Abstract:Violent conduct in society is a major health concern, and therefore one of the major aims in forensic mental healthcare is the assessment of the risk for (future) violence. The prediction of violence risk is traditionally done by using violence risk assessment tools in the form of actuarial instruments and structured professional judgments. However, research has shown that the risk assessment tools that are currently being used, are not as accurate in predicting future risk as one would hope for. Therefore, some scholars advocate a paradigm shift by incorporating a biopsychosocial model, including neurobiological measures in risk assessment models. The current review aimed to give answer to the question whether neuro-imaging measurements add in the accuracy of predicting aggression and violent (criminal) recidivism, and to discuss what next steps need to be taken to move neuro-imaging informed risk assessment forward in forensic clinical practice. The results of the current review showed that findings regarding 'neuroprediction' are mainly based on cross-sectional studies and that studies using rigorous methods for determining the incremental predictive accuracy of neural parameters are lacking. Future studies need to rely on prospective and longitudinal data that address the added value of neural measures over traditional risk factors specifically, taken different ethical and judicial considerations into account. Studies that take a personalized approach, focusing on neurobiologically informed risk assessment at the individual level to translate this into forensic clinical practice, is in our view the next step that will move this forward.

Neuroprediction of violence and criminal behavior using neuro-imaging data: From innovation to considerations for future directions