Skip to main content
Browse Subject Areas

Click through the PLOS taxonomy to find articles in your field.

For more information about PLOS Subject Areas, click here.

  • Loading metrics

Human Fear Conditioning and Extinction in Neuroimaging: A Systematic Review

  • Christina Sehlmeyer,

    Affiliations Department of Psychiatry, University of Muenster, Muenster, Germany, Interdisciplinary Center for Clinical Research (IZKF), University of Muenster, Muenster, Germany

  • Sonja Schöning,

    Affiliations Department of Psychiatry, University of Muenster, Muenster, Germany, Interdisciplinary Center for Clinical Research (IZKF), University of Muenster, Muenster, Germany

  • Pienie Zwitserlood,

    Affiliation Department of Psychology, University of Muenster, Muenster, Germany

  • Bettina Pfleiderer,

    Affiliation Department of Clinical Radiology, University of Muenster, Muenster, Germany

  • Tilo Kircher,

    Affiliation Department of Psychiatry und Psychotherapy, Philipps-University Marburg, Marburg, Germany

  • Volker Arolt,

    Affiliation Department of Psychiatry, University of Muenster, Muenster, Germany

  • Carsten Konrad

    Affiliations Department of Psychiatry, University of Muenster, Muenster, Germany, Interdisciplinary Center for Clinical Research (IZKF), University of Muenster, Muenster, Germany, Department of Psychiatry und Psychotherapy, Philipps-University Marburg, Marburg, Germany


Fear conditioning and extinction are basic forms of associative learning that have gained considerable clinical relevance in enhancing our understanding of anxiety disorders and facilitating their treatment. Modern neuroimaging techniques have significantly aided the identification of anatomical structures and networks involved in fear conditioning. On closer inspection, there is considerable variation in methodology and results between studies. This systematic review provides an overview of the current neuroimaging literature on fear conditioning and extinction on healthy subjects, taking into account methodological issues such as the conditioning paradigm.

A Pubmed search, as of December 2008, was performed and supplemented by manual searches of bibliographies of key articles. Two independent reviewers made the final study selection and data extraction. A total of 46 studies on cued fear conditioning and/or extinction on healthy volunteers using positron emission tomography or functional magnetic resonance imaging were reviewed. The influence of specific experimental factors, such as contingency and timing parameters, assessment of conditioned responses, and characteristics of conditioned and unconditioned stimuli, on cerebral activation patterns was examined. Results were summarized descriptively. A network consisting of fear-related brain areas, such as amygdala, insula, and anterior cingulate cortex, is activated independently of design parameters. However, some neuroimaging studies do not report these findings in the presence of methodological heterogeneities. Furthermore, other brain areas are differentially activated, depending on specific design parameters. These include stronger hippocampal activation in trace conditioning and tactile stimulation. Furthermore, tactile unconditioned stimuli enhance activation of pain related, motor, and somatosensory areas.

Differences concerning experimental factors may partly explain the variance between neuroimaging investigations on human fear conditioning and extinction and should, therefore, be taken into serious consideration in the planning and the interpretation of research projects.


Fear conditioning is an ability that is vital for the detection of danger, initiation of self-protection mechanisms, and for survival of a species. Disorders in humans associated with increased anxiety and fear levels, such as posttraumatic stress disorder, phobias, or panic disorder, exemplify how misguided fear conditioning might render originally innocuous stimuli fear-inducing and threatening. In addition, extinction of these associations is also hampered in these disorders. A life time prevalence of anxiety disorders of about 16,6% [1] highlights the substantial clinical and socioeconomic relevance of fear conditioning and extinction.

The term conditioning refers to the process of learning the association between two previously unrelated stimuli [2]. In a typical differential fear conditioning design, a previously neutral conditioned stimulus (CS+) is associated with an aversive and fear-inducing unconditioned stimulus (US) and becomes intrinsically aversive, while another neutral stimulus remains unpaired (CS-) [3]. Two main types of conditioning designs can be distinguished, which differ in the temporal relationship between CS+ and US, hence in the temporal contiguity. In trace conditioning, a time interval ranging from for example 500 milliseconds [4] to 10 seconds [5] separates the presentation of the CS+ from presentation of the US. The expression “trace conditioning” stems from the idea that a memory trace needs to bridge the gap between CS+ and the delayed US to form an association, therefore working-memory processes are more strongly involved in trace conditioning. In contrast, in delay conditioning the CS+ overlaps or is immediately followed by the US. A repeated exposure of the originally neutral stimulus without presenting the aversive stimulus gradually eliminates the fear reaction and is defined as extinction. In the past, extinction was regarded as a process of forgetting this association. However, the phenomena of spontaneous recovery, renewal, rapid acquisition, and reinstatement after extinction, suggest that fear extinction is “an active learning process that is distinct from acquisition and requires additional training to develop” [6].

Fear conditioning has proven to be an extremely robust, rapid, and precise experimental approach for studying the neurobiological substrates of fear [2], [7][15], while fear extinction most probably represents the main therapeutic ingredient of exposure-based psychotherapies. Numerous studies have investigated fear conditioning and extinction in animals and humans, resulting in a core neural network involved in conditioning and extinction (see e.g. [16][18]).

While the literature on animals has been summarized in several review articles (see e.g. [3], [6]), there has been no such approach in the current functional neuroimaging literature on human fear conditioning. So far, only a few reviews have been published and they focus on special topics such as extinction of conditioned fear [19], [20], or socio-cultural and cognitive influences on learning [21]. Büchel et al.'s (2000) review compared event-related fear conditioning studies to block-design studies and positron emission tomography (PET) studies. This important review was one of the first to identify a common core network for human aversive conditioning, including the amygdala and anterior cingulate cortex (ACC) [22]. Other reviews concentrated on cellular and synaptic mechanisms, or on plasticity within this neuroanatomical circuitry [3], [23].

Even though a core network for fear conditioning has consistently been reported in most imaging studies, results obtained from modern neuroimaging techniques differ in many respects, for example, in the number or the type of activated areas.

Therefore, the main aim of this review is to identify consistent and common findings on aversive conditioning and extinction in humans, as assessed by PET and functional magnetic resonance imaging (fMRI), and to present them in a structured manner. The second aim is to look at the differences between neuroimaging studies with respect to neuroimaging results and design parameters. We therefore identify and evaluate typical experimental factors that may influence brain activation patterns and may thereby contribute to the heterogeneity of neuroimaging results. Overall, this review is intended to facilitate the interpretation of seemingly contradictory neuroimaging findings, as well as the selection of an appropriate conditioning design for specific research purposes. Therefore, this review is relevant both to clinicians seeking for a state-of-the-art overview and to researchers investigating fear conditioning or extinction by means of neuroimaging.

The main results of the reviewed studies will be briefly summarized first, followed by an evaluation of specific consequences on activation patterns of critical factors concerning conditioning paradigms, measures of conditioning success, stimuli, and their timing. The review concludes with a critical discussion of these factors and an evaluation of their impact on past and future research.


Literature Search

To identify relevant neuroimaging studies on human fear conditioning and extinction, a computerized database search of journal articles via Pubmed was conducted for the years 1994–2008. This Pubmed search, as of December 2008, used combinations of the keywords “conditioning”, “extinction”, “aversive”, “fear”, “fMRI”, “neuroimaging”, “PET” and “humans”. No truncations and language restrictions were applied. We screened the abstracts for relevant literature based on the literature search criteria and additionally examined the references sections of articles and reviews for potentially useful studies.

Selection criteria

Studies were included if they were: (1) PET or fMRI studies, (2) performed on healthy volunteers, (3) focused on cued fear conditioning and/or extinction. Furthermore, exclusion criteria were: (1) pharmacological modulation, (2) subliminal or masked presentation, (3) context conditioning, (4) combination of fear conditioning with other experimental tasks, such as cognitive-demanding working-memory tasks. Inclusion criteria were applied independently by two reviewers. Specific experimental designs for fear conditioning in fMRI and PET were compared, focusing on the impact of critical experimental variables, such as timing parameters, the contingency rate, or characteristics of the stimuli, on neuroimaging results.

Data Extraction

Data were extracted by the first author (CS) and double-checked independently by the second author (SS). The discrepancies were resolved by consensus and the senior author (CK) was consulted if needed. The following variables were extracted and presented in Table 1: 1) demographic characteristics (number of participants, gender, and age), 2) study design (delay, trace, and extinction), 3) neuroimaging technique (fMRI, PET), 4) characteristics of the stimuli (modality of CS and US), 5) independent assessments of the conditioning process (e.g. heart rate), and 6) neuroimaging results. In the data analysis, the outcomes of interest were brain areas activated during conditioning and extinction. Therefore, we extracted the neuroimaging data presented by each study as the main results. Finally, we extracted those contrasts of interest that represent the conditioning or extinction effect (e.g. CS+>CS-).

Table 1. Forty-six studies on aversive conditioning and/or extinction, with forty studies on delay conditioning (including seven studies on extinction), two studies on trace, and four studies on delay and trace conditioning, with focus on main results of acquisition and/or extinction of conditioned responses (in alphabetic order).

Data Analysis

The review provides a qualitative summary of neuroimaging findings on fear conditioning and extinction of the included empirical studies. These studies were classified according to the type of study design (delay, trace, and extinction), the modality of the CS and US, the contingency rate, and the independent assessment of the conditioned response. For each category, we extracted the absolute frequency of activated brain areas for the contrasts of interest. Moreover, we attempted to identify common and divergent activations across individual study results. Studies, reporting additional or different activation from those described in the core fear network, were examined for the following variables to shed light on reasons for the discrepant findings: conditioning design (delay, trace, and extinction), contingency rate, and characteristics of the CS and US. We refrained from statistically combining results from the studies due to the differences in their design.


Included studies

Based on the literature search strategies, 147 citations were retrieved from the Pubmed database. Among these, we identified 33 relevant studies. Additionally, we examined the references of relevant articles and reviews. Thirteen citations met the selection criteria. As a whole, we reviewed 46 articles on human fear conditioning and/or extinction. Figure 1 shows the search and selection process. Forty studies exclusively used a delay conditioning paradigm during the acquisition phase (Table 1; No. 1–3, 5, 7, 9, 11–18, 20–23, 26, 28–29, 31–33, 35–36, 38–40, 42–45) [4], [5], [24][65]. Only two studies investigated solely trace conditioning during acquisition (Table 1; No. 4, 34) [66], [67], whereas four other studies used both delay and trace conditioning protocols (Table 1; No. 6, 8, 10, 25) [4], [5], [27], [28]. Extinction of learned fear was additionally reported by seven of the 40 delay conditioning studies (Table 1; No. 19, 24, 27, 30, 37, 41, 46) [37], [42], [44], [47], [51], [56], [60]. Thirty-two of the 46 studies are fMRI studies, 14 are PET studies. Table 1 contains information on empirical study characteristics and corresponding neuroimaging results.

Figure 1. QUOROM flow chart used to identify studies for review.

Summary of Findings

Brain regions involved in delay fear conditioning.

As a major and stable result, the amygdala, the ACC and the insular cortex turned out to be crucial structures in the acquisition of aversive delay conditioning, independent of general design characteristics. Twenty-five of the 44 delay conditioning studies reported amygdala activation, with results varying with respect to the laterality of activation. While nine studies reported bilateral amygdala activation (e.g.[31], [49], [64]), eight studies detected left-lateralized (e.g. [27], [61], [65]), and eight right-lateralized activations (e.g.[53], [62]). Methodologically, nineteen of the 25 studies additionally tested for temporal interactions of amygdala activation or split up the acquisition phase into an early and late phase, in order to assess the temporal gradation in the signal intensity of the amygdala. Seventeen of these studies reported learning-related responses of the amygdala (e.g. [45], [48], [57], [66]): fourteen studies found initial increase and rapid decrease of activation during repeated exposure to unpleasant stimuli (e.g. [33], [44]), whereas three studies only reported increases of amygdala activation during the acquisition phase [40], [51], [58]. The remaining 19 delay conditioning studies did not report activation of the amygdala. Seventeen of them did not test for temporal aspects of amygdala activation (e.g. [39]). Sixteen delay conditioning studies found activation of the ACC (e.g. [25], [26]), five of the posterior cingulate cortex (PCC) (e.g. [30]), and two reported activation of the cingulate cortex [33], [57]. Sixteen studies detected insular activities (e.g [39], [54], [65]). These areas are all part of the classical key fear network as described previously [22], [23]. Activation of brain areas such as the hippocampus, the cerebellum, the thalamus, the striatum or the sensory cortices has been reported by fewer delay conditioning studies, underlining the considerable variability in neuroimaging findings. Hippocampal activity, mostly lateralized, was found for example by ten studies (e.g. [34]). Twelve studies showed activation of the striatum (including putamen, accumbens nucleus, caudate nucleus) (e.g. [39], [61]), whereas thalamic activity (including pulvinar, geniculate nucleus) was reported by twelve delay conditioning studies (e.g. [48]) (for details, see Figure 2). As argued below, we believe such differences in results to be methodological in origin [14].

Figure 2. Brain areas involved in aversive conditioning and/or extinction.

Different brain areas (with at least unilateral activation during aversive conditioning and/or extinction) are plotted against the x-axis. The number of studies out of 46 studies per brain region is plotted against the y-axis, taking into account the conditioning design which is delay conditioning in 40, trace conditioning in two, delay and trace conditioning in four, and extinction in seven studies.

Brain regions involved in trace fear conditioning.

So far, only two fMRI studies have employed solely trace conditioning [66], [67] and four fMRI studies were conducted on both delay and trace conditioning [4], [5], [27], [28] (for details, see Table 1), all with either auditory, visual or tactile aversive stimulation. Again, the amygdala and the medial temporal lobe (MTL) were predominantly activated during the acquisition of trace conditioning in five studies (e.g. [4], [27], [66]). Activation of the ACC was apparent in three studies [5], [66], [67] and of the PCC in one study [5]. The hippocampus was bilaterally activated in three trace conditioning studies [5], [27], [66], and two studies showed additional activation of the insula [66], [67]. These fear-related structures such as the amygdala, the hippocampus, the ACC, the insula and the MTL were active independently of US-modality. Furthermore, activation was observed in different areas of the frontal cortex such as the dorsolateral prefrontal cortex (DLPFC) (e.g. [66]) or the middle frontal gyrus [5], [27] in four trace conditioning studies. Activation of other brain areas, such as the cerebellum, was reported in one study, of the motor cortices in three studies (e.g. [4], [5]) (for details, see Figure 2). Again, this variability in study results may be due to critical design characteristics, which will be discussed below.

Brain areas involved in fear extinction.

Although extinction is very relevant in therapeutic settings, only seven studies with focus on extinction met criteria for our review [37], [42], [44], [47], [51], [56], [60]. All seven used a classical delay conditioning design during acquisition. Six studies used a tactile US (e.g. [51]), and one an olfactory US [37]. Three of the seven studies reported major activation foci in the amygdala [37], [42], [44], two in the ACC [51], [60], one study in the PCC [47] and three in the insula (e.g. [37], [47]), whereas four studies observed activation in frontal regions such as the prefrontal cortex (PFC), and the ventromedial prefrontal cortex (VMPFC) (e.g. [60]). Activation of the hippocampus was found in only one study [42] (for details, see Figure 2).

Although consensus exists that the amygdala again plays an important role in extinction, a closer look reveals that the details about amygdala activation vary. As with acquisition, four of the seven studies reported habituation of the amygdala response during extinction [37], [42], [44], [51]. To assess the temporal gradation in signal intensity of the amygdala, two of them split up the extinction phase into an early and late phase [44], [51], and one study tested for time×condition interaction [37]. Knight and co-workers (2004) reported an increase of right amygdala and a decrease of left amygdala activation during extinction, by t-test comparison [42]. Three other studies that did not analyze temporal activation patterns failed to find amygdala activation [47], [56], [60].

The influence of CS-US-contingency.

Contingency describes the rate of pairing between the previously neutral CS+ and the aversive US, and therefore the predictability of the US in relation to the CS. In some cases, the CS is paired with the US on every trial (continuous pairing), whereas in other conditioning designs, CS and US are paired intermittently.

Contingency rates in neuroimaging studies cited here are quite heterogeneous. Twenty-five studies used 100% contingency (e.g. [29], [43], [44], [63]), two employed an 80% or a 90% pairing rate [41], [46], six included a 50% partial reinforcement procedure (e.g. [37]), and eight described lower contingencies of 40%, 33%, 25% or 0% (e.g. [33], [48], [51], [52], [65]). Three studies used 100% and 50% contingency rates during different phases of the experiment [31], [32], [60]. Another study employed a continuous pairing design during trace conditioning and a 50% pairing rate during delay conditioning [27]. One study did not report any contingency rates [30]. Results of the studies cited here indicate that activation of the amygdala seems to be independent of contingency rate: While thirteen studies employing continuous (100%) pairing, eight studies using 50% reinforcement and six studies with 0%, 25%, 33%, 40% and 80% all reported amygdala activation, (e.g. [26], [29], [37], [49], [66]), others with the same pairing rates did not (e.g. [25], [60]).

Awareness about this CS-US-contingency, mediated by conscious US-expectancies or by explicit instruction about the CS-US-contingency, also influences brain activation. Participants were explicitly informed about the CS-US pairing before the experiment in some studies (e.g. [33]), but not in others (e.g. [44]).

Finally, the choice of contingency rates is related to a problem specific to neuroimaging studies: the choice of contrasts between conditions. In a continuous pairing paradigm where the CS+ is always presented with the US, contrasts may be calculated between CS+ and CS- (e.g. [58]), between paired und unpaired subjects (e.g. [42]), or between conditioned and pseudo-conditioned phases - in which CS and US are not correlated in time (e.g. [25]). In a partial-reinforcement design, CS+ may be paired or unpaired with the US. Here, contrasts are mainly calculated between CS+unpaired and CS- (e.g. [26]).

Characteristics of the CS and US.

Neuroimaging studies on fear conditioning have used different types of conditioned and unconditioned stimuli. Conditioned stimuli were presented visually, acoustically or olfactory. Thirty-one studies used a visual cue as CS: five studies used coloured lights (e.g. [28], [42], [53]), one study photographs [30], and four videotapes (e.g. [34], [35]). Seven studies, however, employed photographs of human faces (e.g. [24], [26], [37], [49], [64]), and 14 used geometrical figures (e.g. [51], [58]). Fourteen investigations used auditory conditioned stimuli (e.g. [31], [41], [46], [66]), whereas only one study employed odours [45]. Again, activation of the amygdala was independent of CS-modality: five studies with auditory CS (e.g. [66]), 21 using a visual CS (e.g. [57]) and one study which employed an olfactory CS [45] reported amygdala activation.

Unconditioned stimuli differ in modality (auditory, olfactory, tactile, and visual), in salience, as well as in unpleasantness, factors that may all influence the neurobiology of fear learning. Twenty-four studies used electric shocks (e.g. [5], [27][29], [33], [44]). The intensity of the shock is often assessed and adjusted to an individual level described as “unpleasant but not painful”, such that voltage varied from 40 V to 70 V between participants (e.g. [30], [37], [43], [44], [61], [68]). Electrical stimuli were administered to different areas, such as the wrist (e.g. [44], [51], [52]), shin (e.g. [58]), foot (e.g. [27]), or finger (e.g. [33], [39]). Further tactile stimulations, such as air blasts are reported in eight studies [4], [25], [46], [53], [56], [60], thermal stimulation with hot water in one study [54], and painful phasic esphageal distention in another study [60]. Nine studies cited here included auditory US, such as loud unpleasant tones [26], [66], or loud white noises (e.g. [31], [41], [48], [49]) at intensities of 95dB to 100dB, for 500–1000 ms. A verbal stimulus, a human scream, was presented as unconditioned stimulus in one study [24]. Another study used an olfactory unconditioned stimulus in human fear conditioning, such as “rotten eggs” and “sweaty socks” [37]. Finally, pictures (IAPS; International Affective Picture System [69]) or aversive videotapes were presented as aversive stimuli in three studies [30], [40], [67].

Again, activation of the fear network was observed to be independent of US-modality. In spite of different USs, activations of the amygdala, ACC and insula were reported for every stimulus type. Of the 33 studies with tactile stimulation, fifteen found activation of the amygdala (e.g. [29]), ten of the ACC (e.g. [35]), and ten of the insular cortex (e.g. [39]). Other main activation foci for tactile stimuli concern the thalamus in seven (e.g. [46]), and the striatum in ten studies (e.g. [52]). Other regions such as the occipital cortex, motor or somatosensory cortices are also activated during tactile conditioning in 16 studies (e.g. [27], [35]). By contrast, the nine studies on auditory fear conditioning mainly report activation of the fear network, with emphasis on amygdala in seven (e.g. [31]), on ACC in five (e.g. [48]), and on insula in five studies (e.g. [41]). Moreover, activations of the motor or sensory cortices (e.g. auditory, occipital) are also apparent in five studies (e.g. [26]). The one study on olfactory conditioning mainly reports activations in amygdala, insula and orbitofrontal cortex (OFC) [37], areas that are also associated with the perception of disgust [70], [71]. All three studies on visual aversive conditioning reported activation of key fear areas such as the amygdala and ACC or the PCC [30], [40], [67]. Activation of the insula was found in two of the studies (e.g. [67]). Furthermore, activations of the DLPFC, OFC, thalamus, nucleus accumbens and the occipital cortex are apparent in these visual conditioning studies (e.g. [40]) (for details, see Figure 3).

Figure 3. Brain areas involved in aversive conditioning according to the modality of the US.

Different brain areas (with at least unilateral activation during aversive conditioning) are plotted against the x-axis. The number of studies out of 46 studies per brain region is plotted against the y-axis, taking into account US modality, which is tactile in 33 studies (such as electrical shocks), auditory in nine studies (such as noise), olfactory in one study (such as odors), or visual in three studies (such as aversive pictures).

Our review reveals that 38 of the reviewed studies employed different modalities of US and CS. Only five studies chose an auditory CS paired with an auditory US (e.g. [31], [41], [66]), and three were conducted on visual CS and US (e.g. [40]). Again, research is needed to quantify this effect of common CS-US-modality on neuroimaging results.

Independent assessment of the conditioning process.

A control procedure to ensure that a physiological response towards the CS+ has actually occurred, with data from dependent variables other than brain activation, was used in 41 of the 46 studies cited here. Autonomous, endocrine, or behavioral responses, such as skin-conductance responses, heart rate, verbal responses (ratings of the CS, US-expectancy ratings, or CS-US-contingency assessment), reaction times, or eye-blink reflex qualify as parameters of successful conditioning. The majority of the studies employed independent measures online during scanning. Autonomous measures, such as heart rate, were applied in two [35], [63], skin-conductance responses in 26 (e.g. [4], [32], [50], [62], [64], [65]), and eye-blink startle response in eight studies (e.g. [24], [25], [46], [56]). Only three studies used SCR outside the scanner: before and after conditioning [38] or in an additional experiment [44], [57]. Online assessments of verbal responses, such as CS-ratings, were used in one study [30], and US-expectancy ratings in seven studies (e.g. [28], [31], [62]). Two studies compared ratings of the CS before and after scanning [57], [64]. Twelve studies employed CS-US-contingency ratings and three studies CS-ratings post experimentally (e.g. [27], [40]). To conclude, twenty-three studies combined different measurements of the conditioned response (e.g. [27], [44]) (for details, see Table 1, Figure 4). To summarize, objective measurements are necessary when studying conditioning, to verify that conditioned learning has indeed occurred.

Figure 4. Number of studies employing an independent assessment of the conditioning process.

Different independent assessments of the conditioning process which may be autonomous (such as skin-conductance responses), or behavioral (such as verbal ratings), are plotted against the x-axis. The number of studies out of 46 studies per technique is plotted against the y-axis taking into account if the technique is applied online during scanning, offline after scanning or offline before and after scanning (pre/post).


This review deals with the neural correlates of human fear conditioning in current fMRI and PET studies. Our analysis indicates that neuroimaging studies on human fear conditioning and extinction activate a common core fear network which is in accordance with evidence from other sources (e.g. [17]). Some neuroimaging studies do not find these activations. This heterogeneity is not surprising taking into account the large methodological variety in imaging and design parameters. Methodological differences were found a) in the conditioning protocol (delay, trace), b) in the contingency rate (100%, 80%, 50% or less) and awareness, c) in the modality of CS and US (tactile, auditory, visual, olfactory), and d) with respect to the further assessment of the conditioned response (e.g. psycho-physiological measurements, verbal ratings).

Neuroimaging studies have substantially extended our understanding of fear conditioning and extinction, adding in vivo evidence from humans to previous electrophysiological and lesion studies from animals [3], [72]. Consistent with comparative animal data, neuroimaging investigations have corroborated the finding of a neural fear network activated during fear conditioning. Within this core fear network, key structures for the acquisition and the extinction of conditioned fear have been identified, although there is considerable methodological heterogeneity between studies, with some of them not reporting these activations. Furthermore, it turned out that anatomical regions relevant in fear conditioning are also involved in the extinction of fear memories. In conformity with animal and lesion data, our review indicates that the amygdala, as one principal structure of the limbic system, is one of the key regions involved in fear conditioning and extinction. Amygdala activation occurs in response to emotional stimuli and is therefore regarded as the gate keeper funnelling emotionally relevant information into different processing channels. This region is activated during conditioned-fear acquisition as well as during the expression of learned fear (see for an overview [23]). Furthermore, amygdala activation undergoes rapid habituation during acquisition and extinction that should be taken into account in neuroimaging studies (e.g. [26]). This typical response profile of the amygdala may not be detected by categorical comparisons of e.g. CS+ and CS-, as this contrast reflects time-invariant neural responses. Consequently, some studies carried out an analysis that tested for this type of time-dependent response profile. They set up a statistical model that allows characterizing the activation of the amygdala by a time by condition interaction. Therefore, we suppose that testing for interactions between conditions and time may reveal conditioning results that otherwise remain hidden, such as amygdala activation.

Furthermore, some brain regions, especially the MTL, are difficult to assess using echo-planar imaging (EPI) because they are highly vulnerable to susceptibility artifacts [73]. These differences may cause image distortion and signal dropout [73][75]. This might be another reason why some studies did not find amygdala activation during conditioning. Activation of the insula, another central structure for emotion processing, was also shown in 40% of the neuroimaging studies. Phelps and co-workers (2001) assume that the insula cortex conveys a cortical representation of fear to the amygdala [52], and that uncertainty about the advent of the aversive stimulus during intermittent pairing is reflected by insula and dorsal prefrontal cortex activation [31], [32], [76]. Another region belonging to the core fear conditioning and extinction network described by the majority of the cited neuroimaging studies is the ACC (for an overview, see [17], [26]). The ACC plays an important role in approach and avoidance learning [77] as well as in fear learning [78]. The frontal cortex is particularly crucial for emotional regulation and therefore for the extinction of conditioned fear. Although extinction is the essential process in therapeutic settings, only seven studies have so far focused on extinction. From both animal data and theoretical considerations, it is evident that fear extinction involves mainly interactions between cortical and subcortical structures, such as the PFC and the amygdala or the hippocampus (see for an overview [79]). As one of the principal structures of the brain's extinction circuitry, the PFC regulates the expression of fear by inhibiting the amygdala, such that the fear-conditioned stimulus is prevented from causing a conditioned fear response [79][81]. In this review, only one study reported hippocampal activation during extinction. This is surprising, because from other studies is known that the hippocampus and the VMPFC seem specifically important during late phases of extinction, and therefore for the retention of extinction [51], [82].

There are, however, considerable variances and discrepancies between studies. Whereas some studies only report activation of the core network, others do not find these activations or observe activation within additional brain regions, such as the hippocampus, striatum, sensory cortices or thalamus. The choice of conditioning protocol, CS-US contingency, and modality of the US seem to be very important factors modifying brain-activation patterns in fear conditioning studies.

Our review indicates that of these factors, the conditioning protocol has great impact on brain activation. Delay conditioning leads to more rapid learning of the CS-US association than trace conditioning [83][85]. Thus, from the experimental point of view, delay conditioning has the advantage of a shorter acquisition time, fewer trials, and a more rapid conditioning process than trace conditioning. Additionally, delay conditioning designs are known to extinguish associations faster than those established during trace conditioning [86]. Therefore, all studies that investigated extinction employed delay conditioning in advance. By contrast, in trace conditioning, CS is separated from the US by a temporal gap, resulting in prolonged acquisition times and a larger number of trials being required to form an association. The length of the temporal gap and its distance to the subsequent stimulus also exerts a strong influence. When the US is followed immediately by the next CS, backward conditioning (US-CS associations) or contextual conditioning can occur. In backward conditioning, the US is associated with the next CS, so that no conditioned response is established [87]. Contextual conditioning describes the association of the CS with contextual cues [88], [89]. Hence, there is no contiguity in trace conditioning. While in general, delay and trace conditioning involve comparable fear-related networks, activation of the hippocampus is typical of trace conditioning. In trace conditioning, hippocampal activation is required to bridge the gap between CS and US, retaining a memory trace which is needed to form an association between CS and US [90]. The hippocampus is involved in trace conditioning irrespective of the length of trace interval. However, animal data show that some neurons in the hippocampus encode the duration of trace interval [91]. Thus we assume that the level of hippocampal activation may be enhanced by increasing the length of trace interval.

Another important variable contributing to heterogeneity of neuroimaging results is the CS-US pairing or contingency rate. Effects of CS-US-contingency on conditioning have been repeatedly described in the psychological and behavioral literature [31], [32], [92][94]. Contingency rates determine how fast conditioned responses are acquired, and regulate extinction processes. Our review reveals that the activation of the core fear network consisting of amygdala, insula and ACC is independent of pairing rate, but the time courses of neural responses and the degree of activation may be influenced by contingency. In general, a predictable US is less aversive than an unpredictable US. Therefore, the continuous (100%) pairing of CS+ and US reduces fear responses and activity in fear-related brain areas [31], [32], and promotes the habituation of the amygdala [26], [44], [57], [58], relative to intermittent pairing. Nevertheless, the majority of the studies cited in this review employed a continuous pairing paradigm. In intermittent procedures, US expectancy and response frequency is decreased, which slows conditioning and prolongs the extinction phase [31], [32], [51]. The choice of these pairing parameters has important implications for analysis of imaging data. First, in the light of habituation processes, analysis of time by condition interactions may well improve the detection of amygdala activation. Second, the choice of the contingency rate influences the definition of contrasts of interest between test and control conditions. The cited studies differ in their contrasts of interest which may also influence resulting activation and complicate comparing studies even further. For example, in a 100% pairing design resulting differences in neural responses may be confounded by US-induced BOLD changes. In contrast, in a partial reinforcement design, differences in neural responses are only due to the anticipation of the US.

Our review also illustrates that there is an ongoing controversy on the role of contingency awareness. It seems clear that awareness of the CS-US contingency bridges the CS-US gap in trace conditioning [95], [96]. Therefore, it may be very important for trace conditioning, but less so for delay conditioning. Still, this topic requires further investigation. While some researchers found autonomic fear reactions only in contingency-aware subjects, others reported activation of the fear-network independently of contingency awareness [40], [97], [98]. For example, Phelps et al. (2001) showed that instructions alone can induce fear and that activation of the amygdala can occur without direct experience of the aversive event [52]. Tabbert et al. (2006) explicitly investigated the effect of contingency awareness. They either informed their subjects about the relationship of CS and US or prevented contingency detection by employing a distracter figure or a working-memory task. Amygdala and the OFC were only activated in the unaware group [98], but Klucken et al. (2008) found activation of fear-related areas independent of awareness [40]. However, robust conditioned skin-conductance responses have been observed only in aware participants who acquired a cognitive representation of CS-US-contingencies, and who were able to recall the correct contingency [97]. At this moment, concrete advice as to whether participants should be informed about contingency to obtain faster conditioning responses, is premature.

Concerning the modality of the US and CS, 33 of the 46 studies employed a tactile US, making it the most frequently applied US. Only nine studies used auditory aversive stimuli which may be due to the surrounding and interfering scanner noise. To the best of our knowledge, the problem of scanner noise as being aversive itself has not been discussed so far. The activation of the key fear network including amygdala, ACC and insula seems to be independent of the applied stimuli (auditory, olfactory, tactile, and visual). Nevertheless, many studies do not show activation of the key fear network or observe modality-specific activations. In fear conditioning with tactile US, activation of the thalamus, the striatum, somatosensory and of motor cortices is often reported. These areas are also associated with the nociceptive system, pain anticipation and perception (e.g. [99][102]). The nociceptive system includes the somatosensory cortices, ACC, insula, prefrontal and parietal cortices [103]. Koyama et al. (2005) showed that ACC activation increases with the magnitude of expected pain, and pain-intensity [104]. The thalamus, a major relay site for nociceptive inputs to cortical and subcortical structures, is thought to be responsible for the onset plasticity in the amygdala during fear conditioning [105]. Therefore, we suggest that a “pain-fear network” may be activated during tactile fear conditioning. The one study on olfactory conditioning reported mainly activations of amygdala, ACC and OFC [37]. Odour perception is more often related to disgust than to fear. Disgust and fear are basic emotions with different elicitors and expressions, and appear to be mediated by different neuronal circuits [70], [71], [106], [107]. Therefore, further research is needed to clarify if olfactory conditioning activates a “disgust-fear-network” rather than a mere “fear-network”. To conclude, it seems likely that odours, visual or acoustic stimuli may weaken conditioning effects and may cause activations in different brain regions than electrical stimuli. But to the best of our knowledge, this has never been tested directly in neuroimaging studies. Again, research is needed to quantify the effect of common CS-US-modality on neuroimaging results.

Concerning the modality of the CS, the majority of the studies used visual stimuli as CS, especially photographs of human faces. Faces as CS might be more emotionally relevant to human subjects than tones or coloured lights [108]. However, there seems to be a gender-related effect that needs to be considered in neuroimaging studies. For example, in women, the presentation of faces leads to stronger and persisting amygdala activation, while amygdala activation in men decreases rapidly [109]. Moreover, it is known that the amount of preexposure influences the outcome of aversive learning. These phenomena, so called “latent inhibition” and “US-preexposure effect”, emphasize that novel and unknown CS and US produce more robust conditioning effects than familiar stimuli [31], [110]. The disadvantage of unfamiliar stimuli is the mixing of novelty effects and conditioning effects.

Finally, it is very important to ensure that conditioning really takes place by sampling a second psycho-physiological or behavioral measure to avoid contamination of successful conditioning with unsuccessful trials. Skin-conductance responses as measures of autonomic responses have been widely investigated and are well validated [41]. Classifying subjects as “responders” and “non-responders”, or classifying single trials as “successful” or “not successful” conditioning based on autonomous measures has proven extremely useful, to exclude erroneous trials or subjects from further analysis (e.g. [28], [51]). However, technical issues in the scanner environment have to be solved. Measurement of skin-conductance responses may well prolong the experiment beyond critical time values for such experimental designs. On the other hand, verbal ratings may easily be influenced and consciously manipulated. Alternatives are the assessment of heart rate, or of the startle reflex, which is an elegant measure if an eye-tracker or electromyography is available. In all, the combination of different psycho-physiological and behavioral methods has proven valuable to assure that conditioning has really taken place.

Strengths and Limitations

To the best of our knowledge, this review is the first summarizing current literature on neuroimaging fear conditioning and extinction and providing an overview on similarities and heterogeneities between study results. In this review, we focused on discussing experimental factors that are typical for conditioning paradigms, such as the design (delay, trace), the contingency rate, the contrasts of interests, or the stimuli (CS, US), and that may contribute to the reported heterogeneity in neuroimaging results. Other experimental factors that may influence fear conditioning and fMRI-studies are, for example, the MR-sequence (e.g. [74], [75]), the sample size, gender of participants (e.g. [111], [112]), genetic variables (e.g. [17], [113][115]), or personality factors (e.g. [116][124]). These variables may also contribute to the diversity of neuroimaging results. Another limitation is that our search did not include conditioning studies that were conducted on context conditioning, on patients, on pharmacological interventions, or that included another experimental task. However, we excluded these studies to limit the number of potential influencing variables.


This review provides an overview of 46 current neuroimaging studies on fear conditioning and extinction. Neuroimaging yields new in-vivo evidence with respect to humans revealing and corroborating a consistent pattern of key areas in aversive conditioning and extinction. These structures encompass the amygdala, ACC, and insular cortex for both associative conditioning and extinction. This confirms previous electrophysiological or lesion studies on animals. The key fear-related brain areas, such as amygdala, ACC and insula, are activated independently of specific design parameters. However, some studies still do not report these findings or observe additional modality-specific activations. We pinpointed a number of methodological differences between the functional imaging studies and conclude that these may contribute to the observed variance between results. Prime candidate factors for modifying brain activation patterns are the choice of conditioning protocol, CS-US contingency, and modality of the US. Thus, the contingency and timing parameters, the modality of the CS and US, as well as the assessment of conditioned responses are important for conducting and interpreting neuroimaging studies on fear conditioning and extinction.

Author Contributions

Wrote the paper: CS CK. Analysis and interpretation of data: CK CS SS. Revised the article for intellectual content: CK SS PZ BP TK VA. Final approval of the completed article: CK CS. Performed the literature search: CS. Provided comments on the manuscript: SS PZ BP TK VA.


  1. 1. Somers JM, Goldner EM, Waraich P, Hsu L (2006) Prevalence and incidence studies of anxiety disorders: a systematic review of the literature. Can J Psychiatry 51: 100–113.
  2. 2. Pavlov IP (1927) Conditioned reflexes. Oxford, England: Oxford University Press.
  3. 3. Maren S (2001) Neurobiology of Pavlovian fear conditioning. Annu Rev Neurosci 24: 897–931.
  4. 4. Cheng DT, Disterhoft JF, Power JM, Ellis DA, Desmond JE (2008) Neural substrates underlying human delay and trace eyeblink conditioning. Proc Natl Acad Sci U S A 105: 8108–8113.
  5. 5. Knight DC, Cheng DT, Smith CN, Stein EA, Helmstetter FJ (2004a) Neural substrates mediating human delay and trace fear conditioning. J Neurosci 24: 218–228.
  6. 6. Myers KM, Davis M (2002) Behavioral and neural analysis of extinction. Neuron 36: 567–584.
  7. 7. Pape HC, Stork O (2003) Genes and mechanisms in the amygdala involved in the formation of fear memory. Ann N Y Acad Sci 985: 92–105.
  8. 8. Pape HC, Driesang RB, Heinbockel T, Laxmi TR, Meis S, et al. (2001) Cellular processes in the amygdala: gates to emotional memory? Zoology (Jena) 104: 232–240.
  9. 9. Pape HC (2005) GABAergic neurons: gate masters of the amygdala, mastered by dopamine. Neuron 48: 877–879.
  10. 10. Pape HC, Munsch T, Budde T (2004) Novel vistas of calcium-mediated signalling in the thalamus. Pflugers Arch 448: 131–138.
  11. 11. Pape HC, Narayanan RT, Smid J, Stork O, Seidenbecher T (2005) Theta activity in neurons and networks of the amygdala related to long-term fear memory. Hippocampus 15: 874–880.
  12. 12. Pape HC, Pare D, Driesang RB (1998) Two types of intrinsic oscillations in neurons of the lateral and basolateral nuclei of the amygdala. J Neurophysiol 79: 205–216.
  13. 13. Anderson KC, Insel TR (2006) The promise of extinction research for the prevention and treatment of anxiety disorders. Biol Psychiatry 60: 319–321.
  14. 14. Lissek S, Powers AS, McClure EB, Phelps EA, Woldehawariat G, et al. (2005) Classical fear conditioning in the anxiety disorders: a meta-analysis. Behav Res Ther 43: 1391–1424.
  15. 15. Rachman S (1977) The conditioning theory of fear-acquisition: a critical examination. Behav Res Ther 15: 375–387.
  16. 16. Fendt M, Fanselow MS (1999) The neuroanatomical and neurochemical basis of conditioned fear. Neurosci Biobehav Rev 23: 743–760.
  17. 17. Stoppel C, Albrecht A, Pape HC, Stork O (2006) Genes and neurons: molecular insights to fear and anxiety. Genes Brain Behav 5: Suppl 234–47.
  18. 18. LeDoux JE (2000) Emotion circuits in the brain. Annu Rev Neurosci 23: 155–184.
  19. 19. Barad M, Gean PW, Lutz B (2006) The role of the amygdala in the extinction of conditioned fear. Biol Psychiatry 60: 322–328.
  20. 20. Sotres-Bayon F, Bush DE, LeDoux JE (2007) Acquisition of fear extinction requires activation of NR2B-containing NMDA receptors in the lateral amygdala. Neuropsychopharmacology 32: 1929–1940.
  21. 21. Delgado MR, Olsson A, Phelps EA (2006) Extending animal models of fear conditioning to humans. Biol Psychol 73: 39–48.
  22. 22. Buchel C, Dolan RJ (2000) Classical fear conditioning in functional neuroimaging. Curr Opin Neurobiol 10: 219–223.
  23. 23. Kim JJ, Jung MW (2006) Neural circuits and mechanisms involved in Pavlovian fear conditioning: a critical review. Neurosci Biobehav Rev 30: 188–202.
  24. 24. Anders S, Lotze M, Wildgruber D, Erb M, Grodd W, et al. (2005) Processing of a simple aversive conditioned stimulus in a divided visual field paradigm: an fMRI study. Exp Brain Res 162: 213–219.
  25. 25. Blaxton TA, Zeffiro TA, Gabrieli JD, Bookheimer SY, Carrillo MC, et al. (1996) Functional mapping of human learning: a positron emission tomography activation study of eyeblink conditioning. J Neurosci 16: 4032–4040.
  26. 26. Buchel C, Morris J, Dolan RJ, Friston KJ (1998) Brain systems mediating aversive conditioning: an event-related fMRI study. Neuron 20: 947–957.
  27. 27. Carter RM, O'Doherty JP, Seymour B, Koch C, Dolan RJ (2006) Contingency awareness in human aversive conditioning involves the middle frontal gyrus. Neuroimage 29: 1007–1012.
  28. 28. Cheng DT, Knight DC, Smith CN, Helmstetter FJ (2006) Human amygdala activity during the expression of fear responses. Behav Neurosci 120: 1187–1195.
  29. 29. Cheng DT, Knight DC, Smith CN, Stein EA, Helmstetter FJ (2003) Functional MRI of human amygdala activity during Pavlovian fear conditioning: stimulus processing versus response expression. Behav Neurosci 117: 3–10.
  30. 30. Doronbekov TK, Tokunaga H, Ikejiri Y, Kazui H, Hatta N, et al. (2005) Neural basis of fear conditioning induced by video clip: positron emission tomography study. Psychiatry Clin Neurosci 59: 155–162.
  31. 31. Dunsmoor JE, Bandettini PA, Knight DC (2007) Impact of continuous versus intermittent CS-UCS pairing on human brain activation during Pavlovian fear conditioning. Behav Neurosci 121: 635–642.
  32. 32. Dunsmoor JE, Bandettini PA, Knight DC (2008) Neural correlates of unconditioned response diminution during Pavlovian conditioning. Neuroimage 40: 811–817.
  33. 33. Fischer H, Andersson JL, Furmark T, Fredrikson M (2000) Fear conditioning and brain activity: a positron emission tomography study in humans. Behav Neurosci 114: 671–680.
  34. 34. Fischer H, Andersson JL, Furmark T, Wik G, Fredrikson M (2002) Right-sided human prefrontal brain activation during acquisition of conditioned fear. Emotion 2: 233–241.
  35. 35. Fredrikson M, Wik G, Fischer H, Andersson J (1995) Affective and attentive neural networks in humans: a PET study of Pavlovian conditioning. Neuroreport 7: 97–101.
  36. 36. Furmark T, Fischer H, Wik G, Larsson M, Fredrikson M (1997) The amygdala and individual differences in human fear conditioning. Neuroreport 8: 3957–3960.
  37. 37. Gottfried JA, Dolan RJ (2004) Human orbitofrontal cortex mediates extinction learning while accessing conditioned representations of value. Nat Neurosci 7: 1144–1152.
  38. 38. Hugdahl K, Berardi A, Thompson WL, Kosslyn SM, Macy R, et al. (1995) Brain mechanisms in human classical conditioning: a PET blood flow study. Neuroreport 6: 1723–1728.
  39. 39. Jensen J, McIntosh AR, Crawley AP, Mikulis DJ, Remington G, et al. (2003) Direct activation of the ventral striatum in anticipation of aversive stimuli. Neuron 40: 1251–1257.
  40. 40. Klucken T, Kagerer S, Schweckendiek J, Tabbert K, Vaitl D, et al. (2008) Neural, electrodermal and behavioral response patterns in contingency aware and unaware subjects during a picture-picture conditioning paradigm. Neuroscience.
  41. 41. Knight DC, Nguyen HT, Bandettini PA (2005) The role of the human amygdala in the production of conditioned fear responses. Neuroimage 26: 1193–1200.
  42. 42. Knight DC, Smith CN, Cheng DT, Stein EA, Helmstetter FJ (2004) Amygdala and hippocampal activity during acquisition and extinction of human fear conditioning. Cogn Affect Behav Neurosci 4: 317–325.
  43. 43. Knight DC, Smith CN, Stein EA, Helmstetter FJ (1999) Functional MRI of human Pavlovian fear conditioning: patterns of activation as a function of learning. Neuroreport 10: 3665–3670.
  44. 44. LaBar KS, Gatenby JC, Gore JC, LeDoux JE, Phelps EA (1998) Human amygdala activation during conditioned fear acquisition and extinction: a mixed-trial fMRI study. Neuron 20: 937–945.
  45. 45. Li W, Howard JD, Parrish TB, Gottfried JA (2008) Aversive learning enhances perceptual and cortical discrimination of indiscriminable odor cues. Science 319: 1842–1845.
  46. 46. Logan CG, Grafton ST (1995) Functional anatomy of human eyeblink conditioning determined with regional cerebral glucose metabolism and positron-emission tomography. Proc Natl Acad Sci U S A 92: 7500–7504.
  47. 47. Molchan SE, Sunderland T, McIntosh AR, Herscovitch P, Schreurs BG (1994) A functional anatomical study of associative learning in humans. Proc Natl Acad Sci U S A 91: 8122–8126.
  48. 48. Morris JS, Dolan RJ (2004) Dissociable amygdala and orbitofrontal responses during reversal fear conditioning. Neuroimage 22: 372–380.
  49. 49. Morris JS, Friston KJ, Dolan RJ (1997) Neural responses to salient visual stimuli. Proc Biol Sci 264: 769–775.
  50. 50. Morris JS, Friston KJ, Dolan RJ (1998) Experience-dependent modulation of tonotopic neural responses in human auditory cortex. Proc Biol Sci 265: 649–657.
  51. 51. Phelps EA, Delgado MR, Nearing KI, LeDoux JE (2004) Extinction learning in humans: role of the amygdala and vmPFC. Neuron 43: 897–905.
  52. 52. Phelps EA, O'Connor KJ, Gatenby JC, Gore JC, Grillon C, et al. (2001) Activation of the left amygdala to a cognitive representation of fear. Nat Neurosci 4: 437–441.
  53. 53. Pine DS, Fyer A, Grun J, Phelps EA, Szeszko PR, et al. (2001) Methods for developmental studies of fear conditioning circuitry. Biol Psychiatry 50: 225–228.
  54. 54. Ploghaus A, Tracey I, Gati JS, Clare S, Menon RS, et al. (1999) Dissociating pain from its anticipation in the human brain. Science 284: 1979–1981.
  55. 55. Schreurs BG, Bahro M, Molchan SE, Sunderland T, McIntosh AR (2001) Interactions of prefrontal cortex during eyeblink conditioning as a function of age. Neurobiol Aging 22: 237–246.
  56. 56. Schreurs BG, McIntosh AR, Bahro M, Herscovitch P, Sunderland T, et al. (1997) Lateralization and behavioral correlation of changes in regional cerebral blood flow with classical conditioning of the human eyeblink response. J Neurophysiol 77: 2153–2163.
  57. 57. Straube T, Weiss T, Mentzel HJ, Miltner WH (2007) Time course of amygdala activation during aversive conditioning depends on attention. Neuroimage 34: 462–469.
  58. 58. Tabbert K, Stark R, Kirsch P, Vaitl D (2005) Hemodynamic responses of the amygdala, the orbitofrontal cortex and the visual cortex during a fear conditioning paradigm. Int J Psychophysiol 57: 15–23.
  59. 59. Timmann D, Kolb FP, Baier C, Rijntjes M, Muller SP, et al. (1996) Cerebellar activation during classical conditioning of the human flexion reflex: a PET study. Neuroreport 7: 2056–2060.
  60. 60. Yaguez L, Coen S, Gregory LJ, Amaro E Jr, Altman C, et al. (2005) Brain response to visceral aversive conditioning: a functional magnetic resonance imaging study. Gastroenterology 128: 1819–1829.
  61. 61. Carlsson K, Andersson J, Petrovic P, Petersson KM, Ohman A, et al. (2006) Predictability modulates the affective and sensory-discriminative neural processing of pain. Neuroimage 32: 1804–1814.
  62. 62. Cheng DT, Richards J, Helmstetter FJ (2007) Activity in the human amygdala corresponds to early, rather than late period autonomic responses to a signal for shock. Learn Mem 14: 485–490.
  63. 63. Dimitrova A, Kolb FP, Elles HG, Maschke M, Gerwig M, et al. (2004) Cerebellar activation during leg withdrawal reflex conditioning: an fMRI study. Clin Neurophysiol 115: 849–857.
  64. 64. Petrovic P, Kalisch R, Pessiglione M, Singer T, Dolan RJ (2008) Learning affective values for faces is expressed in amygdala and fusiform gyrus. Soc Cogn Affect Neurosci 3: 109–118.
  65. 65. Schiller D, Levy I, Niv Y, LeDoux JE, Phelps EA (2008) From fear to safety and back: reversal of fear in the human brain. J Neurosci 28: 11517–11525.
  66. 66. Buchel C, Dolan RJ, Armony JL, Friston KJ (1999) Amygdala-hippocampal involvement in human aversive trace conditioning revealed through event-related functional magnetic resonance imaging. J Neurosci 19: 10869–10876.
  67. 67. Nitschke JB, Sarinopoulos I, Mackiewicz KL, Schaefer HS, Davidson RJ (2006) Functional neuroanatomy of aversion and its anticipation. Neuroimage 29: 106–116.
  68. 68. Neumann DL, Waters AM (2006) The use of an unpleasant sound as an unconditional stimulus in a human aversive Pavlovian conditioning procedure. Biol Psychol 73: 175–185.
  69. 69. Lang PJ, Bradley MM, Cuthbert BN (1990) International Affective Picture System (IAPS): instrcuction manual and affective ratings, technical report A4. Gainsville: Center of Research in Psychophysiology, University of Florida.
  70. 70. Schienle A, Stark R, Walter B, Blecker C, Ott U, et al. (2002) The insula is not specifically involved in disgust processing: an fMRI study. Neuroreport 13: 2023–2026.
  71. 71. Stark R, Zimmermann M, Kagerer S, Schienle A, Walter B, et al. (2007) Hemodynamic brain correlates of disgust and fear ratings. Neuroimage 37: 663–673.
  72. 72. Blair HT, Sotres-Bayon F, Moita MA, Ledoux JE (2005) The lateral amygdala processes the value of conditioned and unconditioned aversive stimuli. Neuroscience 133: 561–569.
  73. 73. Morawetz C, Holz P, Lange C, Baudewig J, Weniger G, et al. (2008) Improved functional mapping of the human amygdala using a standard functional magnetic resonance imaging sequence with simple modifications. Magn Reson Imaging 26: 45–53.
  74. 74. Bellgowan PS, Bandettini PA, van Gelderen P, Martin A, Bodurka J (2006) Improved BOLD detection in the medial temporal region using parallel imaging and voxel volume reduction. Neuroimage 29: 1244–1251.
  75. 75. Stocker T, Kellermann T, Schneider F, Habel U, Amunts K, et al. (2006) Dependence of amygdala activation on echo time: results from olfactory fMRI experiments. Neuroimage 30: 151–159.
  76. 76. Volz KG, Schubotz RI, von Cramon DY (2003) Predicting events of varying probability: uncertainty investigated by fMRI. Neuroimage 19: 271–280.
  77. 77. Freeman JH Jr, Cuppernell C, Flannery K, Gabriel M (1996) Limbic thalamic, cingulate cortical and hippocampal neuronal correlates of discriminative approach learning in rabbits. Behav Brain Res 80: 123–136.
  78. 78. Buchanan SL, Powell DA (1982) Cingulate cortex: its role in Pavlovian conditioning. J Comp Physiol Psychol 96: 755–774.
  79. 79. Sotres-Bayon F, Cain CK, LeDoux JE (2006) Brain mechanisms of fear extinction: historical perspectives on the contribution of prefrontal cortex. Biol Psychiatry 60: 329–336.
  80. 80. Quirk GJ, Garcia R, Gonzalez-Lima F (2006) Prefrontal mechanisms in extinction of conditioned fear. Biol Psychiatry 60: 337–343.
  81. 81. Quirk GJ, Likhtik E, Pelletier JG, Pare D (2003) Stimulation of medial prefrontal cortex decreases the responsiveness of central amygdala output neurons. J Neurosci 23: 8800–8807.
  82. 82. Milad MR, Wright CI, Orr SP, Pitman RK, Quirk GJ, et al. (2007) Recall of Fear Extinction in Humans Activates the Ventromedial Prefrontal Cortex and Hippocampus in Concert. Biol Psychiatry 62: 446–454.
  83. 83. Prokasy WF, Whaley FL (1963) Intertrial interval range shift in classical eyelid conditioning. Psychological Reports 12: 55–58.
  84. 84. Gibbon J, Balsam P (1981) Spreading association in time. In: Locurto CM, Terrace HS, Gibbon J, editors. Autoshaping and conditioning theory. New York: Academy Press. pp. 219–253.
  85. 85. Jenkins HM, Barnes R, Barrera FJ (1981) Why autoshaping depends on trial spacing. In: Locurto CM, Terrace HS, Gibbon J, editors. Autoshaping and conditioning theory. New York: Academy Press. pp. 255–284.
  86. 86. Shors TJ (2004) Memory traces of trace memories: neurogenesis, synaptogenesis and awareness. Trends Neurosci 27: 250–256.
  87. 87. Hall JF (1984) Backward conditioning in Pavlovian type studies. Reevaluation and present status. Pavlov J Biol Sci 19: 163–168.
  88. 88. Marchand AR, Luck D, Di Scala G (2004) Trace fear conditioning: a role for context? Arch Ital Biol 142: 251–263.
  89. 89. Marschner A, Kalisch R, Vervliet B, Vansteenwegen D, Buchel C (2008) Dissociable roles for the hippocampus and the amygdala in human cued versus context fear conditioning. J Neurosci 28: 9030–9036.
  90. 90. Bangasser DA, Waxler DE, Santollo J, Shors TJ (2006) Trace conditioning and the hippocampus: the importance of contiguity. J Neurosci 26: 8702–8706.
  91. 91. McEchron MD, Tseng W, Disterhoft JF (2003) Single neurons in CA1 hippocampus encode trace interval duration during trace heart rate (fear) conditioning in rabbit. J Neurosci 23: 1535–1547.
  92. 92. Svartdal F (2003) Extinction after partial reinforcement: predicted vs. judged persistence. Scand J Psychol 44: 55–64.
  93. 93. Leonard DW (1975) Partial reinforcement effects in classical aversive conditioning in rabbits and human beings. J Comp Physiol Psychol 88: 596–608.
  94. 94. Schurr BC, Runquist WN (1973) Acquisition and extinction of human eyelid conditioned response as a function of schedule of reinforcement and unconditioned stimulus intensity under two masked conditioning procedures. J Exp Psychol 101: 398–401.
  95. 95. Knight DC, Nguyen HT, Bandettini PA (2006) The role of awareness in delay and trace fear conditioning in humans. Cogn Affect Behav Neurosci 6: 157–162.
  96. 96. Weike AI, Schupp HT, Hamm AO (2007) Fear acquisition requires awareness in trace but not delay conditioning. Psychophysiology 44: 170–180.
  97. 97. Hamm AO, Weike AI (2005) The neuropsychology of fear learning and fear regulation. Int J Psychophysiol 57: 5–14.
  98. 98. Tabbert K, Stark R, Kirsch P, Vaitl D (2006) Dissociation of neural responses and skin conductance reactions during fear conditioning with and without awareness of stimulus contingencies. Neuroimage 32: 761–770.
  99. 99. Coghill RC, Talbot JD, Evans AC, Meyer E, Gjedde A, et al. (1994) Distributed processing of pain and vibration by the human brain. J Neurosci 14: 4095–4108.
  100. 100. Porro CA, Baraldi P, Pagnoni G, Serafini M, Facchin P, et al. (2002) Does anticipation of pain affect cortical nociceptive systems? J Neurosci 22: 3206–3214.
  101. 101. Rainville P, Duncan GH, Price DD, Carrier B, Bushnell MC (1997) Pain affect encoded in human anterior cingulate but not somatosensory cortex. Science 277: 968–971.
  102. 102. Bornhovd K, Quante M, Glauche V, Bromm B, Weiller C, et al. (2002) Painful stimuli evoke different stimulus-response functions in the amygdala, prefrontal, insula and somatosensory cortex: a single-trial fMRI study. Brain 125: 1326–1336.
  103. 103. Schnitzler A, Ploner M (2000) Neurophysiology and functional neuroanatomy of pain perception. J Clin Neurophysiol 17: 592–603.
  104. 104. Koyama T, McHaffie JG, Laurienti PJ, Coghill RC (2005) The subjective experience of pain: where expectations become reality. Proc Natl Acad Sci U S A 102: 12950–12955.
  105. 105. Quirk GJ, Armony JL, LeDoux JE (1997) Fear conditioning enhances different temporal components of tone-evoked spike trains in auditory cortex and lateral amygdala. Neuron 19: 613–624.
  106. 106. Stein DJ, Arya M, Pietrini P, Rapoport JL, Swedo SE (2006) Neurocircuitry of disgust and anxiety in obsessive-compulsive disorder: a positron emission tomography study. Metab Brain Dis 21: 267–277.
  107. 107. Schafer A, Schienle A, Vaitl D (2005) Stimulus type and design influence hemodynamic responses towards visual disgust and fear elicitors. Int J Psychophysiol 57: 53–59.
  108. 108. Vuilleumier P, Pourtois G (2007) Distributed and interactive brain mechanisms during emotion face perception: evidence from functional neuroimaging. Neuropsychologia 45: 174–194.
  109. 109. Williams LM, Barton MJ, Kemp AH, Liddell BJ, Peduto A, et al. (2005) Distinct amygdala-autonomic arousal profiles in response to fear signals in healthy males and females. Neuroimage 28: 618–626.
  110. 110. Mineka S, Oehlberg K (2008) The relevance of recent developments in classical conditioning to understanding the etiology and maintenance of anxiety disorders. Acta Psychol (Amst) 127: 567–580.
  111. 111. Butler T, Pan H, Epstein J, Protopopescu X, Tuescher O, et al. (2005) Fear-related activity in subgenual anterior cingulate differs between men and women. Neuroreport 16: 1233–1236.
  112. 112. Guimaraes FS, Hellewell J, Hensman R, Wang M, Deakin JF (1991) Characterization of a psychophysiological model of classical fear conditioning in healthy volunteers: influence of gender, instruction, personality and placebo. Psychopharmacology (Berl) 104: 231–236.
  113. 113. Finger EC, Marsh AA, Buzas B, Kamel N, Rhodes R, et al. (2007) The impact of tryptophan depletion and 5-HTTLPR genotype on passive avoidance and response reversal instrumental learning tasks. Neuropsychopharmacology 32: 206–215.
  114. 114. Garpenstrand H, Annas P, Ekblom J, Oreland L, Fredrikson M (2001) Human fear conditioning is related to dopaminergic and serotonergic biological markers. Behav Neurosci 115: 358–364.
  115. 115. Kamprath K, Marsicano G, Tang J, Monory K, Bisogno T, et al. (2006) Cannabinoid CB1 receptor mediates fear extinction via habituation-like processes. J Neurosci 26: 6677–6686.
  116. 116. Gallinat J, Kunz D, Lang UE, Neu P, Kassim N, et al. (2007) Association between cerebral glutamate and human behaviour: the sensation seeking personality trait. Neuroimage 34: 671–678.
  117. 117. Hooker CI, Verosky SC, Miyakawa A, Knight RT, D'Esposito M (2008) The influence of personality on neural mechanisms of observational fear and reward learning. Neuropsychologia 46: 2709–2724.
  118. 118. Keightley ML, Seminowicz DA, Bagby RM, Costa PT, Fossati P, et al. (2003) Personality influences limbic-cortical interactions during sad mood induction. Neuroimage 20: 2031–2039.
  119. 119. Most SB, Chun MM, Johnson MR, Kiehl KA (2006) Attentional modulation of the amygdala varies with personality. Neuroimage 31: 934–944.
  120. 120. O'Gorman RL, Kumari V, Williams SC, Zelaya FO, Connor SE, et al. (2006) Personality factors correlate with regional cerebral perfusion. Neuroimage 31: 489–495.
  121. 121. Otto MW, Leyro TM, Christian K, Deveney CM, Reese H, et al. (2007) Prediction of “fear” acquisition in healthy control participants in a de novo fear-conditioning paradigm. Behav Modif 31: 32–51.
  122. 122. Rauch AV, Ohrmann P, Bauer J, Kugel H, Engelien A, et al. (2007) Cognitive coping style modulates neural responses to emotional faces in healthy humans: a 3-T FMRI study. Cereb Cortex 17: 2526–2535.
  123. 123. Rauch SL, Milad MR, Orr SP, Quinn BT, Fischl B, et al. (2005) Orbitofrontal thickness, retention of fear extinction, and extraversion. Neuroreport 16: 1909–1912.
  124. 124. Sugiura M, Kawashima R, Nakagawa M, Okada K, Sato T, et al. (2000) Correlation between human personality and neural activity in cerebral cortex. Neuroimage 11: 541–546.