Extinction, or the reduction of conditioned responses, occurs after multiple presentations of previously-conditioned stimuli (CS) without the presence of the unconditioned stimuli (UCS) (Milad & Quirk, 2002). Considerable evidence exists that extinction does not reflect unlearning, but instead indicates the meaning of the stimulus becomes obfuscated or ambiguous through learning of additional meaning(s) (Hartley & Phelps, 2009; Milad & Quirk, 2002; Phelps, Delgado, Nearing, & LeDoux, 2004). Others point to possible habituation through the loss of associative strength of the CS (Orsini & Maren, 2012). While habituation is an important part of extinction, inhibitory learning is believed to be the key to extinction (Craske, Treanor, Conway, Zbozinek, & Vervliet, 2014).
Agren, Björkstrand, & Fredrikson (2017) argued about the importance of context in post-extinction behavior, including reinstatement, renewal, spontaneous recovery, and reacquisition. In reinstatement, presenting the unconditioned stimulus alone (after extinction training) can reinstate responding to the conditioned stimulus. In Bouton’s (2002) study, reinstatement to the CS occurred only when the UCS was presented in the same context where the extinguished CS was presented. Bouton indicated renewal is possible when extinction training occurs in a different context than conditioning originally occurred, meaning responding to the CS can return in contexts different from extinction training. This suggests that fear reduction may depend on determining the CS is safe in that context (Herry et al., 2008), which has obvious implications for treatment in humans (Craske et al., 2014). In spontaneous recovery, time may be enough to reduce the effects of extinction because “just as extinction is specific to its physical context, it might be specific to the context of time” (Bouton, 2002, p. 978). Spontaneous recovery can also result from testing in a different context from conditioning and extinction training, which suggests conditioning may be less sensitive to context than is extinction. When the CS is again presented in conjunction with the UCS after extinction, reacquisition of learning occurs quickly, which also suggests the original association was not unlearned, but retrieval may have been affected by extinction training (Herry et al., 2008).
The distributed network consisting of the basolateral complex of the amygdala (BLA), the medial prefrontal cortex (mPFC), and the hippocampus is believed to be the site of encoding for bidirectional changes in behavior related to fear during both extinction and renewal (Hartley & Phelps, 2009; Knox, 2016). Cholinergic neurons located in the basal forebrain project to the hippocampus, the anterior cingulate cortex (ACC), the infralimbic cortex (IL), and other areas of the medial prefrontal cortex (mPFC). These cholinergic neurons are associated with contextual fear memory consolidation, the consolidation of contextual conditioning, and modulation of contextual extinction memory (Knox, 2016). Rauch, Shin, & Phelps (2006) reported that amygdala hypersensitivity, in addition to inadequate control by the ventromedial prefrontal cortex, may mediate hyperarousal and contribute to extinction deficits.
Knox et al. (2016) described a neural circuit for fear extinction that included the PL, IL, the dorsal and ventral areas of the hippocampus, and the BLA. They reported that single prolonged stress (SPS), a commonly-used model of persistent fear in PTSD, can be useful in understanding how stress can impede extinction training. In addition, they reported SPS leads to decreased spine arborization and increased neural apoptosis and spine density, which can result in increased excitation in the BLA. These factors are believed to decrease inhibitory effects on the IL. Understanding neuroanatomy and the effects of stress can lead to increasingly effective interventions. The maintenance of pathological fear memories is widely-believed to result in disorders of fear, anxiety, PTSD, and specific phobias (Orsini & Maren, 2012). Understanding the properties associated with acquisition and extinction of fear responses is a primary target for clinical treatments, including prolonged exposure (PE; Cooper, Clifton, & Feeny, 2017) and cognitive processing therapy (CPT; Farmer, Mitchell, Parker-Guilbert, & Galovski, 2017).
Amygdala
According to Herry et al. (2008), the amygdala is a vital region in the mediation of defensive behaviors related to anxiety and fear. The amygdala has also been implicated in the regulation of protective behaviors and the attribution of emotional significance to stimuli (Orsini & Maren, 2012). They indicated that amygdalar activation is necessary for fear memory retrieval. They also reported that activation after extinction is believed to be bound by the mPFC through local inhibitory circuits. The BLA, which contains the lateral (LA) and basal (BA) nuclei, is actively involved during fear extinction. Herry et al. conducted an experiment to explore individual BA neuron plasticity by studying the rates of action potentials using two different auditory CS, one paired with a foot shock (CS1) and one that was not (CS-). Two different, non-segregated classes of neurons emerged. One class, called “fear neurons,” experienced an increase in spike firing to CS1 during and post-conditioning that was eliminated during extinction and converted into CS1-evoked inhibition. A different class of “extinction neurons” emerged during extinction training. These extinction neurons did not show increased activity to CS1 during conditioning (instead showed a slight reduction), but did show increased CS1-evoked activity during extinction training. To further determine whether extinction neurons could serve as fear neurons to different stimuli, the researchers conditioned mice to different stimuli, CS1 and CS2, and conducted extinction training with just CS1. They found that extinction neurons responded only to CS1, while fear neurons responded only to the non-extinguished CS2. They further found that activity in extinction neurons began one trial before declines in activity in fear neurons were observed. They report that extinction neurons connect to the mPFC and hippocampus reciprocally, while fear neurons were found to project to the mPFC exclusively. After experimenting with BA inactivation, the researchers discovered that extinction acquisition and context-dependent renewal relies on the BA.
The sensory interface for cortical and thalamic sensory afferents is the LA (Herry et al., 2008), which is believed to encode CS- and UCS-related signals (Hartley & Phelps, 2009; Repa et al., 2001). Herry et al. reported no findings of increased firing to CS1 during extinction in the LA, which may indicate extinction neurons are indigenous to the BA. The central nucleus (CE) of the amygdala projects to the hypothalamus, intercalated cells, and the midbrain and may modulate the passive expression of fear CR (Hartley & Phelps, 2009; Milad & Quirk, 2002). Herry et al. indicated the LA also projects to the BA, which sends information to the striatal nucleus accumbens (NA). The striatum is believed to be involved with active coping through the integration of motivation and action, and in contributing to decision-making and reward-based learning. This belief was supported by findings that indicate reduced striatal activation in individuals diagnosed with PTSD (Hartley & Phelps, 2009). Repa et al. reported that transiently plastic cells were located most in the dorsal LA and tended to receive direct thalamic projection, while long-term plastic cells tended to be localized in the ventral subnucleus and received more intra-amygdalar and cortico-amygdala projections. They suggest that these findings may “differentially account for the initial learning and subsequent maintenance of the long-term memory of the training experience” and may suggest dorsal LA processing contribute to, but may not be essential for, learning fear responses (p. 727).
Bouton (2002) indicated that both conditioning and extinction depend on amygdalar N-methyl-D-aspartate (NMDA) glutamate receptors. Further, Milad and Quirk (2002) suggest long –term fear extinction memory requires NMDA, but short-term extinction memory may be independent of NMDA. Hartley and Phelps (2009) stated that blocking NMDA or glutamate receptors in the BLA reduces extinction learning and blocking the activity of BLA mitogen-activated kinase (MAPk) can completely hinder extinction acquisition.
Using functional magnetic resonance imaging (fMRI) to explore blood-oxygen-level dependent contrast imaging, or BOLD imaging, Phelps et al. (2004) reported increased BOLD responding in the amygdala during acquisition training to the CS and decreased responding during extinction. They reported that while activity was found in both the right and left amygdala, the right amygdala consistently showed greater BOLD responding and was the only side with significant results. These findings were supported by Rauch et al.’s (2006) study of responses to fearful faces.
Hippocampus
Reciprocally connected to the BLA is the hippocampus. The hippocampus is engaged in fear conditioning, fear extinction, and renewal through the processing of contextual stimuli (Herry et al., 2008). Herry et al. also reported the hippocampus is a site responsible for contextual information processing related to both extinction and expression of conditioned fear behavior. They indicated the hippocampus provides input for fear neurons, but not to extinction neurons, which may help override extinction memory retrieval, allowing for the expression of fear after extinction. Though the hippocampus is involved with contextual memory initially, memories are ultimately transferred and made independent of the hippocampus (Orsini & Maren, 2012).
Orsini & Maren (2012) report that while the hippocampus is vital for contextual learning of fear, other brain structures can compensate in the absence of hippocampus functionality. They report the ventral hippocampus is required for auditory and contextual fear acquisition and has reciprocal amygdala connection and robust projections to the NA. Orsini and Maren further suggest that the pairing of a weak CS stimulus with a strong UCS results in both pathways becoming stronger through long-term potentiation. They point to findings that inactivation of dorsal and ventral hippocampus blocked renewal, which they suggest is evidence the hippocampus uses contextual clues to disambiguate CS meanings.
Lang et al. (2009) reported posterior hippocampus BOLD responding was found in early acquisition, while dorsal hippocampus responding occurred mostly during late acquisition (it was found, yet nonsignificant in early activation). Interestingly, Phelps et al. reported no significant hippocampal responses in extinction learning, which supports their network hypothesis. Additionally, decreased hippocampal and ventromedial prefrontal cortex activity is implicated in reduced ability for explicit memory and the identification of contextual safety (Rauch et al., 2006). Additionally, Rauch et al. found higher regional cerebral blood flow (rCBF) in the right ACC, paralimbic regions, and the right amygdala during reminders of traumatic experiences.
Frontal, Prefrontal, and Medial Prefrontal Cortices
The mPFC contains the prelimbic (PL) and infralimbic (IL) cortices and has been implicated in long-term extinction memory (Milad & Quirk, 2002), extinction memory consolidation (Phelps et al., 2004), the suppression of expression of conditioned fear in the amygdala (Sotres-Bayon, 2004), and the renewal of fear (Orsini & Maren, 2012). Both the IL and PL are major targets of projections from the hippocampus and IL input from the hippocampus has been described as a major source of brain-derived neurotropic factor (BDNF) that is required for conditioned fear suppression (Orsini & Maren, 2012). Milad and Quirk recorded data from 74 neurons in the PL, IL, and medial orbital (MO) cortices. They found no IL activity during habituation and conditioning, which suggests the IL differs from dorsal mPFC regions that signal fear conditioning acquisition and may instead reflect extinction memory. Neither the PL or MO cortices responded to extinction tones, which signals “ a high degree of anatomical specificity in the ability of mPFC to signal extinction memory” (Milad & Quirk, 2002). The IL receives excitatory projections from the BLA, and may be a site for potentiation during extinction consolidation (Hartley & Phelps, 2009). IL stimulation both reduces CR rates and CE responsiveness (Quirk, Likhtik, Pelletier, & Paré, 2003). The IL is also believed to be involved in extinction consolidation (Orsini & Maren, 2012).
Lang et al. (2009) found BOLD responses in the mPFC and in both orbitofrontal and supplementary motor cortices (as well as the hippocampus and left amygdala). These researchers found significant BOLD response in the dorsal anterior cingulate cortex (dACC) exclusively during extinction. The left medial frontal gyrus and the left precuneus (in the superior parietal lobe, rostral to the occipital lobe) were also found to show BOLD responses only during extinction. The precuneus is believed to be involved in episodic memory recall. In the early stages of acquisition, Lang et al. (2009) reported “functional coupling between the left posterior hippocampus (seed region) and the bilateral orbitofrontal cortex, the right rostral anterior cingulate, the bilateral precuneus, and the contralateral hippocampus” (p. 827). The frontal cortex is implicated in context conditioning. Along with the insula (which is involved with emotional and arousal processing of stimuli) and parietal cortex, these findings further suggest context-related conditional may involve a network of brain regions (Lang et al., 2009).
Clinical Implications
According to Scheveneels, Boddez, Vervliet, & Hermans (2016), the extinction of fear is amongst the most effective models in experimental psychopathology. It also serves as the foundation for exposure-based therapies (Cooper et al., 2017; Craske et al., 2014). Internal cognitive mechanisms are a primary vehicle for emotional regulation by cognitive-behavioral interventions and have shown treatment efficacy for fear and myriad other psychological disorders (A-Tjak et al., 2015; Craske et al., 2014; Hartley & Phelps, 2009). Agren et al. (2017) pointed to reconsolidation as an effective therapeutic point of intervention as activation of a fear memory makes said memory sensitive to modification. They stated that new data that is incongruent with the fear memory can weaken the fear response. They focused primarily on imaginal exposure, where the client’s fear memories were verbally activated and processed to re-assess the negative perceptions of the event(s), as more studies have investigated in vivo exposure. The researchers fear conditioned the entire sample, then compared in vivo and imaginal exposure one day after acquisition. After random assignment to four treatment groups, two groups (one imaginal group and one in vivo group) participated in exposure ten minutes after reactivation of the fear memory. The other two groups (one imaginal group and one in vivo group) participated in exposure training six hours after reactivation, which they suggested is outside the reconsolidation window. Their findings suggested that reduction of conditioned responses can occur after both imaginal and in vivo exposure. They also reported differences in stimulus discrimination between the ten-minute and six-hour groups. Agren et al. indicated that stimulus discrimination was eliminated in both types of exposure groups exposed ten minutes after activation, but not in the six-hour exposure groups. They suggested this was the result of mediation by increased response to CS- in lieu of decreased CS+ response. They also posit these findings point to new learned safety memories. After analyzing exposure types, they reported that reduced responding to CS+ in the ten-minute group was found only for those who received in vivo exposure, evidence that the exposure types are effective thought not equivalent.
A primary challenge faced by therapists is the longevity of fear memories and their ability for cross-contextual generalizability (Orsini & Maren, 2012). Hartley and Phelps (2009) posit that using cognitive regulation strategies like reappraisal, suppression, and selective attention can help reinterpret event significance, direct attention to less distressing situational elements, or change event engagement. They point to the association between decreased amygdalar response and increased prefrontal cortex response during cognitive regulation as evidence for an inhibitory relationship, which is vital to fear control. Further, Bryant et al. (2007) found the success of cognitive-behavioral treatment for PTSD can be predicted by fearful face responses in the amygdala and subgenual ACC using fMRI activation. Active coping, or committed action, is among the most commonly used strategy in emotional regulation and involves deliberate engagement in activities that are likely to result in positive outcomes and avoidance of activities with a high likelihood of negative outcomes (Hartley & Phelps, 2009; Hayes, Pistorello, & Levin, 2012).
Bouton (2002) suggested that therapy may involve context-specific learning that might result in fears that are reduced in one context returning in others. Bouton also suggests that while extinction learning is context-specific, conditioning tends to be less sensitive to context and more stable, which may explain the persistence of behavior disorders. One possible explanation includes that extinction is not the first learning experience with the CS, which suggests that extinction results from conditional or context-specific exceptions to initial learning. These findings suggest that extinction would benefit most if training occurs over time and in multiple contexts, especially within the same contexts where the symptoms are most problematic for the client, like in vitro exposure (Cooper et al., 2017; Craske et al., 2014).
Results from neuroimaging studies indicate extinction through inhibitory learning shares similar neural mechanisms with exposure-based interventions. Future research may expand the usage of multi-sensory stimuli to increase common factors to extinction training and exposure-based interventions (Scheveneels et al., 2016). This becomes especially important as a considerable percentage of clients fail to experience clinically-significant levels of symptom reduction (Craske et al., 2014).
References
A-Tjak, J. G. L., Davis, M. L., Morina, N., Powers, M. B., Smits, J. A. J., & Emmelkamp, P. M. G. (2015). A meta-analysis of the efficacy of acceptance and commitment therapy for clinically relevant mental and physical health problems. Psychotherapy and Psychosomatics, 84(1), 30–36. http://doi.org/10.1159/000365764
Agren, T., Björkstrand, J., & Fredrikson, M. (2017). Disruption of human fear reconsolidation using imaginal and in vivo extinction. Behavioural Brain Research, 319, 9–15. http://doi.org/10.1016/j.bbr.2016.11.014
Bouton, M. E. (2002). Context, ambiguity, and unlearning: Sources of relapse after behavioral extinction. Biological Psychiatry, 52(10), 976–986. http://doi.org/10.1016/S0006-3223(02)01546-9
Bryant, R. A., Felmingham, K., Kemp, A., Das, P., Hughes, G., Peduto, A., & Williams, L. (2007). Amygdala and ventral anterior cingulate activation predicts treatment response to cognitive behaviour therapy for post-traumatic stress disorder. Psychological Medicine, 38(04), 255–7. http://doi.org/10.1017/S0033291707002231
Cooper, A. A., Clifton, E. G., & Feeny, N. C. (2017). An empirical review of potential mediators and mechanisms of prolonged exposure therapy. Clinical Psychology Review, 56, 106–121. http://doi.org/10.1016/j.cpr.2017.07.003
Craske, M. G., Treanor, M., Conway, C. C., Zbozinek, T., & Vervliet, B. (2014). Maximizing exposure therapy: An inhibitory learning approach. Behaviour Research and Therapy, 58(c), 10–23. http://doi.org/10.1016/j.brat.2014.04.006
Farmer, C. C., Mitchell, K. S., Parker-Guilbert, K., & Galovski, T. E. (2017). Fidelity to the Cognitive Processing Therapy Protocol: Evaluation of Critical Elements. Behavior Therapy, 48(2), 195–206. http://doi.org/10.1016/j.beth.2016.02.009
Hartley, C. A., & Phelps, E. A. (2009). Changing fear: The neurocircuitry of emotion regulation. Neuropsychopharmacology, 35(1), 136–146. http://doi.org/10.1038/npp.2009.121
Hayes, S. C., Pistorello, J., & Levin, M. E. (2012). Acceptance and commitment therapy as a unified model of behavior change. The Counseling Psychologist, 40(7), 976–1002. http://doi.org/10.1177/0011000012460836
Herry, C., Ciocchi, S., Senn, V., Demmou, L., Müller, C., & Lüthi, A. (2008). Switching on and off fear by distinct neuronal circuits. Nature, 454(7204), 600–606. http://doi.org/10.1038/nature07166
Knox, D. (2016). The role of basal forebrain cholinergic neurons in fear and extinction memory. Neurobiology of Learning and Memory, 133(C), 39–52. http://doi.org/10.1016/j.nlm.2016.06.001
Knox, D., Stanfield, B. R., Staib, J. M., David, N. P., Keller, S. M., & DePietro, T. (2016). Neural circuits via which single prolonged stress exposure leads to fear extinction retention deficits. Learning & Memory, 23(12), 689–698. http://doi.org/10.1101/lm.043141.116
Lang, S., Kroll, A., Lipinski, S. J., Wessa, M., Ridder, S., Christmann, C., et al. (2009). Context conditioning and extinction in humans: Differential contribution of the hippocampus, amygdala and prefrontal cortex. European Journal of Neuroscience, 29(4), 823–832. http://doi.org/10.1111/j.1460-9568.2009.06624.x
Milad, M. R., & Quirk, G. (2002). Neurons in medial prefrontal cortex signal memory for fear extinction. Nature, 420(6911), 70–74. http://doi.org/10.1038/nature01144
Orsini, C. A., & Maren, S. (2012). Neural and cellular mechanisms of fear and extinction memory formation. Neuroscience and Biobehavioral Reviews, 36(7), 1773–1802. http://doi.org/10.1016/j.neubiorev.2011.12.014
Phelps, E. A., Delgado, M. R., Nearing, K. I., & LeDoux, J. E. (2004). Extinction learning in humans: role of the amygdala and vmPFC. Neuron, 43(6), 897–905. http://doi.org/10.1016/j.neuron.2004.08.042
Quirk, G. J., Likhtik, E., Pelletier, J. G., & Paré, D. (2003). Stimulation of medial prefrontal cortex decreases the responsiveness of central amygdala output neurons. The Journal of Neuroscience, 23(25), 8800–8807.
Rauch, S. L., Shin, L. M., & Phelps, E. A. (2006). Neurocircuitry Models of posttraumatic stress disorder and extinction: Human neuroimaging research—Past, present, and future. Biological Psychiatry, 60(4), 376–382. http://doi.org/10.1016/j.biopsych.2006.06.004
Repa, J. C., Muller, J., Apergis, J., Derochers, T. M., Zhou, Y., & LeDoux, J. E. (2001). Two different lateral amygdala cell populations contribute to the initiation and storage of memory. Nature Neuroscience, 4(7), 724–731.
Scheveneels, S., Boddez, Y., Vervliet, B., & Hermans, D. (2016). The validity of laboratory-based treatment research: Bridging the gap between fear extinction and exposure treatment. Behaviour Research and Therapy, 86(C), 87–94. http://doi.org/10.1016/j.brat.2016.08.015
Sotres-Bayon, F. (2004). Emotional perseveration: An update on prefrontal-amygdala interactions in fear extinction. Learning & Memory, 11(5), 525–535. http://doi.org/10.1101/lm.79504
