pulse width is larger creating a thermal event and subsequent coagulation necrosis of tissue – vessels, neurons, and all tissues in the ablated region are destroyed. Then we wanted to compare if the tissue damage in the epileptic hippocampus after NTIRE ablations is reduced compared to epileptic mice receiving a thermal ablation of the same region. To better understand this, the parameters of non-thermal ablations are discussed in detail in Figure 2 and Supplementary Figure 1.

The primary parameters of interest in a thermal versus non-thermal ablation are power and electric field. The electric field must be sufficiently high to cause the irreversible poration of cell membranes. An ablation is performed with identical electric field values, high enough to cause irreversible poration, here 1,750 V/cm (Supplementary Figure 1). However, in NTIRE ablation, the associated power is only 200 mW, creating a non-thermal event. As can be seen in the image, the cortex immediately after the NTIRE event is visibly unchanged. For TIRE, the associated power is 11 W, creating a very thermal event. As can be seen in the image, the cortex immediately after the TIRE event shows a increase in bleeding. Because of absence of bleeding in the NTIRE process, mice can be monitored for longer periods of time without significant reduction in imaging quality of the two photon visualization. Consequently, mice were monitored up to 37 days after NTIRE ablation, with no visual destruction of vessels seen for the duration of the monitoring (Figure 2B). As seen in Figure 2C, blood plasma flow slows, with red blood cells seen individually in the vessels (panel, t = 1 min 45 s) as the speed of blood plasma flow has decreased. Flow recovers for vessels of all sizes after several minutes, however, (right panel) the difference between large and small vessels is simply the percentage of contraction – with no discernable long-term effects. Indeed, it was the short-term effects on vessels which were quite interesting immediately after NTIRE, namely the leakage and the percentage of contraction. Then, the contraction and leakage as been analyzed as a function of diameter. In Figure 2D, a section of cortex can be seen with vessels down to a depth of approximately 400 μm. Vessels of numerous sizes can be seen before and after, without any damage to the vessels. Vessel diameters due to the NTIRE have completely recovered. On the panel E, data showing typical contractions in a section of tissue is grouped into large vessels (greater than 20 μm diameter) and small vessels (less than 20 μm diameter). The percentage of vasocontraction is more dramatic for smaller vessels than larger vessels (t_test: p-value = 0.02622). Interestingly, this corresponds well to leakage, which is measured as a change in fluorescence measured 10 μm outside of vessels. So, three example vessels are shown, with 60, 44, and 10 μm diameters. The change in fluorescence outside of all vessels of 60 μm in diameter or greater is negligible, which is an important result as damage and leaking to such large vessels would create a significant hemorrhage. The change in fluorescence outside of vessels of large-medium vessels 30–40 μm in diameter (40 μm in the example) is small and relatively fast (Figure 2F). The interesting measurements are from vessels less than 20 μm, (10 μm diameter in the example) where fluorescence increases substantially. Although the vessels are not destroyed, leakage will likely allow a small amount of blood into the surrounding tissue. The result is that leakage from a vessel due to NTIRE seems to scale with the percentage of vasoconstriction of a given vessel, with an increased percentage in vasoconstriction from smaller vessels correlating with increased leakage. However, independent of size, vessels remain intact over time with contraction and leakage limited to the initial time immediately after NTIRE. In contrast, the consequences for neurons due to NTIRE is not temporary, as seen in Figures 3A,B and Supplementary Figure 4.

As depicted in Figure 3, ablation of neurons scales very closely with the electric field. In Figure 3A a section of cortex directly between the two micro electrodes can be seen. Figure 3B shows a loss of neurons can be seen in the region above 400 μm which, as shown in Figure 3C, correspond to a field strength of ∼800 V/cm; this correlates well with the threshold for irreversible electroporation. The electric field below the depth of 400 μm is not zero, but not high enough to create the irreversible poration and ablation. Although the region of cortex is ablated, meaning the neurons have been destroyed, clearly no evidence of coagulation or thermal events can be seen in any of the tissue slices. Around 70% of the neurons in the neocortex are eliminated in the first 200 μm with a decrease in the number of ablated cell with depth (Supplementary Figure 4).

Unlike the thermal ablation which destroys vessels and neurons immediately and leaves the brain with a coagulated region, NTIRE does not show a complete cell loss right after the stimulation (corresponding well with histology). However, already at 7 min, a cell and dendrite loss already started (Figure 4A). The panel B, shows a section of cortex highlighting an individual pair of neurons in the field between the two microelectrodes. Before (top left panel) and several minutes after (middle panel) the NTIRE event is performed, the neurons are clearly visible. Indeed, neurons can often be visible for more than 24 h after NTIRE. The peak clearance of neurons is at approximately day 3; any neurons visible at or after day 3 will not be removed. Interestingly, as seen in the Figure 4B (right panel) for the same section of cortex, although the neurons have died and been cleared away, enervation of fibers from a region below the ablated cortex can be seen. This is a very interesting result, as this is not the case in a thermally coagulated section of neural tissue, where there is induction of axonal sprouting and synchronous activity (Carmichael and Chesselet, 2002). This could indicate an accelerated time-line for recovery, compared to the case of a thermal ablation, avoiding also the possible generation of epilepsy following cortical injury (Prince et al., 2009). Day 3, and also day 4, corresponds to a definite and visible increase in dendritic cells, specifically around the vessels in a region which has experienced the NTIRE event (Supplementary Figure 5). Day 7 seems to be the peak of dendritic cell concentration. In general, however, due to the lack of a thermal coagulated necrosis of tissue, it seems that the immune response and recovery is much similar to a mild stroke, with recovery on a much faster time-scale compared to thermal ablation. This is likely related to the fact that vasculature and extracellular matrix is completely in place, and there is no sign of tissue scarring.

Figure 5 shows the results of the NTIRE ablation applied to epileptic foci in the mouse hippocampus. As seen in Figure 5A, a two-electrode implant is placed at the border of CA3 in the mouse hippocampus. Animals had previously been rendered epileptic with a standard kindling protocol at this location (detail in the Supplementary Material). Animals under continuous video and electrophysiological supervision displayed behavioral epileptic state and pathological electrophysiological activity. As shown in Figure 5B, all the animals were kindled and were having burst of interictal spikes. For the ones treated with an NTIRE ablation no longer displayed behavioral seizures and no longer showed pathological electrophysiological activity (representative of behavioral epileptic state available in Supplementary Material). The ablation rendering the animal seizure-free is important, but the lack of coagulated tissue in the NTIRE ablation has perhaps additional and positive consequences regarding recovery after ablation. In order to demonstrate this impact on hippocampal oscillation, a ratio of the detected interictal spikes (spikes detection after treatment/spikes detection before treatment) as been performed. The No treatment group has a ratio ∼1 meaning that there is no decrease of the number of spikes detection with time. But NTIRE ablation induced a significative reduction of the interictal spiking (p value = 0.02597) is a technique of ablation that can also be performed at depth like in an epileptic focus located in the hippocampus, similar cell death is seen with a corresponding reduced tissue damage and lack of thermal coagulation (Supplementary Figure 1C).