Novelty, conditioning and attentional bias to sexual rewards (2015)

COMMENTS: New Cambridge University brain study. Subjects were carefully screened porn addicts. Compared with controls, they habituated faster to sexual images. That is, their brains became less activated seeing the same image…they were more quickly bored. Thus, the novelty of internet porn drives the addiction to it, creating a circular spiral of needing more novelty to overcome faster habituation. But this desire for novelty in porn addicts was NOT pre-existing. That is, the ‘chicken’ is porn use and the ‘egg’ is novelty-seeking.

Press Release. November 23, 2015

• Link to full study

People who show compulsive sexual behaviour – sex addiction – are driven to search more for new sexual images than their peers, according to new research led by the University of Cambridge. The findings may be particularly relevant in the context of online porn, which potentially provides an almost endless source of new images.

In a study published in the Journal of Psychiatric Research, researchers also report that sex addicts are more susceptible to environment ‘cues’ linked to sexual images than to those linked to neutral images.

Sex addiction – when an individual has difficulty controlling their sexual thoughts, feelings or behaviour – is relatively common, affecting as many as one in 25 young adults. It is heavily stigmatised and can lead to a sense of shame, affecting an individual’s family and social life as well as their work. There is no formal definition of the condition to help with diagnosis.

In previous work led by Dr Valerie Voon from the Department of Psychiatry at the University of Cambridge, scientists found that three brain regions were more active in sex addicts compared with the healthy volunteers. Significantly, these regions – the ventral striatum, dorsal anterior cingulate and amygdala – were regions that are also activated in drug addicts when shown drug stimuli.

In the new study, funded by the Wellcome Trust, Dr Voon and colleagues studied the behaviour of 22 sex addicts and 40 ‘healthy’ male volunteers undergoing tasks. In the first task, individuals were shown a series of images in pairs, including naked women, clothed women and furniture. They were then shown further image pairs, including familiar and new images, and asked to choose an image to ‘win £1’ – although the participants were unaware of the odds, the probability of winning for either images was 50%.

The researchers found that sex addicts were more likely to choose the novel over the familiar choice for sexual images relative to neutral object images, whereas healthy volunteers were more likely to choose the novel choice for neutral human female images relative to neutral object images.

“We can all relate in some way to searching for novel stimuli online – it could be flitting from one news website to another, or jumping from Facebook to Amazon to YouTube and on,” explains Dr Voon. “For people who show compulsive sexual behaviour, though, this becomes a pattern of behaviour beyond their control, focused on pornographic images.”

In a second task, volunteers were shown pairs of images – an undressed woman and a neutral grey box – both of which were overlaid on different abstract patterns. They learned to associate these abstract images with the images, similar to how the dogs in Pavlov’s famous experiment learnt to associate a ringing bell with food. They were then asked to select between these abstract images and a new abstract image.

This time, the researchers showed that sex addicts were more likely to choose cues (in this case the abstract patterns) associated with sexual and monetary rewards. This supports the notion that apparently innocuous cues in an addict’s environment can ‘trigger’ them to seek out sexual images.

“Cues can be as simple as just opening up their internet browser,” explains Dr Voon. “They can trigger a chain of actions and before they know it, the addict is browsing through pornographic images. Breaking the link between these cues and the behaviour can be extremely challenging.”

The researchers carried out a further test where 20 sex addicts and 20 matched healthy volunteers underwent brain scans while being shown a series of repeated images – an undressed woman, a £1 coin or a neutral grey box.

They found that when the sex addicts viewed the same sexual image repeatedly, compared to the healthy volunteers they experienced a greater decrease of activity in the region of the brain known as the dorsal anterior cingulate cortex, known to be involved in anticipating rewards and responding to new events. This is consistent with ‘habituation’, where the addict finds the same stimulus less and less rewarding – for example, a coffee drinker may get a caffeine ‘buzz’ from their first cup, but over time the more they drink coffee, the smaller the buzz becomes.

This same habituation effect occurs in healthy males who are repeatedly shown the same porn video. But when they then view a new video, the level of interest and arousal goes back to the original level. This implies that, to prevent habituation, the sex addict would need to seek out a constant supply of new images. In other words, habituation could drive the search for novel images.

“Our findings are particularly relevant in the context of online pornography,” adds Dr Voon. “It’s not clear what triggers sex addiction in the first place and it is likely that some people are more pre-disposed to the addiction than others, but the seemingly endless supply of novel sexual images available online helps feed their addiction, making it more and more difficult to escape.”

More information: Paula Banca et al. Novelty, conditioning and attentional bias to sexual rewards, Journal of Psychiatric Research (2016). DOI: 10.1016/j.jpsychires.2015.10.017

THE STUDY

Paula Banca, Laurel S. Morris, Simon MitchellNeil A. Harrison, Marc N. Potenza, Valerie Voon (Dr)

DOI: http://dx.doi.org/10.1016/j.jpsychires.2015.10.017

Abstract

The Internet provides a large source of novel and rewarding stimuli, particularly with respect to sexually explicit materials. Novelty-seeking and cue-conditioning are fundamental processes underlying preference and approach behaviors implicated in disorders of addiction. Here we examine these processes in individuals with compulsive sexual behaviors (CSB), hypothesizing a greater preference for sexual novelty and stimuli conditioned to sexual rewards relative to healthy volunteers. Twenty-two CSB males and forty age-matched male volunteers were tested in two separate behavioral tasks focusing on preferences for novelty and conditioned stimuli. Twenty subjects from each group were also assessed in a third conditioning and extinction task using functional magnetic resonance imaging. CSB was associated with enhanced novelty preference for sexual, as compared to control images, and a generalized preference for cues conditioned to sexual and monetary versus neutral outcomes compared to healthy volunteers. CSB individuals also had greater dorsal cingulate habituation to repeated sexual versus monetary images with the degree of habituation correlating with enhanced preference for sexual novelty. Approach behaviors to sexually conditioned cues dissociable from novelty preference were associated with an early attentional bias to sexual images. This study shows that CSB individuals have a dysfunctional enhanced preference for sexual novelty possibly mediated by greater cingulate habituation along with a generalized enhancement of conditioning to rewards. We further emphasize a dissociable role for cue-conditioning and novelty preference on the early attentional bias for sexual cues. These findings have wider relevance as the Internet provides a wide range of novel and potentially rewarding stimuli.

Keywords: novelty, cue-conditioning, sexual reward, dorsal cingulate habituation, addiction, attentional bias

Introduction

Why is online surfing so compulsively engaging to many individuals? The Internet provides a large source of novel and potentially rewarding stimuli. Novelty-seeking, attentional bias and cue-conditioning are fundamental processes that may drive unconscious preference and approach decisions in daily life. These processes may also contribute to the development and maintenance of disorders of addiction.

Novelty-seeking may be both a predictor and consequence of disorders of addiction. This trait, which is often assessed using the Zuckerman’s sensation-seeking scale, has been repeatedly found elevated in diverse behavioral and substance addictions (Belin et al., 2011, Redolat et al., 2009). A suggested explanation for this strong relationship relies on the hypothesis that exposure to novelty may activate, at least in part, the same neural machinery that mediates the rewarding effects of drugs of abuse (Bardo et al., 1996). In rodent studies, novelty preference predicts a transition towards compulsive cocaine-seeking behaviors (Belin and Deroche-Gamonet, 2012). In human studies, sensation-seeking is associated prospectively with binge drinking in adolescents (Conrod et al., 2013).

Conditioned signals or cues in our environment may also significantly influence behavior. The smell of cigarettes, places or friends associated with drug use, or the sight of money may act as conditioned cues and may enhance reactivity and trigger cravings, urges and relapse in disorders of addiction (for review see (Childress et al., 1993)). These cues are neutral stimuli that may inadvertently acquire motivational significance through the process of conditioning with repeated pairing with either drug rewards or other biologically relevant natural rewards such as food (Jansen, 1998) or sex (Pfaus et al., 2001, Toates, 2009).

The processing of novelty and learning has been proposed to include a functional polysynaptic loop involving the hippocampus, ventral striatum, and midbrain dopaminergic region (Lisman and Grace, 2005). Detection of novelty, long-term memory encoding and learning involves dopaminergic activity that enhances hippocampal synaptic plasticity which, through glutamatergic projections to the ventral striatum, relays information to the ventral tegmental area (VTA) which then projects directly back to the hippocampus (Knight, 1996, Lisman and Grace, 2005). With repeated exposure, the hippocampus and midbrain dopaminergic responses to novelty decrease, conferring habituation when stimuli become familiar (Bunzeck and Duzel, 2006, Bunzeck et al., 2013). Converging primate and human studies also show that phasic dopaminergic activity encodes prediction error, a comparison between the actual and expected outcomes indicating an unexpected salient outcome, acting as a teaching signal underlying conditioning processes (Schultz et al., 1997). Mesolimbic dopaminergic cell bodies in the midbrain project to a network including the striatum, dorsal anterior cingulate cortex (dACC) and hippocampus (Williams and Goldman-Rakic, 1998). The dACC is implicated in attentional responding to novel and salient events and in processing reward anticipation and prediction error (Ranganath and Rainer, 2003, Rushworth et al., 2011).

In addition to novelty-seeking and cue-conditioning influences, the tendency to preferentially process cues related to the object of addiction (attentional biases) is also an important feature that characterizes disorders of addiction (Ersche et al., 2010, van Hemel-Ruiter et al., 2013, Wiers et al., 2011). The influence of emotional stimuli on attentional processes is widely reported both in healthy and clinical samples (Yiend, 2010). Attentional biases towards substance-related stimuli have been found in substance-use disorders for alcohol, nicotine, cannabis, opiates and cocaine (Cox et al., 2006). Moreover, a direct association between highly arousing sexual pictures and attentional interference has also been found in healthy individuals, which seems to be influenced by sexuality-related attitudes and sexual motivation (Kagerer et al., 2014, Prause et al., 2008). We have previously extended these findings to individuals with compulsive sexual behaviour (CSB) using a dot-probe task (Mechelmans et al., 2014).

With increasing access to the Internet, there is growing concern regarding the potential for excessive use. A study that assessed the predictive power of several types of Internet applications (gaming, gambling, email, etc.) on the development of compulsive Internet use suggested that online sexually explicit stimuli have the highest potential for addictive/compulsive use (Meerkerk et al., 2006). Online explicit stimuli are vast and expanding, and this feature may promote escalation of use in some individuals. For instance, healthy males viewing repeatedly the same explicit film have been found to habituate to the stimulus and find the explicit stimulus as progressively less sexually arousing, less appetitive and less absorbing (Koukounas and Over, 2000). However, subsequent exposure to a novel explicit film segment increases the levels of sexual arousal and absorption to the same previous levels prior to habituation, suggesting important roles for novelty and habituation. Imaging studies have identified a specific network for the neural processing of sexual stimuli in healthy humans, comprising the hypothalamus, nucleus accumbens, orbitofrontal, occipital and parietal areas (Wehrum et al., 2013, Wehrum-Osinsky et al., 2014). This neural network, which is independent of general emotional arousal, is found both in men and women although men show overall stronger activations than women, which might be indicative of a stronger sexual responsivity in men. The same neural network activates for stimuli conditioned sexual arousal, with a gender effect in the same direction (Klucken et al., 2009).

In our study, we assess novelty, attentional bias and cue-conditioning to online explicit sexual material in individuals with CSB. These processes are highly relevant to substance-use disorders and may also be relevant to CSB. Online sexually explicit stimuli have significant potential for compulsive use, and CSB is relatively common, occurring in 2 to 4% in community and college-based young adults and in psychiatric inpatients (Grant et al., 2005, Odlaug and Grant, 2010, Odlaug et al., 2013). CSB is associated with significant distress, feelings of shame and psychosocial dysfunction. Although a working group for the 11^th edition of the International Classification of Diseases is currently proposing to include CSB as an impulse-control disorder (Grant et al., 2014), CSB was not included in DSM-5, albeit with some controversy (Toussaint and Pitchot, 2013), largely due to limited data. Thus, further studies are needed. Understanding the similarities and differences between CSB and other psychiatric disorders, particularly impulse-control disorders and addictions, may help with classification efforts as well as with the development of improved prevention and treatment approaches.

We have previously found that individuals with CSB demonstrate greater regional brain activation in response to explicit sexual cues in the ventral striatum, dorsal anterior cingulate cortex (dACC) and amygdala, regions implicated in drug cue reactivity and craving in disorders of addiction (Voon et al., 2014). Functional connectivity of this network, and particularly the dACC, was associated with greater sexual desire or motivation to explicit stimuli. We further observed that individuals with CSB, as compared to those without, demonstrate an early attentional bias towards sexually explicit cues (Mechelmans, Irvine, 2014). This early attentional bias was proposed to reflect facilitatory mechanisms underlying the motivational effect of cues conditioned to sexual outcomes. Here, we deepen our research focus investigating the mechanisms underlying the development of enhanced attentional bias and cue reactivity in CSB by assessing both behavioral and neural responses to novelty and to cue-conditioning in response to explicit sexual stimuli.

We conducted two behavioral tasks outside of the scanner to assess choice preference for novel versus familiar sexual stimuli and choice preference for cues conditioned to Sexual, Monetary and Neutral stimuli. We hypothesized that CSB individuals relative to healthy volunteers (HVs) would have greater choice preference to novel relative to familiar images in the Sexual condition but not in the control condition. We further hypothesized that CSB subjects would have greater choice preference to the conditioned cues in the Sexual condition but not in the Monetary condition.

Participants also performed a functional magnetic resonance imaging (fMRI) conditioning and extinction task involving conditioning to Sexual, Monetary and Neutral images. Two neutral stimuli were randomly paired with different sexual images shown repeatedly during conditioning. In the outcome phase of the conditioning arm, neural habituation to the sexual images was assessed by evaluating the change in neural activity of each different sexual image over time focusing on the repeated exposure thereby dissociating the analysis of the conditioning and outcome phases. We hypothesized that CSB subjects relative to HVs would show enhanced neural activity to the Sexual versus Neutral conditioned stimuli particularly in the dACC and striatum, regions previously identified in sexual cue reactivity in CSB subjects (Voon, Mole, 2014). We further hypothesized that CSB subjects compared to HVs would show greater neural habituation to Sexual compared to Neutral stimuli.

Method

Recruitment

The recruitment has been extensively described elsewhere (Voon, Mole, 2014). CSB subjects were recruited via Internet-based advertisements and therapist referrals. HVs were recruited from community-based advertisements in East Anglia. CSB subjects were interviewed by a psychiatrist to confirm they fulfilled diagnostic criteria for CSB (proposed diagnostic criteria for Hypersexual Disorder; criteria for sexual addiction) (Carnes et al., 2001, Kafka, 2010, Reid et al., 2012), focusing on compulsive use of online sexually explicit material.

All CSB subjects and comparably aged HVs were male and heterosexual given the nature of the cues. HVs were matched in a 2:1 ratio with CSB subjects to increase statistical power. Exclusionary criteria included being under 18 years of age, history of substance-use disorders, current regular user of illicit substances (including cannabis), and having a serious psychiatric disorder, including current moderate-severe major depression (Beck Depression Inventory > 20) or obsessive-compulsive disorder, or history of bipolar disorder or schizophrenia (Mini International Neuropsychiatric Inventory) (Sheehan et al., 1998). Other compulsive or behavioral addictions were exclusions, which were assessed by a psychiatrist, including problematic use of online gaming or social media, pathological gambling or compulsive shopping and binge-eating disorder.

Subjects completed the UPPS-P Impulsive Behavior Scale (Whiteside and Lynam, 2001), Beck Depression Inventory (Beck et al., 1961), State Trait Anxiety Inventory (Spielberger et al., 1983) and the Alcohol-Use Disorders Identification Test (AUDIT) (Saunders et al., 1993). The National Adult Reading Test (Nelson, 1982) was used to obtain an index of IQ.

Two CSB subjects were taking antidepressants and had comorbid generalized anxiety disorder and social phobia: social phobia (N=1) and a childhood history of ADHD (N=1).

Written informed consent was obtained and the study was approved by the University of Cambridge Research Ethics Committee. Subjects were paid for their participation.

Behavioral tasks

Twenty-two CSB subjects and 40 comparably aged male volunteers were tested in a novelty-preference task and two conditioning-preference tasks reported here, and an attentional bias task (dot-probe task) reported elsewhere (Mechelmans, Irvine, 2014). The tasks were carried out after the fMRI experiment, in a counterbalanced order.

Novelty preference

Subjects were familiarized to three categories of stimuli (Sexual images, Neutral human images and Neutral object images) and then performed a choice-discrimination testing phase, choosing between novel versus familiar stimuli matched within each category (Figure 1A). In the familiarization phase, six images were shown to the participant: 2 images of undressed women (Sexual condition), 2 images of dressed women (Control1) and 2 images of a piece of furniture (Control2) (2 images per condition). The 6 images were randomly presented in pairs to the participants, in a total of 48 trials (16 trials each condition). The duration of each trial was 5 sec. To ensure engagement with the task, subjects were instructed to carefully study the images because they would be asked questions during the familiarization phase. Simple questions about the images were randomly posed during the task in the inter-trial interval (e.g., to indicate which woman had her arms crossed using the right or left arrow: ‘Arms crossed’). Each question was relevant to the previously viewed pair of images, hence ensuring that subjects maintained attention to each pair of images.

In the testing phase, subjects viewed three image-pairs consisting of a familiarized image and a novel image matched for each experimental condition. Six images were used: 3 familiar, selected from the previous familiarization phase (one for each of the three conditions) and 3 new images (one novel for each condition). The image-pair was shown for 2.5 seconds followed by a 1-second feedback (win £1 or win nothing). A total of 60 trials (20 trials each condition) were presented. The probability of winning for any of the images was random at p=0.50. The subject was instructed to choose one of the stimuli from the pair with the goal to make as much money as possible and told that they would receive a proportion of their earnings. They were instructed that the first trial would be a guess but that one of the stimuli would be associated with a greater likelihood of winning. The primary outcome measure was the proportion of novel choices across trials for each condition. Since the learning contingencies used here were purely random (p=0.50), the outcome measure exclusively indicates a stimuli preference. After the study, subjects were asked to rate attractiveness of the female subjects on a scale of 1 to 10 following testing. Task duration was 8 minutes (4 min for the training and 3.5 min for the testing phase).

Conditioning preference

Subjects were tested on two conditioning preference tasks in a counterbalanced order, both consisting of a conditioning phase and a testing phase (Figure 1B). Both tasks had the same design but one focused on sexual and the other on monetary conditioning.

In one training phase, two visual patterns (CS+Sex, CS-), presented for 2 seconds, were conditioned to an image of an undressed female or a neutral grey box (1-second outcome), respectively. This was followed by an inter-trial interval of 0.5 to 1 second. Sixty trials were presented in total (30 CS⁺ and 30 CS^–). To ensure task engagement, subjects were instructed to keep track of the number of times they saw a red square around the outcome image, and they reported this number at the end of the training phase.

The training phase was followed by a testing phase in which the CS+Sex and CS- stimuli were each paired with a novel visual-pattern stimulus (e.g. Image A or Image B respectively). Subjects were asked to choose one of the stimuli from the stimulus pair (e.g. CS+Sex or Image A; CS- or Image B; duration 2.5 seconds), which was followed by feedback of win £1 or win nothing (duration 1 second). The CS+Sex and CS- had a greater probability of winning (p=0.70 win £1/p=0.30 win nothing) relative to the novel paired stimulus (p=0.70 win nothing/p=0.30 win £1). Subjects were tested for 40 trials in total (20 trials per condition) and were told the goal was to make as much money as possible and that they would receive a proportion of their earnings. They were instructed that the first trial would be a guess but that one of the stimuli would be associated with a greater likelihood of winning.

In the second training and testing task, a similar task design was used paired with monetary outcomes: a different set of visual patterns was conditioned (CS+Money, CS-) to the image of £1 or a neutral grey box. Subjects were told they would win a proportion of the money they viewed. A similar testing phase followed.

As the CS+ and CS- stimuli were associated with higher probabilities of winning, we assessed novelty choice preference of the first trial to assess initial approach behaviors and the proportion of times the CS+ and CS- stimuli were selected across all trials to assess the influence of choice preference of the cue on instrumental learning. Each task lasted approximately 7 minutes (4 min for the training and 2.5 min for the testing phases).

Imaging task

Twenty CSB subjects and 20 matched HVs were scanned performing a conditioning and extinction task (Figure 3A ). In the conditioning phase, six images (colored patterns) were used as conditioned stimuli (CS+) paired with the unconditioned stimulus (US) image of an undressed female (CS+sex), £1 (CS+money) or a neutral grey box (CS-). Two CS+ were paired per outcome. Five different images of undressed females were used for the sexual outcomes and repeated 8 times over the course of conditioning. The CS+ duration was 2000 msec; at 1500 msec, the US was overlaid for 500 msec and followed by a response block with a central fixation point, which ranged from 500 to 2500 msec. To maintain attention to the task, subjects pressed the left button for the money outcome, the right button for the person outcome, and either button for the neutral outcome during the fixation period. Subjects viewed a total of 120 trials (20 per CS+ or 40 per condition) in the conditioning phase. The conditions were randomly presented. In the extinction phase, each CS+ was shown for 2000 msec without the US for a total of 90 trials (15 per CS+ or 30 per condition) followed by a fixation point (500 to 2500 msec). Thus, at 1500 msec, subjects would expect an outcome, which was unexpectedly omitted. Prior to the study, subjects were trained outside of the scanner on 20 trials of a similar design with different CS+ and images of females, money and neutral items to practice the motor response during the response block. During the practice, subjects viewed images of dressed females but were told that in the scanner, they might see explicit stimuli. All tasks were programmed using E-prime professional v2.0 software.