Virtual Reality and Behavioural Sciences: a discussion with Prof. Patrice Renaud

I had the pleasure to discuss with Prof. Patrice Renaud about his work inside the ARVIPL laboratory at the Philippe-Pinel Institute, and his contribution to “A pilot development of virtual stimuli depicting affective dispositions for penile plethysmography assessment of sex offenders”. This study, published in 2014, laid the basis for an assessment protocol taking into account the emotional state of the victim [1].

Prof. Patrice Renaud, director of the ARVIPL at the Philippe-Pinel Institute in Montreal.

Prof. Renaud describes the ARVIPL (“Applications de la Réalité Virtuelle en Psychiatrie Légale”) as a laboratory specialized in the applications of virtual reality and other simulation techniques to the study of mental illness, the development of assessment protocols and the treatment of psychopathologies such as sexual deviancy or empathy disorders.

His early background is somewhat typical for a psychologist, but during his PhD he specialized in ergonomy and human-machine interactions. After a clinical internship in a correctional facility where he worked on forensic profiling of aggressors, he started working at the University of Quebec in Ottawa (UQO) and started developing his particular line of work combining ergonomy and forensic psychology.

Entering the simulation: how it all began

As Prof. Renaud points out, things were radically different when he started working: assessment of sexual deviancy was performed using real photographs of victims retrieved by the Royal Canadian Mounted Police (RCMP) or the FBI during their investigations. Given the sensitivity of the matter and the ethical questions raised by this method, the use of real photographs was ultimately prohibited. For this reason, his team started using audio stimuli, a technique much more limited given the high visual component of masculine sexuality [2]. But in 1996, two or three years after the Rothbaum laboratory started using virtual reality to evaluate acrophobia (i.e. fear of heights), the idea of replacing audio stimuli by “synthetic” visual stimuli in the assessment of sexual deviancy emerged and quickly flourished. The idea was to design a virtual reality environment in which the subject would be confronted with different 3D models, while monitoring various physiological responses linked with sexual arousal, in order to probe his or her sexual preferences.

From virtual to reality: assessment protocols for justice

Twenty years later, virtual reality has improved a lot and is now a key component of the assessment of sexual deviancy for the judicial system, in Canada and around the world. Using penile plethysmography (PPG) coupled with occulometry and electroencephalography (EEG), sexual interest for various 3D models is monitored and the conclusions serve as a basis for the court to determine the dangerousness of the individual. For its assessments and its clinical studies, the ARVIPL has always made sure to use state of the art virtual reality equipment and software, even though virtual reality technologies are perpetually improving.

Example of setup used for the study, combining visual and audio stimuli and monitored via occulometry, EEG, and PPG (not shown).

Measuring sexual arousal: the question of PPG

As Prof. Renaud points out, PPG is the gold standard in terms of sexual arousal assessment and it has been used for nearly 50 years, but it is a technique that can be tricked with simple cognitive processes such as associations (with a bad memory or a displeasing mental image). Indeed, around 80% of individuals are able to control their erectile function with such techniques. Thus, PPG has to be coupled with other measures such as occulometry, that allows the examiners to determine if the subject is averting his eyes, and even to recognize recently identified patterns associated with inhibitory cognitive techniques. Though the existence of recognizable inhibition patterns haven’t yet been proved when brain activity is measured by EEG, Prof. Renaud seems confident that it will happen soon. He explained that the validation of those new measures is a really complicated process, that requires comparison with already validated and well-established measures such as PPG.

Occulometry setup used in tandem with PPG.

(1) active Nuvision 60GX stereoscopic glasses (2) InterSense IS-900 motion tracker (3) occulomotor tracking system (ASL model H6) (4) wired frame virtual object (5) a virtual measurement point (VMP) (6) a gaze radial angular deviation (GRAD) from a VMP.

As I walk through the uncanny valley I will fear no evil: improved realism and contextualization

In an article entitled “A pilot development of virtual stimuli depicting affective dispositions for penile plethysmography assessment of sex offenders”, Elissa Dennis and her colleagues laid the basis of a new virtual reality protocol adding basic behavioural traits (fear, interest, anger, lasciviousness, etc) to the 3D models [1]. Indeed, behavioural traits play a key role in sexual arousal, and especially help in detecting deviant sexual proclivities. As it turns out, this new protocol has been validated, leading the way to a more realistic virtual environnement with which the subjects interact. The study shows that the displayed emotions and ages of the 3D models are correctly recognized by the subjects, and that the participants respond to the 3D models as expected.

Examples of 3D models used for assessment of sexual preferences and deviances, exhibiting differences in gender, age and posture.

As Prof. Renaud explains it, realism is important for a variety of reasons, and among them the necessity of an external validity on which the court can base a verdict. The example he gave is that, when evaluating a pedophile, it is necessary to specify the prefered age of the victims and the intensity of this preference, something which requires a great developmental and biological realism of the 3D models. For this reason, the models are developed according to Tanner’s developmental criteria for the caucasian mesomorphic body type. These criteria define developmental stages in terms of sexual attributes and pilosity, and are essential for the creation of accurate models. The team uses the BehaVR Software, a virtual reality software used for various behavioural health applications such as PTSD (Post-Traumatic Stress Disorder), anxiety and phobias.

Example of basic behavioural traits implemented into the 3D models

(a) sadness (b) shame (c) neutral (d) fear.

As we were discussing the realism of the models, I raised the paradoxical problem of the uncanny valley : the more globally realistic the simulation, the more the unrealistic details appear disturbing [3]. Prof. Renaud insisted that anatomical realism isn’t required to obtain sexual arousal (taking the example of japanese pornography which tends to unrealistically alter body proportions), so as of now the level of realism they have is sufficient to attain the goal of assessing sexual deviancy. There is no doubt that with the improvement of simulation the problem of the uncanny valley will arise, but never in a way that would drastically impair the diagnosis.

As I evoked the study of Goode et al in 2010, that pointed out the importance of situational factors in the selection of victims by sexual abusers, Prof. Renaud described his more recent work: A Walk In The Park [4]. This time incorporating context as a way to improve realism, this protocol presents the subject with various possibilities of aggression. The 3D models, based on the motion capture of both real victims displaying “victimization” attitudes and more confident individuals, are presented to the subject and contextualized during a stroll in a park. The goal is to assess the “criminal awareness” of the subject, a notion comparable to the situational awareness of soldiers. Situational awareness is defined as an acute perception of the environment and a cognitive projection of the future status of this environment after some variable has changed.

“Context is very important, but always depends on the goal of the study”, Prof. Renaud said, “that’s why virtual reality is so interesting: we can create an infinity of criminogenic situations and test how much psychological and/or physiological responses are modified by the environmental characteristics”.

Beyond the pleasure principle: next steps

When I asked Prof. Renaud in what ways he would like to orient the use of virtual reality, he said that “the key resides in the combination of virtual reality and augmented reality”. In contrast with virtual reality in which the environment is entirely virtual, augmented reality only adds few virtual components to the real environment. Indeed, the ultimate goal is to obtain the individual signature of the offender with a classic virtual reality assessment protocol, in order to treat the subject using a virtual reality environment and finally give him a way to continue his treatment outside of the carceral or psychiatric institution. For example, this could be about using augmented reality to give the subject feedback on his own behaviour and physiological responses, while monitoring said behaviour in order to keep the individual in check. “Evidently this is not Brave New World.”, said Prof. Renaud, “The idea is to give the individual feedback, to improve his comprehension of his case and evaluate his own forces and limitations”.

Another step would be to apply virtual reality to the study, assessment and treatment of other psychopathologies such as empathy disorders associated with psychopathy and other violent behaviours. That’s why A Walk In The Park is conducted with Prof. J.P. Guay from the department of criminology at UdeM. Indeed, Prof. Renaud explained that “as we know in the litterature, a good proportion of criminality with interpersonal violence or even acquisitive violence is based on an incorrect interpretation of other people’s emotions, or an interpretation without any inhibition of criminal and aggressive tendencies”. Thus, developing virtual reality and assessment protocols might help in understanding more psychopathologies involving interpretation of emotions, contextualization, and the notion of empathy. “Evaluating the potential for empathy and treating it could be really useful in our line of work”, concluded Prof. Renaud.

References:

[1] E.Dennis, J.L.Rouleau, P. Renaud, K. Nolet, C. Saumur (2014). A pilot development of virtual stimuli depicting affective dispositions for penile plethysmography assessment of sex offenders, The Canadian Journal of Human Sexuality 23(3), 2014, pp. 200–208; doi:10.3138/cjhs.2529

[2] H. A. Rupp, K.Wallen (2008). Sex Differences in Response to Visual Sexual Stimuli: A Review. Archives of Sexual Behavior37(2), 206–218. http://doi.org/10.1007/s10508-007-9217-9

[3] M. Cheetham, (2017). Editorial: The Uncanny Valley Hypothesis and beyond. Frontiers in Psychology8, 1738. http://doi.org/10.3389/fpsyg.2017.01738

[4] Goode, S.D. (2010). Understanding and addressing adult sexual attraction to children: A study of paedophiles in contemporary society. New York: Routledge.

Leave a Reply

Your email address will not be published. Required fields are marked *