Advertisement

Transdiagnostic fear and anxiety: Prospective prediction using the NPU threat task

  • Kayla A. Wilson
    Correspondence
    Corresponding Author: Kayla A. Wilson, Department of Psychological and Brain Sciences, Texas A&M University, 4235 TAMU, College Station, TX 77843
    Affiliations
    Department of Psychological and Brain Sciences, Texas A&M University, College Station, TX
    Search for articles by this author
  • Annmarie MacNamara
    Affiliations
    Department of Psychological and Brain Sciences, Texas A&M University, College Station, TX

    Institute for Neuroscience, Texas A&M University, College Station, TX
    Search for articles by this author
Open AccessPublished:October 25, 2022DOI:https://doi.org/10.1016/j.bpsgos.2022.10.004

      Abstract

      Background

      Fear and anxiety are distinct dimensions of psychopathology that may be characterized by differences in dimensional threat reactivity. Heightened response to predictable threat is hypothesized to underlie fear symptomatology, whereas increased response to unpredictable threat may underlie anxiety. Despite widespread acceptance of this model, these purported associations have rarely been tested, and the prognostic value of predictable and unpredictable threat responding is unclear. Here, we examined multi-level indicators of predictable and unpredictable threat response as cross-sectional correlates and prospective predictors of transdiagnostic fear and anxiety.

      Methods

      52 individuals with varying levels of internalizing psychopathology (31 female) performed the NPU (no-threat, predictable threat, and unpredictable threat) task. Transdiagnostic fear and anxiety were assessed at baseline (Time 1) and approximately 1.5 years later (Time 2). We used event-related potential, the stimulus-preceding negativity (SPN) as a measure of threat anticipation, and startle eyeblink as a measure of defensive reactivity during the NPU task. These probes were assessed as cross-sectional correlates and prospective predictors of fear and anxiety.

      Results

      Participants with larger Time 1 SPNs to predictable threat were characterized by greater Time 1 fear. Larger Time 1 SPNs to unpredictable threat were associated with greater increases in Time 2 anxiety. Heightened Time 1 startle to predictable threat predicted larger increases in Time 2 fear.

      Conclusions

      Results validate predictable and unpredictable threat responding as dimensional correlates of transdiagnostic fear versus anxiety, and suggest that psychophysiological measures of predictable and unpredictable threat response hold promise as prospective predictors for trajectories of fear and anxiety.

      Keywords

      Introduction

      Animal work has suggested that defensive behaviors can be organized into phasic fear versus sustained anxiety. For example, cues that predictably signal an upcoming aversive stimulus elicit short-term fear behaviors in rodents, whereas uncertain threat cues activate a sustained, anxiety-like state (
      • Davis M.
      • Walker D.L.
      • Miles L.
      • Grillon C.
      Phasic vs Sustained Fear in Rats and Humans: Role of the Extended Amygdala in Fear vs Anxiety [no. 1].
      ). Theoretical distinctions between fear and anxiety have been adapted into a transdiagnostic model for conceptualizing anxiety disorders (
      • Grillon C.
      • Baas J.P.
      • Lissek S.
      • Smith K.
      • Milstein J.
      Anxious Responses to Predictable and Unpredictable Aversive Events.
      ,
      • Hamm A.O.
      Fear, anxiety, and their disorders from the perspective of psychophysiology.
      ,
      • Robinson O.J.
      • Pike A.C.
      • Cornwell B.
      • Grillon C.
      The translational neural circuitry of anxiety.
      ,
      • Hur J.
      • Smith J.F.
      • DeYoung K.A.
      • Anderson A.S.
      • Kuang J.
      • Kim H.C.
      • et al.
      Anxiety and the Neurobiology of Temporally Uncertain Threat Anticipation.
      ). These distinctions are also reflected in the Research Domain Criteria (RDoC; 5), which includes constructs of acute/predictable threat and potential/unpredictable threat. Here, we tested this framework by assessing associations between neurobiological response to predictable and unpredictable threat and transdiagnostic dimensions of fear and anxiety in a mixed, internalizing sample. To determine whether these constructs represent liabilities for the development of future psychopathology, we assessed whether predictable and unpredictable threat reactivity prospectively predict transdiagnostic fear and anxiety.
      The NPU task was designed to probe predictable and unpredictable threat reactivity. Originally designed for use with eyeblink startle, the task involves three different trial types: no-threat (no aversive stimulus delivered), predictable threat (cues predict aversive stimulus delivery) and unpredictable threat (aversive stimulus might be delivered). Increased startle response to predictable and unpredictable threat in this task have been hypothesized to underlie fear-based psychopathology versus anxiety, respectively (
      • Schmitz A.
      • Grillon C.
      Assessing fear and anxiety in humans using the threat of predictable and unpredictable aversive events (the NPU-threat test).
      ). Despite the popularity of the NPU task (
      • Grillon C.
      • O’Connell K.
      • Lieberman L.
      • Alvarez G.
      • Geraci M.
      • Pine D.S.
      • Ernst M.
      Distinct Responses to Predictable and Unpredictable Threat in Anxiety Pathologies: Effect of Panic Attack.
      ,
      • Kaye J.T.
      • Bradford D.E.
      • Curtin J.J.
      Psychometric properties of startle and corrugator response in NPU, affective picture viewing, and resting state tasks.
      ,
      • MacNamara A.
      • Barley B.
      Event-related potentials to threat of predictable and unpredictable shock.
      ,
      • Radoman M.
      • Phan K.L.
      • Gorka S.M.
      Neural correlates of predictable and unpredictable threat in internalizing psychopathology.
      ) and widespread acceptance of distinctions between fear versus anxiety in internalizing disorders (
      • Radoman M.
      • Phan K.L.
      • Gorka S.M.
      Neural correlates of predictable and unpredictable threat in internalizing psychopathology.
      ,
      • Gorka S.M.
      • Lieberman L.
      • Klumpp H.
      • Kinney K.L.
      • Kennedy A.E.
      • Ajilore O.
      • et al.
      Reactivity to unpredictable threat as a treatment target for fear-based anxiety disorders.
      ,
      • Grillon C.
      • Pine D.S.
      • Lissek S.
      • Rabin S.
      • Bonne O.
      • Vythilingam M.
      Increased Anxiety During Anticipation of Unpredictable Aversive Stimuli in Posttraumatic Stress Disorder but not in Generalized Anxiety Disorder.
      ), few studies have tested associations with these constructs transdiagnostically. Results to-date using the NPU task can be conceptualized within a fear versus anxiety framework when considering genetic and epidemiological data that has divided internalizing disorders into those characterized by: a) anxious misery/distress – i.e., generalized anxiety disorder (GAD), panic disorder (PD), agoraphobia, and post-traumatic stress disorder (PTSD); b) fear – i.e., the specific phobias (SP); and c) somewhere in between – i.e., social anxiety disorder (SAD; 14–16).
      Most work using the NPU task in clinical samples has focused on PD, which, although dominated by worry about uncertain threat (e.g., future panic attacks; 17), is also characterized by phasic fear (e.g., during panic attacks; 3,4). Results support that individuals with PD are characterized by heightened startle response to unpredictable threat (
      • Grillon C.
      • O’Connell K.
      • Lieberman L.
      • Alvarez G.
      • Geraci M.
      • Pine D.S.
      • Ernst M.
      Distinct Responses to Predictable and Unpredictable Threat in Anxiety Pathologies: Effect of Panic Attack.
      ,
      • Gorka S.M.
      • Nelson B.D.
      • Shankman S.A.
      Startle response to unpredictable threat in comorbid panic disorder and alcohol dependence.
      ,
      • Gorka S.M.
      • Lieberman L.
      • Shankman S.A.
      • Phan K.L.
      Startle potentiation to uncertain threat as a psychophysiological indicator of fear-based psychopathology: An examination across multiple internalizing disorders.
      ,
      • Lieberman L.
      • Gorka S.M.
      • Shankman S.A.
      • Phan K.L.
      Impact of Panic on Psychophysiological and Neural Reactivity to Unpredictable Threat in Depression and Anxiety.
      ). A few studies have found that individuals with PD/elevated panic symptoms are characterized by increased startle potentiation to both unpredictable and predictable threat, in line with the notion that PD may be characterized by both sustained anxiety and phasic fear (
      • Gorka S.M.
      • Liu H.
      • Sarapas C.
      • Shankman S.A.
      Time course of threat responding in panic disorder and depression.
      ,
      • Shankman S.A.
      • Nelson B.D.
      • Sarapas C.
      • Robison-Andrew E.J.
      • Campbell M.L.
      • Altman S.E.
      • et al.
      A psychophysiological investigation of threat and reward sensitivity in individuals with panic disorder and/or major depressive disorder.
      ). Neuroimaging work suggests PD is associated with hyperactivation in the ventromedial prefrontal cortex to all cues in the NPU task, suggesting overgeneralization and excessive regulation of response to both threat and safety cues (
      • Klahn A.L.
      • Klinkenberg I.A.
      • Lueken U.
      • Notzon S.
      • Arolt V.
      • Pantev C.
      • et al.
      Commonalities and differences in the neural substrates of threat predictability in panic disorder and specific phobia.
      ). Panic symptoms have also been associated with increased reactivity to unpredictable threat cues in the dorsal anterior cingulate cortex (
      • Lieberman L.
      • Gorka S.M.
      • Shankman S.A.
      • Phan K.L.
      Impact of Panic on Psychophysiological and Neural Reactivity to Unpredictable Threat in Depression and Anxiety.
      ), brainstem (
      • Radoman M.
      • Phan K.L.
      • Gorka S.M.
      Neural correlates of predictable and unpredictable threat in internalizing psychopathology.
      ), and bed nucleus of the stria terminalis (BNST; 24), a brain region involved in mediating response to uncertain threat (25, for review, see 26). Few studies have examined PTSD, which has typically been conceptualized as an anxiety (not fear) disorder (
      • Chantarujikapong S.I.
      • Scherrer J.F.
      • Xian H.
      • Eisen S.A.
      • Lyons M.J.
      • Goldberg J.
      • et al.
      A twin study of generalized anxiety disorder symptoms, panic disorder symptoms and post-traumatic stress disorder in men.
      ,
      • Scherrer J.F.
      • True W.R.
      • Xian H.
      • Lyons M.J.
      • Eisen S.A.
      • Goldberg J.
      • et al.
      Evidence for genetic influences common and specific to symptoms of generalized anxiety and panic.
      ), as evidenced by associations with increased startle response to unpredictable threat cues (
      • Grillon C.
      • Pine D.S.
      • Lissek S.
      • Rabin S.
      • Bonne O.
      • Vythilingam M.
      Increased Anxiety During Anticipation of Unpredictable Aversive Stimuli in Posttraumatic Stress Disorder but not in Generalized Anxiety Disorder.
      ,
      • Gorka S.M.
      • Kreutzer K.A.
      • Petrey K.M.
      • Radoman M.
      • Phan K.L.
      Behavioral and neural sensitivity to uncertain threat in individuals with alcohol use disorder: Associations with drinking behaviors and motives.
      ), as well as sustained BNST activation to anticipation of unpredictable aversive stimuli (
      • Brinkmann L.
      • Buff C.
      • Neumeister P.
      • Tupak S.V.
      • Becker M.P.I.
      • Herrmann M.J.
      • Straube T.
      Dissociation between amygdala and bed nucleus of the stria terminalis during threat anticipation in female post-traumatic stress disorder patients.
      ). GAD has also been associated with greater BNST activation to unpredictable threat cues (
      • Buff C.
      • Brinkmann L.
      • Bruchmann M.
      • Becker M.P.I.
      • Tupak S.
      • Herrmann M.J.
      • Straube T.
      Activity alterations in the bed nucleus of the stria terminalis and amygdala during threat anticipation in generalized anxiety disorder.
      ). Therefore, prior findings from categorical studies suggest that anxiety symptomatology is characterized by increased peripheral response to unpredictable threat and greater engagement of brain regions involved in mediating these responses.
      Fewer studies have examined predictable and unpredictable threat reactivity in the fear disorders. SP has been associated with greater startle reactivity to unpredictable threat cues (
      • Gorka S.M.
      • Lieberman L.
      • Shankman S.A.
      • Phan K.L.
      Startle potentiation to uncertain threat as a psychophysiological indicator of fear-based psychopathology: An examination across multiple internalizing disorders.
      ,
      • Nelson B.D.
      • Hajcak G.
      Anxiety and Depression Symptom Dimensions Demonstrate Unique Relationships with the Startle Reflex in Anticipation of Unpredictable Threat in 8 to 14 Year-Old Girls.
      ), which is not in keeping with the hypothesized fear-predictable threat and anxiety-unpredictable threat associations. Results have also shown mixed findings in individuals with SAD: some studies demonstrated an association with increased startle response to predictable threat (
      • Grillon C.
      • O’Connell K.
      • Lieberman L.
      • Alvarez G.
      • Geraci M.
      • Pine D.S.
      • Ernst M.
      Distinct Responses to Predictable and Unpredictable Threat in Anxiety Pathologies: Effect of Panic Attack.
      ), whereas others found increased startle to unpredictable threat cues (
      • Gorka S.M.
      • Lieberman L.
      • Shankman S.A.
      • Phan K.L.
      Startle potentiation to uncertain threat as a psychophysiological indicator of fear-based psychopathology: An examination across multiple internalizing disorders.
      ). Neuroimaging work has found that SAD is associated with decreased BNST-amygdala connectivity to unpredictable threat cues, suggesting less coordinated response to unpredictable threat (
      • Clauss J.A.
      • Avery S.N.
      • Benningfield M.M.
      • JU Blackford
      Social anxiety is associated with BNST response to unpredictability.
      ).
      Prospective and longitudinal examination of predictable and unpredictable threat responding is rare. One study found that increased startle reactivity to both unpredictable and predictable threat cues predicted worse functional impairment approximately one year later in individuals with current and past internalizing disorders (
      • Stevens E.S.
      • Lieberman L.
      • Funkhouser C.J.
      • Correa K.A.
      • Shankman S.A.
      Startle during threat longitudinally predicts functional impairment independent of DSM diagnoses.
      ). Another study found that cognitive behavioral therapy reduced startle reactivity to unpredictable threat in individuals with PD, SAD, and PTSD (
      • Gorka S.M.
      • Lieberman L.
      • Klumpp H.
      • Kinney K.L.
      • Kennedy A.E.
      • Ajilore O.
      • et al.
      Reactivity to unpredictable threat as a treatment target for fear-based anxiety disorders.
      ). While limited, this work suggests that predictable and unpredictable threat reactivity might account for variance in outcomes beyond what can be explained by categorical diagnoses alone, and could play a causal role in internalizing psychopathology.
      Given the substantial heterogeneity characterizing the current diagnostic categories (
      • Insel T.
      • Cuthbert B.
      • Garvey M.
      • Heinssen R.
      • Pine D.S.
      • Quinn K.
      • et al.
      Research Domain Criteria (RDoC): Toward a New Classification Framework for Research on Mental Disorders.
      ), it would be ambitious to expect that hypothesized associations between disorders categorized as “fear” versus “anxiety” would consistently show expected associations with predictable and unpredictable threat reactivity. That is, PD, SAD and SP may be characterized to some degree by both elevated fear and anxiety. Substantial comorbidity between categorical disorders means that participants often meet criteria for multiple diagnoses, or may be characterized by heightened, but subthreshold, fear/anxiety psychopathology. In sum, transdiagnostic fear and anxiety are potentially more homogeneous constructs than the current categorical diagnoses, and might more closely track variation in neurobiological response to predictable and unpredictable threat.
      Here, we used the NPU task to assess predictable and unpredictable threat responding as cross-sectional correlates and prospective predictors of fear versus anxiety symptomatology, in a mixed, internalizing sample. While the majority of work using the NPU task has used startle potentiation to probe defensive response during predictable and unpredictable threat, an event-related potential (ERP), the stimulus-preceding negativity (SPN), provides a central autonomic measure of threat anticipation (
      • Böcker K.B.E.
      • Baas J.M.P.
      • Kenemans J.L.
      • Verbaten M.N.
      Stimulus-preceding negativity induced by fear: a manifestation of affective anticipation.
      ). The SPN is a negative-going, frontally distributed ERP component that grows larger as threatening stimuli approach (
      • Tanovic E.
      • Joormann J.
      Anticipating the unknown: The stimulus-preceding negativity is enhanced by uncertain threat.
      ) and is sensitive to stimulus probability (

      Catena A, Perales JC, Megías A, Cándido A, Jara E, Maldonado A (2012): The Brain Network of Expectancy and Uncertainty Processing ((C. Soriano-Mas, editor)). PLoS ONE 7: e40252.

      ,
      • Lin H.
      • Gao H.
      • You J.
      • Liang J.
      • Ma J.
      • Yang N.
      • et al.
      Larger N2 and smaller early contingent negative variation during the processing of uncertainty about future emotional events.
      ), making it an ideal measure for the NPU task (
      • MacNamara A.
      • Barley B.
      Event-related potentials to threat of predictable and unpredictable shock.
      ). Here, we measured the SPN to threat cues, as well as startle eyeblink, to provide a link with prior work. We expected that individuals higher in transdiagnostic fear would show increased SPNs and heightened startle potentiation to predictable threat, and that individuals higher in transdiagnostic anxiety would show increased SPNs and heighted startle potentiation in response to unpredictable threat (
      • Davis M.
      • Walker D.L.
      • Miles L.
      • Grillon C.
      Phasic vs Sustained Fear in Rats and Humans: Role of the Extended Amygdala in Fear vs Anxiety [no. 1].
      ,
      • Grillon C.
      • Baas J.P.
      • Lissek S.
      • Smith K.
      • Milstein J.
      Anxious Responses to Predictable and Unpredictable Aversive Events.
      ). We also expected to observe these associations prospectively – i.e., increased SPNs and startle to predictable threat cues at baseline would predict greater self-reported fear at follow-up approximately 1.5 years later, and increased SPNs and startle to unpredictable threat cues at baseline would predict greater self-reported anxiety at follow-up (
      • Grillon C.
      • Baas J.P.
      • Lissek S.
      • Smith K.
      • Milstein J.
      Anxious Responses to Predictable and Unpredictable Aversive Events.
      ,
      • Robinson O.J.
      • Pike A.C.
      • Cornwell B.
      • Grillon C.
      The translational neural circuitry of anxiety.
      ). If confirmed, these hypotheses would support neurobiological distinctions between fear versus anxiety and would suggest predictable and unpredictable threat responding might represent unique liabilities for the development and worsening of fear versus anxiety psychopathology, respectively.

      Methods and Materials

      Participants

      Fifty-six individuals who were part of a larger study participated in data collection at baseline (Time 1) and returned at follow-up approximately 1.5 years later (Time 2). From this initial sample, four participants were excluded for having poor quality data recordings at Time 1, leaving a final sample size of n = 52 (31 female; M age = 24.46 years, SD = 9.33). There were two outliers excluded for the late SPN leaving 50 participants for late SPN analyses (30 female; M age = 24.60 years, SD = 9.48). One outlier was excluded for startle, and six participants did not have a sufficient number of startle trials for analyses, leaving 45 participants (26 female; M age = 23.49 years, SD = 7.78) for startle analyses. Details concerning sample size determinations, bad data exclusion criteria, and outliers are presented in the Supplemental Information.
      Clinical characteristics of the sample are presented in Table 1. Participants were initially recruited as part of a larger study, for which they were selected to fall into a psychiatrically healthy group (no current or prior psychiatric diagnoses; n = 20) or an anxiety group. Although participant recruitment was structured this way to ensure variability in the sample, planned analyses (both in the parent study and in the current study) were dimensional. Participants in the anxiety group were required to meet criteria for a focal fear diagnosis at Time 1 (specific phobia [n = 16] or performance-only social anxiety [n = 20]), but were permitted to vary in levels of additional, comorbid internalizing psychopathology (e.g., major/persistent depressive disorder [MDD/PDD], GAD, generalized SAD). Exclusionary criteria are included in the Supplemental Information. Participants were not engaged in psychiatric treatment of any kind. Diagnoses were made at Time 1 according to the Structured Clinical Interview for the Diagnostic and Statistical Manual of Mental Disorders, DSM-5 (SCID; 30). Study procedures were in compliance with the Helsinki Declaration of 1975 (as revised in 1983), and were approved by the Texas A&M University institutional review board.
      Table 1Clinical characteristics of participants at Time 1 and Time 2
      Time 1Time 2
      M (SD)M (SD)
      PANAS-X Fear11.02 (4.26)10.81 (4.59)
      STAI trait43.10 (12.76)45.55 (11.39)
      PSWQ47.98 (16.24)49.00 (14.23)
      SPIN18.98 (17.28)20.10 (14.80)
      # of current diagnoses2.27 (1.93)-
      n (%)n (%)
      Current Diagnosis
      Focal fear36 (68)-
      SAD (Generalized)19 (
      • Lin H.
      • Gao H.
      • You J.
      • Liang J.
      • Ma J.
      • Yang N.
      • et al.
      Larger N2 and smaller early contingent negative variation during the processing of uncertainty about future emotional events.
      )
      -
      GAD11(21)-
      MDD/PDD6 (
      • Gorka S.M.
      • Lieberman L.
      • Klumpp H.
      • Kinney K.L.
      • Kennedy A.E.
      • Ajilore O.
      • et al.
      Reactivity to unpredictable threat as a treatment target for fear-based anxiety disorders.
      )
      -
      PTSD4 (
      • Grillon C.
      • O’Connell K.
      • Lieberman L.
      • Alvarez G.
      • Geraci M.
      • Pine D.S.
      • Ernst M.
      Distinct Responses to Predictable and Unpredictable Threat in Anxiety Pathologies: Effect of Panic Attack.
      )
      -
      PMDD3 (
      • Insel T.
      • Cuthbert B.
      • Garvey M.
      • Heinssen R.
      • Pine D.S.
      • Quinn K.
      • et al.
      Research Domain Criteria (RDoC): Toward a New Classification Framework for Research on Mental Disorders.
      )
      -
      Agoraphobia2 (
      • Robinson O.J.
      • Pike A.C.
      • Cornwell B.
      • Grillon C.
      The translational neural circuitry of anxiety.
      )
      -
      Anorexia nervosa2 (
      • Robinson O.J.
      • Pike A.C.
      • Cornwell B.
      • Grillon C.
      The translational neural circuitry of anxiety.
      )
      -
      Substance use disorder1 (
      • Grillon C.
      • Baas J.P.
      • Lissek S.
      • Smith K.
      • Milstein J.
      Anxious Responses to Predictable and Unpredictable Aversive Events.
      )
      -
      Note: PANAS-X, Positive and Negative Affect Schedule – Expanded Form; STAI, State Trait Anxiety Inventory; PSWQ, Penn State Worry Questionnaire; SPIN, Social Phobia Inventory; SAD, social anxiety disorder; GAD, generalized anxiety disorder; MDD, major depressive disorder; PDD, persistent depressive disorder; PTSD, posttraumatic stress disorder; PMDD, premenstrual dysphoric disorder. Focal fear included specific phobia (n = 16) and performance-only social anxiety (n = 20).

      Materials

      Dimensional psychopathology and internalizing symptoms were assessed using the Positive and Negative Affect Schedule-Expanded Form (PANAS-X; 31), the Penn State Worry Questionnaire (PSWQ; 32), the State Trait Anxiety Inventory, trait version (STAI; 33), and the Social Phobia Inventory (SPIN; 34). Details on each of these measures are provided in Supplemental Information.
      We used the PANAS-X Fear subscale as a continuous measure of transdiagnostic fear, computed separately at each timepoint (Time 1 Fear, Time 2 Fear). Transdiagnostic anxiety was operationalized as a composite of averaged z-scored STAI, PSWQ, and SPIN scores, computed separately at each timepoint. This provided a broad measure of anxiety (Time 1 Anxiety, Time 2 Anxiety) that was not specific to a particular diagnosis, in keeping with the characteristics of our mixed, internalizing sample (
      • Banica I.
      • Sandre A.
      • Shields G.S.
      • Slavich G.M.
      • Weinberg A.
      The error-related negativity (ERN) moderates the association between interpersonal stress and anxiety symptoms six months later.
      ).

      Procedure

      Time 1

      Participants’ shock levels were set using standard procedures, in order to control for individual differences in shock sensitivity (
      • MacNamara A.
      • Barley B.
      Event-related potentials to threat of predictable and unpredictable shock.
      ,

      Bradford DE, Magruder KP, Korhumel RA, Curtin JJ (2014): Using the threat probability task to assess anxiety and fear during uncertain and certain threat. J Vis Exp JoVE 51905.

      ). While EEG was recorded, participants performed the NPU task used in our prior work (
      • MacNamara A.
      • Barley B.
      Event-related potentials to threat of predictable and unpredictable shock.
      ) and adapted from Kaye and colleagues (
      • Kaye J.T.
      • Bradford D.E.
      • Curtin J.J.
      Psychometric properties of startle and corrugator response in NPU, affective picture viewing, and resting state tasks.
      ). Participants were asked to view colored shape “cues” (blue circle, red square, green triangle). Each shape cue indicated whether the participant would definitely receive a shock (predictable), possibly receive a shock (unpredictable) or would never receive a shock (no-threat). Additional task parameters are detailed in the Supplemental Information.
      Acoustic startle probes were delivered binaurally (40 ms, 90 dB white noise with near instantaneous rise time) during cues and interstimulus intervals (ISIs). Three initial startle probes were presented before the task to allow for stabilization of the startle response (
      • Blumenthal T.D.
      • Cuthbert B.N.
      • Filion D.L.
      • Hackley S.
      • Lipp O.V.
      • Boxtel A.V.
      Committee report: Guidelines for human startle eyeblink electromyographic studies.
      ); this data was not analyzed. Startle probes were presented a minimum of 12500 ms after shock or other startle probes, with the serial position of startle probes across each condition balanced within subjects. Two different, counterbalanced orders of startle probe serial position were used. See Supplemental Information for additional task details.

      Time 2

      Approximately 1.5 years following their initial visit (M = 1.68 years; SD = 0.68), participants completed the same set of questionnaires completed at Time 1; due to the COVID-19 pandemic, questionnaires were completed online and participants did not complete the SCID or receive an EEG again at this time.

      Electroencephalographic (EEG) Data Acquisition - Time 1

      Continuous EEG recordings were collected using an ActiCap and the ActiCHamp amplifier system (Brain Products GmbH, Gilching Germany) at Time 1. Thirty-two electrode sites were used based on the 10/20 system. The electrooculogram (EOG) was recorded from four facial electrodes (see Supplemental Information). The EEG data were digitized at 24-bit resolution and a sampling rate of 1000 Hz. EEG data reduction is described in the Supplemental Information.
      Based on visual inspection of grand-averaged waveforms and topographic maps, the SPN was scored at Fz as the mean activity during an early window (1000-2000 ms post-cue onset) and a late window (3500-4500 ms post-cue onset; 39); as in our prior work (
      • MacNamara A.
      • Barley B.
      Event-related potentials to threat of predictable and unpredictable shock.
      ), time windows were chosen to avoid shock delivery (which was at 2000 ms or 4800 ms on U trials and at 4800 ms on P trials) and startle probe delivery (at 4500 ms).

      Electromyographic (EMG) Data Acquisition - Time 1

      Startle eyeblink EMG activity was recorded from two 4-mm diameter electrodes placed over the orbicularis oculi muscle under the left eye and using the ActiCHamp amplifier system. Data were digitized at 24-bit resolution and a sampling rate of 1000 Hz. EMG data reduction is described in the Supplemental Information.

      Data Analyses

      To assess condition effects and group differences, cue-locked SPN amplitudes were submitted to separate 2 (group: control, anxious) X 3 (condition: no-threat, predictable threat, unpredictable threat) between-within analyses of variance (ANOVAs) and startle amplitudes were submitted to a 2 (group: control, anxious) X 2 (cue, ISI) X 3 (condition: no-threat, predictable threat, unpredictable threat) between-within ANOVA. Greenhouse-Geisser corrections were applied as necessary when the assumption of sphericity was violated. Significant effects were followed up using dependent and independent samples t-tests, as appropriate.
      To examine associations between psychophysiological measures and fear and anxiety at Time 1, we conducted separate linear regressions for each of the SPN and startle as predictors of Time 1 Fear and Time 1 Anxiety. For each regression, SPN/startle response to no-threat, predictable threat and unpredictable threat were entered as simultaneous predictors, along with the other Time 1 dimension (i.e., Fear or Anxiety).
      To examine associations between psychophysiological measures and Time 2 symptoms, we conducted separate linear regressions for each of the SPN and startle elicited during no-threat, predictable threat and unpredictable threat, as predictors of Time 2 Fear and Time 2 Anxiety, controlling for Time 1 Fear and Time 1 Anxiety and the other experimental conditions (i.e., SPN/startle response to no-threat, predictable threat and/or unpredictable threat). In addition, because the time between Time 1 and Time 2 visits varied somewhat across participants, we also controlled for time passed between visits.
      We used bootstrapped regression analyses (using 2000 bootstraps) which yielded bootstrapped p values and 95% confidence intervals (
      • Westfall P.H.
      On Using the Bootstrap for Multiple Comparisons.
      ). Bootstrapping is a non-parametric resampling method that can produce more accurate Type 1 error rate and higher statistical power than the single sample parametric method (e.g., testing mediation effects; 48). Beta weights were considered significant when both the bootstrapped p < 0.05 and the confidence interval did not include zero (
      • Grady C.L.
      • Luk G.
      • Craik F.I.M.
      • Bialystok E.
      Brain Network Activity in Monolingual and Bilingual Older Adults.
      ). Analyses of categorical diagnosis, analyses that included outliers, and tests of condition specificity within each regression are presented in the Supplemental Information. Supplemental analyses demonstrated that startle findings were robust to inclusion of outliers, however, the late SPN findings were not. Analyses were performed using SPSS statistical software version 26.0 (IBM, Armonk, NY).

      Results

      Table 2 presents means and standard deviations for all psychophysiological measures, shown separately for each condition (no-threat, predictable threat, unpredictable threat). Table 3 presents regression results for both Time 1 and Time 2.
      Table 2ERP and startle means (standard deviations) for each condition
      No-threat (μV)Predictable threat (μV)Unpredictable threat (μV)
      Early SPN1.59 (5.93)-1.63 (6.77)2.00 (7.27)
      Late SPN1.87 (8.10)-2.33 (9.22)1.65 (11.29)
      Cue Startle49.86 (51.69)82.08 (62.78)89.06 (65.93)
      ISI Startle53.32 (55.84)76.78 (71.22)85.60 (74.85)
      Note: ISI, interstimulus interval.
      Table 3Regression results
      Outcome: Time 1 Fear
      Early SPNLate SPNStartle
      N0.0110.040-0.018
      P-0.004-0.120*0.009
      U-0.0280.0250.007
      Time 1 Anxiety3.490*3.278*3.542*
      Outcome: Time 1 Anxiety
      N0.007-0.0010.001
      P-0.0040.0060.000
      U-0.001-0.0050.000
      Time 1 Fear0.153*0.157*0.173*
      Outcome: Time 2 Fear
      N0.070-0.022-0.008
      P0.0330.0080.056*
      U-0.129-0.035-0.039
      Time 1 Fear0.700*0.698*0.830*
      Time 1 Anxiety0.1630.315-0.438
      Time Passed (years)0.0040.0030.003
      Outcome: Time 2 Anxiety
      N-0.009-0.0180.001
      P0.0230.0110.008
      U-0.032*-0.016*-0.005
      Time 1 Fear0.0400.0500.044
      Time 1 Anxiety0.556*0.533*0.511*
      Time Passed (years)0.000-0.0000.000
      Note: Columns represent separate regression models with analogous ERPs or startle entered as predictors of fear or anxiety. Regression coefficients are presented as bootstrapped, unstandardized beta weights. The SPN is a negative-going ERP component; therefore, negative beta-weights indicate that larger SPNs were associated with increased fear and/or anxiety. *bootstrapped p < .05
      Time 1

      Electrocortical activity

      Early SPN. There was a significant effect of condition, F(2, 100) = 4.45, p = .01, ηp2 = .08: predictable cues elicited larger (more negative) SPNs compared to no-threat cues, t(
      • Cohen J.
      The Cost of Dichotomization.
      ) = 2.64, p = .01 and compared to unpredictable cues, t(
      • Cohen J.
      The Cost of Dichotomization.
      ) = 3.73, p < .001. The early SPN to unpredictable and no-threat cues did not differ significantly, p = .75. The effect of group and the interaction between group X condition failed to reach significance, ps > .20. Dimensional analyses showed no significant associations between the early SPN to no-threat, predictable threat, or unpredictable threat and Time 1 Fear (ps > .58) or Time 1 Anxiety (ps > .69).
      Late SPN. There was a significant effect of condition, F(2, 96) = 3.56, p = .03, ηp2 = .07: predictable cues elicited larger (more negative) SPNs compared to no-threat cues, t(
      • Grady C.L.
      • Luk G.
      • Craik F.I.M.
      • Bialystok E.
      Brain Network Activity in Monolingual and Bilingual Older Adults.
      ) = 2.90, p = .01, and compared to unpredictable cues, t(
      • Grady C.L.
      • Luk G.
      • Craik F.I.M.
      • Bialystok E.
      Brain Network Activity in Monolingual and Bilingual Older Adults.
      ) = 2.73, p = .01. The late SPN to unpredictable and no-threat cues did not differ, p = .91. The effect of group and the interaction between group X condition failed to reach significance, ps > .06.
      Dimensional analyses showed that larger late SPNs to predictable threat cues were associated with increased Time 1 Fear, B = -.120, CI: -.216, -.033, p = .03 (Figure 1, Figure 2). The late SPN to no-threat and unpredictable threat cues was not associated with Time 1 Fear, ps > .49. The other regression model, which aimed to predict Time 1 Anxiety found no association with the SPN to no-threat, predictable threat or unpredictable threat cues, ps > .62.
      Figure thumbnail gr1
      Figure 1Time 1 late SPN to predictable threat and Time 1 Fear. Time 1 grand-averaged waveforms at Fz where the late SPN was scored, shown separately for no-threat (top), predictable threat (middle), and unpredictable threat (bottom), and for participants with high Time 1 Fear and low Time 1 Fear; positive is plotted downwards. Headmaps depict the voltage distributions for predictable threat cues, shown separately for participants with high Time 1 Fear and low Time 1 Fear. Note: high Time 1 Fear (upper third) and low Time 1 Fear (lower third) groups were created for illustrative purposes only, all analyses were continuous.
      Figure thumbnail gr2
      Figure 2Scatterplot depicting the association (as unstandardized residuals after controlling for covariates) between Time 1 late SPN to predictable threat and Time 1 Fear.

      Startle

      There was a significant effect of condition, F(2, 78) = 24.02, p < .001, ηp2 = .38: predictable threat (averaged across cue and ISI) elicited larger startle responses compared to no-threat, t(

      Bradford DE, Magruder KP, Korhumel RA, Curtin JJ (2014): Using the threat probability task to assess anxiety and fear during uncertain and certain threat. J Vis Exp JoVE 51905.

      ) = 5.46, p < .001. Additionally, unpredictable threat elicited larger startle responses compared to no-threat, t(

      Bradford DE, Magruder KP, Korhumel RA, Curtin JJ (2014): Using the threat probability task to assess anxiety and fear during uncertain and certain threat. J Vis Exp JoVE 51905.

      ) = 6.33, p < .001, and predictable threat, t(

      Bradford DE, Magruder KP, Korhumel RA, Curtin JJ (2014): Using the threat probability task to assess anxiety and fear during uncertain and certain threat. J Vis Exp JoVE 51905.

      ) = 2.68, p = .01. No other effects reached significance at the omnibus level, ps > .13. Dimensional analyses showed no significant associations between startle to no-threat, predictable threat, or unpredictable threat and continuous symptoms of Time 1 Fear (ps > .38) or Time 1 Anxiety (ps > .83).
      Time 2

      Electrocortical activity

      Early SPN. Larger early SPNs to unpredictable threat cues predicted increased Time 2 Anxiety, B = -.032, CI: -.054, -.011, p = .01 (Figure 3, Figure 4). Early SPNs to no-threat and predictable threat cues were not associated with Time 2 Anxiety, ps > .06. The other regression model, which aimed to predict Time 2 Fear found no association with the SPN to no-threat, predictable threat or unpredictable threat cues, ps > .054.
      Figure thumbnail gr3
      Figure 3Time 1 early and late SPN to unpredictable threat and Time 2 Anxiety. Time 1 grand-averaged waveforms at Fz where the early SPN and late SPN were scored, shown separately for no-threat (top), predictable threat (middle), and unpredictable threat (bottom) for participants with high ΔAnxiety and low ΔAnxiety; positive is plotted downwards. Headmaps depict the voltage distributions for unpredictable threat cues, shown separately for participants with high ΔAnxiety and low ΔAnxiety for both the early and late time windows. Note: ΔAnxiety = Time 2 Anxiety – Time 1 Anxiety; high ΔAnxiety (upper third) and low ΔAnxiety (lower third) groups were created for illustrative purposes only, all analyses were continuous.
      Figure thumbnail gr4
      Figure 4Scatterplot depicting the association (as unstandardized residuals after controlling for covariates) between Time 1 early SPN to unpredictable threat cues and Time 2 Anxiety.
      Late SPN. Larger late SPNs to unpredictable threat cues predicted increased Time 2 Anxiety, B = -.016, CI: -.030, -.002, p = .04 (Figure 3, Figure 5). Late SPNs to no-threat and predictable threat cues were not associated with Time 2 Anxiety, ps > .15. The other regression model, which aimed to predict Time 2 Fear found no association with the SPN to no-threat, predictable threat, or unpredictable threat cues, ps > .41.
      Figure thumbnail gr5
      Figure 5Scatterplot depicting the association (as unstandardized residuals after controlling for covariates) between Time 1 late SPN to unpredictable threat cues and Time 2 Anxiety.

      Startle

      Larger startle responses to predictable threat predicted increased Time 2 Fear, B = .056, CI: .007, .098, p = .039 (Figure 6). Startle to no-threat and unpredictable threat was not associated with Time 2 Fear, ps > .054. The other regression model, which aimed to predict Time 2 Anxiety found no association with startle to no-threat, predictable threat or unpredictable threat, ps > .051.
      Figure thumbnail gr6
      Figure 6Scatterplot depicting the association (as unstandardized residuals after controlling for covariates) between Time 1 startle to predictable threat cues and Time 2 Fear.

      Discussion

      Heightened predictable and unpredictable threat reactivity have been hypothesized to underlie transdiagnostic fear and anxiety in internalizing disorders, but these associations had not been tested. Here, we found that individuals with greater fear symptomatology at baseline were characterized by increased anticipation of predictable threat (SPN). Moreover, participants who showed greater defensive reactivity to predictable threat at baseline (startle) went on to show greater increases in fear symptomatology approximately 1.5 years later. On the other hand, greater anticipation of unpredictable threat (SPN) at baseline uniquely and prospectively predicted increased anxiety. Results suggest mechanistic distinctions between transdiagnostic fear versus anxiety, implicating predictable and unpredictable threat reactivity as risk factors for the development of fear and anxiety psychopathology, respectively.
      Prior work failed to find evidence of an association between predictable threat reactivity and diagnosis of a quintessential fear disorder - SP (e.g., 18, 26). All else being equal, dimensional analyses offer more power than categorical analyses (
      • Meyer T.J.
      • Miller M.L.
      • Metzger R.L.
      • Borkovec T.D.
      Development and validation of the penn state worry questionnaire.
      ,

      Spielberger CD (1983): Manual for the State-Trait Anxiety Inventory (STAI). Palo Alto, CA: Consulting Psychologists Press.

      ); moreover, transdiagnostic fear is likely a more cohesive construct than categorical diagnosis of SP. Therefore, dimensional assessment of fear – as in the current study - might more accurately “carve nature at its joints”, and could explain why we observed an association between anticipation of predictable threat and transdiagnostic fear, where prior work had not.
      When examining prospective associations, greater defensive reactivity (i.e., startle eye blink) to predictable threat cues predicted larger increases in fear symptoms over time. Proximal threat is associated with sudden increases in autonomic arousal (i.e., fight or flight response), thoughts of immediate danger (e.g., upcoming shock), and escape behaviors (
      • Hamm A.O.
      Fear, anxiety, and their disorders from the perspective of psychophysiology.
      ,
      • Robinson O.J.
      • Pike A.C.
      • Cornwell B.
      • Grillon C.
      The translational neural circuitry of anxiety.
      ). Startle eye blink may track the neurobiological pre-disposition to respond excessively to proximal threat, putting individuals at risk for greater fear symptomatology over time. As such, our results suggest that increased startle reactivity to certain threat might be a viable target for early intervention or prevention efforts aimed at reducing transdiagnostic fear.
      Greater sustained anticipation (SPN) of unpredictable threat predicted larger increases in anxiety symptoms over time. These results are broadly in line with prior work, which had found evidence of cross-sectional associations between anxiety and increased defensive responding to uncertain threat (
      • Grillon C.
      • O’Connell K.
      • Lieberman L.
      • Alvarez G.
      • Geraci M.
      • Pine D.S.
      • Ernst M.
      Distinct Responses to Predictable and Unpredictable Threat in Anxiety Pathologies: Effect of Panic Attack.
      ,
      • Grillon C.
      • Pine D.S.
      • Lissek S.
      • Rabin S.
      • Bonne O.
      • Vythilingam M.
      Increased Anxiety During Anticipation of Unpredictable Aversive Stimuli in Posttraumatic Stress Disorder but not in Generalized Anxiety Disorder.
      ,
      • Gorka S.M.
      • Nelson B.D.
      • Shankman S.A.
      Startle response to unpredictable threat in comorbid panic disorder and alcohol dependence.
      ,
      • Gorka S.M.
      • Lieberman L.
      • Shankman S.A.
      • Phan K.L.
      Startle potentiation to uncertain threat as a psychophysiological indicator of fear-based psychopathology: An examination across multiple internalizing disorders.
      ,
      • Lieberman L.
      • Gorka S.M.
      • Shankman S.A.
      • Phan K.L.
      Impact of Panic on Psychophysiological and Neural Reactivity to Unpredictable Threat in Depression and Anxiety.
      ,
      • Gorka S.M.
      • Kreutzer K.A.
      • Petrey K.M.
      • Radoman M.
      • Phan K.L.
      Behavioral and neural sensitivity to uncertain threat in individuals with alcohol use disorder: Associations with drinking behaviors and motives.
      ). Results observed here suggest that heightened anticipation of uncertain threat may be present before the development and/or worsening of anxiety symptoms. Excessive anticipation of uncertain threat could underlie the development of behaviors such as avoidance, which can lead to increases in anxiety (
      • Grupe D.W.
      • Nitschke J.B.
      Uncertainty and anticipation in anxiety: an integrated neurobiological and psychological perspective.
      ). Early intervention or prevention efforts targeting anticipation of uncertain threat could be useful in combatting the development or worsening of transdiagnostic anxiety.
      The unique associations we observed involving threat anticipation (SPN) versus defensive responding (startle) indicate that these measures may provide different information about prospective fear versus anxiety psychopathology, with future fear predicted by startle (to predictive threat) and future anxiety predicted by the SPN (to unpredictable threat). The SPN provides a measure of protracted and cognitively-mediated threat anticipation (
      • Tanovic E.
      • Joormann J.
      Anticipating the unknown: The stimulus-preceding negativity is enhanced by uncertain threat.
      ). Our prospective results could be interpreted as indicating that sustained, future-oriented attention to unpredictable threat is a risk factor for increased anxiety. This aligns with the notion that anxiety is characterized by heightened assessment of the probability and extent of threatening events (
      • Grupe D.W.
      • Nitschke J.B.
      Uncertainty and anticipation in anxiety: an integrated neurobiological and psychological perspective.
      ), and with prior work, which found that for individuals at risk for future anxiety, excessive attention to threat over a period lasting several seconds was uniquely predictive of increased anxiety one year later (
      • Bardeen J.R.
      • Daniel T.A.
      Anxiety sensitivity and attentional bias to threat interact to prospectively predict anxiety.
      ).
      On the other hand, startle is a subcortically-mediated measure of reflexive responding to threat (
      • Kuhn M.
      • Wendt J.
      • Sjouwerman R.
      • Büchel C.
      • Hamm A.
      • Lonsdorf T.B.
      The Neurofunctional Basis of Affective Startle Modulation in Humans: Evidence From Combined Facial Electromyography and Functional Magnetic Resonance Imaging.
      ). Therefore, our finding that increases in Time 2 Fear were predicted by heightened startle response at Time 1 suggests that increased “bottom up” response to fear-provoking stimuli and/or failure to inhibit bottom-up responding might serve as a risk factor for the development of fear-based psychopathology (

      Peng Y, Knotts JD, Young KS, Bookheimer SY, Nusslock R, Zinbarg RE, et al. (2022): Threat Neurocircuitry Predicts the Development of Anxiety and Depression Symptoms in a Longitudinal Study. Biol Psychiatry Cogn Neurosci Neuroimaging. https://doi.org/10.1016/j.bpsc.2021.12.013

      ,
      • Feng P.
      • Chen Z.
      • Becker B.
      • Liu X.
      • Zhou F.
      • He Q.
      • et al.
      Predisposing Variations in Fear-Related Brain Networks Prospectively Predict Fearful Feelings during the 2019 Coronavirus (COVID-19) Pandemic.
      ). Our cross-sectional results suggest, however, that after acquiring fear symptomatology, it might manifest in greater elaborated anticipation of predictable threat in the seconds prior to its delivery, as indicated by the association between the SPN to predictable threat and Time 1 Fear (
      • Michalowski J.M.
      • Pané-Farré C.A.
      • Löw A.
      • Hamm A.O.
      Brain dynamics of visual attention during anticipation and encoding of threat- and safe-cues in spider-phobic individuals.
      ). In sum, different neurobiological markers may be best suited to tracking cross-sectional versus prospective risk for fear and anxiety. In the context of the NPU task, ERPs and startle appear to work well-together to provide insight into multiple processes that may uniquely portend risk for or covary with current fear versus anxiety. Nonetheless, more work is needed to increase confidence in the specificity of these findings.
      Our results also provide initial support for distinct associations between response to predictable and unpredictable threat and transdiagnostic fear and anxiety. Continued investigation mapping neurobiological response to transdiagnostic fear and anxiety – constructs that may be more homogeneous and more closely tied to mechanism than the categorical disorders – may lead to improved classification and treatment of internalizing disorders. For example, targeting interventions at unpredictable or predictable threat responding might prove more effective than interventions that focus more generally on overall threat reactivity. Along these lines, prior work has shown that two weeks of selective serotonin reuptake inhibitors (SSRIs) modulates startle response to unpredictable threat cues in healthy adults (
      • Grillon C.
      • Chavis C.
      • Covington M.F.
      • Pine D.S.
      Two-week treatment with the selective serotonin reuptake inhibitor citalopram reduces contextual anxiety but not cued fear in healthy volunteers: a fear-potentiated startle study.
      ), suggesting why SSRIs might be more beneficial for individuals with sustained anxiety versus phasic fear. Given that most patients will manifest with symptoms of both fear and anxiety, accurate mapping of the relative contribution of abnormalities in predictable versus unpredictable threat responding to these dimensions may facilitate personalized treatment protocols. Greater specification of both treatment targets and their intended effects (i.e., more homogeneous dimensions of psychopathology) may help ensure that viable treatments are not discarded because they are targeted at more general operationalizations of threat reactivity and/or effects are measured in terms of heterogeneous diagnostic categories.
      Together, the current results provide support at multiple neurobiological levels for the theoretical distinction between fear and anxiety symptomatology, and link these dimensions to exaggerated predictable and unpredictable threat reactivity, respectively. Results support consideration of predictable and unpredictable threat reactivity as potential prognostic indicators of fear and anxiety symptomatology, which could lead to earlier and more targeted clinical care.

      Uncited reference

      • Hettema J.M.
      • Prescott C.A.
      • Myers J.M.
      • Neale M.C.
      • Kendler K.S.
      The Structure of Genetic and Environmental Risk Factors for Anxiety Disorders in Men and Women.
      ,
      • Kendler K.S.
      • Prescott C.A.
      • Myers J.
      • Neale M.C.
      The Structure of Genetic and Environmental Risk Factors for Common Psychiatric and Substance Use Disorders in Men and Women.
      ,
      • Barlow D.H.
      Unraveling the mysteries of anxiety and its disorders from the perspective of emotion theory.
      ,
      • Brinkmann L.
      • Buff C.
      • Feldker K.
      • Tupak S.V.
      • Becker M.P.I.
      • Herrmann M.J.
      • Straube T.
      Distinct phasic and sustained brain responses and connectivity of amygdala and bed nucleus of the stria terminalis during threat anticipation in panic disorder.
      ,
      • Goode T.D.
      • Ressler R.L.
      • Acca G.M.
      • Miles O.W.
      • Maren S.
      Bed nucleus of the stria terminalis regulates fear to unpredictable threat signals.
      ,

      Hulsman AM, Terburg D, Roelofs K, Klumpers F (2021): Chapter 28 - Roles of the bed nucleus of the stria terminalis and amygdala in fear reactions. In: Swaab DF, Kreier F, Lucassen PJ, Salehi A, Buijs RM, editors. Handbook of Clinical Neurology, vol. 179. Elsevier, pp 419–432.

      ,

      First MB, Williams JBW, Karg RS, Spitzer RL (2015): Structured Clinical Interview for DSM 5-Research Version (SCID-5 for DSM-5, Research Version; SCID-5-RV). Arlington, VA: American Psychiatric Association.

      ,

      Watson D, Clark LA (1994): The PANAS-X: Manual for the Positive and Negative Affect Schedule-Expanded Form. University of Iowa. Retrieved from http://www.psychology.uiowa.edu/Faculty/Watson/PANAS-X.pdf

      ,
      • Connor K.M.
      • Davidson J.R.
      • Churchill L.E.
      • Sherwood A.
      • Foa E.
      • Weisler R.H.
      Psychometric properties of the Social Phobia Inventory (SPIN). New self-rating scale.
      ,
      • Morton D.L.
      • Brown C.A.
      • Watson A.
      • El-Deredy W.
      • Jones A.K.P.
      Cognitive changes as a result of a single exposure to placebo.
      ,

      MacKinnon D (2008): Introduction to Statistical Mediation Analysis. New York, NY: Routledge.

      ,
      • Chmura Kraemer H.
      • Noda A.
      • O’Hara R.
      Categorical versus dimensional approaches to diagnosis: methodological challenges.
      .

      Acknowledgements

      This work was supported by National Institute of Mental Health grants, K23MH105553 and R01MH125083 (to AM). Thanks to Oi-Man Kwok for consulting on statistical procedures, as well Blake Barley, Elizabeth A. Bauer and Jared Ruchensky for their roles in data collection.

      References

        • Davis M.
        • Walker D.L.
        • Miles L.
        • Grillon C.
        Phasic vs Sustained Fear in Rats and Humans: Role of the Extended Amygdala in Fear vs Anxiety [no. 1].
        Neuropsychopharmacology. 2010; 35: 105-135
        • Grillon C.
        • Baas J.P.
        • Lissek S.
        • Smith K.
        • Milstein J.
        Anxious Responses to Predictable and Unpredictable Aversive Events.
        Behav Neurosci. 2004; 118: 916-924
        • Hamm A.O.
        Fear, anxiety, and their disorders from the perspective of psychophysiology.
        Psychophysiology. 2020; 57e13474
        • Robinson O.J.
        • Pike A.C.
        • Cornwell B.
        • Grillon C.
        The translational neural circuitry of anxiety.
        J Neurol Neurosurg Psychiatry. 2019;
        • Hur J.
        • Smith J.F.
        • DeYoung K.A.
        • Anderson A.S.
        • Kuang J.
        • Kim H.C.
        • et al.
        Anxiety and the Neurobiology of Temporally Uncertain Threat Anticipation.
        Neuroscience. 2020; https://doi.org/10.1101/2020.02.25.964734
        • Insel T.
        • Cuthbert B.
        • Garvey M.
        • Heinssen R.
        • Pine D.S.
        • Quinn K.
        • et al.
        Research Domain Criteria (RDoC): Toward a New Classification Framework for Research on Mental Disorders.
        Am J Psychiatry. 2010; 167: 748-751
        • Schmitz A.
        • Grillon C.
        Assessing fear and anxiety in humans using the threat of predictable and unpredictable aversive events (the NPU-threat test).
        Nat Protoc. 2012; 7: 527-532
        • Grillon C.
        • O’Connell K.
        • Lieberman L.
        • Alvarez G.
        • Geraci M.
        • Pine D.S.
        • Ernst M.
        Distinct Responses to Predictable and Unpredictable Threat in Anxiety Pathologies: Effect of Panic Attack.
        Biol Psychiatry Cogn Neurosci Neuroimaging. 2017; 2: 575-581
        • Kaye J.T.
        • Bradford D.E.
        • Curtin J.J.
        Psychometric properties of startle and corrugator response in NPU, affective picture viewing, and resting state tasks.
        Psychophysiology. 2016; 53: 1241-1255
        • MacNamara A.
        • Barley B.
        Event-related potentials to threat of predictable and unpredictable shock.
        Psychophysiology. 2018; 55e13206
        • Radoman M.
        • Phan K.L.
        • Gorka S.M.
        Neural correlates of predictable and unpredictable threat in internalizing psychopathology.
        Neurosci Lett. 2019; 701: 193-201
        • Gorka S.M.
        • Lieberman L.
        • Klumpp H.
        • Kinney K.L.
        • Kennedy A.E.
        • Ajilore O.
        • et al.
        Reactivity to unpredictable threat as a treatment target for fear-based anxiety disorders.
        Psychol Med. 2017; 47: 2450-2460
        • Grillon C.
        • Pine D.S.
        • Lissek S.
        • Rabin S.
        • Bonne O.
        • Vythilingam M.
        Increased Anxiety During Anticipation of Unpredictable Aversive Stimuli in Posttraumatic Stress Disorder but not in Generalized Anxiety Disorder.
        Biol Psychiatry. 2009; 66: 47-53
        • Chantarujikapong S.I.
        • Scherrer J.F.
        • Xian H.
        • Eisen S.A.
        • Lyons M.J.
        • Goldberg J.
        • et al.
        A twin study of generalized anxiety disorder symptoms, panic disorder symptoms and post-traumatic stress disorder in men.
        Psychiatry Res. 2001; 103: 133-145
        • Hettema J.M.
        • Prescott C.A.
        • Myers J.M.
        • Neale M.C.
        • Kendler K.S.
        The Structure of Genetic and Environmental Risk Factors for Anxiety Disorders in Men and Women.
        Arch Gen Psychiatry. 2005; 62: 182-189
        • Kendler K.S.
        • Prescott C.A.
        • Myers J.
        • Neale M.C.
        The Structure of Genetic and Environmental Risk Factors for Common Psychiatric and Substance Use Disorders in Men and Women.
        Arch Gen Psychiatry. 2003; 60: 929-937
        • Barlow D.H.
        Unraveling the mysteries of anxiety and its disorders from the perspective of emotion theory.
        Am Psychol. 2000; 55: 1247-1263
        • Gorka S.M.
        • Nelson B.D.
        • Shankman S.A.
        Startle response to unpredictable threat in comorbid panic disorder and alcohol dependence.
        Drug Alcohol Depend. 2013; 132: 216-222
        • Gorka S.M.
        • Lieberman L.
        • Shankman S.A.
        • Phan K.L.
        Startle potentiation to uncertain threat as a psychophysiological indicator of fear-based psychopathology: An examination across multiple internalizing disorders.
        J Abnorm Psychol. 2017; 126: 8-18
        • Lieberman L.
        • Gorka S.M.
        • Shankman S.A.
        • Phan K.L.
        Impact of Panic on Psychophysiological and Neural Reactivity to Unpredictable Threat in Depression and Anxiety.
        Clin Psychol Sci. 2017; 5: 52-63
        • Gorka S.M.
        • Liu H.
        • Sarapas C.
        • Shankman S.A.
        Time course of threat responding in panic disorder and depression.
        Int J Psychophysiol Off J Int Organ Psychophysiol. 2015; 98: 87-94
        • Shankman S.A.
        • Nelson B.D.
        • Sarapas C.
        • Robison-Andrew E.J.
        • Campbell M.L.
        • Altman S.E.
        • et al.
        A psychophysiological investigation of threat and reward sensitivity in individuals with panic disorder and/or major depressive disorder.
        J Abnorm Psychol. 2013; 122: 322-338
        • Klahn A.L.
        • Klinkenberg I.A.
        • Lueken U.
        • Notzon S.
        • Arolt V.
        • Pantev C.
        • et al.
        Commonalities and differences in the neural substrates of threat predictability in panic disorder and specific phobia.
        NeuroImage Clin. 2017; 14: 530-537
        • Brinkmann L.
        • Buff C.
        • Feldker K.
        • Tupak S.V.
        • Becker M.P.I.
        • Herrmann M.J.
        • Straube T.
        Distinct phasic and sustained brain responses and connectivity of amygdala and bed nucleus of the stria terminalis during threat anticipation in panic disorder.
        Psychol Med. 2017; 47: 2675-2688
        • Goode T.D.
        • Ressler R.L.
        • Acca G.M.
        • Miles O.W.
        • Maren S.
        Bed nucleus of the stria terminalis regulates fear to unpredictable threat signals.
        eLife. 2019; 8e46525
      1. Hulsman AM, Terburg D, Roelofs K, Klumpers F (2021): Chapter 28 - Roles of the bed nucleus of the stria terminalis and amygdala in fear reactions. In: Swaab DF, Kreier F, Lucassen PJ, Salehi A, Buijs RM, editors. Handbook of Clinical Neurology, vol. 179. Elsevier, pp 419–432.

        • Scherrer J.F.
        • True W.R.
        • Xian H.
        • Lyons M.J.
        • Eisen S.A.
        • Goldberg J.
        • et al.
        Evidence for genetic influences common and specific to symptoms of generalized anxiety and panic.
        J Affect Disord. 2000; 57: 25-35
        • Gorka S.M.
        • Kreutzer K.A.
        • Petrey K.M.
        • Radoman M.
        • Phan K.L.
        Behavioral and neural sensitivity to uncertain threat in individuals with alcohol use disorder: Associations with drinking behaviors and motives.
        Addict Biol. 2020; 25e12774
        • Brinkmann L.
        • Buff C.
        • Neumeister P.
        • Tupak S.V.
        • Becker M.P.I.
        • Herrmann M.J.
        • Straube T.
        Dissociation between amygdala and bed nucleus of the stria terminalis during threat anticipation in female post-traumatic stress disorder patients.
        Hum Brain Mapp. 2017; 38: 2190-2205
        • Buff C.
        • Brinkmann L.
        • Bruchmann M.
        • Becker M.P.I.
        • Tupak S.
        • Herrmann M.J.
        • Straube T.
        Activity alterations in the bed nucleus of the stria terminalis and amygdala during threat anticipation in generalized anxiety disorder.
        Soc Cogn Affect Neurosci. 2017; 12: 1766-1774
        • Nelson B.D.
        • Hajcak G.
        Anxiety and Depression Symptom Dimensions Demonstrate Unique Relationships with the Startle Reflex in Anticipation of Unpredictable Threat in 8 to 14 Year-Old Girls.
        J Abnorm Child Psychol. 2017; 45: 397-410
        • Clauss J.A.
        • Avery S.N.
        • Benningfield M.M.
        • JU Blackford
        Social anxiety is associated with BNST response to unpredictability.
        Depress Anxiety. 2019; 36: 666-675
        • Stevens E.S.
        • Lieberman L.
        • Funkhouser C.J.
        • Correa K.A.
        • Shankman S.A.
        Startle during threat longitudinally predicts functional impairment independent of DSM diagnoses.
        Psychiatry Res. 2019; 279: 207-215
        • Böcker K.B.E.
        • Baas J.M.P.
        • Kenemans J.L.
        • Verbaten M.N.
        Stimulus-preceding negativity induced by fear: a manifestation of affective anticipation.
        Int J Psychophysiol. 2001; 43: 77-90
        • Tanovic E.
        • Joormann J.
        Anticipating the unknown: The stimulus-preceding negativity is enhanced by uncertain threat.
        Int J Psychophysiol. 2019; 139: 68-73
      2. Catena A, Perales JC, Megías A, Cándido A, Jara E, Maldonado A (2012): The Brain Network of Expectancy and Uncertainty Processing ((C. Soriano-Mas, editor)). PLoS ONE 7: e40252.

        • Lin H.
        • Gao H.
        • You J.
        • Liang J.
        • Ma J.
        • Yang N.
        • et al.
        Larger N2 and smaller early contingent negative variation during the processing of uncertainty about future emotional events.
        Int J Psychophysiol. 2014; 94: 292-297
      3. First MB, Williams JBW, Karg RS, Spitzer RL (2015): Structured Clinical Interview for DSM 5-Research Version (SCID-5 for DSM-5, Research Version; SCID-5-RV). Arlington, VA: American Psychiatric Association.

      4. Watson D, Clark LA (1994): The PANAS-X: Manual for the Positive and Negative Affect Schedule-Expanded Form. University of Iowa. Retrieved from http://www.psychology.uiowa.edu/Faculty/Watson/PANAS-X.pdf

        • Meyer T.J.
        • Miller M.L.
        • Metzger R.L.
        • Borkovec T.D.
        Development and validation of the penn state worry questionnaire.
        Behav Res Ther. 1990; 28: 487-495
      5. Spielberger CD (1983): Manual for the State-Trait Anxiety Inventory (STAI). Palo Alto, CA: Consulting Psychologists Press.

        • Connor K.M.
        • Davidson J.R.
        • Churchill L.E.
        • Sherwood A.
        • Foa E.
        • Weisler R.H.
        Psychometric properties of the Social Phobia Inventory (SPIN). New self-rating scale.
        Br J Psychiatry J Ment Sci. 2000; 176: 379-386
        • Banica I.
        • Sandre A.
        • Shields G.S.
        • Slavich G.M.
        • Weinberg A.
        The error-related negativity (ERN) moderates the association between interpersonal stress and anxiety symptoms six months later.
        Int J Psychophysiol Off J Int Organ Psychophysiol. 2020; 153: 27-36
      6. Bradford DE, Magruder KP, Korhumel RA, Curtin JJ (2014): Using the threat probability task to assess anxiety and fear during uncertain and certain threat. J Vis Exp JoVE 51905.

        • Blumenthal T.D.
        • Cuthbert B.N.
        • Filion D.L.
        • Hackley S.
        • Lipp O.V.
        • Boxtel A.V.
        Committee report: Guidelines for human startle eyeblink electromyographic studies.
        Psychophysiology. 2005; 42: 1-15
        • Morton D.L.
        • Brown C.A.
        • Watson A.
        • El-Deredy W.
        • Jones A.K.P.
        Cognitive changes as a result of a single exposure to placebo.
        Neuropsychologia. 2010; 48 (–1964): 1958
        • Westfall P.H.
        On Using the Bootstrap for Multiple Comparisons.
        J Biopharm Stat. 2011; 21: 1187-1205
      7. MacKinnon D (2008): Introduction to Statistical Mediation Analysis. New York, NY: Routledge.

        • Grady C.L.
        • Luk G.
        • Craik F.I.M.
        • Bialystok E.
        Brain Network Activity in Monolingual and Bilingual Older Adults.
        Neuropsychologia. 2015; 66: 170-181
        • Chmura Kraemer H.
        • Noda A.
        • O’Hara R.
        Categorical versus dimensional approaches to diagnosis: methodological challenges.
        J Psychiatr Res. 2004; 38: 17-25
        • Cohen J.
        The Cost of Dichotomization.
        Appl Psychol Meas. 1983; 7: 249-253
        • Grupe D.W.
        • Nitschke J.B.
        Uncertainty and anticipation in anxiety: an integrated neurobiological and psychological perspective.
        Nat Rev Neurosci. 2013; 14: 488-501
        • Bardeen J.R.
        • Daniel T.A.
        Anxiety sensitivity and attentional bias to threat interact to prospectively predict anxiety.
        Cogn Behav Ther. 2018; 47: 482-494
        • Kuhn M.
        • Wendt J.
        • Sjouwerman R.
        • Büchel C.
        • Hamm A.
        • Lonsdorf T.B.
        The Neurofunctional Basis of Affective Startle Modulation in Humans: Evidence From Combined Facial Electromyography and Functional Magnetic Resonance Imaging.
        Biol Psychiatry S000632231931563X. 2019;
      8. Peng Y, Knotts JD, Young KS, Bookheimer SY, Nusslock R, Zinbarg RE, et al. (2022): Threat Neurocircuitry Predicts the Development of Anxiety and Depression Symptoms in a Longitudinal Study. Biol Psychiatry Cogn Neurosci Neuroimaging. https://doi.org/10.1016/j.bpsc.2021.12.013

        • Feng P.
        • Chen Z.
        • Becker B.
        • Liu X.
        • Zhou F.
        • He Q.
        • et al.
        Predisposing Variations in Fear-Related Brain Networks Prospectively Predict Fearful Feelings during the 2019 Coronavirus (COVID-19) Pandemic.
        Cereb Cortex. 2022; 32: 540-553
        • Michalowski J.M.
        • Pané-Farré C.A.
        • Löw A.
        • Hamm A.O.
        Brain dynamics of visual attention during anticipation and encoding of threat- and safe-cues in spider-phobic individuals.
        Soc Cogn Affect Neurosci. 2015; 10: 1177-1186
        • Grillon C.
        • Chavis C.
        • Covington M.F.
        • Pine D.S.
        Two-week treatment with the selective serotonin reuptake inhibitor citalopram reduces contextual anxiety but not cued fear in healthy volunteers: a fear-potentiated startle study.
        Neuropsychopharmacol Off Publ Am Coll Neuropsychopharmacol. 2009; 34: 964-971