Table of Contents

Perineuronal Nets in the Rat Medial Prefrontal Cortex Alter Hippocampal–Prefrontal Oscillations and Reshape Cocaine Self-Administration Memories

  1. Barbara A. Sorg1,2
  1. 1R.S. Dow Neurobiology, Legacy Research Institute, Portland, Oregon 97232
  2. 2Program in Neuroscience, Washington State University, Vancouver, Washington 98686
  3. 3Department of Chemical Physiology and Biochemistry, Oregon Health & Science University, Portland, Oregon 97239
  4. 4Program in Neuroscience, Washington State University, Pullman, Washington 99164,
  5. 5Departments of Behavioral Neuroscience, Oregon Health & Science University, Portland, Oregon 97239
  6. 6Psychiatry, Oregon Health & Science University, Portland, Oregon 97239
  7. 7Research Division, VA Portland Health Care System, Portland, Oregon 97239

  1. Author contributions: J.C.W., J.D.R., D.M.H., S.A.A., T.E.B., A.I.A., and B.A.S. designed research; J.C.W., J.D.R., A.E.G., R.M.R., D.M.H., and L.G.B. performed research; J.C.W., J.D.R., S.X.R., D.M.H., S.A.A., L.G.B., T.E.B., and B.A.S. analyzed data; J.C.W., D.M.H., S.A.A., and B.A.S. wrote the paper.

 
Next Section

Abstract

The medial prefrontal cortex (mPFC) is a major contributor to relapse to cocaine in humans and to reinstatement in rodent models of cocaine use disorder. The output from the mPFC is potently modulated by parvalbumin (PV)-containing fast–spiking interneurons, the majority of which are surrounded by perineuronal nets. We previously showed that treatment with chondroitinase ABC (ABC) reduced the consolidation and reconsolidation of a cocaine conditioned place preference memory. However, self-administration memories are more difficult to disrupt. Here we report in male rats that ABC treatment in the mPFC attenuated the consolidation and blocked the reconsolidation of a cocaine self-administration memory. However, reconsolidation was blocked when rats were given a novel, but not familiar, type of retrieval session. Furthermore, ABC treatment prior to, but not after, memory retrieval blocked reconsolidation. This same treatment did not alter a sucrose memory, indicating specificity for cocaine-induced memory. In naive rats, ABC treatment in the mPFC altered levels of PV intensity and cell firing properties. In vivo recordings from the mPFC and dorsal hippocampus (dHIP) during the novel retrieval session revealed that ABC prevented reward-associated increases in high-frequency oscillations and synchrony of these oscillations between the dHIP and mPFC. Together, this is the first study to show that ABC treatment disrupts reconsolidation of the original memory when combined with a novel retrieval session that elicits coupling between the dHIP and mPFC. This coupling after ABC treatment may serve as a fundamental signature for how to disrupt reconsolidation of cocaine memories and reduce relapse.

Previous SectionNext Section

Significance Statement

Powerful memories are associated with drug-taking behavior over extended periods, and these memories can drive relapse to drugs of abuse. The medial prefrontal cortex (mPFC) is a major contributor to relapse in cocaine use disorder. In a well-established rodent model of cocaine use disorder, we identify how the removal of key extracellular matrix structures called perineuronal nets within the mPFC reduces the ability to update a cocaine memory. We further show that the ABC treatment within the mPFC impairs the coupling of oscillations between the mPFC and dorsal hippocampus during the updating of cocaine memory. This impaired communication between mPFC and hippocampal circuitry may act as a signature for disrupting cocaine-related memories to help break the cycle of relapse.

Previous SectionNext Section

Introduction

The medial prefrontal cortex (mPFC) is a key region contributing to relapse to cocaine. The output from the mPFC is critical for cocaine-seeking in humans (Volkow et al., 1991; Grant et al., 1996; Maas et al., 1998; Childress et al., 1999; Zhai et al., 2021) and reinstatement behavior in rodents (McFarland and Kalivas, 2001; W. K. Park et al., 2002; Capriles et al., 2003; Fuchs et al., 2007; Ma et al., 2014). The neuronal output from the mPFC is strongly regulated by parvalbumin (PV)-containing fast–spiking inhibitory interneurons (Markram et al., 2004; A. T. Lee et al., 2014; Sparta et al., 2014). The majority of mPFC PV neurons are enwrapped in extracellular matrix (ECM) structures called perineuronal nets (PNNs), which facilitate the precise firing of PV neurons (Wingert and Sorg, 2021). PV neurons provide dense connectivity to pyramidal neurons in the PFC (Packer and Yuste, 2011) and rapid feedforward inhibition (A. T. Lee et al., 2014). PNNs form in the central nervous system during the critical period of plasticity and stabilize networks during adulthood (Pizzorusso et al., 2002). ABC treatment promotes plasticity in adulthood in several memory tasks (Fawcett et al., 2019, 2022). However, ABC treatment appears to prevent the plasticity induced by strong stimuli, such as drugs of abuse and footshock stress (for review, see Fawcett et al., 2022; Brown and Sorg, 2023).

We previously demonstrated that degradation of PNNs with the enzyme chondroitinase ABC (ABC) in the prelimbic mPFC reduced the acquisition and reconsolidation of cocaine-induced conditioned place preference (CPP) in rats (M. Slaker et al., 2015). Reconsolidation is a memory maintenance process in which memories can be rendered labile upon retrieval (reactivation) and subsequently restabilized or updated. Reactivated memories can be diminished if amnestic agents are present during the reconsolidation period (∼6 h after reactivation), including drug-associated memories that are difficult to disrupt (J. L. C. Lee et al., 2005; Sorg, 2012; Tronson and Taylor, 2013; Exton-McGuinness and Milton, 2018). Some studies have shown that disrupting stronger memories requires novel information to be incorporated into the preexisting memory to induce destabilization by the amnestic agent and reduce expression of the old memory (Diaz-Mataix et al., 2013; Sevenster et al., 2013; Exton-McGuinness et al., 2014; Exton-McGuinness and Lee, 2015). Drug self-administration memories become well established by repeated operant responding for the drug. We therefore tested whether ABC treatment in the mPFC could disrupt the reconsolidation of a cocaine self-administration memory by providing novel information during memory reactivation. Moreover, PNNs are believed to contribute to the maintenance of memories through their network-stabilizing properties (Fawcett et al., 2019; Christensen et al., 2021) and may do so through altering firing properties of their underlying PV neurons during memory reconsolidation. PV neurons regulate neural oscillations, particularly in the gamma range, which are critical to memory (Steullet et al., 2014; Lensjo et al., 2017; Faini et al., 2018; Thompson et al., 2018; Shi et al., 2019). Gamma oscillations are altered after cocaine exposure (Kiyatkin and Smirnov, 2010; Dilgen et al., 2013; McCracken and Grace, 2013) and when PNNs are removed (Cabungcal et al., 2013; Steullet et al., 2014; Lensjo et al., 2017; Faini et al., 2018; Thompson et al., 2018; Shi et al., 2019; Carceller et al., 2020). However, only a few studies have assessed oscillations across brain regions in any addiction model (Sjulson et al., 2018; Guillem and Ahmed, 2020; Li et al., 2020; Muller Ewald et al., 2022; Fattahi et al., 2023), and no studies have examined how disrupted reconsolidation of a drug memory alters network communication in response to cocaine or cocaine cues. Here we demonstrate that ABC treatment in the mPFC blocks cue-induced reinstatement and reduces the reinforcing efficacy of a cocaine self-administration memory, but only when a novel contingency is presented at the time of memory reactivation. We further show that ABC treatment in the mPFC reduces the synchrony of oscillations between the mPFC and the dorsal hippocampus (dHIP) and alters mPFC cross-hemispheric synchrony.

Previous SectionNext Section

Materials and Methods

Drugs

Cocaine hydrochloride (generous gift from the National Institute on Drug Abuse) was dissolved in saline to intravenously deliver 0.5 mg/kg/infusion in 0.1 ml over 6 s during all self-administration sessions with exception of the acquisition study, in which 0.125 mg/kg/infusion cocaine was delivered. For the sucrose self-administration experiment, sucrose pellets (Bio-Serv, F0025) were dispensed for oral consumption during self-administration. Chondroitinase ABC (ABC; Sigma-Aldrich, C3667) degrades glycosaminoglycan side chains of chondroitin sulfate proteoglycans of PNNs and the “loose” ECM (Bruckner et al., 1998). ABC was dissolved in sterile water vehicle (Veh) to a final concentration of 0.09 U/µl (M. Slaker et al., 2015).

Animals

A total of 160 male Sprague Dawley rats were acquired from Simonsen Laboratories and Envigo or bred within the University of Wyoming vivarium. Rats weighed between 275 and 300 g at the beginning of the experiment. Animals were pair-housed up until surgery and were then housed singly in a temperature- and humidity-controlled room under a 12 h reverse light/dark cycle, and all experiments were run during the dark cycle. Animals were given food and water ad libitum throughout the study with exception of the time they were in the self-administration chambers. All experiments were approved by the Washington State University, University of Wyoming, and Legacy Research Institute Institutional Care and Use Committees and were done according to the National Institutes of Health Guide for the Care and Use of Laboratory Animals. All efforts were made to reduce the number of animals used in the experiments and to minimize pain and suffering.

Surgery

Rats were anesthetized with isoflurane followed by either intramuscular injection of zyket (ketamine 87 mg/kg; xylazine 13 mg/kg) or intraperitoneal ketdex (ketamine 75 mg/kg; dexmedetomidine 1 mg/kg). Atipamezole (1 mg/kg) was given subcutaneously to reverse anesthesia postoperatively. Rats for cocaine self-administration experiments were implanted with chronic indwelling intravenous catheters. Catheters were made from silastic tubing (RenaSil Silicone Rubber Tubing, 0.037″ OD, 0.025″ID, Braintree Scientific) cut to a length of 9.0 cm with a slight taper on one end with a suture bead made from medical-grade silicone adhesive (Kwik-Sil, World Precision Instruments) placed 2.8 cm from the tapered side for secure anchoring. The right jugular vein was isolated, and a small incision was made. The tapered end of the catheter was inserted into the vein until the suture bead was flush with the vein incision. The catheter was secured to the vein by tying 3-0 silk suture thread around the vein and catheter on both sides of the suture bead; additionally, the thread on both sides was tied together. Incisions were sutured, and the animal was given 6–7 d to recover. After surgery, catheters were flushed daily with 0.1 ml of heparin (10 U/ml) and cefazolin antibiotic (10 mg/ml) in sterile saline to help protect against infection and catheter occlusion.

The cannula either consisted of 26 gauge, 1.4 mm spaced, bilateral cannula with dummy obturators (Plastics One) or was made by cutting 26 gauge stainless steel hypodermic tubing (Component Supply) to a length of 12 mm with 33 gauge stainless steel tubing (Component Supply) obturators of the same length as the cannula. The cannula was implanted to target 1 mm above the microinjection site within the prelimbic mPFC (+3.2 mm from the bregma, 0.7 mm from midline, −2.5 mm from the skull).

For in vivo electrophysiology, custom cannulated bilateral tetrode microdrives were implanted in the prelimbic PFC targeting the same site as the cannula above. Tetrodes were just caudal to and lying against the microinjection cannula. A small craniotomy and durotomy was performed over the mPFC and 26 gauge stainless steel cannula of the drive were lowered to 2.5 mm, placing the tetrode guide tubes just above the surface of the brain (for drive details, see the section on electrophysiology). Three tetrodes per mPFC hemisphere were driven down 2 mm during surgery and were slowly lowered to reach their final depth equal near to where the microinjection would be given. Using a Nano-Z (White Matter), tetrodes were electroplated using NeuraLynx gold plating solution (NeuraLynx) until all wires had impedances ranging from 200 to 400 kOhms. To simultaneously record field potentials from the dHIP, a 50 micron tungsten wire was lowered into each hemisphere of the dHIP (CA1 region; −4.0 mm from the bregma, 2.2 mm from midline, −3.5 mm from the skull). All wires were pinned in place into a 32-channel electrode interface board (NeuraLynx). The drives were designed to lower three tetrodes per hemisphere to the mPFC and allow for drug delivery and to record from the mPFC. A Kwik-Sil silicone adhesive was used to fill the craniotomy and protect the brain from the dental acrylic used to secure the drive. A stainless steel ground screw was placed above the somatosensory cortex, and a reference screw was placed above the cerebellum.

Microinjections

Custom-made 33 gauge stainless steel microinjection needles were made to project 1 mm below the cannula and were used to deliver 0.4–0.6 µl of either Veh or ABC over a period of 108 s as previously described (M. Slaker et al., 2015) with a 1 min postinjection period to allow for diffusion. Some groups of animals were naive and were microinjected during stereotaxic surgery, and the guide cannula was not fixed in place as for animals that were used for all other experiments. However, the same custom needles, injection volume, and infusion rate were used as with the other animals. We targeted the prelimbic cortex, but some areas surrounding this region were also affected by ABC, including regions dorsal to the prelimbic mPFC (the cingulate cortex and medial-most aspect of the secondary motor cortex) and, in rostral-most regions, the medial orbital and medial portions of the ventral orbital cortex. We did not include animals that had ABC spread into bilateral infralimbic mPFC based on our previous findings that ABC targeted to infralimbic mPFC did not alter cocaine-induced CPP (M. Slaker et al., 2015; representative images are shown in Fig. 4A,B and Extended Data Fig. 7-1).

Experimental design for self-administration experiments

Self-administration training

Animals were trained on one of the two different training protocols, including a fixed-ratio 1 (FR1) or fixed-ratio 3 (FR3) schedule of reinforcement. All animals were initially trained during 2 h sessions on an FR1 schedule in which they were placed in operant chambers (Med Associates) with one active (right lever) and one inactive (left) lever, with a cue light over each lever and a house light on the wall opposite from the levers. Active lever presses were reinforced with a 0.1 ml infusion of cocaine (either 0.125 mg/kg/infusion for the acquisition experiment or 0.5 mg/kg/infusion for all other experiments) or a 45 mg sucrose pellet (Bio-Serv). Reward delivery was accompanied by simultaneous cue light activation for 6 s and house light activation for 20 s that signaled a time-out period during which active lever pressing produced no consequences. Inactive lever presses at any time had no consequences. All active and inactive presses were recorded. To be included in the study, animals had to reach 3 consecutive days with >10 rewards. For FR3 training in a subset of animals, rats began on an FR1 schedule and had to reach the criterion above before being switched to an FR3 schedule in which the animals had to press the active lever three times to be rewarded a single cocaine infusion, cue light, and given the time-out light. FR3-trained animals, in addition to meeting the FR1 to FR3 transition criteria, had to have 3 or more days of FR3 training within the total of 10 training sessions. Three to four hours following the last training session, most cohorts received a microinjection of either Veh or ABC targeted to the prelimbic mPFC. Two additional cohorts received Veh or ABC targeted to the prelimbic mPFC within ∼1 h after the memory reactivation session (see below).

Memory reactivation

Three days after the last training session, FR1-trained animals received either a 30 min FR1 or a novel variable ratio schedule (VR5 reactivation session). VR5 reactivation consisted of animals responding to a VR5 schedule of reinforcement in which the number of active presses needed to receive each reward was randomly selected between one and nine presses with an average of five lever presses. During the memory reactivation session, we chose to allow animals to press for cocaine to avoid extinction processes during this session. In addition, our previous work (Brown et al., 2007) and that of others (Milekic et al., 2006; Valjent et al., 2006) using the CPP model showed that both the unconditioned stimulus (cocaine or morphine) and the drug-paired context were necessary to disrupt reconsolidation.

One set of FR1-trained animals was not microinjected following training but was instead microinjected after the VR5 reactivation session, with the Veh or ABC microinjection occurring within a 5 min to 2 h window for all but one animal, which received a Veh microinjection 2.4 h after the VR5 memory reactivation session. The average time (±standard error of the mean, SEM) for post-VR5 microinjection was 67 ± 10 min after the completion of the 30 min memory reactivation session in this group. Hereafter, this group is referred to as “post-VR5.” The FR3-trained animals received a 30 min FR3 reactivation session. One set of animals, a no-reactivation group, did not receive a reactivation session but received Veh or ABC in their home cages.

Cue-induced reinstatement

The day following the reactivation session, all animals received a 30 min extinction session in which no cues or rewards were delivered. Immediately following the 30 min extinction, animals underwent a 30 min cue-induced reinstatement session in which a single cue light primed animals to continue to press on an FR1 schedule for the cue light above the cocaine lever, but no rewards were delivered (cocaine or sucrose).

Progressive ratio

The day following the extinction and cue reinstatement sessions, most groups of rats received 3 consecutive days of a 3-h-long progressive ratio (PR) self-administration in which the reinforcement schedule needed to obtain a reward increased as previously described (Richardson and Roberts, 1996). Following the PR days, rats were killed with an overdose of sodium pentobarbitol (67 mg/kg, i.p.) and intracardially perfused with 4% paraformaldehyde (PFA). The brains were removed and cryoprotected in 20% sucrose and were frozen with powdered dry ice and stored at −80°C until analysis for cannula placement and confirmation of PNN degradation.

Naive animals not undergoing self-administration received microinjections of Veh or ABC either 1 h or 3 d prior to killing them. These groups are referred to as either “1 h” or “3 d” to temporally match the self-administering experiments in which animals received Veh or ABC 3 d prior to or ∼1 h after memory reactivation. Naive animals subjected to this protocol were used for (1) slice electrophysiology experiments and (2) PV cell intensity inhibitory/excitatory puncta experiments.

BS3 cross-linking and Western blot

For the cross-linking surface protein quantification assay, we followed the procedure of Boudreau et al. (2012), with slight modifications. In brief, following cocaine self-administration training, one group of rats underwent the FR1 reactivation protocol, and the other group underwent the VR5 reactivation protocol described above. Immediately following the 30 min reactivation session, animals were rapidly decapitated, and the brain was extracted. A ∼2 mm coronal section was taken encompassing either the mPFC or the amygdala. The section was then placed on a piece of wetted filter paper placed on an inverted glass petri dish on ice, and the brain region of interest was rapidly dissected with a razor blade. The dissected region was rapidly sliced into strips of ∼400 microns by hand before being transferred to a sample tube containing 1 ml of cross-linking solution. The cross-linking solution was composed of ice-cold artificial cerebral spinal fluid (aCSF) containing the following (in mM): 1.2 CaCl2, 20 HEPES, 147 NaCl, 2.7 KCl, 1 MgCl2, and 100 dextrose in ddH2O, mixed with 52 mM bis(sulfosuccinimidyl)suberate (BS3; Thermo Fisher Scientific, catalog #21580), a membrane-impermeable bifunctional cross-linking agent, in 5 mM sodium citrate buffer (5 mM citric acid and 5 mM sodium citrate), pH 5.0, to bring the final concentration of BS3 to 2 mM in aCSF. Following the transfer to the tube of cross-linking agent, the sample tube was incubated at 4°C for 30 min gently mixing on a rocker. The cross-linking reaction was quenched by adding 100 μl of 1 M glycine to the sample tube, which was then placed on the rocker for an additional 10 min at 4°C. After quenching, the samples were centrifuged at 20,000 × g for 2 min at 4°C. The supernatant was discarded, and the pellet was resuspended in 400 μl of NP-40 lysis buffer, pH 7.4, containing the following (in mM): 25 HEPES, 500 NaCl, 2 EDTA, and 1 DTT, 0.1% Nonidet P-40 (v/v) with phosphatase inhibitors (Pierce, A32957, or Millipore Sigma, 4906845001) and EDTA-free protease inhibitors (Pierce, A329965). Samples were then sonicated for 5 s before centrifuging for 2 min at 20,000 × g at 4°C. The supernatant was aliquoted into 20 μl aliquots and flash frozen on dry ice and stored at −80°C. A BCA protein assay kit (Thermo Fisher Scientific, 23225) was used to determine protein concentration. Forty microgram of the sample protein was diluted with NP-40 buffer and SDS Sample Buffer (Bio-Rad Laboratories) and were then loaded into 10% precast gels (Bio-Rad Laboratories) and electrophoresed with Tris/glycine/SDS running buffer (Bio-Rad Laboratories) before being transferred to PVDF membranes. The blots were then washed three times for 5 min in TBS-T (10 mM Tris/HCL, 15 M NaCL, 0.05% Tween 20), pH 8.0, before being blocked for 1 h in Odyssey Blocking Buffer (LI-COR Biosciences). The blots were then probed with primary antibodies (GluN2B, 1:500, Abcam, 71-8600) in Odyssey Blocking Buffer on a shaker at 4°C overnight. Blots were then washed in TBS-T three times for 10 min before incubating at room temperature on a shaker in HRP-conjugated secondary antibody solution (1:10,000) in a blocking buffer. After washing in TBS-T three more times for 10 min each, the blots were incubated in Western blotting substrate (Thermo Fisher Scientific, 34579) for 2 min and imaged on a ChemiDoc Touch Imaging System (Bio-Rad Laboratories). To normalize to total protein concentration, we incubated the blots in ponceau S staining solution (Sigma-Aldrich) for 5 min to detect protein before being imaged on the ChemiDoc Touch. Each protein band was normalized to the total protein intensity of the entire lane.

Immunohistochemistry

Analysis of PNNs and PV at 1 h and 3 d

To assess the intensity and number of PV and PV/Wisteria floribunda agglutinin (WFA) cells at 1 h and 3 d after ABC microinjection, we anesthetized the animals with pentobarbital (67 mg/kg, i.p.) and perfused intracardially with 4% PFA in phosphate-buffered saline. Brains were removed and stored overnight in 4% PFA at 4°C. The following day, brains were moved to a 20% sucrose solution and 24 h later were frozen with powdered dry ice and stored at −80°C until analysis. Coronal sections (40 µm) through the mPFC were sectioned using a sliding microtome and stained with WFA (1:500; Vector Laboratories; FL-1351; RRID, AB_2336875) and rabbit anti-PV antibody (1:1,000; Novus; NB120-11427; RRID, AB_791498) according to Jorgensen et al. (2020). For each rat, we analyzed 3–5 images/hemisphere/rat (one rat had only one image). PNNs were identified by staining with WFA, a plant lectin that binds to chondroitin sulfate proteoglycan components of the PNNs (Hartig et al., 1992). PNN “stubs” after ABC digestion were identified by staining with mouse anti-chontroitin sulfate (1B5) antibody (1:20, AMS Biotechnology, 270431-CS). Secondary antibodies were Alexa Fluor 405 goat anti-rabbit for PV or Alexa Fluor 594 goat anti-mouse (1:500, Thermo Fisher Scientific, A31556).

The mPFC and dHIP images were taken using a Leica SP8 laser scanning confocal microscope with an HCX PL Apo CS2, dry, 20× objective with 0.75 numerical aperture (NA) or with a water immersion 63× objective with 1.2 NA. Images (0.84 µm/pixel for 20×, 0.23 µm/pixel for 63×) were compiled into summed images using ImageJ macro plug-in Pipsqueak (M. L. Slaker et al., 2016) or PipsqueakTM, scaled, and converted into 8 bit, grayscale tiff files. Pipsqueak (M. L. Slaker et al., 2016) or PipsqueakTM was run in “semiautomatic mode” to select ROIs to identify individual PV cells and PNNs. The plug-in compiles this analysis to identify single, double-labeled, and triple-labeled neurons (https://ai.RewireNeuro.com). Histologies for targeted sites and for PNN loss in ABC-treated rats were identified by at least two investigators blinded to experimental treatment.

Analysis of puncta around PV cells at 1 h and 3 d

Immunohistochemical methods and antibodies used for this study were similar to those described previously (Harkness et al., 2021). Confocal imaging was performed on either a Zeiss LSM 780 or a Zeiss LSM 900 confocal microscope with a 63× 1.4 NA Plan-Apochromat oil objective (Carl Zeiss MicroImaging) due to a change in instrumentation in the Oregon Health and Science University Advanced Light Microscopy Core midway through the study. Coronal mPFC tissue sections from a particular animal were only imaged on one of the microscopes. The imaging strategy was similar to what was described previously (M. L. Slaker et al., 2018; Jorgensen et al., 2020). The dimensions of images captured using the LSM 900 confocal microscope (x-y-z, 126.77 × 126.77 × 0.230 µm) were smaller than those captured on the LSM 780 (x-y-z, 134.95 × 134.95 × 0.384 µm) due to improved pixel resolution in all planes. The change in Z-axis pixel thickness affected the number of optical slices used to create the Z-stack subsets through the middle of each PV neuron. Using the Zen software (Carl Zeiss; RRID, SCR_013672), PV neurons that met our criteria for inclusion in the study were identified. Inclusion criteria for PV neurons included being fully within the boundaries of the field of view and having a visible nucleus within the Z-stack subset. For LSM 780 confocal images, three consecutive 0.384 µm optical slices were used for each Z-stack subset for a total thickness of 1.152 µm through the middle of each PV neuron. For LSM 900 confocal images, five 0.230 µm optical slices were used for each Z-stack subset for a total thickness of 1.15 µm. Image analysis of GAD65/67- and VGluT1-labeled puncta apposing the exterior boundaries of the PV neuron was performed using the Imaris image analysis software as previously described (Jorgensen et al., 2020; Harkness et al., 2021). All parameters to identify puncta with the Imaris Spots segmentation tool were the same, regardless of the confocal microscope used to image that particular tissue section.

Whole-cell patch–clamp electrophysiology

Whole-cell patch–clamp recordings of fast-spiking interneurons (FSIs) were performed, as FSIs have been identified to contain PV (Kawaguchi and Kubota, 1993). Recordings in brain slices from adult [Postnatal Day (P)60–P90] male rats with and without PNNs in Layer 5 of the prelimbic mPFC were done using established procedures (M. L. Slaker et al., 2018; Jorgensen et al., 2020). Rats were microinjected with either Veh (0.6 µl saline) or ABC (0.09 U/µl, 0.6 µl/hemisphere) bilaterally into the prelimbic mPFC (AP, +3.2; DV, −2.5; ML, ±0.8) over 2 min and were given 1 min to be absorbed before removing the cannula. One hour or 3 d after microinjection, rats were perfused and decapitated, and coronal slices (300 µm) containing the mPFC (2.8–3.5 mm from the bregma) were prepared using a vibratome (Leica VT1200S). Brains were perfused and cut in a choline-based cutting solution (in mM): 2.5 KCl, 84.01 NaHCO3, 119 NaH2PO4, 180 D-glucose, 0.5 CaCl2, 7 MgCl2, 110 choline Cl, 3 Na-pyruvate, 11.6 Na-ascorbate, 300–310 mOsm, pH 7.2–7.3. After cutting, slices were incubated for 10 min in aCSF at 32°C followed by a minimum of 1 h in aCSF prior to recording (in mM): 120 NaCl, 2.5 KCl, 26.2 NaHCO3, 1.25 NaH2PO4, 11 Glucose, 2.5 CaCl2, 1.3 MgCl2, 300–310 mOsm, pH 7.2–7.3.

Prior to recording, slices were incubated with WFA (1 µg/ml) in aCSF for 5 min to stain for or verify reduction of PNNs. Fluorescing neurons positive for WFA in the vehicle group were identified in Layer 5 of the prelimbic mPFC using the cellSens software (Olympus) and subsequently patched. For rats microinjected with ABC, WFA stain was used to verify a reduction in positive WFA neurons, and FSIs were visually and electrophysiologically identified. For intrinsic experiments, 1.5 mm borosilicate glass (World Precision Instruments) pipettes were filled with an intracellular solution containing the following(in mM): 145 K gluconate, 3 KCl, 2 MgCl2, 2 Na+–ATP, 10 HEPES, 0.2 EGTA, pH 7.2–7.3, adjusted using potassium hydroxide, and 290–300 mOsm. FSIs in Layer 5 of the prelimbic mPFC were identified as described above and with their response to current injection (−100 to 500 pA, 100 pA increments, 250–500 ms) using previously established criteria (M. L. Slaker et al., 2018). Elicited action potentials (APs) were recorded, counted, and analyzed using pClamp11.0.2 (Clampfit, Axon Instruments). Neuronal excitability was assessed by quantifying the number of APs generated in response to each depolarizing current injection. Intrinsic properties were collected from the first AP elicited at 200 pA current injection. Input resistance was determined by the change in voltage from −100 pA current injection. Rheobase is the minimum amount of current to elicit an AP. The AP threshold was determined using the maximum second derivative method (Sekerli et al., 2004). The peak amplitude (difference between the AP peak and threshold) and amplitude of the afterhyperpolarization (AHP) were measured from the calculated AP threshold. Capacitance was measured from the membrane properties within pClamp. The reported capacitance values have been estimated through the utilization of pClamp's membrane property analysis. The interspike intervals (ISIs) were calculated using Clampfit 11.2 using the threshold search function, with the baseline set at −70 mV and the threshold for spike detection set at 0 mV. The values for ISIs from the 200 and 500 pA traces were taken, and the mean and standard deviation were calculated. The coefficient of variation was then calculated by dividing the standard deviation by the mean for each trace from each animal.

In vivo electrophysiology

Recording local field potentials (LFPs)

For recording, 32-channel headstages (Intan Technologies) were connected to a motorized commutator (Plexon) and an Open Ephys data acquisition system. All channels were referenced to a screw electrode placed over the cerebellum (reference shorted to ground), filtered between 1 and 7,500 Hz, multiplexed, and digitized at 30 kHz. A Med Associates transistor-transistor logic (TTL) output card was used to deliver TTL pulses to the digital inputs of the Open Ephys data acquisition system to timestamp lever presses, cue lights, and houselights used in the self-administration paradigm.

LFP analyses

All electrophysiology analyses were performed using custom Python scripts. Signals were referenced to a screw electrode over the cerebellum, downsampled from 30 to 2 kHz, and bandpass filtered between 1 and 300 Hz with a Butterworth bandpass filter using butter and filtfilt functions from the signal module of Python's scientific library scipy. Lever press timestamps were recorded via TTL pulses, and 2.5 s sections of data centered around each TTL pulse were sliced and considered as trials. A 2.7 root mean square (rms) threshold was set for artifact detection, and any section of data which exceeded this threshold was considered artifactual. Trials containing any artifacts were automatically rejected. Following artifact detection, trials were manually curated by visual inspection of each trial; trials with sections that were clearly artifactual but remained undetected by rms thresholding were removed.

Spectrograms were calculated via Morlet wavelet convolution from 1 to 120 Hz, implemented using the fft and ifft functions included in Python's numerical processing library, numpy. The number of wavelet cycles varied from three to nine, increasing as the frequency of the wavelets increased from 1 to 120 Hz, linearly. Power was extracted by taking the square of the real component of the analytic signal. A baseline period was defined from −1,000 to −100 ms, where each press occurred at 0 ms. Each trial was normalized at every frequency by dividing each sample by the mean power of the baseline period. Spectrograms were then generated by taking the average of all trials. Spectrograms were quantified at various frequency bands by defining a time–frequency window; then for each trial, power was averaged across the time–frequency window, and statistical significance was assessed via two-way analyses of variance (ANOVAs).

Phase–amplitude coupling (PAC) was assessed by calculating both mean vector length (MVL) as described (Canolty et al., 2006), and modulation index (MI) as described (Tort et al., 2008). To generate comodulograms, we extracted the phase angles from 3 to 13 Hz and amplitude from 20 to 100 Hz via Morlet wavelet convolution for each trial as described above. MVL was calculated from the concatenated trials for every pair of phase–amplitude frequencies. Each MVL was then z-scored by permutation testing with 10,000 permutations, where each permutation was generated by time-shifting the amplitude time series of each trial by a randomly generated number of samples before concatenation. The MVL was calculated for each permutation and the observed MVL was standardized to the distribution of the permuted MVLs with a significance threshold set at z = 1.96. To confirm the coupling between frequencies indicated by the z-scored comodulograms, gamma power and theta phase time series were extracted via a filter-Hilbert method using a Butterworth bandpass filter and the Hilbert function from Python's scipy library. Trials were concatenated, and the phase was binned into 30 equally sized bins. The mean normalized power in each phase bin was taken to generate a distribution of power over phase. The MI is calculated from this distribution and then z-scored by permutation testing with 10,000 permutations, where each permutation was generated by randomly shuffling trial IDs before concatenation. The MI was calculated for each permutation, and the observed MI was standardized to the distribution of the permuted MIs with a significance threshold set at z = 1.65.

Lag analysis and directionality

The direction of signaling across different brain sites can be assessed directly by calculating the cross-correlation of amplitude envelopes between the two sites at varying lags in a sliding window fashion, as described (Adhikari et al., 2010). First, envelopes of instantaneous theta amplitude (4–12 Hz) from the signals of two different sites were computed by bandpass filtering signals from 4 to 12 Hz, applying the Hilbert transform, and then taking the magnitude of the real component of the resulting analytic signal. The envelope from one site was incrementally shifted from −100 to +100 ms. At each increment, Spearman's rho was calculated between the envelopes of the two sites. Correlation coefficients were normalized to the minimum and maximum observed correlation within each trial, and the mean normalized correlation ± SEM was plotted for each lag. For each trial, the lag at which peak cross-correlation occurs was taken to generate a distribution of peak lags. Statistical significance between Veh and ABC treatment was assessed with the Kolmogorov–Smirnov test. To determine whether a significant lead was observed, for each trial, we averaged the normalized correlation coefficients across all negative (Neg) lags and all positive (Pos) lags (zero excluded), and we assessed the statistical significance with a two-way ANOVA. If the mean normalized correlation coefficient of all Neg lags is significantly different from the mean normalized correlation coefficient of all Pos lags, then it must be the case that signals from one region lead the other and directionality is denoted by the sign of the lag. In the current work, a Neg lag denotes that either the dorsal hippocampus leads the mPFC or, in the case of cross-hemispheric mPFC signals, a Neg lag denotes that the left hemisphere of the mPFC leads the right hemisphere of the mPFC.

Like lag cross-correlation analysis, Granger causality (GC) testing can also be used to assess the directionality of the flow of information between two areas. Given two time series, <mml:math display="inline"><mml:mi>X</mml:mi><mml:mo stretchy="false">(</mml:mo><mml:mi>t</mml:mi><mml:mo stretchy="false">)</mml:mo></mml:math>X(t) and <mml:math display="inline"><mml:mi>Y</mml:mi><mml:mo stretchy="false">(</mml:mo><mml:mi>t</mml:mi><mml:mo stretchy="false">)</mml:mo></mml:math>Y(t), if information about the past of <mml:math display="inline"><mml:mi>X</mml:mi><mml:mo stretchy="false">(</mml:mo><mml:mi>t</mml:mi><mml:mo stretchy="false">)</mml:mo></mml:math>X(t) and <mml:math display="inline"><mml:mi>Y</mml:mi><mml:mo stretchy="false">(</mml:mo><mml:mi>t</mml:mi><mml:mo stretchy="false">)</mml:mo></mml:math>Y(t) is more useful in making predictions about future values of <mml:math display="inline"><mml:mi>X</mml:mi><mml:mo stretchy="false">(</mml:mo><mml:mi>t</mml:mi><mml:mo stretchy="false">)</mml:mo></mml:math>X(t) than information about the past of <mml:math display="inline"><mml:mi>X</mml:mi><mml:mo stretchy="false">(</mml:mo><mml:mi>t</mml:mi><mml:mo stretchy="false">)</mml:mo></mml:math>X(t) alone, then <mml:math display="inline"><mml:mi>Y</mml:mi><mml:mo stretchy="false">(</mml:mo><mml:mi>t</mml:mi><mml:mo stretchy="false">)</mml:mo></mml:math>Y(t) Granger causes <mml:math display="inline"><mml:mi>X</mml:mi><mml:mo stretchy="false">(</mml:mo><mml:mi>t</mml:mi><mml:mo stretchy="false">)</mml:mo></mml:math>X(t) in the <mml:math display="inline"><mml:mi>Y</mml:mi><mml:mo stretchy="false">→</mml:mo><mml:mi>X</mml:mi></mml:math>Y→X direction. Mathematically, we compared two models representing these data: a univariate autoregressive (AR) model built only on data from <mml:math display="inline"><mml:mi>X</mml:mi><mml:mo stretchy="false">(</mml:mo><mml:mi>t</mml:mi><mml:mo stretchy="false">)</mml:mo></mml:math>X(t) and a bivariate AR model built on data from both <mml:math display="inline"><mml:mi>X</mml:mi><mml:mo stretchy="false">(</mml:mo><mml:mi>t</mml:mi><mml:mo stretchy="false">)</mml:mo></mml:math>X(t) and <mml:math display="inline"><mml:mi>Y</mml:mi><mml:mo stretchy="false">(</mml:mo><mml:mi>t</mml:mi><mml:mo stretchy="false">)</mml:mo></mml:math>Y(t). If the error term from the bivariate AR model <mml:math display="inline"><mml:msub><mml:mi>e</mml:mi><mml:mrow><mml:mi>X</mml:mi><mml:mi>Y</mml:mi></mml:mrow></mml:msub></mml:math>eXY is smaller than the error term of the univariate AR model <mml:math display="inline"><mml:msub><mml:mi>e</mml:mi><mml:mi>X</mml:mi></mml:msub></mml:math>eX, that is, information from both time series is a better predictor of <mml:math display="inline"><mml:mi>X</mml:mi><mml:mo stretchy="false">(</mml:mo><mml:mi>t</mml:mi><mml:mo stretchy="false">)</mml:mo></mml:math>X(t) than just information from <mml:math display="inline"><mml:mi>X</mml:mi><mml:mo stretchy="false">(</mml:mo><mml:mi>t</mml:mi><mml:mo stretchy="false">)</mml:mo></mml:math>X(t) alone, then we may consider that <mml:math display="inline"><mml:mi>Y</mml:mi><mml:mo stretchy="false">(</mml:mo><mml:mi>t</mml:mi><mml:mo stretchy="false">)</mml:mo></mml:math>Y(t) Granger causes <mml:math display="inline"><mml:mi>X</mml:mi><mml:mo stretchy="false">(</mml:mo><mml:mi>t</mml:mi><mml:mo stretchy="false">)</mml:mo></mml:math>X(t), and so GC is given by <mml:math display="inline"><mml:mi>G</mml:mi><mml:mi>C</mml:mi><mml:mo>=</mml:mo><mml:mi>l</mml:mi><mml:mi>n</mml:mi><mml:mrow><mml:mo>(</mml:mo><mml:mrow><mml:mrow><mml:mfrac><mml:mrow><mml:msub><mml:mi>e</mml:mi><mml:mrow><mml:mi>X</mml:mi><mml:mi>y</mml:mi></mml:mrow></mml:msub></mml:mrow><mml:mi>x</mml:mi></mml:mfrac></mml:mrow></mml:mrow><mml:mo>)</mml:mo></mml:mrow></mml:math>GC=ln(eXyx) (Seth et al., 2013).

To ensure stationarity, we first detrended and normalized the signals by subtracting ERPs and z-scoring. Next, univariate and bivariate AR models were generated by a custom Python implementation of MATLAB's armorf function. Models were built on data from a 400 ms time window, with a model order of 6. Optimal model order was selected by minimizing Bayesian information criterion. This process was done in a sliding window fashion at increments of 0.01 s to calculate GC across time for each trial. Time domain GC was normalized to a baseline period, and the statistical significance was assessed via two-way ANOVA (for a one-tailed test, p ≤ 0.05 if z ≥ 1.65; for two-tailed test, p < 0.05 if z ≥ 1.96).

Code accessibility

Custom Python code for analysis of in vivo electrophysiology data can be accessed at the following site: https://github.com/jonny-ramos/LRI/tree/main/Novel%20Memory%20Reactivation%20LFP%20Analysis.

Statistical analyses

Statistical data, with some exceptions (see in vivo electrophysiology studies above and the three-way, repeated-measure (RM) ANOVA to assess behavior), were analyzed using GraphPad Prism 9.5.1. Behavioral data were analyzed with either a two-way ANOVA, a two-way ANOVA with RM over time or day, or a three-way RM ANOVA with a Greenhouse–Geisser correction followed by Šídák's post hoc correction, with significance set at p < 0.05. In the case of a three-way RM ANOVA for behavior, R (v 4.2.3) was used. For the acquisition of cocaine self-administration, we also used a linear mixed-effect model correction followed by a Šídák's to assess the first day of acquisition that was different from Day 1 and not different from Day 20. In the case of two comparisons between treatment groups, a Student's unpaired, two-tailed t test was conducted. For analysis of WFA or PV intensity in the immunohistochemical studies and for analysis of relative power in the spectrograms, due to non-normal distribution, data were log transformed and then subjected to analysis using a two-way ANOVA. Puncta data were analyzed using a two-way ANOVA followed by a Šídák's post hoc test. For analysis of neuronal excitability, data were analyzed with a three-way, RM ANOVA. For analysis of firing variability and intrinsic properties, data were analyzed with a two-way RM ANOVA followed by a Šídák's post hoc test to compare significant interactions. All results are summarized as mean ± SEM unless otherwise stated. Differences were considered significant at p < 0.05, and trends are reported at p < 0.06. Statistics used for in vivo electrophysiology are described separately within each paragraph under “EEG Recordings” above. All statistics are reported in Extended Data Figures 1-1–9-1, and all statistically significant differences are reported in Results. Total (cumulative) active lever presses for all 14 treatment groups are shown in Extended Data Figure 1-1.

Figure 1-1

Cocaine or sucrose self-administration total active and inactive lever presses and total rewards. Cumulative lever presses and rewards for 14 treatment groups. Download Figure 1-1, DOCX file.

Figure 1-2

Statistical comparisons for acquisition of cocaine self-administration. Download Figure 1-2, XLSX file.

Figure 2-1

Rewards and inactive lever presses. Rewards and inactive lever presses for all phases of cocaine self-administration in Veh and ABC groups given FR1, Pre-VR5, or Post-VR5 reactivation session. Download Figure 2-1, XLSX file.

Figure 2-2

Statistical comparisons for FR1, Pre-VR5 and Post-VR5 groups. Download Figure 2-2, XLSX file.

Figure 3-1

Rewards and inactive lever presses. Rewards and inactive lever presses for all phases of cocaine self-administration in control groups, including No React, FR3, and Sucrose VR5 groups. Download Figure 3-1, XLSX file.

Figure 3-2

Statistical comparisons for No React, FR3, and Sucrose VR5 groups. Download Figure 3-2, XLSX file.

Figure 4-1

Statistical comparisons for PV cell intensity and excitatory and inhibitory puncta. Download Figure 4-1, XLSX file.

Figure 5-1

Statistical comparisons for slice electrophysiology. Download Figure 5-1, XLSX file.

Figure 6-1

Statistical comparisons for western blots. Download Figure 6-1, XLSX file.

Figure 7-1

Location of electrodes in mPFC for Veh and ABC treatment. (A) Sites of Veh injection and electrodes. Larger holes show electrode sites after application of current for identification of location. Microinjection sites were just caudal and below electrode sites. (B) Magnification of injection sites for each of the Veh-treated rats. (C) Sites of ABC injection and electrodes. (D) Magnification of injection sites for each of the ABC-treated rats. Rectangles show areas magnified in images in the right-hand panels; arrows indicate areas near electrode/cannulae sites showing WFA staining or lack of staining for PNNs. Scale bar: 1  mm. Download Figure 7-1, TIF file.

Figure 7-2

Statistical comparisons for spectrograms. Comparisons are shown for last training day of self-administration vs novel VR5 reactivation in Veh group and on novel VR5 reactivation in Veh vs ABC groups. Download Figure 7-2, XLSX file.

Figure 8-1

Statistical comparisons on comodulograms, Granger prediction, and cross-correlation lags for dHIP and mPFC during rewarded lever presses. Download Figure 8-1, XLSX file.

Figure 9-1

Statistical comparisons of comodulograms and cross-correlation lags for dHIP and mPFC and cross-hemispheric mPFC during unrewarded lever presses. Download Figure 9-1, XLSX file.

Previous SectionNext Section

Results

ABC in the mPFC blunts acquisition of cocaine self-administration

We tested whether ABC treatment in the mPFC would suppress the acquisition of cocaine self-administration using a low dose of cocaine previously used for acquisition studies (Goeders et al., 1998). Figure 1A shows the experimental timeline, and the sites targeted for Veh or ABC microinjection are shown in Figure 1E. Rats increased the number of active lever presses and cocaine infusions across the 20 d of training. However, ABC treatment in the mPFC reduced the number of active presses and infusions compared with the Veh group (Fig. 1B,C; Extended Data Figs. 1-1, 1-2). For active lever presses, a two-way RM ANOVA showed a main effect of treatment (F(1, 12) = 11.10; p = 0.006) and a main effect of day (F(3.22, 38.58) = 8.815; p < 0.0001). We also tested when the first day of acquisition was different from Day 1 and also not significantly different from Day 20. Compared with Day 1, Veh rats significantly increased self-administration beginning on Day 6 (p = 0.0005), while ABC rats did not reach these criteria until Day 12 (p = 0.0054). For infusions, a two-way RM ANOVA showed a main effect of treatment (F(1, 12) = 9.167; p = 0.0105) and a main effect of day (F(2.90, 31.14) = 10.71; p < 0.0001). Compared with Day 1, Veh rats significantly increased self-administration beginning on Day 6 (p = 0.0031), while ABC rats did not reach these criteria until Day 11 (p = 0.277). There were no effects of ABC treatment for inactive lever presses (Fig. 1D; Extended Data Figs. 1-1, 1-2). Thus, the acquisition of a cocaine-associated memory appeared to be delayed and blunted by ABC treatment in the mPFC (all statistical data are shown in Extended Data Fig. 1-2).

Figure 1.
View larger version:
Figure 1.

ABC treatment in the mPFC delays and blunts acquisition of cocaine self-administration. A, Timeline of experiment. B, Active lever presses during acquisition are reduced after ABC treatment compared with Veh treatment. For Veh rats, the first day of active lever pressing significantly higher than Day 1 was at Day 6, while for ABC rats, the first day higher than Day 1 was Day 12. C, The number of cocaine infusions during acquisition are reduced after ABC treatment. For Veh rats, the first day of the number of infusions significantly higher than Day 1 was at Day 6, while for ABC rats, the first day higher than Day 1 was Day 11. D, Inactive lever presses during acquisition were not different between groups. E, Targeted microinjection sites in the dorsal PFC. Gray dots, Veh; red dots, ABC. Data are mean ± SEM. Veh N = 8; ABC N = 6. *p < 0.05 compared between groups on the same day. Extended Data Figure 1-1 shows the total number of active and inactive lever presses and rewards for all groups. Extended Data Figure 1-2 shows statistical comparisons for Figure 1.

ABC in the mPFC blocks cocaine cue-induced reinstatement when given prior to a novel VR5 reactivation session

We next tested whether ABC treatment in the mPFC after training, but either prior to or after memory reactivation, would alter reactivation or subsequent reinstatement. Because it was previously found that cocaine self-administration memories may not be updated unless a novel contingency is introduced during memory reactivation (Exton-McGuinness and Lee, 2015), we also examined the effect of ABC treatment when the reactivated memory was updated with a novel VR5 schedule. Figure 2A shows the experimental timeline, and Figure 2, L and M, shows the sites for Veh and ABC microinjection, which spanned prelimbic and dorsal anterior cingulate regions. Figure 2, B and C, shows there were no late training differences among groups in active lever presses (Fig. 2B), number of cocaine infusions (Fig. 2C), or inactive lever presses (Extended Data Figs. 2-1, 2-2), as determined by a three-way RM ANOVA. Rats received Veh or ABC microinjection into the mPFC either 3 d prior to or within ∼1 h after the memory reactivation session (pre-VR5 or post-VR5, respectively).

Figure 2.
View larger version:
Figure 2.

ABC treatment in the mPFC blocks cocaine cue-induced reinstatement when given prior to a novel VR5 memory reactivation session. A, A timeline of experiment. B, Active lever presses during the last 3 d of training were not different among groups. C, The number of cocaine infusions during the last 3 d of training were not different among groups. D, Active lever presses during memory reactivation were higher in the VR5 compared with FR1 groups, with no difference between Veh and ABC groups. E, The number of cocaine infusions during memory reactivation was higher in the FR1 versus pre-VR5 group, with no difference between Veh and ABC groups. F–H, Extinction of active lever presses in FR1, pre-VR5, and post-VR5 groups, respectively. There was decreased responding across the session for all three groups, with no differences between Veh and ABC groups. I, Active lever presses during cue reinstatement. The pre-VR5 ABC group had decreased lever pressing versus the Veh group and versus the post-VR5 ABC group. J, Number of cue rewards during cue reinstatement. The pre-VR5 ABC group had a decreased number of cue rewards versus the Veh group, versus the FR1 ABC group, and versus the post-VR5 ABC group. K, Active lever presses during PR schedule of reinforcement over 3 consecutive days in FR1 and pre-VR5 groups. There was decreased responding in the pre-VR5 versus the FR1 groups across all days. L, M, Microinjection sites for FR1 or pre-VR5 (L) or post-VR5 (M). Gray dots, Veh; red dots, ABC. Data are mean ± SEM. FR1, Veh N = 9; ABC N = 6; pre-VR5, Veh N = 15 (10 for PR); ABC N = 10 (7 for PR); post-VR5, Veh N = 9; ABC N = 7. *p < 0.05. Extended Data Figure 2-1 shows the number of inactive lever presses and rewards from animals shown in Figure 2. Extended Data Figure 2-2 shows statistical comparisons for Figure 2.

During the memory reactivation session, there was no effect of ABC treatment on active lever presses, infusions, or inactive lever presses, irrespective of FR1 or VR5 schedules (Fig. 2D,E; Extended Data Figs. 2-1, 2-2). As expected, however, active lever presses during memory reactivation were higher in the VR5 compared with FR1 groups due to the requirement for a higher number of lever presses to receive a cocaine infusion. A two-way ANOVA showed a main effect of reactivation type (F(2, 50) = 23.98; p < 0.0001) but no main effect of ABC treatment. In a subsequent one-way ANOVA and Šídák's post hoc test, active lever presses were lower in FR1 versus pre- and post-VR5 groups (p < 0.0001 for both groups). In addition, for infusions, a two-way ANOVA showed a main effect of reactivation type (F(2, 50) = 8.136; p = 0.009), and a one-way ANOVA followed by a Šídák's post hoc test showed that infusions were higher in the FR1 versus pre-VR5 group (p = 0.0003). There were no differences among groups for inactive lever presses (Extended Data Figs. 2-1, 2-2).

During extinction, there were no differences in active lever presses between Veh- and ABC-treated rats in any group (Fig. 2F–H), with decreased responding over session time in all three groups (FR1 reactivation, F(3.196, 41.54) = 7.467; p = 0.0003; pre-VR5 reactivation, F(3.583, 82.40) = 9.779; p < 0.0001; post-VR5 F(3.081, 43.13) = 13.79; p < 0.0001).

During cue reinstatement, ABC treatment prior to VR5 (but not FR1) memory reactivation reduced active lever presses and associated rewards during a 30 min cue reinstatement session. For active lever presses, a two-way ANOVA showed a main effect of reactivation type (F(2, 50) = 9.303; p = 0.0004) and a treatment × reactivation interaction (F(2, 50) = 6.145; p = 0.0041). Pre-VR5 ABC–treated rats had decreased lever pressing versus Veh controls (p = 0.0079) and versus post-VR5 ABC rats (p < 0.0001). For rewards, there was a main effect of the reactivation type (F(2, 50) = 8.612; p = 0.0006) and a treatment × reactivation interaction (F(2, 50) = 4.266; p = 0.0195). Pre-VR5 ABC–treated rats had a decreased number of cue rewards versus Veh controls (p = 0.0233) versus FR1 ABC-treated rats (p = 0.0167) and versus post-VR5 ABC–treated rats (p = 0.0002). For inactive lever presses, there was only a main effect of reactivation (F(2, 50) = 4.461; p = 0.0165; Extended Data Figs. 2-1, 2-2). ABC treatment after memory reactivation had no effect, irrespective of FR1 or VR5 schedule (Fig. 2I,J).

Rats were subsequently tested on a PR schedule of reinforcement to determine whether (1) reduced responding in pre-VR5 ABC–treated animals would persist for additional days and (2) the reinforcing efficacy of cocaine was altered (Richardson and Roberts, 1996) when rats could once again obtain cocaine infusions. For active lever presses, a three-way RM ANOVA showed a main effect of reactivation (F(1, 28) = 4.665; p = 0.039) across days, with a higher reactivation in the FR1 versus pre-VR5 group (the post-VR5 group was not tested; Fig. 2K). There was a similar effect for infusions, with a main effect of reactivation (F(1, 28) = 7.438; p = 0.011), and for inactive lever presses, with a main effect of reactivation (F(1, 28) = 8.844; p = 0.006; Extended Data Figs. 2-1, 2-2). Thus, ABC treatment reduced cue responding only when a novel VR5 reactivation session was combined with ABC treatment in the mPFC prior to VR5 memory reactivation, suggesting that PNNs are necessary for updating or maintaining stability of the cocaine memory (all statistical data are shown in Extended Data Figs. 2-1, 2-2).

ABC in the mPFC does not block cocaine cue-induced reinstatement in the absence of memory reactivation, nor does it block sucrose cue-induced reinstatement

We examined three additional control groups given Veh or ABC treatment. The first group was a “no-reactivation” group that did not receive memory reactivation. The second group self-administered cocaine on an FR3 schedule of reinforcement to test whether higher active lever pressing during the VR5 memory reactivation session (vs the FR1 group; Fig. 2D) could account for better memory recall during reactivation and therefore reduce subsequent cue reinstatement in the ABC treatment group. The third group self-administered sucrose rather than cocaine to determine whether ABC combined with VR5 reactivation would alter responding for a nondrug reward. We used the same timeline as in Figure 2A, with Veh or ABC given 3 d prior to the reactivation session. Brain sites for microinjection are shown in Figure 3, R and S. There were no late training differences among Veh or ABC groups for active lever presses (Fig. 3A,C). As expected, there was an increase in lever pressing in the FR3 versus the no-react group trained on an FR1 schedule. A three-way RM ANOVA showed a main effect of the reactivation type (F(1, 20) = 9.426; p = 0.006), with increased lever pressing in the FR3 group versus no-react groups (trained on an FR1). There were no differences in infusions or inactive lever presses among groups during training days (Fig. 3B,D; Extended Data Figs. 3-1, 3-2). There were also no differences among groups during reactivation (Fig. 3E–H; Extended Data Figs. 3-1, 3-2). During extinction, in all three groups, there was a main effect of time, with reduced responding over the 30 min extinction period, but no effect of ABC treatment (Fig. 3I–K; for the no-react group, active lever, time, F(2.426, 26.69) = 14.01; p < 0.0001; for inactive lever, time, F(2.862, 31.49) = 4.316; p = 0.0127; for the FR3 group, active lever, time, F(2.085, 18.77) = 21.25; p < 0.0001; for the sucrose VR5 group, active lever, time, F(3.630, 43.56) = 10.22; p < 0.0001). There were no effects of ABC or reactivation group on cue-induced reinstatement in any group (Fig. 3L–O). PR responding showed only a react × day interaction when comparing the NR and FR3 groups for inactive lever presses (react × day, F(1.95, 38.97) = 3.362; p = 0.046; Fig. 3P; Extended Data Figs. 3-1, 3-2). For the sucrose VR5 groups, there was a main effect of day for active lever presses (day, F(1.752, 21.02) = 10.87; p = 0.0008) and for sucrose rewards (day, F(1.923, 23.08) = 12.17; p = 0.0003; Fig. 3Q; Extended Data Figs. 3-1, 3-2). Overall, treatment with ABC did not alter any parameters in the three control groups we tested (all statistical data are shown in Extended Data Figs. 3-1, 3-2).

Figure 3.
View larger version:
Figure 3.

ABC treatment in the mPFC does not block cocaine cue-induced reinstatement in the absence of memory reactivation, nor does it block sucrose cue-induced reinstatement after a novel VR5 memory reactivation session. A, Active lever presses during the last 3 d of training in rats trained on an FR1 schedule (no-react and sucrose groups) or an FR3 schedule (FR3; Fig. 2). As expected, there was an increase in lever pressing in the FR3 versus no-react group trained on an FR1 schedule. B, The number of cocaine infusions was not different across the last 3 d of training. C, The number of active lever presses for sucrose pellets was not different across the last 3 d of training (trained on an FR1, reactivated on a VR5). D, The number of sucrose rewards was not different across the last 3 d of training. E–H, During memory reactivation, rats trained on the FR3 schedule were also given an FR3 schedule for memory reactivation, while rats trained for sucrose pellets were given a VR5 schedule for memory reactivation. In the FR3 groups, active lever presses (E) or cocaine infusions (F) were not different between Veh and ABC groups, and in the sucrose VR5 groups, neither active lever presses (G) nor number of sucrose pellets (H) were different between Veh and ABC groups. I–K, Active lever presses during extinction in no react (I), FR3 (J), and sucrose VR5 (K) groups. There was a decrease in responding across the session for all three groups, with no differences between Veh and ABC groups. L–O, During cue reinstatement, there were no differences between Veh and ABC groups in active lever presses (L) or cue rewards (M) in the no-react and FR3 groups. There were also no differences between Veh and ABC groups in active lever presses for sucrose (N) or sucrose pellets (O) in the sucrose VR5 group. P, Q, Active lever presses during PR schedule over 3 consecutive days in no-react, FR3 (P), or sucrose VR5 (Q) groups. There was only a decrease over day for active lever presses and sucrose rewards in the sucrose VR5 group. R, S, Microinjection sites for no-react, FR3 (R), or sucrose VR5 (S). Gray dots, Veh; red dots, ABC. Data are mean ± SEM. No react, Veh N = 7; ABC N = 6; FR3, Veh N = 5; ABC N = 6; sucrose VR5, Veh N = 9; ABC N = 5. Extended Data Figure 3-1 shows the number of inactive lever presses and rewards from animals shown in Figure 2. Extended Data Figure 3-2 shows statistical comparisons for Figure 3.

ABC treatment reduces PV cell immunolabeling intensity

Given the dependence of ABC treatment prior to memory reactivation but not after reactivation on subsequent reduction in cue reinstatement, experiments were done to identify whether ABC treatment at different timepoints would lead to changes in PV neurons. We determined in naive rats whether there were changes in PV staining intensity levels when ABC was delivered 3 d versus 1 h prior to killing animals that might contribute to the differences in cue reinstatement we observed in the pre- versus post-VR5 reactivation groups. Figure 4, A and B, shows WFA-labeled PNNs in naive rats given unilateral microinjections of Veh or ABC into the mPFC. ABC decreased WFA staining intensity levels of not only PNNs but also of WFA-stained loose ECM. Figure 4C shows a double-labeled PV/WFA cell, and Figure 4D shows a double-labeled PV/stub antibody cell. Stub antibody detects remnants of PNNs digested with ABC (Caterson, 2012). Thus, we could measure PV-surrounded neurons with either WFA in Veh animals or with the stub antibody that labeled the “stubs” of PNNs after ABC treatment. Since not all PV cells have PNNs, we assessed the intensity and number of total PV cells, PV cells that were surrounded by WFA after Veh treatment, and PV cells that were surrounded by stub antibody after ABC treatment. When all PV cells were assessed (those with and without PNNs), a two-way ANOVA showed a main effect of treatment (F(1, 2,420) = 36.78; p < 0.0001), with decreases in PV intensity after ABC treatment at both time points (Fig. 4E), while the total number of PV cells was not altered (Fig. 4F). For PV/WFA double-labeled cells, a two-way ANOVA showed a main effect of treatment (F(1, 826) = 4.171; p = 0.0414) and a treatment × time interaction (F(1, 826) = 4.653; p = 0.0313). At 3 d after ABC treatment, PV intensity in PV/WFA double-labeled cells was reduced (p = 0.0082), with no change 1 h after ABC treatment (Fig. 4G).

Figure 4.
View larger version:
Figure 4.

ABC treatment in the mPFC reduces PV cell intensity but does not alter inhibitory/excitatory puncta on PV cells. A, B, Confocal image of WFA-stained brain (green) showing dorsal PFC area after unilateral Veh or ABC treatment at 1 h (A) or 3 d (B) postinjection. Bottom panels show a 20× magnification of the white boxed areas in panels A and B. C, D, Confocal micrographs of single PNN identified by WFA staining (green) surrounding a PV neuron (blue) from Veh side (C) and labeling of “stubs” after ABC treatment identified with 1B5 stub antibody (red) (D). Scale bars for panels A and B, 100 µm, and C and D, 50 µm. E, Intensity of all PV cells was reduced after ABC at both 1 h and 3 d. Number of rats (number of cells) for each group, 1 h, Veh N = 4 (746); ABC N = 4 (546); 3 d, Veh N = 3 (541); ABC N = 3 (561). F, Number of all PV cells was not different after ABC treatment. Number of rats for each group, 1 h, Veh N = 4; ABC N = 4; 3 d, Veh N = 3; ABC N = 3. G, Intensity of PV/WFA or PV/Stub double-labeled cells. Number of rats (number of cells) for each group: 1 h, Veh N = 4 (296); ABC N = 4 (206); 3 d, Veh N = 3 (162); ABC N = 3 (166). H, The number of PV/WFA or PV/stub double-labeled cells was not different across days or treatment. Number of rats for each group, 1 h, Veh N = 4; ABC N = 4; 3 d, Veh N = 3; ABC N = 3. For intensity, violin plots show the median and quartiles; for cell number, data are mean ± SEM. I, Confocal micrograph of a representative PV neuron (red) surrounded by a WFA-labeled PNN (blue). This PV neuron receives appositions from both glutamatergic (VGluT1, magenta) and GABAergic (GAD65/67, green) puncta. J, The Imaris Spots segmentation tool was used to analyze GAD65/67 (green dots) and VGLUT1 (magenta dots) puncta apposing the outer boundaries of the PV neuron (gray). Scale bar, 5 µm. K, GAD 65/67 puncta apposing PV neurons. There was a reduced number of puncta in both Veh and ABC groups at 3 d versus 1 h. L, vGLUT1 puncta apposing PV neurons. There was a reduced number of vGLUT1 puncta in both Veh and ABC groups at 3 d versus 1 h. M, Ratio of GAD65/67 to vGLUT1 puncta apposing PV neurons. There was a reduced ratio in both Veh and ABC groups at 1 h versus 3 d. Data are mean ± SEM. Number of rats (number of cells) for each group, 1 h, Veh N = 4 (16); ABC N = 4 (22); 3 d, Veh N = 3 (18); ABC N = 3 (31). *p < 0.05. Extended Data Figure 4-1 shows statistical comparisons for Figure 4.

Together, these results indicate that 3 d after ABC treatment, PV intensity was more affected in cells that were not originally surrounded by PNNs. The number of PV cells surrounded by PNNs showed a strong trend toward a decrease at both timepoints (p = 0.0625; Fig. 4H). Altogether, these results show that ABC decreased the intensity of PV in neurons surrounded by PNNs 3 d later, but not 1 h later, consistent with others demonstrating a decrease in PV intensity after ABC treatment (Yamada et al., 2015; A. E. Gonzalez et al., 2022; Wang et al., 2023; all statistical data are shown in Extended Data Fig. 4-1).

We also determined whether ABC treatment altered excitatory or inhibitory puncta surrounding PV cells (Fig. 4I–M). In a subset of the animals used for analysis of PV intensity and cell number (above), we quantified the number of GAD 65/67 and vGLUT1-labeled puncta apposing PV neurons on the Veh-treated side and the ABC-treated side. Note that puncta were assessed around PV neurons with WFA in the case of Veh treatment and PV neurons without WFA in the case of ABC treatment; thus, a subset of PV neurons in the latter group may not have been originally surrounded by PNNs. For GAD 65/67 puncta, there was a main effect of time (F(1, 83) = 4.817; p = 0.031), with a reduced number of puncta in both Veh and ABC groups at 3 d versus 1 h (Fig. 4K). Similarly, for vGLUT1 puncta, there was a main effect of time (F(1, 83) = 23.22; p < 0.0001), with a reduced number of vGLUT1 puncta in both Veh and ABC groups at 3 d versus 1 h (Fig. 4L). For the ratio of GAD65/67 to vGLUT1 apposing PV neurons, there was a main effect of time (F(1, 78) = 10.52; p = 0.0017), with a reduced ratio in both Veh and ABC groups at 1 h versus 3 d (Fig. 4M; Extended Data Fig. 4-1). Overall, these findings suggest that the microinjection itself may have altered the number of excitatory and inhibitory puncta over time, possibly due to physical disruptions caused by local microinjections into the brain region. Since we did not assess naive animals without microinjection in this study, we cannot determine the directionality of that change at 1 h postinjection. Therefore, it is unclear whether there was an acute effect of microinjection on the puncta or if more persistent mechanisms, such as inflammatory responses, that led to puncta changes at the later time point (all statistical data are shown in Extended Data Fig. 4-1).

ABC treatment in the mPFC alters intrinsic properties of FSIs

To examine the impact of degrading PNNs on neuronal excitability and intrinsic properties of FSIs, we employed whole-cell electrophysiology techniques on brain slices from naive rats. Figure 5A shows the timeline for assessing the intrinsic properties of FSIs (which contain PV; Kawaguchi and Kubota, 1993) within Layer 5 of the prelimbic mPFC at 1 h or 3 d following in vivo Veh or ABC treatment in the mPFC. Several properties were evaluated, including capacitance, input resistance, resting membrane potential (RMP), AP half-width, AP threshold, AP amplitude, AHP amplitude, first ISI, and rheobase (Fig. 5B). As expected, neuronal excitability in FSIs increased as current steps increased (Fig. 5C), and a three-way RM ANOVA showed a main effect of the current step (F(2.819, 138.1) = 628.7; p < 0.0001) and a time × current step interaction (F(6, 294) = 2.565; p = 0.0195). There was only a trend for an interaction between treatment and current injection on neuronal excitability (p = 0.0995). Figure 5D shows representative traces at 500 pA for the four experimental groups. We found no differences in firing variability among groups (Fig. 5E,F). ABC treatment increased capacitance at both time points, with a main effect of treatment (F(1, 49) = 7.415; p = 0.0089; Fig. 4G), consistent with a previous report (Tewari et al., 2018). In addition, the timing of slice preparation after surgery affected some FSI intrinsic properties. There was higher input resistance at 3 d compared with 1 h after both Veh and ABC treatment, with a main effect of time (F(1, 49) = 13.73; p = 0.0005; Fig. 5H). There were no changes in RMP (Fig. 5I) or AP half-width (Fig. 5J), but ABC treatment decreased the AP threshold at both 1 h and 3 d (Fig. 5K), with a main effect of treatment (F(1, 49) = 5.984; p = 0.0181), suggesting that PNNs may alter the response of PV neurons to incoming signals. For AP amplitude, there was an increase for both groups at 3 d (Fig. 5L), with a main effect of time (F(1, 49) = 5.170; p = 0.0274). We found no increases in AHP amplitude (Fig. 5M), but there was an increase in the first ISI at 1 h, but not 3 d, after ABC treatment (Fig. 5N). A two-way ANOVA showed a time × treatment interaction (F(1, 49) = 9.689; p = 0.0031), with post hoc analysis showing an effect of ABC at 1 h (Veh vs ABC, p = 0.0046) and an effect across time in ABC-treated rats (p = 0.0093), suggesting that ABC treatment delays firing. Rheobase current was reduced at 3 d (Fig. 5O), with a main effect of time (F(1, 49) = 12. 99; p = 0.0007). Collectively, these findings show that ABC treatment altered some firing properties of FSIs that may alter the precise timing of these cells needed for normal circuit function (all statistical data are shown in Extended Data Fig. 5-1).

Figure 5.
View larger version:
Figure 5.

ABC treatment in the mPFC alters intrinsic properties of FSIs. A, Timeline of treatment. B, Description of parameters. C, ABC treatment attenuated neuronal excitability as current steps increased. Number of rats (number of cells), 1 h, Veh N = 7 (15); ABC N = 4 (12); 3 d, Veh N = 10 (17); ABC N = 7 (9). D, Representative traces at 500 pA for the four experimental groups. E, There were no differences in firing variability at 1 h in either treatment group. Number of rats (number of cells), 200 pA, Veh N = 7 (13); ABC N = 4 (11); 500 pA, Veh N = 7 (15); ABC N = 4 (12). F, There were no differences in firing variability at 3 d in either treatment group. Number of rats (number of cells), 200 pA, Veh N = 10 (17); ABC N = 7 (10); 500 pA, Veh N = 10 (17); ABC N = 7 (10). G, ABC treatment increased membrane capacitance at both 1 h and 3 d. H, Input resistance was increased at 3 d versus 1 h. I, There were no differences in RMP at 1 h or 3 d in either treatment group. J, There were no differences in the AP half-width at 1 h or 3 d in either treatment group. K, ABC treatment reduced the AP threshold at both 1 h and 3 d. L, The AP amplitude was increased at 3 d versus 1 h in both groups. M, There were no differences in the AHP amplitude at 1 h or 3 d in either treatment group. N, ABC treatment increased the first ISI time at 1 h, but not at 3 d. O, Rheobase current was reduced at 3 d versus 1 h. Data are mean ± SEM. For (H–O) the number of rats (number of cells), 1 h, Veh N = 7 (15); ABC N = 4 (12); 3 d, Veh N = 10 (17); ABC N = 7 (9). *p < 0.05. Extended Data Figure 5-1 shows statistical comparisons for Figure 5.

ABC treatment does not alter GluN2B receptors in the mPFC or amygdala after novel VR5 reactivation session

The blockade of GluN2B receptors prevents the ability to render labile a reactivated memory for disruption by amnestic agents (Mamou et al., 2006; Yu et al., 2016; Radiske et al., 2021). More recent evidence suggests that GluN2B receptors may only play a role in destabilizing more recent (nonreward) memories and may not support novelty-induced destabilization (Wideman et al., 2020). Nevertheless, we determined whether FR1 or novel VR5 memory reactivation in our cocaine-exposed rats altered the level of surface or intracellular GluN2B subunits in the mPFC and amygdala (Mamou et al., 2006) immediately after the 30 min memory reactivation session (Fig. 6A–E). We found no differences in surface or intracellular GluN2B subunits in either brain area, nor did we find differences between treatment groups in the ratio of surface:intracellular GluN2B levels (data not shown). GluN2B receptor subunit trafficking may therefore not be sensitive to the differences in FR1 versus VR5 memory reactivation, or the changes may be too small to measure when assessing tissue punches (all statistical data are shown in Extended Data Fig. 6-1).

Figure 6.
View larger version:
Figure 6.

GluN2B levels are not altered in mPFC or amygdala immediately after novel VR5 memory reactivation session. A, mPFC Western blots showing surface and intracellular GluN2B immediately after the 30 min FR1 or VR5 memory reactivation session. B, mPFC surface and intracellular mGluN2B in the absence or presence of cross-linking with BS3. C, Quantification of mPFC surface and intracellular GluN2B levels after the FR1 or VR5 memory reactivation session. D, Amygdala Western blots showing surface and intracellular GluN2B immediately after the 30 min FR1 or VR5 memory reactivation session. E, Quantification of amygdala GluN2B levels after the FR1 or VR5 reactivation session. For FR1 N = 6; for VR5 N = 7. Extended Data Figure 6-1 shows statistical comparisons for Figure 6.

Increased mPFC beta and gamma power after novel VR5 reactivation session is prevented by ABC in the mPFC during rewarded lever presses

To determine the extent to which the novel VR5 reactivation session altered LFPs within the mPFC or coupling with the hippocampus and whether ABC would modify this coupling, we measured LFPs within the mPFC and dHIP on the last day of self-administration training and during the VR5 reactivation session in a subset of animals. All animals were given ABC 3 d prior to the VR5 reactivation as in Figure 2A (pre-VR5 group). Figure 7A shows the raw and filtered traces for LFPs in the mPFC. We analyzed time–frequency windows based on the first several hundred milliseconds after the lever press/cue/reward presentation to assess early effects of processing in which stimulus awareness occurs, starting at ∼200 ms (Ehret and Romand, 2022). We chose the time window cutoff of 600 ms to examine the impact of lever press/cue/reward events independent of elevations in brain cocaine levels, which take longer to appear than the period we examined (Wise and Kiyatkin, 2011). In addition, we wished to separate beta and gamma activity based on previous literature demonstrating the different frequencies associated with novelty or other information processing (Marco-Pallares et al., 2015; Franca et al., 2021; Han et al., 2021).

Figure 7.
View larger version:
Figure 7.

ABC treatment in the mPFC prevents increases in beta/gamma power for rewarded lever presses during novel VR5 memory reactivation session. A, Raw and filtered LFP traces from mPFC. B, Spectrograms of mPFC power in the 10–100 Hz range (top panel) and 0–25 Hz range (bottom panel) after rewarded lever presses on the last training day. C, Spectrograms of mPFC power in Veh rats in the 10–100 Hz range (top panel) and 0–25 Hz range (bottom panel) after rewarded lever presses on the VR5 reactivation session. D, Spectrograms of mPFC power in ABC rats in the 10–100 Hz range (top panel) and 0–25 Hz range (bottom panel) after rewarded lever presses on the VR5 reactivation session. For B–D, color bar is relative power. E, Quantification of 60–90 Hz power (top white square in 5C, top panel) in the mPFC 200–600 ms after rewarded lever presses during last day training versus VR5 reactivation session in Veh rats. VR5 reactivation did not alter 60–90 Hz power compared with the last training day. Note that ABC was not delivered until after the last training session, and therefore only Veh animals are shown for the last training day for direct comparison with the Veh group on the VR5 reactivation session. F, Quantification of 30–60 Hz power (middle white rectangle in 5C, top panel) in the mPFC 200–600 ms after rewarded lever presses showed no differences in power compared with the last training day. G, Quantification of 15–30 Hz power (lower white rectangle in 5C, top panel) in the mPFC 200–600 ms after rewarded lever presses showed VR5 reactivation increased 15–30 Hz power compared with the last training day. There were no differences in the 2–4 Hz range at 200–600 ms after rewarded lever presses among groups. H, Quantification of 60–90 Hz power (top white square in 5C, top panel) in the mPFC 200–600 ms after rewarded and unrewarded lever presses. Gamma power was higher in unrewarded versus rewarded lever presses. I, Quantification of 30–60 Hz power (middle white rectangle in 5C, top panel) in the mPFC 200–600 ms after rewarded and unrewarded lever presses. Rewarded lever presses were accompanied by increased power in Veh- versus ABC-treated rats. J, Quantification of 15–30 Hz power (bottom white rectangle in 5C, top panel) in the mPFC 200–600 ms after rewarded and unrewarded lever presses showed no differences between Veh- and ABC-treated rats. K, Location of electrodes in a representative mPFC (top-left panel) and dHIP (top-right panel). Location of all electrodes in the mPFC (bottom-left panel) and location of all electrodes in the hippocampus, with exception of one electrode site that was not identified in one ABC rat (bottom-right panel). Gray dots, Veh; red dots, ABC. Number of rats (number of trials), Veh N = 5 (16 rewarded, 70 unrewarded); ABC N = 3 (17 rewarded, 71 unrewarded). *p < 0.05; #p < 0.06. Extended Data Figure 7-1 shows electrode/microinjection sites for Veh- and ABC-treated animals used for in vivo electrophysiology studies shown in Figure 7. Extended Data Figure 7-2 shows statistical comparisons for Figure 7.

Spectrograms from the mPFC show that novel VR5 reactivation did not alter high gamma (60–90 Hz; Fig. 7B,C, top panels, top rectangle; Fig. 7E) or low gamma activity (30–60 Hz; Fig. 7B,C, top panels, middle rectangle; Fig. 7F) after active lever presses compared with the familiar last day of self-administration training. However, the novel VR5 reactivation session increased high beta power during rewarded lever presses (15–30 Hz; p = 0.0244; Fig. 7B,C, top panels, bottom rectangle; Fig. 7G). In addition, during the novel VR5 reactivation session, high gamma power was increased after unrewarded versus rewarded lever presses, with a main effect of lever press type (F(1, 170) = 4.486; p = 0.0356; Fig. 7H). For low gamma power, a two-way ANOVA showed a significant treatment × lever press interaction (F(1, 170) = 7.412; p = 0.0072; Fig. 7I), and subsequent post hoc analyses revealed that low gamma power was decreased in the ABC compared with Veh group following rewarded (but not unrewarded) presses (p = 0.0313). There were no changes in high beta power between Veh and ABC rats for rewarded or unrewarded lever presses (Fig. 7J). The visually apparent increase in delta power (2–4 Hz) from Veh-treated animals after VR5 reactivation (Fig. 7C, bottom panel) was not significantly different among any groups and was due to a few trials showing high delta power in the Veh group. The locations of electrodes in a representative mPFC and dHIP are shown in Figure 7K, and all Veh and ABC microinjection sites are shown in Extended Data Figure 7-1. Overall, a novel VR5 reactivation elevated beta activity in response to rewarded lever presses compared with the familiar last day of training, and ABC treatment prevented an elevation in low gamma activity during rewarded lever presses on the novel VR5 reactivation session. Thus, distinct oscillations may carry different information about expected reward and reward processing (all statistical data are shown in Extended Data Fig. 7-2).

Increased dHIP/mPFC coupling during novel VR5 reactivation session is reduced by ABC treatment in the mPFC

Given that PV neurons coordinate long-range communication within and across other brain regions through coupling of theta and gamma oscillations (Buzsaki and Draguhn, 2004), we also examined cross-frequency PAC between dHIP theta phase and mPFC gamma amplitude. Comodulograms 200–600 ms after rewarded lever presses are shown for Veh rats during the last day of training (Fig. 8A), for Veh rats during the VR5 reactivation session (Fig. 8B), and for ABC rats during the VR5 reactivation session (Fig. 8C). Coupling was apparent between dHIP 8–10 Hz theta phase and mPFC 50–70 Hz gamma frequency in the Veh group during the VR5 reactivation session (Fig. 8B). This same apparent coupling was absent on the last day of training in the same Veh animals (Fig. 8A). In addition, on the VR5 reactivation session, ABC reduced this coupling (Fig. 8C). Based on the comodulograms, we tested for changes in the distribution of gamma power (50–70 Hz) over theta phase (8–10 Hz). During the last training day, there was no significant dHIP theta phase modulation of mPFC gamma power 200–600 ms after rewarded lever presses (MI = 1.92 × 10−4; z = −0.77; Fig. 8D). In contrast, during the novel VR5 session, there was significant dHIP theta phase modulation of mPFC gamma power 200–600 ms after rewarded lever presses in Veh rats (MI = 6.59 × 10−3; z = 2.12; Fig. 8E). ABC treatment appeared to reduce, but not completely prevent, significant dHIP modulation of mPFC gamma power compared with Veh controls (MI = 3.43 × 10−3; z = 1.89; Fig. 8F). Thus, the novel VR5 reactivation session was associated with coupling between the dHIP and mPFC, and ABC treatment dampened this coupling. The reduced coupling in ABC-treated rats was not likely due to differences in lever-pressing activity in general, since the two groups did not differ in lever pressing during the VR5 reactivation session (Fig. 2D).

Figure 8.
View larger version:
Figure 8.

Increased dHIP/mPFC coupling and directional flow between the mPFC and dHIP during novel VR5 memory reactivation is prevented by ABC treatment in the mPFC. A, Veh rats on last training day, comodulogram showing dHIP theta phase and mPFC amplitude 200–600 ms after rewarded lever presses. B, Veh rats on VR5 reactivation session, comodulogram showing dHIP theta phase and mPFC amplitude 200–600 ms after rewarded lever presses. Note increased coupling of dHIP theta in the 8–10 Hz range with mPFC gamma in the 50–70 Hz range. C, ABC rats on VR5 reactivation session, comodulogram showing dHIP theta phase and mPFC amplitude 200–600 ms after rewarded lever presses. Note the absence of coupling observed in Veh rats. The color bar is z-score; comodulograms are floored at p < 0.05. D, Veh rats on the last training day, the plot of mPFC gamma power (50–70 Hz) binned by dHIP theta phase (8–10 Hz) 200–600 ms after rewarded lever presses. There was no significant modulation of mPFC gamma power by dHIP theta. E, Veh rats on VR5 reactivation session, the plot of mPFC gamma power (50–70 Hz) binned by dHIP theta phase (8–10 Hz) 200–600 ms after rewarded lever presses. dHIP theta significantly modulated mPFC gamma power after rewarded lever presses. F, ABC rats on VR5 reactivation session, the plot of mPFC gamma power (50–70 Hz) binned by dHIP theta phase (8–10 Hz) 200–600 ms after rewarded lever presses. dHIP theta significantly modulated mPFC gamma power. G, GC analysis computed from the raw LFP signal in the dHIP and mPFC in 500 ms bins for rewarded lever presses prior to lever press (−750 to −250 ms, baseline), surrounding lever press (−250 to 250 ms), and after lever press (250–750 ms) during VR5 memory reactivation session. H, GC normalized to its own averaged baseline for Veh or ABC groups during rewarded lever press and postlever press. The mPFC leads the dHIP around rewarded lever presses but disappears by 250–750 ms after lever press. I, GC analysis as in G for unrewarded lever presses. J, Normalized GC to its own averaged baseline showing no changes between Veh and ABC groups. K, L, Cross-correlation between mPFC/dHIP theta (4–12 Hz) amplitude envelopes at varying lags in Veh (K) and ABC (L) groups. Stars show peak correlation (mean ± SEM, Veh = 43.5 ± 0.03 ms; ABC = −0.5 ± 0.003 ms). The line is mean; shading represents SEM. M, N, Normalized heat maps of theta amplitude cross-correlation by trial from rewarded trials in Veh- and ABC-treated rats in Veh (M) and ABC (N) groups. The darker red color indicates higher cross-correlation. O, Lag at peak correlation, with a significant difference in distributions between the Veh and ABC groups. Box plot shows median and 95% confidence intervals. P, Normalized correlation for Neg and Pos lag, with Pos lag indicating mPFC leading the dHIP. There was a significant lead by the mPFC in the Veh but not in the ABC group. Number of rats (number of trials), Veh N = 4 (12 rewarded, 59 unrewarded); ABC N = 3 (17 rewarded, 71 unrewarded). *p < 0.05. Extended Data Figure 8-1 shows statistical comparisons for Figure 8.

Functional interaction between the dHIP and mPFC was assessed using GC analysis to measure the direction and extent to which information flowed between the mPFC and dHIP during rewarded and unrewarded lever presses on the VR5 session (Fig. 8G–J). GC analysis suggested a significant flow of information from the mPFC to the dHIP during rewarded lever presses in the Veh but not in the ABC group. A two-way RM ANOVA shows a main effect of treatment (F(1, 27) = 4.892; p = 0.0356), time (F(1, 27) = 6.818; p = 0.0146), and a treatment × time interaction (F(1, 27) = 13.20; p = 0.0012), and post hoc analysis shows a decrease in ABC versus Veh rats surrounding the time of rewarded lever presses (p = 0.0010; Fig. 8G,H). In contrast, there were no differences during unrewarded lever presses (Fig. 8I,J). This mPFC lead appeared just prior to the rewarded lever press and disappeared within 250 ms after the lever press. To confirm these findings, we computed cross-correlation between mPFC/dHIP theta (4–12 Hz) amplitude envelopes at varying lags in Veh and ABC groups (Fig. 8K,L). Figure 8, M and N, shows normalized heat maps of theta amplitude cross-correlations by trial from rewarded trials in Veh- and ABC-treated rats, and Figure 8O shows the distribution of lag at peak correlation, which was different between the ABC and Veh groups (p = 0.0111). Figure 8P shows a higher cross-correlation for the Pos lag (mPFC lead) compared with the Neg lag (dHIP lead). There was a main effect of lag (F(1, 54) = 5.241; p = 0.0260) and a significant treatment × lag interaction (F(1, 54) = 6.477; p = 0.0138). Post hoc analysis showed a significant lead by the mPFC in the Veh (p = 0.0052), but not in the ABC group. Thus, consistent with GC analysis, ABC disrupted the ability of the mPFC to lead the dHIP during rewarded lever presses. These findings suggest that PNNs influence directional connectivity from the mPFC to dHIP during cocaine-seeking/cocaine-taking behavior (all statistical data are shown in Extended Data Fig. 8-1).

We also assessed dHIP theta/mPFC gamma coupling during unrewarded lever presses. Unlike for rewarded lever presses, there was no significant modulation of mPFC 50–70 Hz gamma by dHIP 8–10 Hz theta in either the Veh or ABC group (Fig. 9A–D). A plot of mPFC gamma power binned by dHIP theta phase revealed that there was no significant dHIP modulation of mPFC gamma power in either Veh or ABC rats (MI = 4.89 × 10−4; z = 0.24 and MI = 1.50 × 10−4; z = −0.86, respectively). The normalized correlation for Neg and Pos lag (with Pos lag indicating mPFC leading the dHIP) showed a main effect of treatment (F(1, 256) = 9.736; p = 0.0020), with a lower correlation in the Veh compared with ABC groups (Fig. 9E–G), but there was no significant lead by the mPFC or dHIP in either group. Lag analysis of mPFC/dHIP theta amplitude cross-correlation indicated that there was a significant difference in distributions of peak theta lag between the Veh and ABC groups (p = 0.0028; Fig. 9H). Thus, unlike for rewarded lever presses, we did not see dHIP/mPFC coupling or directional flow from the mPFC to dHIP for unrewarded lever presses. All statistical data for Figure 9 are shown in Extended data Figure 9-1.

Figure 9.
View larger version:
Figure 9.

Lack of dHIP modulation of mPFC for unrewarded lever presses during novel VR5 memory reactivation, but ABC alters interhemispheric theta/gamma coupling across the mPFC. A, Vehicle rats, comodulogram showing dHIP theta phase and mPFC amplitude 200–600 ms after unrewarded lever presses. B, ABC rats, comodulogram showing dHIP theta phase and mPFC amplitude 200–600 ms after unrewarded lever presses. The color bar is z-score; comodulograms are floored at p < 0.05. C, D, Vehicle (C) and ABC (D) rats, plot of mPFC gamma power (50–70 Hz) binned by dHIP theta phase (8–10 Hz) 200–600 ms after unrewarded lever presses. dHIP theta did not modulate mPFC gamma power in either Veh or ABC rats. E, F, Cross-correlation between mPFC/dHIP theta (4–12 Hz) amplitude envelopes at varying lags in Veh (E) and ABC (F) groups during VR5 memory reactivation session. Stars show peak correlation (mean ± SEM, Veh = 34.5 ± 0.005 ms; ABC = 1.5 ± 0.003 ms). Number of rats (number of trials), Veh N = 4 (59); ABC N = 3 (71). G, Normalized correlation for Neg and Pos lag, with Pos lag indicating mPFC leading the dHIP. There was a decreased correlation in the Veh group, but no lead by mPFC in either group. H, Lag at peak correlation, with a significant difference in distributions between the Veh and ABC groups. I, Veh rats, comodulogram showing left mPFC theta phase and right mPFC amplitude 200–600 ms after unrewarded lever presses. J, ABC rats, comodulogram showing left mPFC theta phase and right mPFC amplitude 200–600 ms after unrewarded lever presses. The color bar is z-score; comodulograms are floored at p < 0.05. K, Vehicle rats, plot of mPFC gamma power (70–90 Hz) binned by dHIP theta phase (8–10 Hz) 200–600 ms after unrewarded lever presses. Left mPFC theta significantly modulates right mPFC gamma power. L, ABC rats, the plot of mPFC gamma power (70–90 Hz) binned by dHIP theta phase (8–10 Hz) 200–600 ms after unrewarded lever presses. Left mPFC theta does not modulate right mPFC gamma power after lever presses. M, Veh rats, the plot of mPFC gamma power (70–90 Hz) binned by dHIP theta phase (5–7 Hz) 200–600 ms after unrewarded lever presses. Left mPFC theta does not modulate right mPFC gamma power. N, ABC rats, the plot of mPFC gamma power (70–90 Hz) binned by dHIP theta phase (5–7 Hz) 200–600 ms after unrewarded lever presses. Left mPFC theta modulates right mPFC gamma power. O, P, Cross-correlation between left mPFC/right mPFC theta (4–12 Hz) amplitude envelopes at varying lags in Veh (O) and ABC (P) groups. Stars show peak correlation (mean ± SEM, Veh = −5.5 ± 0.004 ms; ABC = −4.5 ± 0.003 ms). The line is mean; shading represents SEM. Q, Normalized correlation for Neg and Pos lag, with Pos lag indicating left mPFC leading the right mPFC. There was no effect of ABC treatment. R, Lag at peak correlation, with no difference in distributions between Veh and ABC groups. The box plot shows median and 95% confidence intervals. *p < 0.05. Extended Data Figure 9-1 shows statistical comparisons for Figure 9.

ABC treatment alters interhemispheric mPFC theta/gamma coupling after unrewarded lever presses during novel VR5 reactivation session

Given that PV interneurons support cross-hemispheric coupling via direct callosal projections (Cho et al., 2020), we also examined the effect of ABC treatment on interhemispheric theta gamma coupling across the mPFC during the novel VR5 session. The most apparent differences in coupling were found after unrewarded lever presses. Figure 9, I and J, shows comodulograms during unrewarded lever presses from Veh and ABC groups, respectively. During the VR5 reactivation session, there was significant cross-hemispheric mPFC 8–10 Hz theta phase modulation of mPFC 70–90 Hz gamma power 200–600 ms after unrewarded lever presses in Veh rats during the VR5 session (MI = 2.38 × 10−3; z = 3.18; Fig. 9K), and ABC blocked modulation in this frequency range (MI = 1.67 × 10−4; z = −0.71; Fig. 9L). However, we also found significant cross-hemispheric mPFC modulation of gamma by theta in the 5–7 Hz range in the ABC group (MI = 3.04 × 10−3; z = 4.81) but not in the Veh group (MI = 1.62 × 10−3; z = 0.41; Fig. 9M,N). Lag analysis of cross-hemispheric mPFC theta amplitude envelope correlations revealed a main effect of lag (F(1, 232) = 10.57; p = 0.0013), with no effect of ABC treatment (Fig. 9O–Q). There were no differences in the distribution of lag at peak correlation between Veh- and ABC-treated groups (Fig. 9R). These results indicate that unrewarded lever presses may be associated with interhemispheric mPFC coupling, with a different theta frequency modulating gamma activity in Veh- compared with ABC-treated rats. All statistical data are shown in Extended Data Figure 9-1.

Previous SectionNext Section

Discussion

These studies established the impact of ABC treatment on both neuron- and network-level changes in the mPFC associated with cocaine memory reconsolidation. Our major findings are that ABC treatment in the mPFC prior to training for cocaine self-administration attenuated the acquisition of cocaine self-administration. Most strikingly, in rats already trained for cocaine self-administration, ABC treatment prior to, but not after, a novel VR5 reactivation session strongly blocked cocaine cue reinstatement, and its impact was specific for cocaine without affecting sucrose cue reinstatement. Moreover, in vivo recordings during the novel VR5 reactivation session revealed that rewarded lever presses increased high-frequency oscillations in the mPFC and the coupling of oscillations between the mPFC and dHIP, which were reduced after ABC treatment. Finally, the directional flow of signaling from the mPFC to the dHIP during memory reactivation was also impaired after ABC treatment. Overall, our findings suggest that mPFC/dHIP interregional communication is associated with restabilizing networks during cocaine memory updating and may serve as a signature for memory reconsolidation.

ABC treatment in the dorsal mPFC attenuates acquisition of cocaine self-administration, blocks cue-induced reinstatement, and reduces reinforcing efficacy of cocaine

Similar to our previous study using cocaine-induced CPP (M. Slaker et al., 2015), here we found that ABC treatment reduced the acquisition of cocaine self-administration in rats. We also had previously shown that removing PNNs with ABC in the mPFC prior to memory reactivation blunted the reconsolidation of a cocaine–CPP memory. However, a self-administration memory entails a much greater number of pairings between the conditioned and unconditioned stimulus than does CPP, and memory reconsolidation is difficult to disrupt in well-established memories (Alberini, 2005; Frankland et al., 2006). Indeed, in the present study, we found no effect of ABC treatment in animals trained and given a memory reactivation session under the same reinforcement schedule. Of note, over several weeks, PNNs slowly return in number after ABC treatment (M. Slaker et al., 2015) but not in intensity (M. Slaker et al., 2015; Lensjo et al., 2017) and therefore are not likely to resume normal functioning even several weeks later (Romberg et al., 2013).

To disrupt stronger memories, exposure to novel information and subsequent memory updating may be necessary to destabilize and interfere with preexisting memories (Diaz-Mataix et al., 2013; Sevenster et al., 2013; Exton-McGuinness et al., 2014). A novel component during memory reactivation may update memories by creating a prediction error (Sara, 2000; Winters et al., 2009; Dudai, 2012; Sevenster et al., 2013; J. L. C. Lee et al., 2017). Previous work showed that reconsolidation of well-learned memories for cocaine self-administration were disrupted by changing the reinforcement schedule during memory reactivation (Exton-McGuinness and Lee, 2015). An unpredictable component may be key to updating strong memories (Dunbar and Taylor, 2017; Xue et al., 2017; Zhu et al., 2018; M. C. Gonzalez et al., 2021). A. J. Park et al. (2021) showed that exposure to novelty resulted in the initial decoupling of the ventral hippocampus from the mPFC and subsequent repotentiation of this circuit. ABC treatment in the current study may similarly prevent one or both of these processes. In support of this, novel VR5 memory reactivation increased coupling between the dHIP and mPFC that was absent on the last day of self-administration, when the memory was familiar. Thus, coupling between these brain regions may be a marker for the circuit impact of novelty and subsequent destabilization and updating of cocaine-associated memories.

The lack of effect of PNN depletion on responding during the memory reactivation session suggests that ABC treatment does not rely on the mPFC for initial retrieval, but only for subsequent postreactivation retrieval. The presence of cocaine during the reactivation session is expected to increase the firing of mPFC neurons (West et al., 2014; Moorman and Aston-Jones, 2023), including PV neurons (Tseng et al., 2006), which can offset reduced GABAergic inhibition that is needed to stabilize the network (Lew and Tseng, 2014). Thus, increased dopamine may have promoted retrieval by overriding the impact of PNN/ECM removal on the retrieval process. Importantly, the effect of ABC and VR5 reactivation persisted under a PR schedule of reinforcement in which cocaine was available, indicating that ABC treatment after a novel VR5 reactivation likely reduces the reinforcing efficacy of cocaine and that the diminished responding is not likely due to state-dependent effects. Our findings are in agreement with those from a study in which ABC treatment in the entorhinal cortex disrupted the grid cell network for spatial memory, but when animals were placed into a novel environment to form a new memory, ABC treatment degraded spatial representation of both the new and old memories (Christensen et al., 2021). Similar to their findings and others (Pizzorusso et al., 2002; Lensjo et al., 2017), ABC treatment in the mPFC did not cause changes in network function (in our studies assessed by reinstatement behavior) until the system was required to process new information. Thus, processing of new information during the VR5 session when PNNs are absent may recruit additional neurons into new neural ensembles, as PNN removal increases the number of neurons recruited into an ensemble, leading to the loss of specific, sparse ensembles underlying engrams (Ramsaran et al., 2023). The recruitment of additional neurons may interfere with the original ensemble supporting the familiar memory or the ability to retrieve that memory. The lack of an effect of ABC treatment in animals self-administering sucrose is consistent with the idea that cocaine induces a state of metaplasticity (Neuhofer and Kalivas, 2018) that sucrose does not and that ABC treatment impacts only cocaine-induced metaplasticity when examined in this paradigm.

The observation that ABC treatment prior to but not after memory retrieval reduced cue reinstatement suggests that ABC treatment altered activity-dependent plasticity of PV cells and/or the circuit in which they are embedded through changes in PV cell activity (Donato et al., 2015; Favuzzi et al., 2017). Our slice electrophysiology findings support the notion that PV neuron function was altered after ABC treatment, with an increased capacitance, in agreement with previous studies (Tewari et al., 2018; Hanssen et al., 2023). Notably, while we observed only a trend toward reduced AP frequency of FSIs after ABC treatment, we assessed PV neuron function in naive rats with no history of cocaine exposure. We previously reported that ABC treatment in the mPFC profoundly suppressed AP frequency in FSIs 2 h after cocaine-induced reinstatement of CPP (A. E. Gonzalez et al., 2022). Thus, memory reactivation in the presence of cocaine in the current study may have reduced PV neuron firing. Decreased PV intensity levels at 3 d prior to but not 1 h after ABC treatment suggests the possibility of suboptimal PV functioning due to imprecise firing of these neurons. As PV neurons and gamma oscillations critically contribute to mPFC output during rule switching (Sohal et al., 2009; Cho et al., 2015), their inability to optimally function may have compromised updating of the cocaine memory. Retrieval under novel conditions, potentially combined with the presence of cocaine, which destabilizes the PFC network (Borzou et al., 2024), may create instability in the mPFC circuit required for reconsolidation, which we show is dependent on intact mPFC PNNs at the time of memory reactivation.

Interregional coherence and reconsolidation

Several studies have shown coupling between the dHIP and mPFC in a variety of tasks that involve spatial or working memory (Jones and Wilson, 2005; Siapas et al., 2005; Benchenane et al., 2010; Tang et al., 2021; Domanski et al., 2023). In the current study, strong coupling between dHIP theta and mPFC gamma did not appear necessary for memory retrieval, since ABC treatment decreased coupling but did not alter behavior during VR5 memory reactivation. Coupling between dHIP/mPFC may instead be associated with destabilization of memory (A. J. Park et al., 2021). Our data support the idea that Veh controls were able to process the VR5 session as novel, whereas rats given ABC likely did not: Veh animals showed elevated beta and gamma activity during rewarded lever presses during the VR5 session, and elevated high-frequency activity has been described for responses to novelty or unexpected rewards (Marco-Pallares et al., 2015; Franca et al., 2021). The direction of information flow from the mPFC to dHIP during rewarded lever presses was also lost after ABC treatment. In addition, ABC treatment altered theta/gamma coupling across mPFC hemispheres during unrewarded lever presses, with a different theta frequency modulating gamma amplitude in Veh- versus ABC-treated rats. Overall, our findings with a novel VR5 reactivation session in which delivery of the reward is less predictable suggest that increased synchrony between the mPFC and dHIP during rewarded lever presses and increased specific theta frequency modulation of gamma across mPFC hemispheres during unrewarded lever presses may serve as error signals, which are reduced in ABC-treated rats.

Limitations

ABC degrades chondroitin sulfate glycosaminoglycans on chondroitin sulfate proteoglycans (Yamagata et al., 1968) but also degrades the loose ECM (Deepa et al., 2006). In our studies, we cannot rule out any potential effects of ABC on the removal of loose ECM (ECM not condensed to form PNNs), and it is possible that the loose ECM mediates decreased cue reinstatement. For example, loose ECM contributes to the tetrapartite synapse (Dityatev et al., 2010) to regulate long-term potentiation, the molecular correlate of memory (Nagy et al., 2006; Dityatev and Rusakov, 2011). The ECM is modified in part by matrix metalloproteinases (MMPs), which modify spine morphology during memory formation (Szklarczyk et al., 2002) and which importantly contribute to drug-associated behaviors, including cocaine-induced behaviors (Brown et al., 2007, 2008; Shen et al., 2014; Smith et al., 2014; Mulholland et al., 2016). Some of these effects appear to occur through integrin signaling to alter transient synaptic potentiation in the nucleus accumbens core (Garcia-Keller et al., 2019), a key target for the mPFC output. Our findings could therefore alter mPFC function locally through changes in MMP/integrin signaling independent of PNNs. Although the vast majority of PNN functional studies have used the ABC enzymatic approach to reduce PNNs, future studies need to use more targeted strategies to impair PNN function while leaving the loose ECM intact; we are currently working on these strategies. Nevertheless, some studies show that PNNs appear to be a major contributor to altered plasticity after ABC in the handful of studies conducted in knock-out mice with depletion of PNN components; these mice still show enhanced plasticity (Apostolova et al., 2006; Rowlands et al., 2018), and ABC treatment in knock-out mice lacking the PNN-specific protein Crtl1 showed no further changes in object recognition memory (Romberg et al., 2013).

An additional limitation is that we used only male rats. Acquisition and reinstatement in cocaine self-administering rats is higher in females (Lynch and Carroll, 1999; Becker and Koob, 2016), and we and others have reported sex differences in PNN intensity in reward-associated brain regions (Griffiths et al., 2019; Dingess et al., 2020; Gildawie et al., 2020; Ciccarelli et al., 2021; Zhang et al., 2021).

Previous SectionNext Section

Conclusions

Our findings strongly suggest that ABC treatment in the mPFC with ABC prevents updating of a cocaine-associated memory, but only when the reactivated memory is destabilized by introducing a novel component to the experience. If PNNs are removed after memory reactivation, reconsolidation is still successful. The novel reactivation session increases beta and gamma activity in the mPFC, and disrupted synchrony between the mPFC and dHIP may prevent memory updating in animals lacking PNNs in the mPFC. Collectively, our findings support the idea that a novel component presented during cocaine memory reactivation increases high-frequency oscillations in the mPFC, the coupling of these oscillations between the dHIP and mPFC, and that a PNN-dependent stable network is needed for updating that memory.

Previous SectionNext Section

Footnotes

    • Received March 11, 2024.
    • Revision received June 20, 2024.
    • Accepted June 25, 2024.

  • We thank Ryan Todd and Drs. Jordan Blacktop, Megan Slaker, and Lynn Churchill for their assistance with the acquisition studies. This work was supported by National Institutes of Health DA 040965 (B.A.S., T.E.B., and S.A.A.) and DA 055645 (B.A.S., T.E.B., S.A.A., and A.I.A.), Oregon Health and Science University Physician Scientist Award, Portland Veterans Affairs Research Foundation (A.I.A.), and the Good Samaritan Foundation of Legacy Health (B.A.S.).

  • The authors declare no competing financial interests.

  • ↵*J.C.W. and J.D.R. are equal first authors.

  • Correspondence should be addressed to Barbara A. Sorg at bsorg{at}downeurobiology.org.
Previous Section
 

References

  1. ↵
    1. Adhikari A,
    2. Sigurdsson T,
    3. Topiwala MA,
    4. Gordon JA
    (2010) Cross-correlation of instantaneous amplitudes of field potential oscillations: a straightforward method to estimate the directionality and lag between brain areas. J Neurosci Methods 191:191–200. https://doi.org/10.1016/j.jneumeth.2010.06.019 pmid:20600317
    CrossRefMedline
  2. ↵
    1. Alberini CM
    (2005) Mechanisms of memory stabilization: are consolidation and reconsolidation similar or distinct processes? Trends Neurosci 28:51–56. https://doi.org/10.1016/j.tins.2004.11.001
    CrossRefMedline
  3. ↵
    1. Apostolova I,
    2. Irintchev A,
    3. Schachner M
    (2006) Tenascin-R restricts posttraumatic remodeling of motoneuron innervation and functional recovery after spinal cord injury in adult mice. J Neurosci 26:7849–7859. https://doi.org/10.1523/JNEUROSCI.1526-06.2006 pmid:16870730
    Abstract/FREE Full Text
  4. ↵
    1. Becker JB,
    2. Koob GF
    (2016) Sex differences in animal models: focus on addiction. Pharmacol Rev 68:242–263. https://doi.org/10.1124/pr.115.011163 pmid:26772794
    Abstract/FREE Full Text
  5. ↵
    1. Benchenane K,
    2. Peyrache A,
    3. Khamassi M,
    4. Tierney PL,
    5. Gioanni Y,
    6. Battaglia FP,
    7. Wiener SI
    (2010) Coherent theta oscillations and reorganization of spike timing in the hippocampal- prefrontal network upon learning. Neuron 66:921–936. https://doi.org/10.1016/j.neuron.2010.05.013
    CrossRefMedline
  6. ↵
    1. Borzou A,
    2. Miller SN,
    3. Hommel JD,
    4. Schwarz JM
    (2024) Cocaine diminishes functional network robustness and destabilizes the energy landscape of neuronal activity in the medial prefrontal cortex. PNAS Nexus 3:pgae092. https://doi.org/10.1093/pnasnexus/pgae092 pmid:38476665
    Medline
  7. ↵
    1. Boudreau AC,
    2. Milovanovic M,
    3. Conrad KL,
    4. Nelson C,
    5. Ferrario CR,
    6. Wolf ME
    (2012) A protein cross-linking assay for measuring cell surface expression of glutamate receptor subunits in the rodent brain after in vivo treatments. Curr Protoc Neurosci Chapter 5:Unit 5.30.1-19. https://doi.org/10.1002/0471142301.ns0530s59
    Google Scholar
  8. ↵
    1. Brown TE,
    2. Sorg BA
    (2023) Net gain and loss: influence of natural rewards and drugs of abuse on perineuronal nets. Neuropsychopharmacology 48:3–20. https://doi.org/10.1038/s41386-022-01337-x pmid:35568740
    CrossRefMedline
  9. ↵
    1. Brown TE,
    2. Forquer MR,
    3. Cocking DL,
    4. Jansen HT,
    5. Harding JW,
    6. Sorg BA
    (2007) Role of matrix metalloproteinases in the acquisition and reconsolidation of cocaine-induced conditioned place preference. Learn Mem 14:214–223. https://doi.org/10.1101/lm.476207 pmid:17353546
    Abstract/FREE Full Text
  10. ↵
    1. Brown TE,
    2. Forquer MR,
    3. Harding JW,
    4. Wright JW,
    5. Sorg BA
    (2008) Increase in matrix metalloproteinase-9 levels in the rat medial prefrontal cortex after cocaine reinstatement of conditioned place preference. Synapse 62:886–889. https://doi.org/10.1002/syn.20562
    CrossRefMedline
  11. ↵
    1. Bruckner G,
    2. Bringmann A,
    3. Hartig W,
    4. Koppe G,
    5. Delpech B,
    6. Brauer K
    (1998) Acute and long-lasting changes in extracellular-matrix chondroitin-sulphate proteoglycans induced by injection of chondroitinase ABC in the adult rat brain. Exp Brain Res 121:300–310. https://doi.org/10.1007/s002210050463
    CrossRefMedline
  12. ↵
    1. Buzsaki G,
    2. Draguhn A
    (2004) Neuronal oscillations in cortical networks. Science 304:1926–1929. https://doi.org/10.1126/science.1099745
    Abstract/FREE Full Text
  13. ↵
    1. Cabungcal JH,
    2. Steullet P,
    3. Morishita H,
    4. Kraftsik R,
    5. Cuenod M,
    6. Hensch TK,
    7. Do KQ
    (2013) Perineuronal nets protect fast-spiking interneurons against oxidative stress. Proc Natl Acad Sci U S A 110:9130–9135. https://doi.org/10.1073/pnas.1300454110 pmid:23671099
    Abstract/FREE Full Text
  14. ↵
    1. Canolty RT,
    2. Edwards E,
    3. Dalal SS,
    4. Soltani M,
    5. Nagarajan SS,
    6. Kirsch HE,
    7. Berger MS,
    8. Barbaro NM,
    9. Knight RT
    (2006) High gamma power is phase-locked to theta oscillations in human neocortex. Science 313:1626–1628. https://doi.org/10.1126/science.1128115 pmid:16973878
    Abstract/FREE Full Text
  15. ↵
    1. Capriles N,
    2. Rodaros D,
    3. Sorge RE,
    4. Stewart J
    (2003) A role for the prefrontal cortex in stress- and cocaine-induced reinstatement of cocaine seeking in rats. Psychopharmacology 168:66–74. https://doi.org/10.1007/s00213-002-1283-z
    CrossRefMedline
  16. ↵
    1. Carceller H,
    2. Guirado R,
    3. Ripolles-Campos E,
    4. Teruel-Marti V,
    5. Nacher J
    (2020) Perineuronal nets regulate the inhibitory perisomatic input onto parvalbumin interneurons and gamma activity in the prefrontal cortex. J Neurosci 40:5008–5018. https://doi.org/10.1523/JNEUROSCI.0291-20.2020 pmid:32457072
    Abstract/FREE Full Text
  17. ↵
    1. Caterson B
    (2012) Fell-muir lecture: chondroitin sulphate glycosaminoglycans: fun for some and confusion for others. Int J Exp Pathol 93:1–10. https://doi.org/10.1111/j.1365-2613.2011.00807.x pmid:22264297
    CrossRefMedline
  18. ↵
    1. Childress AR,
    2. Mozley PD,
    3. McElgin W,
    4. Fitzgerald J,
    5. Reivich M,
    6. O'Brien CP
    (1999) Limbic activation during cue-induced cocaine craving. Am J Psychiatry 156:11–18. https://doi.org/10.1176/ajp.156.1.11 pmid:9892292
    CrossRefMedline
  19. ↵
    1. Cho KKA,
    2. Hoch R,
    3. Lee AT,
    4. Patel T,
    5. Rubenstein JL,
    6. Sohal VS
    (2015) Gamma rhythms link prefrontal interneuron dysfunction with cognitive inflexibility in Dlx5/6(+/-) mice. Neuron 85:1332–1343. https://doi.org/10.1016/j.neuron.2015.02.019 pmid:25754826
    CrossRefMedline
  20. ↵
    1. Cho KKA,
    2. Davidson TJ,
    3. Bouvier G,
    4. Marshall JD,
    5. Schnitzer MJ,
    6. Sohal VS
    (2020) Cross-hemispheric gamma synchrony between prefrontal parvalbumin interneurons supports behavioral adaptation during rule shift learning. Nat Neurosci 23:892–902. https://doi.org/10.1038/s41593-020-0647-1 pmid:32451483
    CrossRefMedline
  21. ↵
    1. Christensen AC,
    2. Lensjo KK,
    3. Lepperod ME,
    4. Dragly SA,
    5. Sutterud H,
    6. Blackstad JS,
    7. Fyhn M,
    8. Hafting T
    (2021) Perineuronal nets stabilize the grid cell network. Nat Commun 12:253. https://doi.org/10.1038/s41467-020-20241-w pmid:33431847
    CrossRefMedline
  22. ↵
    1. Ciccarelli A,
    2. Weijers D,
    3. Kwan W,
    4. Warner C,
    5. Bourne J,
    6. Gross CT
    (2021) Sexually dimorphic perineuronal nets in the rodent and primate reproductive circuit. J Comp Neurol 529:3274–3291. https://doi.org/10.1002/cne.25167
    CrossRefMedline
  23. ↵
    1. Deepa SS,
    2. Carulli D,
    3. Galtrey C,
    4. Rhodes K,
    5. Fukuda J,
    6. Mikami T,
    7. Sugahara K,
    8. Fawcett JW
    (2006) Composition of perineuronal net extracellular matrix in rat brain: a different disaccharide composition for the net-associated proteoglycans. J Biol Chem 281:17789–17800. https://doi.org/10.1074/jbc.M600544200
    Abstract/FREE Full Text
  24. ↵
    1. Diaz-Mataix L,
    2. Ruiz Martinez RC,
    3. Schafe GE,
    4. LeDoux JE,
    5. Doyere V
    (2013) Detection of a temporal error triggers reconsolidation of amygdala-dependent memories. Curr Biol 23:467–472. https://doi.org/10.1016/j.cub.2013.01.053 pmid:23453952
    CrossRefMedline
  25. ↵
    1. Dilgen JE,
    2. Tompa T,
    3. Saggu S,
    4. Naselaris T,
    5. Lavin A
    (2013) Optogenetically evoked gamma oscillations are disturbed by cocaine administration. Front Cell Neurosci 7:213. https://doi.org/10.3389/fncel.2013.00213 pmid:24376397
    CrossRefMedline
  26. ↵
    1. Dingess PM,
    2. Zhang Z,
    3. Sorg BA,
    4. Ferrario CR,
    5. Brown TE
    (2020) Sex and region-specific effects of high fat diet on PNNs in obesity susceptible rats. Physiol Behav 222:112963. https://doi.org/10.1016/j.physbeh.2020.112963 pmid:32416158
    CrossRefMedline
  27. ↵
    1. Dityatev A,
    2. Rusakov DA
    (2011) Molecular signals of plasticity at the tetrapartite synapse. Curr Opin Neurobiol 21:353–359. https://doi.org/10.1016/j.conb.2010.12.006 pmid:21277196
    CrossRefMedline
  28. ↵
    1. Dityatev A,
    2. Seidenbecher CI,
    3. Schachner M
    (2010) Compartmentalization from the outside: the extracellular matrix and functional microdomains in the brain. Trends Neurosci 33:503–512. https://doi.org/10.1016/j.tins.2010.08.003
    CrossRefMedline
  29. ↵
    1. Domanski APF,
    2. Kucewicz MT,
    3. Russo E,
    4. Tricklebank MD,
    5. Robinson ESJ,
    6. Durstewitz D,
    7. Jones MW
    (2023) Distinct hippocampal-prefrontal neural assemblies coordinate memory encoding, maintenance, and recall. Curr Biol 33:1220–1236.e4. https://doi.org/10.1016/j.cub.2023.02.029 pmid:36898372
    CrossRefMedline
  30. ↵
    1. Donato F,
    2. Chowdhury A,
    3. Lahr M,
    4. Caroni P
    (2015) Early- and late-born parvalbumin basket cell subpopulations exhibiting distinct regulation and roles in learning. Neuron 85:770–786. https://doi.org/10.1016/j.neuron.2015.01.011
    CrossRefMedline
  31. ↵
    1. Dudai Y
    (2012) The restless engram: consolidations never end. Annu Rev Neurosci 35:227–247. https://doi.org/10.1146/annurev-neuro-062111-150500
    CrossRefMedline
  32. ↵
    1. Dunbar AB,
    2. Taylor JR
    (2017) Garcinol blocks the reconsolidation of multiple cocaine-paired cues after a single cocaine-reactivation session. Neuropsychopharmacology 42:1884–1892. https://doi.org/10.1038/npp.2017.27 pmid:28169286
    CrossRefMedline
  33. ↵
    1. Ehret G,
    2. Romand R
    (2022) Awareness and consciousness in humans and animals - neural and behavioral correlates in an evolutionary perspective. Front Syst Neurosci 16:941534. https://doi.org/10.3389/fnsys.2022.941534 pmid:35910003
    Medline
  34. ↵
    1. Exton-McGuinness MTJ,
    2. Lee JL
    (2015) Reduction in responding for sucrose and cocaine reinforcement by disruption of memory reconsolidation(1,2,3). eNeuro 2:ENEURO.0009-15.2015. https://doi.org/10.1523/ENEURO.0009-15.2015 pmid:26464973
    CrossRefMedline
  35. ↵
    1. Exton-McGuinness MTJ,
    2. Milton AL
    (2018) Reconsolidation blockade for the treatment of addiction: challenges, new targets, and opportunities. Learn Mem 25:492–500. https://doi.org/10.1101/lm.046771.117 pmid:30115771
    Abstract/FREE Full Text
  36. ↵
    1. Exton-McGuinness MTJ,
    2. Patton RC,
    3. Sacco LB,
    4. Lee JL
    (2014) Reconsolidation of a well-learned instrumental memory. Learn Mem 21:468–477. https://doi.org/10.1101/lm.035543.114 pmid:25135195
    Abstract/FREE Full Text
  37. ↵
    1. Faini G,
    2. Aguirre A,
    3. Landi S,
    4. Lamers D,
    5. Pizzorusso T,
    6. Ratto GM,
    7. Deleuze C,
    8. Bacci A
    (2018) Perineuronal nets control visual input via thalamic recruitment of cortical PV interneurons. Elife 7:e41520. https://doi.org/10.7554/eLife.41520 pmid:30561327
    CrossRefMedline
  38. ↵
    1. Fattahi M,
    2. Modaberi S,
    3. Eskandari K,
    4. Haghparast A
    (2023) A systematic review of the local field potential adaptations during conditioned place preference task in preclinical studies. Synapse 77:e22277. https://doi.org/10.1002/syn.22277
    Google Scholar
  39. ↵
    1. Favuzzi E,
    2. Marques-Smith A,
    3. Deogracias R,
    4. Winterflood CM,
    5. Sanchez-Aguilera A,
    6. Mantoan L,
    7. Maeso P,
    8. Fernandes C,
    9. Ewers H,
    10. Rico B
    (2017) Activity-dependent gating of parvalbumin interneuron function by the perineuronal net protein brevican. Neuron 95:639–655.e10. https://doi.org/10.1016/j.neuron.2017.06.028
    CrossRefMedline
  40. ↵
    1. Fawcett JW,
    2. Oohashi T,
    3. Pizzorusso T
    (2019) The roles of perineuronal nets and the perinodal extracellular matrix in neuronal function. Nat Rev Neurosci 20:451–465. https://doi.org/10.1038/s41583-019-0196-3
    CrossRefMedline
  41. ↵
    1. Fawcett JW,
    2. Fyhn M,
    3. Jendelova P,
    4. Kwok JCF,
    5. Ruzicka J,
    6. Sorg BA
    (2022) The extracellular matrix and perineuronal nets in memory. Mol Psychiatry 27:3192–3203. https://doi.org/10.1038/s41380-022-01634-3 pmid:35760878
    CrossRefMedlineGoogle Scholar
  42. ↵
    1. Franca ASC,
    2. Borgesius NZ,
    3. Souza BC,
    4. Cohen MX
    (2021) Beta2 oscillations in hippocampal-cortical circuits during novelty detection. Front Syst Neurosci 15:617388. https://doi.org/10.3389/fnsys.2021.617388 pmid:33664653
    CrossRefMedline
  43. ↵
    1. Frankland PW,
    2. Ding HK,
    3. Takahashi E,
    4. Suzuki A,
    5. Kida S,
    6. Silva AJ
    (2006) Stability of recent and remote contextual fear memory. Learn Mem 13:451–457. https://doi.org/10.1101/lm.183406 pmid:16882861
    Abstract/FREE Full Text
  44. ↵
    1. Fuchs RA,
    2. Eaddy JL,
    3. Su ZI,
    4. Bell GH
    (2007) Interactions of the basolateral amygdala with the dorsal hippocampus and dorsomedial prefrontal cortex regulate drug context-induced reinstatement of cocaine-seeking in rats. Eur J Neurosci 26:487–498. https://doi.org/10.1111/j.1460-9568.2007.05674.x
    CrossRefMedline
  45. ↵
    1. Garcia-Keller C,
    2. Neuhofer D,
    3. Bobadilla AC,
    4. Spencer S,
    5. Chioma VC,
    6. Monforton C,
    7. Kalivas PW
    (2019) Extracellular matrix signaling through beta3 integrin mediates cocaine cue-induced transient synaptic plasticity and relapse. Biol Psychiatry 86:377–387. https://doi.org/10.1016/j.biopsych.2019.03.982 pmid:31126696
    CrossRefMedline
  46. ↵
    1. Gildawie KR,
    2. Honeycutt JA,
    3. Brenhouse HC
    (2020) Region-specific effects of maternal separation on perineuronal net and parvalbumin-expressing interneuron formation in male and female rats. Neuroscience 428:23–37. https://doi.org/10.1016/j.neuroscience.2019.12.010
  47. ↵
    1. Goeders NE,
    2. Peltier RL,
    3. Guerin GF
    (1998) Ketoconazole reduces low dose cocaine self-administration in rats. Drug Alcohol Depend 53:67–77. https://doi.org/10.1016/S0376-8716(98)00108-2
    CrossRefMedline
  48. ↵
    1. Gonzalez AE,
    2. Jorgensen ET,
    3. Ramos JD,
    4. Harkness JH,
    5. Aadland JA,
    6. Brown TE,
    7. Sorg BA
    (2022) Impact of perineuronal net removal in the rat medial prefrontal cortex on parvalbumin interneurons after reinstatement of cocaine conditioned place preference. Front Cell Neurosci 16:932391. https://doi.org/10.3389/fncel.2022.932391 pmid:35966203
    CrossRefMedline
  49. ↵
    1. Gonzalez MC,
    2. Rossato JI,
    3. Radiske A,
    4. Bevilaqua LRM,
    5. Cammarota M
    (2021) Dopamine controls whether new declarative information updates reactivated memories through reconsolidation. Proc Natl Acad Sci U S A 118:e2025275118. https://doi.org/10.1073/pnas.2025275118 pmid:34253612
    Abstract/FREE Full Text
  50. ↵
    1. Grant S,
    2. London ED,
    3. Newlin DB,
    4. Villemagne VL,
    5. Liu X,
    6. Contoreggi C,
    7. Phillips RL,
    8. Kimes AS,
    9. Margolin A
    (1996) Activation of memory circuits during cue-elicited cocaine craving. Proc Natl Acad Sci U S A 93:12040–12045. https://doi.org/10.1073/pnas.93.21.12040 pmid:8876259
    Abstract/FREE Full Text
  51. ↵
    1. Griffiths BB,
    2. Madden AMK,
    3. Edwards KA,
    4. Zup SL,
    5. Stary CM
    (2019) Age-dependent sexual dimorphism in hippocampal cornu ammonis-1 perineuronal net expression in rats. Brain Behav 9:e01265. https://doi.org/10.1002/brb3.1265 pmid:30912298
    CrossRefMedline
  52. ↵
    1. Guillem K,
    2. Ahmed SH
    (2020) Reorganization of theta phase-locking in the orbitofrontal cortex drives cocaine choice under the influence. Sci Rep 10:8041. https://doi.org/10.1038/s41598-020-64962-w pmid:32415278
    CrossRefMedline
  53. ↵
    1. Han C, et al.
    (2021) Multiple gamma rhythms carry distinct spatial frequency information in primary visual cortex. PLoS Biol 19:e3001466. https://doi.org/10.1371/journal.pbio.3001466 pmid:34932558
    CrossRefMedline
  54. ↵
    1. Hanssen KO,
    2. Grodem S,
    3. Fyhn M,
    4. Hafting T,
    5. Einevoll GT,
    6. Ness TV,
    7. Halnes G
    (2023) Responses in fast-spiking interneuron firing rates to parameter variations associated with degradation of perineuronal nets. J Comput Neurosci 51:283–298. https://doi.org/10.1007/s10827-023-00849-9 pmid:37058180
    CrossRefMedline
  55. ↵
    1. Harkness JH, et al.
    (2021) Diurnal changes in perineuronal nets and parvalbumin neurons in the rat medial prefrontal cortex. Brain Struct Funct 226:1135–1153. https://doi.org/10.1007/s00429-021-02229-4 pmid:33585984
    CrossRefMedline
  56. ↵
    1. Hartig W,
    2. Brauer K,
    3. Bruckner G
    (1992) Wisteria floribunda agglutinin-labelled nets surround parvalbumin-containing neurons. Neuroreport 3:869–872. https://doi.org/10.1097/00001756-199210000-00012
    CrossRefMedline
  57. ↵
    1. Jones MW,
    2. Wilson MA
    (2005) Theta rhythms coordinate hippocampal-prefrontal interactions in a spatial memory task. PLoS Biol 3:e402. https://doi.org/10.1371/journal.pbio.0030402 pmid:16279838
    CrossRefMedline
  58. ↵
    1. Jorgensen ET,
    2. Gonzalez AE,
    3. Harkness JH,
    4. Hegarty DM,
    5. Thakar A,
    6. Burchi DJ,
    7. Aadland JA,
    8. Aicher SA,
    9. Sorg BA,
    10. Brown TE
    (2020) Cocaine memory reactivation induces functional adaptations within parvalbumin interneurons in the rat medial prefrontal cortex. Addict Biol 26:e12947. https://doi.org/10.1111/adb.12947 pmid:32750200
    CrossRefMedlineGoogle Scholar
  59. ↵
    1. Kawaguchi Y,
    2. Kubota Y
    (1993) Correlation of physiological subgroupings of nonpyramidal cells with parvalbumin- and calbindinD28k-immunoreactive neurons in layer V of rat frontal cortex. J Neurophysiol 70:387–396. https://doi.org/10.1152/jn.1993.70.1.387
    CrossRefMedline
  60. ↵
    1. Kiyatkin EA,
    2. Smirnov MS
    (2010) Rapid EEG desynchronization and EMG activation induced by intravenous cocaine in freely moving rats: a peripheral, nondopamine neural triggering. Am J Physiol Regul Integr Comp Physiol 298:R285–R300. https://doi.org/10.1152/ajpregu.00628.2009 pmid:19939972
    CrossRefMedline
  61. ↵
    1. Lee AT,
    2. Gee SM,
    3. Vogt D,
    4. Patel T,
    5. Rubenstein JL,
    6. Sohal VS
    (2014) Pyramidal neurons in prefrontal cortex receive subtype-specific forms of excitation and inhibition. Neuron 81:61–68. https://doi.org/10.1016/j.neuron.2013.10.031 pmid:24361076
    CrossRefMedline
  62. ↵
    1. Lee JLC,
    2. Di Ciano P,
    3. Thomas KL,
    4. Everitt BJ
    (2005) Disrupting reconsolidation of drug memories reduces cocaine-seeking behavior. Neuron 47:795–801. https://doi.org/10.1016/j.neuron.2005.08.007
    CrossRefMedline
  63. ↵
    1. Lee JLC,
    2. Nader K,
    3. Schiller D
    (2017) An update on memory reconsolidation updating. Trends Cogn Sci 21:531–545. https://doi.org/10.1016/j.tics.2017.04.006 pmid:28495311
    CrossRefMedline
  64. ↵
    1. Lensjo KK,
    2. Lepperod ME,
    3. Dick G,
    4. Hafting T,
    5. Fyhn M
    (2017) Removal of perineuronal nets unlocks juvenile plasticity through network mechanisms of decreased inhibition and increased gamma activity. J Neurosci 37:1269–1283. https://doi.org/10.1523/JNEUROSCI.2504-16.2016 pmid:28039374
    Abstract/FREE Full Text
  65. ↵
    1. Lew SE,
    2. Tseng KY
    (2014) Dopamine modulation of GABAergic function enables network stability and input selectivity for sustaining working memory in a computational model of the prefrontal cortex. Neuropsychopharmacology 39:3067–3076. https://doi.org/10.1038/npp.2014.160 pmid:24975022
    CrossRefMedline
  66. ↵
    1. Li Y,
    2. Wang X,
    3. Li N,
    4. Qu L,
    5. Wang P,
    6. Ge SN,
    7. Wang XL
    (2020) The NAc lesions disrupted the hippocampus-mPFC theta coherence during intravenous cocaine administration in rats. Neurosci Lett 729:134986. https://doi.org/10.1016/j.neulet.2020.134986
  67. ↵
    1. Lynch WJ,
    2. Carroll ME
    (1999) Sex differences in the acquisition of intravenously self-administered cocaine and heroin in rats. Psychopharmacology 144:77–82. https://doi.org/10.1007/s002130050979
    CrossRefMedline
  68. ↵
    1. Ma YY, et al.
    (2014) Bidirectional modulation of incubation of cocaine craving by silent synapse-based remodeling of prefrontal cortex to accumbens projections. Neuron 83:1453–1467. https://doi.org/10.1016/j.neuron.2014.08.023 pmid:25199705
    CrossRefMedline
  69. ↵
    1. Maas LC,
    2. Lukas SE,
    3. Kaufman MJ,
    4. Weiss RD,
    5. Daniels SL,
    6. Rogers VW,
    7. Kukes TJ,
    8. Renshaw PF
    (1998) Functional magnetic resonance imaging of human brain activation during cue-induced cocaine craving. Am J Psychiatry 155:124–126. https://doi.org/10.1176/ajp.155.1.124
    CrossRefMedline
  70. ↵
    1. Mamou CB,
    2. Gamache K,
    3. Nader K
    (2006) NMDA receptors are critical for unleashing consolidated auditory fear memories. Nat Neurosci 9:1237–1239. https://doi.org/10.1038/nn1778
    CrossRefMedline
  71. ↵
    1. Marco-Pallares J,
    2. Munte TF,
    3. Rodriguez-Fornells A
    (2015) The role of high-frequency oscillatory activity in reward processing and learning. Neurosci Biobehav Rev 49:1–7. https://doi.org/10.1016/j.neubiorev.2014.11.014
    CrossRefMedline
  72. ↵
    1. Markram H,
    2. Toledo-Rodriguez M,
    3. Wang Y,
    4. Gupta A,
    5. Silberberg G,
    6. Wu C
    (2004) Interneurons of the neocortical inhibitory system. Nat Rev Neurosci 5:793–807. https://doi.org/10.1038/nrn1519
    CrossRefMedline
  73. ↵
    1. McCracken CB,
    2. Grace AA
    (2013) Persistent cocaine-induced reversal learning deficits are associated with altered limbic cortico-striatal local field potential synchronization. J Neurosci 33:17469–17482. https://doi.org/10.1523/JNEUROSCI.1440-13.2013 pmid:24174680
    Abstract/FREE Full Text
  74. ↵
    1. McFarland K,
    2. Kalivas PW
    (2001) The circuitry mediating cocaine-induced reinstatement of drug-seeking behavior. J Neurosci 21:8655–8663. https://doi.org/10.1523/JNEUROSCI.21-21-08655.2001 pmid:11606653
    Abstract/FREE Full Text
  75. ↵
    1. Milekic MH,
    2. Brown SD,
    3. Castellini C,
    4. Alberini CM
    (2006) Persistent disruption of an established morphine conditioned place preference. J Neurosci 26:3010–3020. https://doi.org/10.1523/JNEUROSCI.4818-05.2006 pmid:16540579
    Abstract/FREE Full Text
  76. ↵
    1. Moorman DE,
    2. Aston-Jones G
    (2023) Prelimbic and infralimbic medial prefrontal cortex neuron activity signals cocaine seeking variables across multiple timescales. Psychopharmacology 240:575–594. https://doi.org/10.1007/s00213-022-06287-2 pmid:36464693
    CrossRefMedline
  77. ↵
    1. Mulholland PJ,
    2. Chandler LJ,
    3. Kalivas PW
    (2016) Signals from the fourth dimension regulate drug relapse. Trends Neurosci 39:472–485. https://doi.org/10.1016/j.tins.2016.04.007 pmid:27173064
    CrossRefMedline
  78. ↵
    1. Muller Ewald VA,
    2. Kim J,
    3. Farley SJ,
    4. Freeman JH,
    5. LaLumiere RT
    (2022) Theta oscillations in rat infralimbic cortex are associated with the inhibition of cocaine seeking during extinction. Addict Biol 27:e13106. https://doi.org/10.1111/adb.13106 pmid:34672059
    CrossRefMedline
  79. ↵
    1. Nagy V,
    2. Bozdagi O,
    3. Matynia A,
    4. Balcerzyk M,
    5. Okulski P,
    6. Dzwonek J,
    7. Costa RM,
    8. Silva AJ,
    9. Kaczmarek L,
    10. Huntley GW
    (2006) Matrix metalloproteinase-9 is required for hippocampal late-phase long-term potentiation and memory. J Neurosci 26:1923–1934. https://doi.org/10.1523/JNEUROSCI.4359-05.2006 pmid:16481424
    Abstract/FREE Full Text
  80. ↵
    1. Neuhofer D,
    2. Kalivas P
    (2018) Metaplasticity at the addicted tetrapartite synapse: a common denominator of drug induced adaptations and potential treatment target for addiction. Neurobiol Learn Mem 154:97–111. https://doi.org/10.1016/j.nlm.2018.02.007 pmid:29428364
    CrossRefMedline
  81. ↵
    1. Packer AM,
    2. Yuste R
    (2011) Dense, unspecific connectivity of neocortical parvalbumin-positive interneurons: a canonical microcircuit for inhibition? J Neurosci 31:13260–13271. https://doi.org/10.1523/JNEUROSCI.3131-11.2011 pmid:21917809
    Abstract/FREE Full Text
  82. ↵
    1. Park AJ,
    2. Harris AZ,
    3. Martyniuk KM,
    4. Chang CY,
    5. Abbas AI,
    6. Lowes DC,
    7. Kellendonk C,
    8. Gogos JA,
    9. Gordon JA
    (2021) Reset of hippocampal-prefrontal circuitry facilitates learning. Nature 591:615–619. https://doi.org/10.1038/s41586-021-03272-1 pmid:33627872
    CrossRefMedline
  83. ↵
    1. Park WK,
    2. Bari AA,
    3. Jey AR,
    4. Anderson SM,
    5. Spealman RD,
    6. Rowlett JK,
    7. Pierce RC
    (2002) Cocaine administered into the medial prefrontal cortex reinstates cocaine-seeking behavior by increasing AMPA receptor-mediated glutamate transmission in the nucleus accumbens. J Neurosci 22:2916–2925. https://doi.org/10.1523/JNEUROSCI.22-07-02916.2002 pmid:11923456
    Abstract/FREE Full Text
  84. ↵
    1. Pizzorusso T,
    2. Medini P,
    3. Berardi N,
    4. Chierzi S,
    5. Fawcett JW,
    6. Maffei L
    (2002) Reactivation of ocular dominance plasticity in the adult visual cortex. Science 298:1248–1251. https://doi.org/10.1126/science.1072699
    Abstract/FREE Full Text
  85. ↵
    1. Radiske A,
    2. Gonzalez MC,
    3. Noga DA,
    4. Rossato JI,
    5. Bevilaqua LRM,
    6. Cammarota M
    (2021) Glun2b and GluN2A-containing NMDAR are differentially involved in extinction memory destabilization and restabilization during reconsolidation. Sci Rep 11:186. https://doi.org/10.1038/s41598-020-80674-7 pmid:33420399
    CrossRefMedline
  86. ↵
    1. Ramsaran AI, et al.
    (2023) A shift in the mechanisms controlling hippocampal engram formation during brain maturation. Science 380:543–551. https://doi.org/10.1126/science.ade6530
    CrossRef
  87. ↵
    1. Richardson NR,
    2. Roberts DC
    (1996) Progressive ratio schedules in drug self-administration studies in rats: a method to evaluate reinforcing efficacy. J Neurosci Methods 66:1–11. https://doi.org/10.1016/0165-0270(95)00153-0
    CrossRefMedline
  88. ↵
    1. Romberg C,
    2. Yang S,
    3. Melani R,
    4. Andrews MR,
    5. Horner AE,
    6. Spillantini MG,
    7. Bussey TJ,
    8. Fawcett JW,
    9. Pizzorusso T,
    10. Saksida LM
    (2013) Depletion of perineuronal nets enhances recognition memory and long-term depression in the perirhinal cortex. J Neurosci 33:7057–7065. https://doi.org/10.1523/JNEUROSCI.6267-11.2013 pmid:23595763
    Abstract/FREE Full Text
  89. ↵
    1. Rowlands D,
    2. Lensjo KK,
    3. Dinh T,
    4. Yang S,
    5. Andrews MR,
    6. Hafting T,
    7. Fyhn M,
    8. Fawcett JW,
    9. Dick G
    (2018) Aggrecan directs extracellular matrix-mediated neuronal plasticity. J Neurosci 38:10102–10113. https://doi.org/10.1523/JNEUROSCI.1122-18.2018 pmid:30282728
    Abstract/FREE Full Text
  90. ↵
    1. Sara SJ
    (2000) Retrieval and reconsolidation: toward a neurobiology of remembering. Learn Mem 7:73–84. https://doi.org/10.1101/lm.7.2.73
    FREE Full Text
  91. ↵
    1. Sekerli M,
    2. Del Negro CA,
    3. Lee RH,
    4. Butera RJ
    (2004) Estimating action potential thresholds from neuronal time-series: new metrics and evaluation of methodologies. IEEE Trans Biomed Eng 51:1665–1672. https://doi.org/10.1109/TBME.2004.827531
    CrossRefMedline
  92. ↵
    1. Seth AK,
    2. Chorley P,
    3. Barnett LC
    (2013) Granger causality analysis of fMRI BOLD signals is invariant to hemodynamic convolution but not downsampling. Neuroimage 65:540–555. https://doi.org/10.1016/j.neuroimage.2012.09.049
    CrossRefMedline
  93. ↵
    1. Sevenster D,
    2. Beckers T,
    3. Kindt M
    (2013) Prediction error governs pharmacologically induced amnesia for learned fear. Science 339:830–833. https://doi.org/10.1126/science.1231357
    Abstract/FREE Full Text
  94. ↵
    1. Shen HW,
    2. Gipson CD,
    3. Huits M,
    4. Kalivas PW
    (2014) Prelimbic cortex and ventral tegmental area modulate synaptic plasticity differentially in nucleus accumbens during cocaine-reinstated drug seeking. Neuropsychopharmacology 39:1169–1177. https://doi.org/10.1038/npp.2013.318 pmid:24232172
    CrossRefMedline
  95. ↵
    1. Shi W,
    2. Wei X,
    3. Wang X,
    4. Du S,
    5. Liu W,
    6. Song J,
    7. Wang Y
    (2019) Perineuronal nets protect long-term memory by limiting activity-dependent inhibition from parvalbumin interneurons. Proc Natl Acad Sci U S A 116:27063–27073. https://doi.org/10.1073/pnas.1902680116 pmid:31843906
    Abstract/FREE Full Text
  96. ↵
    1. Siapas AG,
    2. Lubenov EV,
    3. Wilson MA
    (2005) Prefrontal phase locking to hippocampal theta oscillations. Neuron 46:141–151. https://doi.org/10.1016/j.neuron.2005.02.028
    CrossRefMedline
  97. ↵
    1. Sjulson L,
    2. Peyrache A,
    3. Cumpelik A,
    4. Cassataro D,
    5. Buzsaki G
    (2018) Cocaine place conditioning strengthens location-specific hippocampal coupling to the nucleus accumbens. Neuron 98:926–934.e5. https://doi.org/10.1016/j.neuron.2018.04.015 pmid:29754750
    CrossRefMedline
  98. ↵
    1. Slaker M,
    2. Churchill L,
    3. Todd RP,
    4. Blacktop JM,
    5. Zuloaga DG,
    6. Raber J,
    7. Darling RA,
    8. Brown TE,
    9. Sorg BA
    (2015) Removal of perineuronal nets in the medial prefrontal cortex impairs the acquisition and reconsolidation of a cocaine-induced conditioned place preference memory. J Neurosci 35:4190–4202. https://doi.org/10.1523/JNEUROSCI.3592-14.2015 pmid:25762666
    Abstract/FREE Full Text
  99. ↵
    1. Slaker ML,
    2. Harkness JH,
    3. Sorg BA
    (2016) A standardized and automated method of perineuronal net analysis using Wisteria floribunda agglutinin staining intensity. IBRO Rep 1:54–60. https://doi.org/10.1016/j.ibror.2016.10.001 pmid:28713865
    CrossRefMedline
  100. ↵
    1. Slaker ML,
    2. Jorgensen ET,
    3. Hegarty DM,
    4. Liu X,
    5. Kong Y,
    6. Zhang F,
    7. Linhardt RJ,
    8. Brown TE,
    9. Aicher SA,
    10. Sorg BA
    (2018) Cocaine exposure modulates perineuronal nets and synaptic excitability of fast-spiking interneurons in the medial prefrontal cortex. eNeuro 5:ENEURO.0221-18.2018. https://doi.org/10.1523/ENEURO.0221-18.2018 pmid:30294670
    Medline
  101. ↵
    1. Smith AC,
    2. Kupchik YM,
    3. Scofield MD,
    4. Gipson CD,
    5. Wiggins A,
    6. Thomas CA,
    7. Kalivas PW
    (2014) Synaptic plasticity mediating cocaine relapse requires matrix metalloproteinases. Nat Neurosci 17:1655–1657. https://doi.org/10.1038/nn.3846 pmid:25326689
    CrossRefMedline
  102. ↵
    1. Sohal VS,
    2. Zhang F,
    3. Yizhar O,
    4. Deisseroth K
    (2009) Parvalbumin neurons and gamma rhythms enhance cortical circuit performance. Nature 459:698–702. https://doi.org/10.1038/nature07991 pmid:19396159
    CrossRefMedline
  103. ↵
    1. Sorg BA
    (2012) Reconsolidation of drug memories. Neurosci Biobehav Rev 36:1400–1417. https://doi.org/10.1016/j.neubiorev.2012.02.004 pmid:22342780
    CrossRefMedline
  104. ↵
    1. Sparta DR,
    2. Hovelso N,
    3. Mason AO,
    4. Kantak PA,
    5. Ung RL,
    6. Decot HK,
    7. Stuber GD
    (2014) Activation of prefrontal cortical parvalbumin interneurons facilitates extinction of reward-seeking behavior. J Neurosci 34:3699–3705. https://doi.org/10.1523/JNEUROSCI.0235-13.2014 pmid:24599468
    Abstract/FREE Full Text
  105. ↵
    1. Steullet P,
    2. Cabungcal JH,
    3. Cuenod M,
    4. Do KQ
    (2014) Fast oscillatory activity in the anterior cingulate cortex: dopaminergic modulation and effect of perineuronal net loss. Front Cell Neurosci 8:244. https://doi.org/10.3389/fncel.2014.00244 pmid:25191228
    CrossRefMedline
  106. ↵
    1. Szklarczyk A,
    2. Lapinska J,
    3. Rylski M,
    4. McKay RD,
    5. Kaczmarek L
    (2002) Matrix metalloproteinase-9 undergoes expression and activation during dendritic remodeling in adult hippocampus. J Neurosci 22:920–930. https://doi.org/10.1523/JNEUROSCI.22-03-00920.2002 pmid:11826121
    Abstract/FREE Full Text
  107. ↵
    1. Tang W,
    2. Shin JD,
    3. Jadhav SP
    (2021) Multiple time-scales of decision-making in the hippocampus and prefrontal cortex. Elife 10:e66227. https://doi.org/10.7554/eLife.66227 pmid:33683201
    CrossRefMedlineGoogle Scholar
  108. ↵
    1. Tewari BP,
    2. Chaunsali L,
    3. Campbell SL,
    4. Patel DC,
    5. Goode AE,
    6. Sontheimer H
    (2018) Perineuronal nets decrease membrane capacitance of peritumoral fast spiking interneurons in a model of epilepsy. Nat Commun 9:4724. https://doi.org/10.1038/s41467-018-07113-0 pmid:30413686
    CrossRefMedline
  109. ↵
    1. Thompson EH,
    2. Lensjo KK,
    3. Wigestrand MB,
    4. Malthe-Sorenssen A,
    5. Hafting T,
    6. Fyhn M
    (2018) Removal of perineuronal nets disrupts recall of a remote fear memory. Proc Natl Acad Sci U S A 115:607–612. https://doi.org/10.1073/pnas.1713530115 pmid:29279411
    Abstract/FREE Full Text
  110. ↵
    1. Tort AB,
    2. Kramer MA,
    3. Thorn C,
    4. Gibson DJ,
    5. Kubota Y,
    6. Graybiel AM,
    7. Kopell NJ
    (2008) Dynamic cross-frequency couplings of local field potential oscillations in rat striatum and hippocampus during performance of a T-maze task. Proc Natl Acad Sci U S A 105:20517–20522. https://doi.org/10.1073/pnas.0810524105 pmid:19074268
    Abstract/FREE Full Text
  111. ↵
    1. Tronson NC,
    2. Taylor JR
    (2013) Addiction: a drug-induced disorder of memory reconsolidation. Curr Opin Neurobiol 23:573–580. https://doi.org/10.1016/j.conb.2013.01.022 pmid:23415831
    CrossRefMedline
  112. ↵
    1. Tseng KY,
    2. Mallet N,
    3. Toreson KL,
    4. Le Moine C,
    5. Gonon F,
    6. O'Donnell P
    (2006) Excitatory response of prefrontal cortical fast-spiking interneurons to ventral tegmental area stimulation in vivo. Synapse 59:412–417. https://doi.org/10.1002/syn.20255 pmid:16485264
    CrossRefMedline
  113. ↵
    1. Valjent E,
    2. Corbille AG,
    3. Bertran-Gonzalez J,
    4. Herve D,
    5. Girault JA
    (2006) Inhibition of ERK pathway or protein synthesis during reexposure to drugs of abuse erases previously learned place preference. Proc Natl Acad Sci U S A 103:2932–2937. https://doi.org/10.1073/pnas.0511030103 pmid:16473939
    Abstract/FREE Full Text
  114. ↵
    1. Volkow ND,
    2. Fowler JS,
    3. Wolf AP,
    4. Hitzemann R,
    5. Dewey S,
    6. Bendriem B,
    7. Alpert R,
    8. Hoff A
    (1991) Changes in brain glucose metabolism in cocaine dependence and withdrawal. Am J Psychiatry 148:621–626. https://doi.org/10.1176/ajp.148.12.1759-b
    CrossRefMedline
  115. ↵
    1. Wang AS,
    2. Wan X,
    3. Storch DS,
    4. Li VY,
    5. Cornez G,
    6. Balthazart J,
    7. Cisneros-Franco JM,
    8. de Villers-Sidani E,
    9. Sakata JT
    (2023) Cross-species conservation in the regulation of parvalbumin by perineuronal nets. Front Neural Circuits 17:1297643. https://doi.org/10.3389/fncir.2023.1297643 pmid:38179221
    CrossRefMedline
  116. ↵
    1. West EA,
    2. Saddoris MP,
    3. Kerfoot EC,
    4. Carelli RM
    (2014) Prelimbic and infralimbic cortical regions differentially encode cocaine-associated stimuli and cocaine-seeking before and following abstinence. Eur J Neurosci 39:1891–1902. https://doi.org/10.1111/ejn.12578 pmid:24690012
    CrossRefMedline
  117. ↵
    1. Wideman CE,
    2. Nguyen J,
    3. Jeffries SD,
    4. Winters BD
    (2020) Fluctuating NMDA receptor subunit levels in perirhinal cortex relate to their dynamic roles in object memory destabilization and reconsolidation. Int J Mol Sci 22:67. https://doi.org/10.3390/ijms22010067 pmid:33374645
    CrossRefMedline
  118. ↵
    1. Wingert JC,
    2. Sorg BA
    (2021) Impact of perineuronal nets on electrophysiology of parvalbumin interneurons, principal neurons, and brain oscillations: a review. Front Synaptic Neurosci 13:673210. https://doi.org/10.3389/fnsyn.2021.673210 pmid:34040511
    CrossRefMedline
  119. ↵
    1. Winters BD,
    2. Tucci MC,
    3. DaCosta-Furtado M
    (2009) Older and stronger object memories are selectively destabilized by reactivation in the presence of new information. Learn Mem 16:545–553. https://doi.org/10.1101/lm.1509909
    Abstract/FREE Full Text
  120. ↵
    1. Wise RA,
    2. Kiyatkin EA
    (2011) Differentiating the rapid actions of cocaine. Nat Rev Neurosci 12:479–484. https://doi.org/10.1038/nrn3043 pmid:21633381
    CrossRefMedline
  121. ↵
    1. Xue YX, et al.
    (2017) Selective inhibition of amygdala neuronal ensembles encoding nicotine-associated memories inhibits nicotine preference and relapse. Biol Psychiatry 82:781–793. https://doi.org/10.1016/j.biopsych.2017.04.017 pmid:28648649
    CrossRefMedline
  122. ↵
    1. Yamada J,
    2. Ohgomori T,
    3. Jinno S
    (2015) Perineuronal nets affect parvalbumin expression in GABAergic neurons of the mouse hippocampus. Eur J Neurosci 41:368–378. https://doi.org/10.1111/ejn.12792
    CrossRefMedline
  123. ↵
    1. Yamagata T,
    2. Saito H,
    3. Habuchi O,
    4. Suzuki S
    (1968) Purification and properties of bacterial chondroitinases and chondrosulfatases. J Biol Chem 243:1523–1535. https://doi.org/10.1016/S0021-9258(18)93574-X
    Abstract/FREE Full Text
  124. ↵
    1. Yu YJ,
    2. Huang CH,
    3. Chang CH,
    4. Gean PW
    (2016) Involvement of protein phosphatases in the destabilization of methamphetamine-associated contextual memory. Learn Mem 23:486–493. https://doi.org/10.1101/lm.039941.115 pmid:27531839
    Abstract/FREE Full Text
  125. ↵
    1. Zhai T,
    2. Salmeron BJ,
    3. Gu H,
    4. Adinoff B,
    5. Stein EA,
    6. Yang Y
    (2021) Functional connectivity of dorsolateral prefrontal cortex predicts cocaine relapse: implications for neuromodulation treatment. Brain Commun 3:fcab120. https://doi.org/10.1093/braincomms/fcab120 pmid:34189458
    CrossRefMedline
  126. ↵
    1. Zhang N, et al.
    (2021) Hypothalamic perineuronal nets are regulated by sex and dietary interventions. Front Physiol 12:714104. https://doi.org/10.3389/fphys.2021.714104 pmid:34393830
    CrossRefMedline
  127. ↵
    1. Zhu H,
    2. Zhou Y,
    3. Liu Z,
    4. Chen X,
    5. Li Y,
    6. Liu X,
    7. Ma L
    (2018) β1-Adrenoceptor in the central amygdala is required for unconditioned stimulus-induced drug memory reconsolidation. Int J Neuropsychopharmacol 21:267–280. https://doi.org/10.1093/ijnp/pyx104 pmid:29216351
    CrossRefMedline

Navigate This Article

  1. Published online before print 11 July 2024, doi: 10.1523/JNEUROSCI.0468-24.2024 The Journal of Neuroscience, 44(34): e0468242024
  1. AbstractFree
  2. » Full Text
  3. Full Text (PDF)
Responses to This Article
Be the first to respond to this article
Submit a Response
View Similar Articles
Similar articles in this journal»
Advertisement