How interindividual differences shape approach-avoidance behavior: Relating self-report and diagnostic measures of interindividual differences to behavioral measurements of approach and avoidance

Highlights

• Interaction of approach-avoidance and interindividual differences often ambiguous.

• Specific phobias can be reliably measured with approach-avoidance tasks.

• No effects of endogenous testosterone on approach-avoidance behavior differences.

• Strong variation in study designs across approach-avoidance literature.

• Important areas of research are currently underrepresented.

Abstract

Responding to stimuli in ambiguous environments is partially governed by approach-avoidance tendencies. Imbalances in these approach-avoidance behaviors are implicated in many mental disorders including anxiety disorders, phobias and substance use disorders. While factors biasing human behavior in approach-avoidance conflicts have been researched in numerous experiments, a much-needed comprehensive overview integrating those findings is missing. Here, we systematically searched the existing literature on individual differences in task-based approach-avoidance behavior and aggregated the current evidence for the effect of self-reported approach/avoidance traits, anxiety and anxiety disorders, specific phobias, depression, aggression, anger and psychopathy, substance use and related disorders, eating disorders and habits, trauma, acute stress and, finally, hormone levels (mainly testosterone, oxytocin). We highlight consistent findings, underrepresented research areas and unexpected results, and detail the amount of controversy between studies. We discuss potential reasons for ambiguous results in some research areas, offer practical advice for future studies and highlight potential variables such as task-related researcher decisions that may influence how interindividual differences and disorders drive automatic approach-avoidance biases in behavioral experiments.

Introduction

One of the essential principles of life is adaptability, i.e., the ability of an organism to respond to the external environment. Behavioral approach and avoidance can be seen as fundamental building blocks allowing us to adapt to an ever-changing environment (Koshland, 2002). For example, we tend to approach stimuli of positive valence such as food when we are hungry and tend to avoid stimuli of negative valence such as pain or predatory threat. Approach and avoidance are complex reactions, shaped by underlying genetics and prior experience, and guided by both reflex and cognitive control (McNaughton et al., 2016). As the stimuli in our environment are often ambiguous, they frequently result in so-called approach-avoidance conflicts. Should one enter a dark alley on the way home or take a detour? Should one take the last piece of cake despite the dietary implications and angry looks of the remaining party guests? Different people display strikingly divergent behaviors when faced with these types of conflicts. However, the reasons for these individual differences remain largely unclear and are the focus of this review.

Previous research has shown the striking interindividual variation in approach-avoidance tendencies to the point that aberrant approach-avoidance behaviors have been interpreted as an integral part or even cause of many pathologies such as excessive approach of addiction-related stimuli in addiction and unreasonable avoidance of social stimuli in social anxiety disorder (e.g., McNaughton et al., 2016). In a more general sense, recent findings highlighted possible associations between personality traits and some psychiatric diagnoses. For instance, individual differences in the fight-flight-freeze system assessed by self-report were found to mediate the relationship of psychopathy and externalizing behaviors (Johnson et al., 2014) and interactions between neuroticism, extraversion and conscientiousness were predictive of depression (Allen et al., 2018). Theoretical work is also done to extend current nosologies with novel models in which personality traits play larger roles in defining psychopathologies (e.g., DeYoung and Krueger, 2018; Kotov et al., 2017). However, the underlying mechanisms of how interindividual differences related to approach and avoidance map onto psychopathology are not fully understood as of yet. Understanding these mechanisms is urgent, considering that despite the unknown background, approach-avoidance trainings are used in clinical research and practice. For example, approach-avoidance tasks have been utilized to train participants to avoid stimuli that are detrimental to their health such as alcohol in problematic drinkers (Eberl et al., 2013) or food-stimuli in bulimia nervosa patients (Brockmeyer et al., 2019). Likewise, other studies trained participants to approach stimuli that they avoid to an unhealthy extent such as contamination-related stimuli in obsessive compulsive disorder (Weil et al., 2017) or social stimuli in social anxiety disorder (Asnaani et al., 2014). While there are reports that training may be effective in reducing alcohol use, cigarette smoking, and unhealthy eating (Kakoschke et al., 2017a), evidence is certainly not found in all studies (e.g., Aulbach et al., 2019; Dickson et al., 2016; Ferentzi et al., 2018; Lester et al., 2015; Sharbanee et al., 2014a) and the reasons for this inconsistency are currently unclear.

Approach-avoidance behavior is theorized to be controlled by multiple motivational systems activated by appetitive and aversive stimuli. One well-established idea, Gray’s revised reinforcement sensitivity theory, postulates that individual differences in sensitivity to reward, punishment and motivation originate from three distinct but interacting brain-behavioral systems (Gray and McNaughton, 2003). According to this theory, the behavioral approach system (BAS) is involved in basic control of approach towards positive stimuli, while the fight-flight-freeze system (FFFS) is activated in case of aversive stimuli to elicit avoidance. The third system is the behavioral inhibition system (BIS), which becomes active in situations of conflict. Thus, the BIS serves as conflict detection system and mediator for conflicts within and between the other two motivational systems. The control of basic approach and avoidance is therefore believed to be handled by separate systems (BAS and FFFS), while the simultaneous activation of both approach and avoidance tendencies will lead to the activation of all three systems. It follows that interindividual differences in behavior can be the result of changes in approach, avoidance or conflict sensitivity, or a combination of these (Gray and McNaughton, 2003).

These different systems involved in approach-avoidance can be affected by individual differences at various levels. For instance, the initial evaluation of a stimulus will depend on the internal value an individual assigns to it and this value varies based on the current level of drive the stimulus elicits and the type of contingency in a given situation (McNaughton et al., 2016). For example, the elicited approach-avoidance reactions of two individuals towards a piece of cake will vary, for instance based on their personality traits, current level of drive (e.g. hunger) and the contingency (e.g. potential gain of a piece of cake or potential loss of a piece of cake warrant different reactions).

Furthermore, rodent experiments have shown that approach and avoidance gradients appear to generally differ in their steepness with approach gradients being shallower and avoidance gradients being steeper. This means that positive stimuli may be approached from a relatively longer distance, while negative stimuli elicit a strong avoidance response in close proximity to the stimulus, which can ‘override’ the shallower approach gradient (McNaughton et al., 2016). The so-called reversal point is located where steep avoidance gradients become comparatively stronger than the shallower approach gradients leading to shifts in the involved approach-avoidance responses.

This phenomenon makes distance a crucial factor in approach-avoidance conflicts. Indeed, manipulating gradients by changing task parameters such as threat level or distance has been shown to affect the approach-avoidance responses of participants on both behavioral and neuronal level (e.g., Fung et al., 2019). Importantly, this study demonstrated that personality traits affect how threat level and distance can influence behavior. The same might hold true for states or other traits. In general, different experimental procedures may lead to different results in terms of approach-avoidance behavior due to being perceived on opposite sides of the reversal point.

The distance at which approach-avoidance conflicts take place will further shift based on individual differences of approach-avoidance gradients’ steepness, i.e. an individuals’ sensitivity to approach and avoidance. In addition, conflict enhances avoidance motivations, which further impacts observed behaviors. Thus, mixed motivations will lead to different approach-avoidance behaviors in human experiments based on where the given stimulus is perceived in space (or time), and this cognitive distance can vary based on experimental procedures and interindividual differences.

Mixed motivations in human experiments can arise not only from the approach-avoidance conflicts introduced by the task, but also by the experimental situation itself. Experimental situations may affect approach-avoidance behavior even prior to the start of the experiment as individual sensitivity for approach and avoidance differs and motivational forces are already at play, for example, in participants sensitive to conflict or anxiety. It is therefore conceivable that tasks which target approach or avoidance tendencies distinctly, still elicit the activation of the BIS due to cognitive conflict introduced by the experimental situation or instructions (e.g. it could be possible that the instruction to approach aversive stimuli leads to activation of the BIS as conflict detector). In summary, individual differences may affect one or several motivational systems of approach-avoidance. Which systems are involved in approach-avoidance tasks may be influenced by researcher choices in task design and various other factors, e.g., individual cognitive distance and the experimental situation itself.

There seems to be some consensus about the neural structures involved in approach-avoidance responses measured in approach-avoidance tendency tasks in humans, particularly in social situations (for a recent review, see Kaldewaij et al., 2017). An important brain region controlling approach-avoidance behaviors in these tasks is the anterior prefrontal cortex (aPFC), where activity increases in cases of approach-avoidance conflict due to incongruency of instructions and action tendencies (Volman et al., 2011a, b). In addition, the amygdala and its functional coupling with the aPFC has been implicated in the control of approach-avoidance decisions as the aPFC suppresses automatic action tendencies incompatible with current goals (Volman et al., 2013, 2016). In more translational approach-avoidance conflict tasks, the anterior hippocampus seems to play a causal role especially in avoidance behaviors (Bach et al., 2014; Korn and Bach, 2019; O’Neil et al., 2015) and may serve as control region, particularly tracking the current level of threat (Bach et al., 2014). For instance, patients with hippocampal lesions demonstrated reduced inhibition and passive avoidance behavior in a foraging task under threat of predation (Bach et al., 2014; O’Neil et al., 2015). Additionally, the amygdala is of key importance for anxiety-like behaviors in these approach-avoidance conflict tasks as well (Korn et al., 2017). However, no increase in amygdala activity was reported in response to conflict in several affective/pain punishment-induced conflict tasks, indicating that different types of approach-avoidance conflicts may activate different brain regions (Kirlic et al., 2017). Moreover, a recent study nicely illustrated that slow threat may involve higher cognitive areas, i.e. hippocampus and PFC, while the response to fast threat may rely on different areas such as the periaqueductal grey (Qi et al., 2018). Further regions that have been implicated in the control of approach-avoidance conflicts and making optimized decisions are the dorsomedial, dorsolateral and ventrolateral prefrontal cortex regions (dmPFC, dlPFC, vlPFC) and the orbitofrontal cortex (OFC; Aupperle et al., 2015; Korn and Bach, 2019; Roelofs et al., 2009a; Volman et al., 2011b). To summarize, PFC areas and anterior hippocampus are implicated in the control of approach-avoidance behaviors, albeit in different types of tasks and not in all reported studies. Interestingly, these brain areas have been implicated in various psychiatric disorders that present with altered approach-avoidance behavior. Approach-avoidance research thus possibly opens a way to better understand how these disorders develop and persist.

Despite the multitude of studies employing approach-avoidance tasks in a variety of research fields such as anxiety or substance use and recent advances in the investigation of neuronal underpinnings, a comprehensive review of the literature addressing interindividual differences in approach-avoidance behaviors is missing. Integrating results from previously conducted studies is important, however, to ensure better, more reliable interpretation of findings and assess implications for, among others, the disorders that have been strongly associated with aberrant approach-avoidance behavior and therapeutic practices. Here, we set out to summarize the current level of knowledge about those interindividual differences that drive disparities in approach-avoidance behavior.

To this end, we focused on behavioral experimental human studies that addressed interindividual differences and examined a distinct approach-avoidance conflict (see methods for inclusion and exclusion criteria). We combined findings pertaining interindividual differences of a certain research field (e.g., anxiety disorders) where possible, i.e., if enough articles were present to allow for an integration of findings. Therefore, research in relation to the following concepts is presented: self-reported approach/avoidance traits, anxiety and anxiety disorders, specific phobias, depression, aggression, anger and psychopathy, substance use and related disorders, eating disorders and habits, trauma and acute stress and, finally, hormone levels (mainly testosterone, oxytocin). Articles that were initially included, but could not be grouped to accumulate evidence are detailed in the supplementary material.

Approach-avoidance decisions have been investigated employing a variety of tasks and populations differing in various states or traits resulting in a large body of research. The majority of the articles investigated here utilizes tasks falling in the broad category of approach-avoidance tendency tasks. These are intended to capture action tendencies, for example by instructing to approach stimuli of positive valence and avoid stimuli of negative valence or the other way around. These tasks thus create an environment where either congruent or incongruent approach-avoidance behaviors are required, and the differences in behavior between those instructions are then analyzed (for detailed descriptions see Box 1Box 1 ). Approach-avoidance tendency tasks target automatic tendencies, but require goal-directed instrumental responding to overcome these tendencies in case of incongruencies and may involve one or more of Gray’s postulated approach-avoidance systems. Other tasks create approach-avoidance conflict within the task by creating ambiguous environments with stimuli that represent positive and negative potential outcomes simultaneously. These tasks clearly target all three approach-avoidance systems (BIS, BAS, FFFS). While we employ a definition of conflict that includes conflict introduced by instruction as seen in approach-avoidance tendency tasks, the tasks are very differently set up and likely activate different neural systems (see Section 1.2). To appreciate the different types of tasks used, the result sub-sections include tables detailing characteristics of the tasks, stimulus types etc.

We hope that this review provides valuable information for fundamental and clinical research by shedding more light on approach-avoidance behaviors in health and disease and discussing their theoretical implications. We identify interindividual differences that reliably drive divergent approach-avoidance responses in commonly used approach-avoidance paradigms but also highlight current inconsistencies. This is of interest in the design of new studies since lacking replicability may warrant higher numbers of participants, clearer separation of participant groups or the use of other tasks or task versions. The review further helps to gain insight into psychiatric disorders in which approach-avoidance behavior is aberrant such as substance use or anxiety disorders and the role the distorted approach-avoidance behavior may play in disease manifestation or treatment. Moreover, we will highlight research areas in which current findings are inconclusive or lacking and more research is required to understand the sources and actions of interindividual differences. Finally, we will discuss translational research that may contribute relevant knowledge to determine the causal factors driving variation of approach-avoidance tendencies.