Pavlovian fear memory induced by activation in the anterior cingulate cortex
© Tang et al; licensee BioMed Central Ltd. 2005
Received: 02 November 2004
Accepted: 09 February 2005
Published: 09 February 2005
Skip to main content
© Tang et al; licensee BioMed Central Ltd. 2005
Received: 02 November 2004
Accepted: 09 February 2005
Published: 09 February 2005
Identifying higher brain central region(s) that are responsible for the unpleasantness of pain is the focus of many recent studies. Here we show that direct stimulation of the anterior cingulate cortex (ACC) in mice produced fear-like freezing responses and induced long-term fear memory, including contextual and auditory fear memory. Auditory fear memory required the activation of N-methyl-D-aspartate (NMDA) receptors in the amygdala. To test the hypothesis that neuronal activity in the ACC contributes to unpleasantness, we injected a GABAA receptor agonist, muscimol bilaterally into the ACC. Both contextual and auditory memories induced by foot shock were blocked. Furthermore, activation of metabotropic glutamate receptors in the ACC enhanced behavioral escape responses in a noxious hot-plate as well as spinal nociceptive tail-flick reflex. Our results provide strong evidence that the excitatory activity in the ACC contribute to pain-related fear memory as well as descending facilitatory modulation of spinal nociception.
Pain in humans is an unpleasant sensory and emotional experience associated with actual or potential tissue damage, or described in terms of such damage . The pain experience contains at least two major components: the first is the encoding and perception of sensory noxious stimulus (e.g., pain intensity); and the second is the encoding of the unpleasantness of the noxious stimuli [2, 3]. Exploration of the centers for pain-related unpleasantness has recently been carried out in human studies using modern imaging techniques [4–8]. Among many central regions investigated, the anterior cingulate cortex (ACC) is believed to be a key structure that contributes to pain affect or unpleasantness. Early human observations showed that surgical ablation of the ACC significantly reduced pain unpleasantness without influencing the ability to detect the intensity or location of the pain [9, 10]. Rainelle et al  reported that specific manipulation of pain unpleasantness produced significant changes in the imaged activity of the ACC, while the manipulation of pain intensity produced changes mainly in the primary somatosensory cortex (S1) [2, 5]. More recently, electrophysiological recordings from the ACC in humans found that some ACC neurons respond to noxious stimuli . More interestingly, a recent study reported that the ACC was also activated during social exclusion . In addition to pain, the ACC has been proposed as the neurobiological substrate for executive control of cognitive and motor processes . Human imaging studies demonstrate that the ACC region is activated by different factors including motivational drive, reward, gain or loss, conflict-monitoring or error prediction, and attention or anticipation [13–23]. The neuronal mechanisms for these different functions within the ACC remain mostly unknown due to the limitation of human studies.
Studies from our group and other investigators, using animal models, provide evidence for the importance of the ACC in behavioural responses related to noxious stimuli [24–32] and the "top-down" descending modulatory effects . Lesion in the medial frontal cortex, including the ACC, significantly reduced the behavioral response to noxious stimuli and aversive memory behaviors [24–26]. Also, electrophysiological recordings demonstrate that neurons within the ACC respond to noxious stimuli [6, 28]. Tissue injury or digit amputation activates immediate early gene expression and triggers long-term potentiation of evoked sensory responses in the ACC [27–29]. In mice genetically modified to over express NMDA NR2B receptors in forebrain areas, including the ACC, behavioral responses to tissue inflammation were significantly enhanced . Behavioural allodynia related to inflammation was reduced by injection of antagonists of NMDA receptors or inhibitors of cAMP-dependent protein kinases [30, 32]. These findings indicate that ACC neurons are clearly involved in the processing of noxious stimuli, and demonstrate activity-dependent long-term plasticity in the ACC after tissue injury. In addition, ACC can also serve as a "top-down" descending modulatory system that regulates spinal nociceptive reflexes. Electrical stimulation or chemical injection of glutamate receptor agonists facilitated a spinal nociceptive tail-flick reflex through a descending facilitatory system relayed to the brainstem rostral ventromedial medulla (RVM) [33–37].
It is difficult to distinguish the role of the ACC in pain-related unpleasantness from its descending pain modulatory effects on sensory transmission in the spinal cord by using behavioral withdrawal responses to noxious stimuli. A recent human imaging study reported that the ACC is activated during placebo analgesia . These results suggest that the ACC may also play roles in placebo analgesia. While the physiological nature of the imaged 'hot' spots (i.e., excitation of excitatory versus inhibitory neurons) remains to be determined, it has been proposed that the ACC may activate endogenous analgesia systems due to its projections to the periaqueductal gray (PAG) in the midbrain [7, 38].
Here we propose that the ACC serves as a region for pain unpleasantness in the brain, and excitation of neurons in the ACC can trigger pain unpleasantness but not analgesia. Because animals will never report human-like 'unpleasantness', we used a classic Pavlovian fear memory to measure the effects of stimulation in the ACC. Our operational definition of 'unpleasantness' in mice is based on the formation of fear memory. If ACC stimulation triggers 'unpleasantness' in mice, we expect to observe behavioural freezing responses in mice receiving paired but not unpaired fear-conditioning training. Another obvious advantage of using the Pavlovian fear memory model is to avoid using behavioural withdrawal responses (e.g., hind paws or tail) that are constantly under descending inhibitory and facilitatory modulatory influences. Experiments were performed in both mice and rats to test the hypothesis.
Pavlovian fear memory occurs when a subject learns to associate a certain conditioned stimulus (CS) or cue with a noxious unconditioned stimulus (US) [40–43]. To determine whether ACC stimulation-induced fear memory results from association between the tone and ACC stimulation, we performed experiments where the tone and ACC stimulation were applied but unpaired in another group of animals (Fig. 2b–c). We predicted that while similar contextual memory may form (since animals received the same amount of ACC stimulation in the same environment), auditory memory, which requires the precise pairing of US-CS, would be blocked. Indeed, mice receiving unpaired training demonstrated clear freezing response in the conditioning context (n = 6 mice; P < 0.05 versus baseline responses), whereas no freezing response was observed during tone presentation in the novel context (P > 0.05, comparing behavioral responses before and during the tone; Fig. 2b-–c).
We present strong evidence that the ACC serves as a critical region for pain unpleasantness in the brains of adult mice. Our results are consistent with previous reports from human imaging that show that ACC is important for pain affect or unpleasantness [2, 3, 5]. Clinical reports show that lesions in the human ACC selectively influence the unpleasant component of pain [9, 10]. In animals, lesions in the ACC blocked formalin-induced conditioned place avoidance . Both animal and human studies indicate that neurons in the ACC respond to peripheral painful stimuli or electric shocks [6, 28], and injury triggered activation of immediate early genes in the ACC [27, 29]. Unlike neurons in the somatosensory cortex, neurons in the ACC tend to have widely diffuse receptive fields, and often contain the whole body of animals, supporting its role in coding unpleasantness of pain. Our results using fear memory to measure pain triggered by ACC stimulation avoid the possible contribution of descending modulation to commonly used behavioural nociceptive responses [33–37]. The present study provides a new approach to study the role of pain in higher brain function.
The ACC is a complex and heterogeneous cortex. Neurophysiological recordings, neuropsychological tests and human imaging studies suggest that the ACC plays a key role in cognitive control and is involved in response conflict monitoring [12, 14, 19]. The ACC may be involved in the neural representation of motivational drives, including sexual desire, hunger and the motivational aspect of pain [13, 22]. However, synaptic and molecular mechanisms of the ACC in these higher order functions are largely unknown, due to the lack of animal models. The different roles of the ACC in various functions require caution when explaining the current results. One possible explanation is that ACC neurons express pain affect, general unpleasantness, aspects of cognition and motor response, and pain analgesia. These different functions might be topographically organized along the extent of the ACC, and through its potential interactions with other cortical areas. Furthermore, possible differences between mouse and human brains certainly contribute to some of the different reports on the functions of human ACC vs mice as presented here. For example, Shidara and Richmond  reported that neural activity in the ACC codes the degree of reward expectancy . It is unclear if any of those neurons are also involved in encoding pain unpleasantness. Due to the important role of the ACC in attention and anticipation, another possibility is that ACC stimulation alters learning by affecting such processes. Specifically, the ACC stimulation produces some anticipatory "state" for unpleasantness rather than one of aversion or unpleasantness. To test this possibility, we performed experiments using a newly developed pain-related hot-plate escaping test. Chemical activation of excitatory mGluRs in the ACC enhanced the escape responses. These results suggest that enhanced escape responses and behavioural freezing responses are most likely due to pain-related unpleasantness, and are unlikely to be explained by simply the cessation of activity or goal directed activity. It is also possible that ACC stimulation only causes pain, and the unpleasantness is coded in other regions of the brain. For fear memory, the projection from the ACC to the amygdala plays an important role. Future experiments are clearly needed to address these possibilities. Interestingly, a recent study from Johansen and Fields  showed that excitatory amino acid microinjection into the ACC in rats during conditioning produced avoidance learning in the absence of a peripheral noxious stimulus. These findings indicate that that neurons in the ACC of adult rats and mice mediate both pain-induced negative affect and a nociceptive aversive teaching signal (see ).
Insular cortex has recently been shown to contribute to pain perception [38, 54]. It was shown that the insular cortex may tonically control spinal nociceptive transmission through descending inhibitory systems . By contrast, it has been reported that stimulation in the ACC produced no antinociceptive effect, and facilitated spinal nociceptive tail-flick reflex . Inhibition of excitatory transmission in the ACC by microinjection of opioids produced powerful analgesic effects in freely moving animals , suggesting that the ACC is unique as a centre for unpleasantness. Recent studies in knockout mice for adenylyl cyclases1 and 8 as well as transgenic mice over expressing NR2B receptors consistently indicate that activity-dependent plastic signaling pathways in the ACC may contribute to persistent pain, a classic condition with long-term pain-related unpleasantness [29, 30]. One possibility is that ACC stimulation may activate endogenous analgesia from the ACC, in the absence of nociception, might itself be aversive. To test this possibility, we performed experiments in awake mice and rats using the spinal nociceptive tail-flick reflex, a classic behavioural test for endogenous analgesic/antinociceptive systems . Activation of mGluRs in the ACC by tACPD actually facilitated behavioural responses in both the tail-flick reflex and hot-plate tests, providing direct evidence that endogenous facilitatory but not inhibitory systems are activated. Moreover, intrathecal injection of a serotonergic receptor antagonist methysergide that blocked descending facilitatory modulation from the RVM [34, 38] completely blocked the facilitatory effects in awakened or anaesthetized rats. These findings thus provide the first evidence, to our knowledge, in freely moving animals, excitatory activity in the ACC exert descending facilitatory modulation on spinal nociception
Experiments using lesions in the central nervous system provide important evidence for the involvement of certain structures in learning and memory . However, when using the lesion technique it is difficult to distinguish between roles of the ACC in the expression of freezing responses, or the formation of fear memory. In the present study, we performed a reversible blockade of ACC activity during fear conditioning and found that fear memory was blocked. Our results indicate that the blockade of fear memory is not simply due to the ACC-dependent expression of freezing. One possible mechanism for the roles of the ACC in fear memory is that inputs from the ACC feedback to the amygdala for the formation of fearful memory. Indeed, many ACC neurons directly project to the amygdala [56, 57]. We found that auditory fear memory induced by pairing the ACC stimulation with a tone was blocked by bilateral injection of the NMDA receptor antagonist AP-5 into the amygdala. Contextual memory was not significantly affected, suggesting that other structures may be involved or required for the expression of contextual fear memory . Our results provide new evidence that cortical input from the ACC (coding unpleasantness) is critical for the formation of fear memory (see Fig. 7e). Due to limitations of the microinjection technique, we cannot completely rule out the possible diffusion of AP-5 into other nuclei within the amygdala. Finally, we believe that mouse genetic models will provide ample opportunities for us to explore the molecular mechanisms for high-order brain functions in experimental conditions. Together with imaging and electrophysiological studies in humans and primates, we hope to reveal new central and molecular targets for the treatment of central pain, phantom pain and the unpleasantness related to various mental disorders.
Animals were adult male C57BL6/J mice or Sprague-Dawley albino rats that were housed individually and maintained on a 12/12 h light/dark cycle. Food and water were provided ad libitum.
Animals were anesthetized with sodium pentobarbital (80 mg/kg, i.p.) and implanted unilaterally with a tungsten electrode in the ACC (0.62 mm anterior, 0.4 mm lateral and 1.7 mm ventral to the Bregma) or S1 (0.62 mm anterior, 2.8 mm lateral and 2.3 mm ventral to the Bregma)  under aseptic conditions. The tungsten wire extended 0.1–0.2 mm of the guide cannula that was used as a reference. Mice were allowed to recover for 2 weeks before the experiments. In preliminary experiments, we performed several experiments to locate electrodes into the mouse ACC. All procedures were in accordance with the Animal Studies Committee at the University of Toronto. At the end of the experiment, using standard histological methods, 30 μm brain sections were stained with cresyl violet and examined by light microscopy for electrode placements or cannula penetrations. We only used animals with stable implantation of electrodes that did not exhibit abnormal behaviour during surgical recovery. Results from stimulation sites outside of the ACC were not used in the current studies.
On the day of conditioning, the electrode assembly was connected to the stimulating hardware under brief isoflurane sedation. Mice were allowed 5 min to recover and habituate in the mouse conditioning chamber located in a sound-attenuating box (Med Associates). A commutator (CRIST INSTRUMENT CO.) was used to handle the connecting wires while mice were moving. Each mouse received 3 pairings of training with 1 min in between. The conditioned stimulus (CS) was a tone (2.8 kHz, 85 dB sound pressure level, 30 s), and the unconditioned stimulus (US) was the 10 s ACC stimulation that co-terminated with the tone. The electrical stimulation parameters are as following: 0.3 mA, 0.2 ms pulse duration, 5 pulses at 100 Hz per train, 200 ms train interval. 1 hour, 1 day and 3 days later, animals were first exposed to the conditioning context without tone for 3 min and then to a novel context without tone for 3 min followed by 3 min with tone presentation. Freezing responses was scored visually and presented as percentage of the total period of time observed . Baseline responses were subtracted in order to evaluate responses to the context or tone. For the unpaired training, tone and ACC stimulation were delivered randomly with 40 s – 80 s interval. In some experiments, tACPD (0.25 μg in 0.5 μl) was injected into the ACC unilaterally. After the injection, mice were placed in the conditioning chamber. Each mouse received a tone (see above) for 10 min (at 0–5 and 10–15 min after the tACPD injection).
Under anesthesia of sodium pentobarbital (80 mg/kg), 25-gauge guide cannulas were implanted bilaterally into the ACC (0.62 mm anterior to Bregma, 0.5 mm lateral from the midline, 0.9 mm beneath to the surface of the skull) or BLAC (1.94 mm anterior to Bregma, 3.5 mm lateral from the midline, 4.2 mm beneath to the surface of the skull). Mice were given at least 2 weeks for recovery after cannula implantation. 30-gauge injection cannula was 0.8 mm lower than the guide. For intra-ACC infusion, 0.5 μl muscimol (1 μg/μl) or saline was delivered bilaterally within 90 s using a pump. 15 min later mice were conditioned by 1 pairing of a tone (2.8 kHz, 85 dB, 30 s) and a foot shock (0.75 mA, 2 s) that terminated simultaneously with the tone. For intra-BLAC infusion, 0.5 μl AP5 (2 μg/μl) or saline was delivered bilaterally within 90 s. 15 min later mice received 3 pairings of tone and ACC stimulation, exactly as above.
Adult mice were trained using a modified thermal hot-plate (10" × 10" heating surface) (Columbus Instruments, Columbus) maintained at 50°C, with an escapable non-thermal platform (8" × 5.5" surface). During the first training trial, the escape platform was blocked until the mice showed signs of nociception (e.g., licking of the hind paws). The escape route was then unblocked, and the mice were then free to explore both platforms. The first time entry to the escape platform was recorded. The total duration of each training trial was 3 min. Mice were returned to the same modified thermal hotplate one day after training, with the escape route remaining open, and the temperature set at a room temperature of 22°C. The first time entry to the escape platform was recorded, and the mice were returned back to their home cage upon escape. The maximum test time was 3 min. Mice were tested a total of eight times with an inter-trial interval of 30 min. Mice that remained on the hotplate for the total test duration were recorded with having an escape time of 3 min.
Mice or rats were used either in the awake or halothane-anesthetized states. The spinal nociceptive tail-flick reflex was evoked by noxious radiant heat provided by a 50 W projector lamp focused on a 1.5 × 10 mm area on the underside of the tail. The latency to reflexive removal of the tail from the heat was measured by a digital photocell timer to the nearest 0.1 s. Baseline TF latency was the mean of 3 trials taken at 3 min intervals. For experiments using anesthesia, rats were anesthetized with 2–3% halothane (Ohio Medical Products) delivered via a specialized gas adapter (Stoleting Instrument; IL) with 30% O2 balanced with nitrogen. Body temperature was maintained at 37 ± 0.5°C by a circulating water, thermostatically-controlled heating pad.
A mGluR agonist, trans-(±)-1-amino-(1S, 3R)-cyclopentanedicarboxylic acid (tACPD), was microinjected into the ACC in a volume of 0.5 μl via an injection cannula (33-gauge, 0.20 mm O.D.) inserted through the 26-gauge guide cannula and also extending 2 mm beyond its tip. Injection of tACPD was monitored by following the movement of an air bubble in a length of calibrated tubing between the syringe and the cannula.
Intrathecal catheters (PE-10 tubing, 8.5 cm in length) were inserted through a small opening in the cisterna magna and extended to the lumbar subarachnoid space. In case of experiments in freely moving rats, animals were recovered for at least two weeks before the testing. Intrathecal drug administration was done in the awake or halothane-anesthetized rats. In rats with ACC drug microinjection, baseline TF latencies were determined at 3, 6 and 9 min before and after tACPD injection. After observing the facilitatory effects at 10 min after the injection, methysergide or saline were injected intrathecally to examine if the facilitatory effects may be blocked. The selection of methysergide and dose used are based on previous studies of descending facilitatory modulation from the RVM to the spinal cord [33, 38].
Mice were anesthetized with sodium pentobarbital (80 mg/kg, i.p.). A concentric stimulating electrode was positioned in the right ACC and another concentric recording electrode was placed to the left ACC (both were in Cg2). Dental cement was used to keep the electrodes in place for the recordings of field potential in freely moving mice. Following surgery, mice were allowed to recovery for at least two weeks. Test responses were elicited by monophasic stimuli (200 μs, 75–190 μA, 1/60 s) at an intensity that evoked 40–50% of the maximal responses. fEPSP potentiation is expressed as percentage change relative to the mean baseline response during the 30 min prior to the single footshock stimulation.
Data are presented as mean ± 1 standard error of mean (S.E.M). Facilitation of the TF reflex or hot-plate test is presented as a percentage of the control TF latency. Results were expressed as mean ± s.e.m. One-way ANOVA or two-way ANOVA with repeated measurements was used to compare the differences between treatments. If not stated otherwise, post-hoc comparisons were made with Tukey test. In all cases, p < 0.05 was considered statistically significant.
Supported by grants from the EJLB-CIHR Michael Smith Chair in Neurosciences and Mental Health, Canada Research Chair, NIH NINDS NS42722 and Canadian Institutes of Health Research.
This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.