This paper examines the development and clinical application of classical fear conditioning, as originally demonstrated by Ivan Pavlov, in understanding and treating anxiety disorders. Beginning with the foundational mechanics of conditioned and unconditioned stimuli and responses, the paper traces the neuroanatomical basis of fear learning—particularly the role of the amygdala, hippocampus, and prefrontal cortex—and reviews molecular mechanisms including NMDA receptors and voltage-gated calcium channels. It surveys theoretical models proposed by Eysenck, Davis, and Orr, evaluates empirical findings on fear acquisition and extinction in anxiety patients versus controls, and discusses pharmacological and behavioral interventions such as D-cycloserine, lamotrigine, pregabalin, and propranolol as emerging treatments grounded in conditioning principles.
The paper demonstrates theory-driven literature synthesis: rather than listing studies chronologically, it organizes evidence around competing theoretical predictions (e.g., Eysenck's incubation hypothesis vs. Davis's safety-signal model) and assesses which predictions are empirically supported. This approach shows readers how to use theory as a scaffolding device for reviewing a body of research.
The paper opens with an introduction defining fear conditioning and its historical role in anxiety research, then devotes a section to Pavlov's original classical conditioning paradigm and its extension to fear (including the Little Albert case). A neuroanatomy section details amygdala circuitry and molecular mechanisms. Theoretical models and their empirical tests occupy the central analytical section. Two specialized sections cover memory consolidation/reconsolidation and extinction processes, including pharmacological facilitators. The conclusion synthesizes imaging findings and neurochemical pathways, underscoring areas requiring further study.
Fear conditioning involves the combination of a neutral stimulus with an aversive unconditioned stimulus (US). Initially, the neutral stimulus produces no emotional response, but after repeated pairing with the US, the neutral stimulus becomes a conditioned stimulus (CS). The CS signals the forthcoming US and elicits nervousness and anticipation. Fear conditioning is generally considered an adaptive form of learning. However, when an aversive response to the CS occurs in the absence of a CS/US contingency, fear conditioning may lead to pathology (Lissek et al., 2005).
For at least eighty years, formal theories have implicated fear conditioning in the pathogenesis of anxiety-related disorders (Pavlov, 1927; Watson & Rayner, 1920). More sophisticated work has followed from the introduction of more complex conditioning models in which both fear and anxiety are distinguishable (e.g., Mineka & Zinbarg, 1996; Gorman, Kent, Sullivan, & Coplan, 2000; Grillon & Morgan, 1999; Pine, 1999). This work was followed by animal research identifying the specific brain circuits involved in fear (reviewed by Blair, Schafe, Bauer, Rodrigues, & LeDoux, 2001), and subsequently by evidence supporting the involvement of related brain regions in human fear learning (e.g., Bechara et al., 1995; LaBar, Gatenby, Gore, LeDoux, & Phelps, 1998). The fear conditioning differences identified within anxiety patients are likely to promote future efforts aimed at identifying the neurobiological loci of clinical anxiety.
The classical conditioning model of anxiety disorders holds that pathological anxiety (neurosis) develops through simple classical conditioning (Pavlov, 1927; Watson & Rayner, 1920). Following the introduction of this theory, other writers extended it by arguing that the classically conditioned model also functions as an instrument that drives avoidance behavior (Eysenck, 1976, 1979; Eysenck & Rachman, 1965; Miller, 1948; Mowrer, 1947, 1960). Other versions of this theory emphasize the origins and intensification of fear (Eysenck, 1979), evolutionarily prepared aversive associations (e.g., Öhman, 1986; Seligman, 1971), failure to inhibit responses to safety signals (Davis, Falls, & Gewirtz, 2000), related learning deficits (Grillon, 2002), stimulus generalization (Mineka & Zinbarg, 1996; Watson & Rayner, 1920), and heightened conditionability in the development and persistence of anxiety disorders (Orr et al., 2000; Peri, Ben Shakhar, Orr, & Shalev, 2000).
Ivan Pavlov (1927) most famously demonstrated the original classical conditioning model. The model begins with a description of certain stimuli considered unconditional, which reliably produce an unconditioned response (UR). Through training, a neutral stimulus paired with the US can come to elicit the same response. Under such conditions, the neutral stimulus is regarded as the conditioned stimulus (CS) and the response it generates is the conditioned response (CR). Pavlov performed an experiment on a dog in which food served as the US and the dog's salivation was the UR (Pavlov, 1927).
The next time Pavlov presented food, he also rang a bell (neutral stimulus); he repeated this pairing many times. Finally, he rang the bell (CS) alone without presenting food, and the mere sound of the bell produced salivation in the dog (conditioned response) (Pavlov, 1927).
When a neutral stimulus is paired with an aversive one, the process is known as Pavlovian fear conditioning. This can be illustrated with the case of a young eleven-month-old boy named Albert. When Albert was first presented with a rat to play with (CS), he showed no fear response. The second time Albert was given the rat, it was accompanied by a loud noise (US), and Albert began to cry (CR). This pairing was repeated several times, and eventually Albert cried (CR) at the mere sight of the rat, even in the absence of any noise (CS). A very small stimulus like a rat could thus produce a strong fear response in the child. While fear responses can serve a valuable protective function against potential dangers, they can also become maladaptive. Under maladaptive conditions, a contextual stimulus can become associated with recurring fear and anxiety — a process known as generalization. In typical fear conditioning models, a mild electric shock produces a freezing reaction, a rise in blood pressure, or an increase in heart rate (CR) (Pavlov, 1927).
The amygdala is the brain region responsible for acquiring fear conditioning (Pare et al., 2004). It is located in the medial temporal lobe. Three of the 13 nuclei present in the amygdala are the basal amygdala (BA), the lateral amygdala (LA), and the central nucleus (CA). These three nuclei are implicated in the fear response pathway (Rosen, 2004). The LA receives input from the sensory thalamus, which is then transferred to the central nucleus via a short-loop pathway. The BA provides a link between the LA and the central nucleus. The LA also receives signals from the sensory cortex, insula, and prefrontal cortex through the long-loop pathway (LeDoux, 2000; Sotres-Bayon et al., 2006). From the LA, information is transferred to the effector regions of the brain and to the hypothalamus, which generates the involuntary signs of fear responses (Cahill et al., 1998).
The LA has been identified as the brain region responsible for memory consolidation and plasticity within fear conditioning (Shumyatsky et al., 2002). Severe memory loss can occur if the LA or CA is damaged (Wallace et al., 2001; Blair et al., 2005; Goosens & Maren, 2001). Evidence also indicates that damage to the BA can adversely affect fear reactions (Anglada-Figueroa & Quirk, 2005). The molecular mechanism through which fear acquisition occurs within the LA is called long-term potentiation (LTP) (Chapman et al., 1990). Memory consolidation takes place when calcium enters cells through N-methyl-D-aspartate (NMDA) receptors and voltage-gated calcium channels (VGCCs) (Bauer et al., 2002). Blockage or damage of VGCCs does not affect long-term memory but significantly disrupts short-term memory. This indicates that NMDA receptors are essential for the pathway to be effective (Rodrigues et al., 2001; Walker & Davis, 2000; Cain et al., 2002).
Research conducted on animals revealed that when the NMDA receptor antagonist L-2-amino-5-phosphonovaleric acid (APV, AP5) blocked the NMDA receptor, acquisition of fear was halted but expression was not disrupted (Roesler et al., 2000; Fendt, 2001; Miserendino et al., 1990). However, more recent studies indicate that both processes are inhibited (Jasnow et al., 2004; Maren et al., 1996; Lee et al., 2001). Genetic studies have revealed high expression of NMDA receptors within the hippocampus as well, signifying the importance of this brain structure in Pavlovian conditioning (Mei et al., 2005). When blockage of these receptors occurs within the amygdala, the result is disrupted fear conditioning responses (Zhao et al., 2005; Melik et al., 2006).
The use of preclinical findings is currently limited in humans. However, NMDA receptor blockers and calcium channel blockers are being used to address memory impairment and to treat symptoms of anxiety. Non-competitive NMDA receptor antagonists are increasingly being used as memory-enhancing agents in patients with moderate to severe Alzheimer's disease, accomplishing this by improving neuronal plasticity and decreasing excitotoxicity within the hippocampus (Parsons et al., 1999). Anxiolytic properties have been observed in some animal studies (Bertoglio et al., 2003; 2004), though not in all (Karcz-Kubicha et al., 1997; Harvey et al., 2005), and these agents have not been considered primary anxiolytic mediators in humans. Zarate and colleagues (Zarate et al., 2006), who conducted a double-blind, placebo-controlled trial, were not successful in treating major depressive disorder, though this opened alternative pathways for mood regulation.
For the treatment of seizures and bipolar disorders, lamotrigine is used. It is a glutamate antagonist that acts by blocking voltage-dependent sodium and calcium channels. Using the conditioned emotional response (CER) model in rats, Mirza and colleagues paired houselights (CS) with electric foot shocks to determine whether this CS would result in reduced lever-pressing for food. Research confirmed that a Na⁺ channel agonist blocked the anxiolytic properties of lamotrigine, while Ca²⁺ channels did not. It was thus concluded that the anxiolytic properties of lamotrigine can be attributed to Na⁺ channels (Mirza et al., 2005). When lamotrigine was used with PTSD patients by Hertzberg and colleagues in a small double-blind trial, it was found to be more effective than placebo and reduced PTSD symptoms substantially, including decreasing the likelihood of recurrence (Hertzberg et al., 1999). Voltage-gated calcium channel (VGCC) inhibitors offer another treatment option for anxiety disorders. Pregabalin is considered a therapeutic agent for generalized anxiety disorder (GAD) (Rickels et al., 2005). Pregabalin is an anticonvulsant that binds to the alpha-2-delta protein subunit to block VGCCs.
The history of research demonstrates that the Pavlovian conditioning model has contributed greatly to understanding the pathophysiology of the fear response. These learning models were initially developed and applied in animals, and the knowledge gained has informed how they may be applied to humans. The models also illuminate our understanding of anxiety disorders. Brain imaging techniques such as PET (positron emission tomography) and MRI (magnetic resonance imaging) have identified specific regions of the brain responsible for the acquisition, maintenance, and unlearning of fear. While different areas play distinct roles, the prefrontal cortex and amygdala are the major structures responsible for fear acquisition and extinction. Furthermore, anxiety disorders are also influenced by neurochemical pathways that remain to be studied in greater detail (Garakani, 2006).
Always verify citation format against your institution’s current style guide requirements.