When Threat Is Near,Get Out of Here

Abstract

When detecting a threat, humans and other animals engage in defensive behaviors and supporting physiological adjustments that vary with threat imminence and potential response options. In the present study, we shed light on the dynamics of defensive behaviors and associated physiological adjustments in humans using multiple psychophysiological and brain measures. When participants were exposed to a dynamically approaching, uncontrollable threat, attentive freezing was augmented, as indicated by an increase in skin conductance, fear bradycardia, and potentiation of the startle reflex. In contrast, when participants had the opportunity to actively avoid the approaching threat, attention switched to response preparation, as indicated by an inhibition of the startle magnitude and by a sharp drop of the probe-elicited P3 component of the evoked brain potentials. These new findings on the dynamics of defensive behaviors form an important intersection between animal and human research and have important implications for understanding fear and anxiety-related disorders.

Keywords

defensive behavior freezing active avoidance psychophysiology startle reflex

Dangers can lurk around every corner, and it is a vital capacity of an organism to quickly and effectively select among a limited set of defensive responses once a threat is encountered. Because environmental conditions can change quickly, being able to dynamically switch between defensive behaviors enhances the probability of survival. Research with animals and humans shows that autonomic responses, protective reflexes, and brain responses change systematically with increasing threat proximity, as outlined in the threat-imminence and defense-cascade models (Fanselow, 1994; Lang, Bradley, & Cuthbert, 1997; Löw, Lang, Smith, & Bradley, 2008). These models conceptualize defensive behavior in three distinct stages. In the first stage (pre-encounter defense), if an organism is in a context in which a threat has been encountered previously but has not yet been detected, preemptive behavior, including threat-nonspecific vigilance, is engaged. Then, as soon as a threat is detected (postencounter defense), attention is selectively allocated to the threat cue, and defensive response output is characterized by fear bradycardia (Campbell, Wood, & Mcbride, 1997), potentiation of the startle reflex gated through the central nucleus of the amygdala (Lang, Davis, & Ohman, 2000), and motor freezing, which depends on projections from the amygdala to the ventral periaqueductal gray (PAG) in the midbrain (Fanselow, 1994; Maren, 2001; Morgan & Carrive, 2001). During the final circa-strike stage (when the threat is most imminent), defensive behavior switches from passive (freezing) to active behavioral strategies (fight or flight); this switching is mediated by the dorsal PAG, which directs the expression of escape behavior (Kim et al., 2013; LeDoux, 2012), and is accompanied by a general discharge of the sympathetic nervous system (Cannon, 1932).

Numerous studies have demonstrated that the whole-body startle reflex (in animals) or the eyeblink component of the startle reflex (in humans) is reliably potentiated when the individual is confronted with a fear-conditioning cue or an anxiety-provoking context (Davis, 2000; Hamm & Weike, 2005; Lang et al., 2000). In early animal research, however, fear-potentiated startle was observed only when the fear cue was paired with foot shocks of mild or moderate intensities (Walker, Cassella, Lee, De Lima, & Davis, 1997). Pairing the same cue with very intense foot shocks resulted in a relative inhibition of the startle reflex and increased motor activity. Notably, lesions of the dorsolateral PAG abolish the inhibition of the startle reflex and increase freezing (De Oca, DeCola, Maren, & Fanselow, 1998). In contrast, damage to the ventrolateral PAG disrupts defensive freezing, and stimulation produces attentive freezing and hyporeactive immobility (Fanselow & Poulos, 2005; for a review, see Benarroch, 2012).

Löw and coworkers (2008) observed this switching in defensive response modulation—depending on the proximity of threat—for the first time in humans. In that study, emotional pictures first appeared at a distance, and potentiation of the startle response was observed during the presentation of threat cues. However, when pictures grew progressively closer and avoiding loss of money depended on reaction speed, startle responses to the same cues were inhibited at the last stage, when the threat was most imminent. In that study, however, there was no approaching threat without the option to avoid the aversive event.

In the current study, we traced the brain and body dynamics of the defensive system in humans, assessing autonomic arousal, protective brain stem reflex activity, and evoked brain potentials in an instructed-fear experiment. In this experiment, the proximity of the threat stimulus increased and the opportunity to actively avoid a painful stimulus was either available or not available, which is typical for fear conditioning (see Fig. 1). First, a cue (colored frame) signaled whether or not participants had the option to avoid the painful stimulus (active vs. passive coping). After 500 ms, another symbol indicated whether or not a threat would occur (safe vs. threat). This cue increased in size in five stages, which signaled the increasing proximity of an aversive electrical stimulus in the threat condition. In the safe condition, no painful stimulus was delivered.

Fig. 1.

Schematic representation of the instructed-fear paradigm. In each of five stages, a colored frame appeared (500 ms), and then a figure appeared in the center of the frame (1,500 ms). The figure increased in size with each stage as though it were approaching the participant. The type of figure (circle or star) signaled whether the trial would potentially end with a painful electric shock (threat condition) or no shock (safe condition). The color of the frame (blue or yellow) indicated whether the electric shock could be avoided by quickly pressing a response button (active condition) or whether the electric shock was inevitable (passive condition). In the active-threat condition, a slow response was followed by a painful shock; in the passive-threat condition, no active coping with the threat was possible, and a painful shock was randomly delivered on 50% of all trials. In the safe condition, a shock was never administered.

We expected attentive freezing and potentiated startle responses with increasing threat imminence when the aversive stimulus was unavoidable. If, however, avoidance was an option, we expected inhibition of the startle amplitude in the circa-strike zone, which would reflect a behavioral switching from freezing to active avoidance probably mediated by midbrain areas, as suggested by animal data. According to previous findings (Löw et al., 2008), we anticipated increased activation of the sympathetic nervous system with increasing threat imminence. Moreover, we expected that increased freezing behavior would be reflected in a clear fear bradycardia in the circa-strike zone (when the aversive stimulus was unavoidable), comparable with findings in prey animals confronted by a predator (Campbell et al., 1997).

In addition to the physiological adjustments during approaching threat, we also wanted to examine the dynamics of selective attention during detection and processing of the approaching threat cue. To study these cognitive processes, we used the P3 component of the event-related potential (ERP) in response to an acoustic startle probe as a measure of attention. Attenuation of the probe P3 component has previously been shown when participants viewed emotional (compared with neutral) foreground pictures, which indicates heightened allocation of attention to the affectively engaging foreground (Bradley, Codispoti, & Lang, 2006). We predicted that more attention would be allocated to the foreground stimulus as the threat grew closer. This increased selective attention, however, should be much stronger if the threat would depend on the behavior, that is, if active avoidance were an option. Thus, focused preparation for active avoidance should be reflected in a reduced P3 amplitude to the task-irrelevant acoustic probe stimulus.

Method

Participants

We aimed for a sample size of 30 participants (on the basis of the results reported by Löw et al., 2008), and data collection was stopped at the end of the semester. At this time, 27 students (22 female, 5 male; mean age = 23.4 years, SD = 3.9 years) from the University of Greifswald had participated in the study for course credit or payment (€12). All participants were right-handed and had normal or corrected-to-normal vision. All provided written informed consent for the protocol approved by the University of Greifswald Institutional Review Board. Twenty-six participants had analyzable heart rate, startle, and electroencephalogram (EEG) data, as 1 participant for each measure had to be removed because of excessive artifacts.

Materials and task

We used an instructed-fear paradigm in which each trial consisted of a cascade of five stages. Each stage began with the presentation of a colored frame (500 ms). This frame remained on screen while a figure (a star or a circle) appeared in the center of the frame for the remainder of that stage (1,500 ms; see Fig. 1). The figure was larger at each stage, which created the impression that it was approaching the participant. The figures served as threat or safety cues, signaling whether the trial would potentially end with a painful electric shock (threat condition) or no shock (safe condition). The color of the frame (blue or yellow) signaled whether the electric shock could be avoided by quickly pressing a response button (active condition) or whether the electric shock was inevitable (passive condition). Geometric figures and colors of the frame were counterbalanced across participants. Each of the five stages lasted for 2 s; thus, one complete cascade lasted for 10 s. In total, 24 trials for each of the four conditions (passive threat, passive safe, active threat, active safe) were presented in randomized order with a randomly varying interval (6.5 to 8.5 s) between each cascade. There was a short break in the middle of the experiment.

Stimuli were presented on a 20-in. LCD monitor with a resolution of 1,024 × 768 pixels. Circles had a diameter of 36, 108, 180, 252, and 335 pixels, respectively, across the five stages. The stars (10-pointed) and circles had identical luminance and area for the corresponding stages. The colored frame had a size of 825 × 625 pixels and a line width of 50 pixels.

In the active-threat condition, an electrical shock could be delivered at the end of the trial. Participants had to press a response button with their right index finger as fast as possible as soon as the last figure of the trial had disappeared from the screen. If the button press was fast enough (i.e., within a certain time window), no shock was delivered. If the button press fell outside the time window, the shock was delivered. In the active-safe condition, the button press had no consequences, but participants were still instructed to respond as fast as possible because response times (RTs) would be recorded. The time window was initially set to 240 ms and adjusted during the experiment depending on the performance of the participant. Once five responses were registered in the active-threat condition, the time window was always set to the median value of those responses so that the task was kept challenging, and in about half of the trials in this condition, a shock was delivered. To keep the number of shocks comparable in the passive-threat condition, we set the probability of a shock to 50% in each trial. Shocks were delivered 1,000 ms after the last stimulus of the sequence.

In each of the four conditions, participants also received acoustic startle probes consisting of a burst of white noise. One startle probe was delivered on 20 of the 24 total trials per condition. Thus, for each condition, there were four startle probes for each of the five stages, and in 16 out of 96 total trials, no startle probe was delivered. Startle probes were presented between 900 and 1,200 ms following the onset of the safety/threat cue.

Experimental procedure

Participants gave their written informed consent before participating in the study. After arrival in the laboratory, they were seated in a reclining chair in a dimly lit, sound-attenuated room, and the sensors were attached for physiological recordings and electrical stimulation. The task was explained, and an example of each type of stimulus was presented. Then, the intensity of the electric stimulation was adjusted by administering five pulses (which participants were warned to expect) to determine a level that was described as “highly annoying, but not painful.” Participants practiced the task, and the experimenter checked whether the instructions had been understood. The geodesic sensor net for EEG recordings was applied according to the specifications of the manufacturer. In-ear headphones (OMX 980; Sennheiser, Weddemark, Germany) were attached, and four startle probes were delivered to familiarize participants with the probes. If participants had no further questions, the meaning of the frame colors and the symbols were repeated before starting the experiment. Six startle probes were delivered with an interstimulus interval of 10 s before the first trial started.

Data recording and analysis

RTs

For each participant and condition, the median response after excluding trials with RTs below 100 ms (2.78% of all trials) was computed.

Heart rate

An electrocardiogram (ECG) was recorded from the forearms (an Einthoven I setup) to measure heart rate. The signal was band-pass filtered (1–13 Hz) and amplified using a Coulborn S75-01 amplifier (Coulborn Instruments, Whitehall, PA). Digital sampling occurred with a frequency of 1,000 Hz. Heart rate was derived from the ECG signal after a visual check for artifacts and correction whenever misplaced R-wave triggers had occurred. Change scores were then computed in half-second bins as proposed by Graham (1978) from a 3-s baseline for each trial and averaged across conditions.

Electrodermal activity

Electrodermal responses were registered by placing Ag/AgCl electrodes 15 mm apart on the hypothenar eminence on the palmar surface of the left hand. A constant voltage of 0.5 V across the sensors was provided by a Coulborn S71-22 skin-conductance coupler. The signal was sampled at 10 Hz. Digital values were converted to microsiemens and averaged across conditions in segments of 16 s (including a 3-s baseline) across trials for each condition. Change scores were then computed in half-second bins from the 3-s baseline.

Startle reflex

The acoustic startle probe was a 103 dB(A) burst of white noise presented binaurally through in-ear headphones for 50 ms. The eyeblink component of the startle reflex was measured using an electromyogram recorded over the left orbicularis oculi muscle. The signal was amplified using a Coulborn V75-01 amplifier, high-pass (30 Hz) and low-pass filtered (400 Hz) before digital sampling at 1,000 Hz. Sampling started 100 ms before onset of the acoustic startle stimulus and lasted for 500 ms. An off-line high-pass filter (60 Hz) was applied before the data were integrated with a time constant of 10 ms. Responses were scored (Globisch, Hamm, Schneider, & Vaitl, 1993) and converted to microvolts. To ensure that each participant contributed equally to the groups’ mean, we standardized the raw startle magnitudes to T scores (50 + (z × 10)). Trials without response were scored as zero magnitudes, and trials with artifacts were removed from the analysis (1.04% of trials).

ERPs

The EEG was sampled at 500 Hz using a 257-channel system (Electrical Geodesics, Eugene, OR); a vertex reference was used, and the signal was filtered on-line (0.1–200 Hz). Bad channels were later interpolated using spherical splines, and the data were filtered off-line at 0.1 through 30 Hz, corrected for blink artifacts (Ille, Berg, & Scherg, 2002), and referenced to the average reference. ERPs evoked by the startle probe were averaged in 700-ms segments (including a 100-ms prestimulus baseline) for each condition and stage of the sequence. On the basis of visual inspection and other studies (e.g., Schupp, Cuthbert, Bradley, Birbaumer, & Lang, 1997), we averaged samples across space (a cluster of 14 sensors around Cz; see Fig. S1 in the Supplemental Material available online) and time (N1 component: 90–120 ms, P3 component: 180–240 ms) for statistical analysis.

Electrical shock

The electrical shock (a 1-ms pulse) was generated by a commercial stimulator (S48K; Grass Instruments, West Warwick, RI) and transmitted via constant current unit (CCU1, Grass Instruments) to a bipolar electrode attached to the participant’s right calf. The intensity of the stimulation was adjusted individually to each participant’s tolerance. On average, the intensity was 10.8 mA (SD = 5.85).

Statistical analysis

For all measures, an overall analysis of variance (ANOVA) with repeated measures of time (24 half-second bins for heart rate and electrodermal activity, 5 data points each for startle reflexes and probe ERPs), defensive behavior (active vs. passive), and threat condition (threat vs. safe) was performed. Separate post hoc ANOVAs were conducted for interactions among time, defensive behavior, and condition at a significance level of p < .1. To substantiate attentive freezing during passive defense in the context of increasing threat imminence, we compared the safe with the threat condition during passive defensive behavior (i.e., when the shock was uncontrollable). To test the impact of the possibility of avoiding the shock, we analyzed defensive responses during the active- and the passive-threat condition for the postencounter phase as well as the circa-strike phase as follows. For heart rate, the factor time consisted of 20 half-second bins for the postencounter phase and the following 4 half-second bins for the circa-strike phase. As electrodermal responses have a delay of at least 1 s (Boucsein et al., 2012), the data points included for postencounter and circa-strike analysis were shifted for 1 s accordingly. For startle reflex and probe ERPs, responses to the probes during the first four stages were used for postencounter analysis, and responses to the last probe (just prior to the offset of the final stimulus that was the cue to respond) were used for circa-strike analysis. Analyses for linear and quadratic trends were performed for significant main effects and interactions including the factor time.

All statistical tests used a significance level of p < .05. Greenhouse-Geisser corrections of degrees of freedom were applied whenever necessary, and effect sizes are reported as partial eta-squared.

Results

Performance

RTs were faster in the threat condition (median = 189.8 ms, 95% confidence interval, or CI = [181.4, 198.1]) than in the safe condition (median = 233.5 ms, 95% CI = [206.9, 260.1]), F(1, 26) = 22.10, p < .001, η_p² = .383.

Skin conductance, heart rate, and startle reflex

The time course of the electrodermal activity, heart rate, and startle-blink reflexes for all conditions are illustrated in Figure 2. Defensive responses varied dynamically with fear imminence as well as whether or not the participants had the opportunity for active avoidance (see Fig. 3), as revealed by three-way interactions among the factors defensive behavior, threat condition, and time—heart rate: F(23, 575) = 5.24, p < .001, η_p² = .173; electrodermal activity: F(23, 598) = 3.23, p < .044, η_p² = .110; startle: F(4, 100) = 2.21, p < .093, η_p² = .081. Follow-up analyses tested the hypotheses that were derived from the theoretically important questions, specifically comparing attentive freezing in the passive-threat and the passive-safe conditions and testing the impact of active avoidance by comparing active and passive defense during approaching threat imminence in the postencounter and circa-strike zone.

Fig. 2.

Mean response magnitude for the four conditions combining the possibilities of a shock (threat vs. safe conditions) with the possibility of avoiding the shock (active vs. passive conditions). Results are shown across the five threat stages and subsequent period of the possible shock, separately for (a) electrodermal activity, (b) heart rate, and (c) fear-potentiated startle. In (a) and (b), changes are shown relative to baseline values.

Fig. 3.

Mean difference in response magnitude between the passive-threat and passive-safe conditions, when a possible shock could not be actively avoided. Results are shown across the five threat stages and subsequent period of the possible shock, separately for (a) electrodermal activity, (b) heart rate, and (c) fear-potentiated startle. In (a) and (b), changes are shown relative to baseline values; in (c), the change in T scores is shown.

When participants had no opportunity for actively avoiding the shock (i.e., in the passive-threat compared with the passive-safe condition), there was a significant interaction between defensive behavior and time, and electrodermal activity increased with increasing imminence of the threat, F(19, 494) = 11.64, p < .001, η_p² = .309, linear trend: p < .001, quadratic trend: n.s. Then, during the immediately following circa-strike phase, electrodermal activity further increased under threat, compared with the safe condition, F(3, 78) = 18.09, p < .001, η_p² = .410, linear trend: p < .001, quadratic trend: n.s. Moreover, strong fear bradycardia occurred in the circa-strike zone when the aversive event was inevitable, F(3, 75) = 28.19, p < .001, η_p² = .530, linear trend: p < .001, quadratic trend: n.s. While heart rate deceleration started prior to stimulus offset in both active conditions (preceding acceleration), heart rate deceleration was only obvious in the threat condition starting at stimulus offset. Such fear bradycardia is typically observed in animals during strong attentive freezing. Startle-reflex magnitudes were significantly larger when evoked during the passive-threat condition compared with the passive-safe condition, F(1, 25) = 21.46, p < .001, η_p² = .462. Fear-potentiated startle increased linearly with increasing proximity of the threat when the shock was uncontrollable, F(4, 100) = 3.84, p > .018, η_p² = .133, linear trend: p < .006, quadratic trend: n.s. (see Fig. 3).

The physiological correlates of defensive behavior changed dramatically when participants were given the opportunity to actively avoid the painful stimulus (Fig. 4), but only in the circa-strike zone. Autonomic electrodermal and cardiac responses as well as startle modulation did not differ between the active and the passive conditions during the postencounter phase, electrodermal activity: F(19, 494) < 1; heart rate: F(19, 475) = 2.12, p < .093, startle: F(3, 75) = 1.45, p = .24. This pattern changed substantially in the circa-strike zone (see Fig. 4). When compared with the passive-threat condition, sympathetic discharge in the active-threat condition was strongly enhanced during preparation for active avoidance, as indicated by a large increase in electrodermal activity, F(3, 78) = 24.25, p < .001, η_p² = .483, linear trend: p < .001, quadratic trend: p < .003, but also by a strong heart rate acceleration, F(1, 25) = 14.78, p ≤ .001, η_p² = .371 (see Fig. 4).

Fig. 4.

Mean difference in response magnitudes between the active- and passive-threat conditions. Results are shown across the five threat stages and subsequent period of the possible shock, separately for (a) electrodermal activity, (b) heart rate, and (c) fear-potentiated startle. In (a) and (b), changes are shown relative to baseline values; in (c), the change in T scores is shown.

In contrast to the passive (freezing) condition, startle magnitudes were abruptly inhibited during preparation for actively avoiding the aversive shock, F(1, 25) = 20.69, p < .001, η_p² = .453. This suggests that protective reflexes are potentiated during attentive freezing but inhibited during preparation for active avoidance, which probably protects the organism from processing distracting cues that might interfere with the effective flight response.

To ensure that the electrical stimulus had no impact on the physiological data, we analyzed the electrodermal-activity and heart rate data by splitting the trials in the threat condition into those with and without an electrical stimulus. Data in the circa-strike phase (compared with a baseline just prior to the circa-strike phase) did not differ depending on whether or not an electrical stimulus was delivered—heart rate: F(1, 25) < 1; electrodermal activity: F(1, 26) < 1.

ERPs

We evaluated two ERP components in response to the startle probes, the N1 component—which mainly indexes preattentive sensory processing (Parasuraman & Beatty, 1980)—and the P3 component, as an index of selective attention (Isreal, Chesney, Wickens, & Donchin, 1980). There was a main effect of time, as evidenced by the early negative peak (N1) increasing linearly with increasing threat imminence, F(4, 100) = 7.873, p < .001, η_p² = .240, linear trend: p < .001, quadratic trend: n.s., with the largest amplitude at the circa-strike zone (final probe position vs. any other: ps < .01; see Fig. 5). This effect was stronger during the threat condition than during the safe condition, F(1, 15) = 4.43, p < .046, η_p² = .150. In contrast to N1 amplitudes, P3 amplitudes decreased significantly with increasing proximity of the threat, which revealed a main effect of time, F(4, 100) = 8.37, p < .001, η_p² = .251, linear trend: p < .002, quadratic trend: p < .001; this was particularly true when individuals had the option to avoid the shock, F(4, 100) = 2.92, p < .033, η_p² = .104, linear trend: p < .05, quadratic trend: p < .02 (Fig. 5). Reduction of the P3 amplitudes was stronger for the threat condition compared with the safe condition, F(1, 25) = 8.5, p < .007, η_p² = .254. Again, the period just prior to active avoidance was critical: P3 amplitudes were not modulated by threat imminence in the passive condition. However, probe-evoked P3 amplitudes just prior to active avoidance were significantly reduced relative to all other stages (all ps < .01), whereas all other pairwise comparisons between stages showed no differences (see also Fig. 5).

Fig. 5.

Event-related potentials evoked by the startle probe. The waveforms (a) show the mean brain potentials evoked by startle probes at each of the five stages. The topographical map shows the voltage distribution at the peak of the N1 (Stage 5). Mean amplitude of the probe-evoked N1 component (b) is shown across the five threat stages. Mean amplitude of the probe-evoked P3 component (c) is shown as a function of threat stage and whether or not the possible shock in the threat condition could be actively avoided.

Discussion

Our data provide direct evidence for a dynamic organization of defensive behavior in humans supporting the threat-imminence model derived from animal research. First, autonomic, somatic, and brain responses varied systematically with the increasing imminence of a potential threat and the behavioral options available. Second, at the imminent circa-strike stage, startle responses were potentiated during passive anticipation of the threat but were inhibited when active coping was possible. This provides a direct measure for the switch from attentive freezing to active avoidance. These behavioral changes were associated with appropriate physiological adjustments to support effective action. Third, the cognitive system also supports effective defensive responding by sensitizing sensory encoding and reducing selective attention to irrelevant stimuli.

Inevitable threat and passive freezing

When there is no possibility to actively avoid the aversive stimulus, humans—as rodents do—show defensive behavior that can be described as attentive freezing. With approaching imminence of the potential threat, sympathetic activation increases linearly, which suggests increased orienting and autonomic arousal to the approaching threat. Heart rate decreased more quickly during the threat than during the safe condition, starting by the last picture of the approaching-threat sequence. Heart rate deceleration has been reliably observed during increased orienting toward external stimuli (Bradley, 2009) and has been associated with facilitated sensory intake (Graham & Clifton, 1966) coupled with somatic inhibition (Obrist, 1981). Heart rate deceleration can become fear bradycardia as attention is focused on the predator, and immobility is a means to avoid discovery (Campbell et al., 1997). Strong heart rate deceleration was observed when the painful stimulus was inevitably approaching.

The cortical data support this pattern of defensive reactivity. The increase in the sensory N1 component of the evoked brain potential to the acoustic probe would also support this interpretation. Again, the N1 component was larger during the threat condition than during the safe condition, which suggests that the general alertness to external stimuli (in this case, the acoustic probe) increased with increasing proximity of the threat. Supporting previous research (for a review, see Hamm & Weike, 2005), our results showed that startle-eyeblink responses were significantly potentiated when elicited during a sequence of cues predicting the occurrence of an aversive event relative to when those cues did not predict an aversive event. Fear-potentiated startle increased linearly with increasing size of the aversive stimulus. These data are in line with findings showing temporal specificity of startle potentiation, with strongest potentiation when the electric shock was expected (Grillon, Ameli, Merikangas, Woods, & Davis, 1993).

Threat and active avoidance

The pattern of defensive reactivity changed substantially when the aversive stimulus was avoidable: a sharp increase in skin conductance, a direct measure of sympathetic arousal, immediately prior to the initiation of the motor response as well as a strong heart rate acceleration when active avoidance was possible. According to the cardiac-somatic-coupling hypothesis (Obrist, 1981), cardiac acceleration is associated with somatic activation in the behavioral context of defense. Notably, a drastic reversal of the fear-potentiated-startle reflex became evident when individuals could actively avoid the electrical stimulus. This pattern of the fear-potentiated startle is remarkably consistent with findings from animal experimentation relating the switch from startle potentiation to startle inhibition to different patterns of defensive behaviors in rodents that are modulated by different subregions of the PAG (Benarroch, 2012; Fanselow, 1991; Walker et al., 1997). The current data suggest that different subregions of the PAG might also regulate adjustments of defensive reflexes in humans, probably actively inhibiting postencounter attentive freezing defense by the dorsolateral PAG during preparation for active avoidance. Data from neuroimaging studies by Mobbs and coworkers support these interpretations. Mobbs et al. (2007, 2009, 2010) found increased activation of the midbrain PAG area and an increased coupling between the midbrain and the mid dorsal anterior cingulate when a threat became imminent, although no further discrimination between different subregions of the PAG depending on different defensive response patterns was possible in these functional MRI studies. Preliminary evidence from our own lab suggests that PAG activation under threat is more pronounced during preparation for active avoidance compared with attentive freezing in humans (Wendt et al., 2014).

Selective attention and defensive action

The P3 component to probes presented during the last picture of the approaching-threat sequence was significantly reduced when active avoidance was possible. In other words, the inhibition of the startle reflex was associated with a reduced P3 amplitude elicited by the same probe stimulus. The P3 amplitudes to secondary acoustic startle probes are also significantly reduced when individuals view emotionally arousing visual stimuli (Bradley et al., 2006; Schupp et al., 1997) or experience unpleasant interoceptive cues (Alius, Pané-Farré, Löw, & Hamm, 2015; Ceunen, Vlaeyen, & Van Diest, 2013), which suggests that more attentional resources are allocated to the visual or interoceptive foreground stimuli that predict the occurrence of the threat and set the stage for effective active avoidance. As a consequence, elaborated processing of the secondary acoustic probe is reduced. This reduction of the P3 component was specific for the active condition, which suggests that the blocking of irrelevant stimuli is relevant only when effective action is required.

Conclusions

In a recent article, LeDoux (2014) argues very convincingly that terms such as “fear conditioning” and “fear system” blur the distinction between processes that give rise to feelings that we call fear and processes that control perception and defensive responses to threat. The current data support this view that defensive behavior is not an entity but is rather dynamically organized depending on the imminence of the threat and the behavioral repertoire at hand. When the threat is inevitable—as in a typical Pavlovian fear-conditioning procedure—the conditioned stimulus elicits defensive attentive freezing characterized by a robust potentiation of the startle reflex that increases with progressive proximity of the approaching threat. Studies that have investigated fear learning have mainly focused on this form of defensive freezing. If the organism, however, has the opportunity to actively avoid the progressing threat, startle responses are inhibited in preparation of active avoidance. These findings are in line with animal data from Moscarello and LeDoux (2013) showing that active avoidance directly opposes the expression of conditioned freezing. Training animals for active avoidance recruits infralimbic prefrontal cortex (ilPFC) to inhibit central-amygdala-mediated expression of conditioned freezing. Pretraining lesions of the ilPFC increased conditioned freezing while causing a corresponding decrease in active avoidance.

Our findings are not only theoretically important for understanding the dynamic organization of defensive systems and circuits but also have substantial clinical implications. In a clinical study, Richter and coworkers (2012) found a strong potentiation of the startle reflex in a high proportion of patients with panic disorder and agoraphobia while being trapped in a small dark room. Notably, for patients who left the small room prematurely (thus showing active avoidance), heart rate increased before escape and startle responses were inhibited, which shows the same blockade of attentive freezing during active avoidance. Indeed, by blocking active avoidance during exposure therapy, we would expect to increase defensive freezing again. A better understanding of these interactions between defensive behaviors and their underlying neural circuits might prove important for improving behavioral treatments of anxiety disorders.

Footnotes

Acknowledgements

A. Löw is now at the Department of Humanities and Social Sciences, Helmut-Schmidt-University/University of the Federal Armed Forces Hamburg, Germany.

Declaration of Conflicting Interests

The authors declared that they had no conflicts of interest with respect to their authorship or the publication of this article.

Funding

This research was funded by the German Research Foundation (Ha 1593/18-1).

Supplemental Material

Additional supporting information can be found at

References

Alius

M. G.

Pané-Farré

C. A.

Löw

Hamm

A. O.

(2015). Modulation of the blink reflex and P3 component of the startle response during an interoceptive challenge. Psychophysiology, 52, 140–148. doi:10.1111/psyp.12295

Benarroch

E. E.

(2012). Periaqueductal gray: An interface for behavioral control. Neurology, 78, 210–217. doi:10.1212/WNL.0b013e31823fcdee

Boucsein

Fowles

D. C.

Grimnes

Ben-Shakhar

Roth

W. T.

Dawson

M. E.

, . . . Society for Psycho-physiological Research Ad Hoc Committee on Electrodermal Measures. (2012). Publication recommendations for electrodermal measurements. Psychophysiology, 49, 1017–1034. doi:10.1111/j.1469-8986.2012.01384.x

Bradley

M. M.

(2009). Natural selective attention: Orienting and emotion. Psychophysiology, 46, 1–11. doi:10.1111/j.1469-8986.2008.00702.x

Bradley

M. M.

Codispoti

Lang

P. J.

(2006). A multi-process account of startle modulation during affective perception. Psychophysiology, 43, 486–497. doi:10.1111/j.1469-8986.2006.00412.x

Campbell

B. A.

Wood

Mcbride

(1997). Origins of orienting and defensive responses: An evolutionary perspective. In Lang

P. J.

Simons

R. F.

Balaban

(Eds.), Attention and orienting: Sensory and motivational processes (pp. 41–68). Mahwah, NJ: Erlbaum.

Cannon

W. B.

(1932). The wisdom of the body. New York, NY: W. W. Norton.

Ceunen

Vlaeyen

J. W. S.

Van Diest

(2013). Atypical modulation of startle in women in face of aversive bodily sensations. International Journal of Psychophysiology, 88, 157–163. doi:10.1016/j.ijpsycho.2013.03.013

Davis

(2000). The role of the amygdala in conditioned and unconditioned fear and anxiety. In Aggleton

(Ed.), The amygdala (2nd ed., pp. 213–287). New York, NY: Oxford.

10.

De Oca

B. M.

DeCola

J. P.

Maren

Fanselow

M. S.

(1998). Distinct regions of the periaqueductal gray are involved in the acquisition and expression of defensive responses. The Journal of Neuroscience, 18, 3426–3432.

11.

Fanselow

M. S.

(1991). The midbrain periaqueductal gray as a coordinator of action in response to fear and anxiety. In Depaulis

Bandler

(Eds.), The midbrain periaqueductal gray matter: Functional, anatomical and immunohistochemical organization (pp. 151–173). New York, NY: Plenum.

12.

Fanselow

M. S.

(1994). Neural organization of the defensive behavior system responsible for fear. Psychonomic Bulletin & Review, 1, 429–438. doi:10.3758/BF03210947

13.

Fanselow

M. S.

Poulos

A. M.

(2005). The neuroscience of mammalian associative learning. Annual Review of Psychology, 56, 207–234. doi:10.1146/annurev.psych.56.091103.070213

14.

Globisch

Hamm

Schneider

Vaitl

(1993). A computer program for scoring reflex eyeblink and electrodermal responses written in Pascal. Psychophysiology, 30, S30.

15.

Graham

F. K.

(1978). Constraints on measuring heart rate and period sequentially through real and cardiac time. Psychophysiology, 15, 492–495.

16.

Graham

F. K.

Clifton

R. K.

(1966). Heart-rate change as a component of the orienting response. Psychological Bulletin, 65, 305–320.

17.

Grillon

Ameli

Merikangas

Woods

S. W.

Davis

(1993). Measuring the time course of anticipatory anxiety using the fear-potentiated startle reflex. Psychophysiology, 30, 340–346. doi:10.1111/j.1469-8986.1993.tb02055.x

18.

Hamm

A. O.

Weike

A. I.

(2005). The neuropsychology of fear learning and fear regulation. International Journal of Psychophysiology, 57, 5–14. doi:10.1016/j.ijpsycho.2005.01.006

19.

Ille

Berg

Scherg

(2002). Artifact correction of the ongoing EEG using spatial filters based on artifact and brain signal topographies. Journal of Clinical Neurophysiology, 19, 113–124.

20.

Isreal

J. B.

Chesney

G. L.

Wickens

C. D.

Donchin

(1980). P300 and tracking difficulty: Evidence for multiple resources in dual-task performance. Psychophysiology, 17, 259–273.

21.

Kim

E. J.

Horovitz

Pellman

B. A.

Tan

L. M.

Richter-Levin

Kim

J. J.

(2013). Dorsal periaqueductal gray-amygdala pathway conveys both innate and learned fear responses in rats. Proceedings of the National Academy of Sciences, USA, 110, 14795–14800. doi:10.1073/pnas.1310845110

22.

Lang

P. J.

Bradley

M. M.

Cuthbert

B. N.

(1997). Motivated attention: Affect, activation and action. In Lang

P. J.

Simons

R. F.

Balaban

(Eds.), Attention and orienting: Sensory and motivational processes (pp. 97–135). Hillsdale, NJ: Erlbaum.

23.

Lang

P. J.

Davis

Ohman

(2000). Fear and anxiety: Animal models and human cognitive psychophysiology. Journal of Affective Disorders, 61, 137–159.

24.

LeDoux

(2012). Rethinking the emotional brain. Neuron, 73, 653–676. doi:10.1016/j.neuron.2012.02.004

25.

LeDoux

J. E.

(2014). Coming to terms with fear. Proceedings of the National Academy of Sciences, USA, 111, 2871–2878. doi:10.1073/pnas.1400335111

26.

Löw

Lang

P. J.

Smith

J. C.

Bradley

M. M.

(2008). Both predator and prey: Emotional arousal in threat and reward. Psychological Science, 19, 865–873. doi:10.1111/j.1467-9280.2008.02170.x

27.

Maren

(2001). Neurobiology of Pavlovian fear conditioning. Annual Review of Neuroscience, 24, 897–931. doi:10.1146/annurev.neuro.24.1.897

28.

Mobbs

Marchant

J. L.

Hassabis

Seymour

Tan

Gray

. . . Frith

C. D.

(2009). From threat to fear: The neural organization of defensive fear systems in humans. The Journal of Neuroscience, 29, 12236–12243. doi:10.1523/JNEUROSCI.2378-09.2009

29.

Mobbs

Petrovic

Marchant

J. L.

Hassabis

Weiskopf

Seymour

. . . Frith

C. D.

(2007). When fear is near: Threat imminence elicits prefrontal-periaqueductal gray shifts in humans. Science, 317, 1079–1083. doi:10.1126/science.1144298

30.

Mobbs

Rowe

J. B.

Eich

FeldmanHall

Dalgleish

(2010). Neural activity associated with monitoring the oscillating threat value of a tarantula. Proceedings of the National Academy of Sciences, USA, 107, 20582–20586. doi:10.1073/pnas.1009076107

31.

Morgan

M. M.

Carrive

(2001). Activation of the ventrolateral periaqueductal gray reduces locomotion but not mean arterial pressure in awake, freely moving rats. Neuroscience, 102, 905–910.

32.

Moscarello

J. M.

LeDoux

J. E.

(2013). Active avoidance learning requires prefrontal suppression of amygdala-mediated defensive reactions. The Journal of Neuroscience, 33, 3815–3823.

33.

Obrist

P. A.

(1981). Cardiovascular psychophysiology: A perspective. New York, NY: Plenum Press.

34.

Parasuraman

Beatty

(1980). Brain events underlying detection and recognition of weak sensory signals. Science, 210, 80–83.

35.

Richter

Hamm

A. O.

Pané-Farré

C. A.

Gerlach

A. L.

Gloster

A. T.

Wittchen

H.-U.

. . . Arolt

(2012). Dynamics of defensive reactivity in patients with panic disorder and agoraphobia: Implications for the etiology of panic disorder. Biological Psychiatry, 72, 512–520. doi:10.1016/j.biopsych.2012.03.035

36.

Schupp

H. T.

Cuthbert

B. N.

Bradley

M. M.

Birbaumer

Lang

P. J.

(1997). Probe P3 and blinks: Two measures of affective startle modulation. Psychophysiology, 34, 1–6.

37.

Walker

D. L.

Cassella

J. V.

Lee

De Lima

T. C.

Davis

(1997). Opposing roles of the amygdala and dorsolateral periaqueductal gray in fear-potentiated startle. Neuroscience & Biobehavioral Reviews, 21, 743–753.

38.

Wendt

Costines

Lindner

Pfannmöller

Hamm

Löw

(2014, June). Time course of periaqueductal gray activation during active and passive coping with threat. Poster presented at the annual meeting of the Organization for Human Brain Mapping, Hamburg, Germany.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.07 MB