Discussion

The relationship between mental illness and radicalisation is a subject of ongoing debate in the literature. In this study, the number of perpetrators with underlying psychopathology was low. While this suggests that mental disorders are not a mainstream common characteristic of perpetrators, it may be partly explained by the fact that the assailants’ possible psychopathological profile was not the main goal of data collection.

In fact, most other reviews fail to demonstrate a significant association between terrorism, radicalisation and mental illnesses, e.g. psychotic disorders. ^{Reference Sarma, Carthy and Cox1–Reference Ho, Quek, Ho and Choo3,Reference Trimbur, Amad, Horn, Thomas and Fovet10–Reference Catapano, Cipolla, De Rosa, Milano, Vozza and Guadagno12} Furthermore, when comparing acts of terrorism and acts of homicide, no significant association with mental illness was found. ^{Reference Theodorou, Sinclair, Ali, Sukhwal, Bassett and Hales13} Still, in some subsets of assailants and terrorist incident characteristics, mental illness issues could be involved. ^{Reference Ho, Quek, Ho and Choo3,Reference Trimbur, Amad, Horn, Thomas and Fovet10–Reference Theodorou, Sinclair, Ali, Sukhwal, Bassett and Hales13}

A recent comparative study by Theodorou and colleagues (2024) on family shooters, mass shooters and terrorists provided evidence of mental illness in only a minority of perpetrators, depression and psychosis being the most frequently reported underlying diagnoses. ^{Reference Theodorou, Sinclair, Ali, Sukhwal, Bassett and Hales13} They concluded that no individual mental illness could be identified as the predominant risk factor in either of the subgroups. Hence, they underlined the importance of the interaction between mental disorders, socioeconomic factors, relationships and loss in determining individual risks.

Analysis of the Columbia (USA) mass murder database revealed that focusing on serious mental illness, e.g. psychotic symptoms, did not support preventive policies for mass murders. ^{Reference Brucato, Appelbaum, Hesson, Shea, Dishy and Lee14}

In our findings, lone actors prevail in terrorist attacks where psychopathology might be involved. Other studies provided similar evidence. Single issues and right-extremist lone actors have been studied most extensively. ^{Reference Schulten, Doosje, Spaaij and Kamphuis2} Autism spectrum disorder, schizophrenia and delusional disorders were more frequent in lone actors than in the general population, but still only a minority of these perpetrators qualified for psychopathological characteristics. ^{Reference Schulten, Doosje, Spaaij and Kamphuis2} Corner and Gill observed that lone actors were 14.3 times more likely to have a history of mental illness than group actors. ^{Reference Corner and Gill15}

A recent French study assessing 90 people in detention for jihadist terrorism confirmed the significant association between lone actors and psychiatric pathology, as 30% had an active psychiatric disease. The most frequent diagnosis was personality disorder, followed by schizophrenia, schizotypal disorders and delusional disorders, neurotic and stress-related disorders and somatoform disorders. ^{Reference Monod, Ambar-Akkaoui, Levasseur and Pham-Scottez16}

Additionally, socio-environmental (family dysfunction, friendships with radicalised individuals, an unstable geopolitical environment and societal polarisation) and neurobiological factors (structural and functional deficits in limbic and frontal brain areas) might be associated with terrorist behaviour. ^{Reference Ho, Quek, Ho and Choo3,Reference Salib17}

The 107 incidents listed in this series, resulted in a total of 99 fatal injuries and 217 NFI. Twenty-nine perpetrators died during the attack.

The possible relation between terrorism and mental illness or addiction is only addressed in the last three decades of the GTD. However, this series is likely an underestimation due to the lack of peripheral data on the incidents in the GTD before 1990, as established in previous GTD-based studies. ^{Reference De Cauwer, Barten, Tin, Mortelmans, Lesaffre and Somville18–Reference De Cauwer, Granholm, Khorram-Manesh, Barten, Tin and Mortelmans20} Changing insight into this matter and a recent focus on the psychopathology of assailants might also be the reason for the growing number of assailants labelled as suffering from a mental disorder.

In this series, the typical attacker profile consisted of a lone actor with a(n assumed) mental illness committing an armed assault. This is in line with the above mentioned studies. ^{Reference Schulten, Doosje, Spaaij and Kamphuis2,Reference Corner and Gill15,Reference Monod, Ambar-Akkaoui, Levasseur and Pham-Scottez16}

Although 29 perpetrators were killed during the incidents, suicide by cop was reported in only five cases.

Suicide by cop is not similar to suicide bombings. In suicide bombings, suicide is merely instrumental, and not necessarily instilled by a psychopathological context. Suicide bombing is only one of many possible modes of attack and is rather driven by group-dynamic indoctrination and political motives. ^{Reference Grimland, Apter and Kerkhof21}

In these five analysed cases, psychopathological factors possibly played a role and were adjudged to take precedence over the perpetrator’s alleged narrative.

Armed assaults were the predominant attack type, followed by infrastructure attack, another difference with suicide bombings. ^{Reference Tin, Galehan, Markovic and Ciottone22} The mean fatal injuries in this series was 0.91, far less than the mean fatal injuries in suicide bombings (9.66 per event). ^{Reference Tin, Galehan, Markovic and Ciottone22}

On the other hand, the most affected targets are similar in both groups: police, the military and private citizens and property, although in different order. ^{Reference Tin, Galehan, Markovic and Ciottone22}

In contrast with a previous study, none of the profiles of the five assailants suspected to have committed suicide by cop corresponded with the typical white male experiencing a romantic relationship conflict who has a significant mental health and criminal history and who is often intoxicated at the time of the offence. ^{Reference Patton and Fremouw23}

Bhui and colleagues found a similar profile in their assessment of sympathies for violent protest and terrorism: having major depression with dysthymia, anxiety, post-traumatic stress disorder, being young (<21 years), White and having previous criminal convictions were positively correlated with extremism. No associations were found with life events, social assets and political engagement. ^{Reference Bhui, Otis, Silva, Halvorsrud, Freestone and Jones24}

Similarly, in our study, none had previous convictions. One assailant did not match the age and race criteria (51-year-old male of Egyptian descent).

The others were definitely younger, but neither of them met the social or psychological profile. Due to the unavailability of forensic investigation results in the GTD, we are unable to thoroughly compare the perpetrator profiles.

Assessing the possible psychiatric background of perpetrators, the targets, the attack type, the motives, the outcome and the judicial aftermath can lead to a better understanding of the goals of such acts of terror. Eventually, this may refine initiatives to prevent terrorist attacks.

Law enforcement officers, social workers, public health authorities and (para)medical caregivers could play a role in preventing incidents, but they need tools to predict and assess imminent events. Treatment protocols have been established for assailants of terrorist attacks due to radicalisation.

Both risk assessment and therapeutic measures seem more efficient and valid in case of radicalisation than in mental illness-related terrorist incidents.

The TRAP-18 Terrorist Radicalisation Assessment Protocol, for instance, is a tool trying to assess the individual threat of targeted violence. Meloy and Gill’s study demonstrated four warning behaviours for successful attackers: leakage, pathway warning behaviour, fixation and identification. ^{Reference Meloy and Gill4} Similarly, the Sympathies for Violent Protest and Terrorism (SVPT)-screening tool measures to what extent individuals sympathise/condemn seven specific actions (ranging from ‘committing a minor crime in political protest’ to ‘the use of suicide bombs to fight against injustices’). ^{Reference Bhui, Otis, Halvorsrud, Freestone and Jones5}

However, practitioners in mental health underline the challenges of risk-assessment procedures. From an ethical point of view, approaching ‘mentally ill’ or ‘radicalised’ people as potential terrorists is not only without scientific foundation but also diametrically opposed to the democratic values underlying our society. ^{Reference Schumann, Brook and Heinze25–Reference Schuurman and Carthy27}

It should be noted that radicalisation is not a mental disorder and that therapy for radicalisation is not always successful. On the other hand, the presence of the mental disorder could make someone more vulnerable to radical ideas or violent behaviour. ^{Reference Schulten6}

Once mentally ill perpetrators are convicted after a successful act of terror or committed in a psychiatric facility, Countering Violent Extremism programmes – added to regular mental health care – could be justified. ^{Reference Schulten6}

To conclude, despite the low prevalence among perpetrators, several attempts have been made in the literature to associate specific psychopathology, either personality traits or mental disorders, with terrorism. Whereas the personality traits ‘paranoid’, ‘antisocial’ and ‘sadomasochistic’ are associated with terrorists in general, ‘dependent’, ‘avoidant’ and ‘impulsive’ traits are more likely to occur in suicide bombers. ^{Reference Ho, Quek, Ho and Choo3} When zooming in on lone actors, the mental disorders most frequently found are schizophrenia, delusions and autism spectrum disorder. ^{Reference Schulten, Doosje, Spaaij and Kamphuis2} The study of Theodorou adds social role disjunction, motive, substance misuse and relevant risk or threat behaviours to the list. ^{Reference Theodorou, Sinclair, Ali, Sukhwal, Bassett and Hales13}

Importantly, the association is, however, not bidirectional: terrorist acts are not associated with mental illness, and radicalisation alone does not equal mental illness. ^{Reference Schuurman and Carthy27} Although counterintuitive from the perspective of humanity, most terrorists are psychologically normal. Of those displaying mental difficulties or disorders, a majority had already developed these problems before they radicalised. ^{Reference de Ridder, Fassaert and Grimbergen28} Instead of having predictive value, mental illnesses can be caused by involvement in terrorism or disable a person from engaging in larger goal-oriented activities. ^{Reference de Ridder, Fassaert and Grimbergen28} Further, half of the perpetrators of mass shootings leaked their plans in advance, suggesting an opportunity for a mental health intervention. ^{Reference Peterson, Erickson, Knapp and Densley29} Since radicalisation is not a static construct independent of individual, temporal and socio-environmental variables, the cause of radicalisation in terrorist violence should not be sought in individual experiences, vulnerability or even pathology; ‘it is the radicalisation process itself that forms the primary risk factor’. ^{Reference Schuurman and Carthy27}

In line with the commonly accepted biopsychosocial model in psychiatry, this paper confirms the importance of not only approaching mental health problems as a continuum, ranging from mental difficulties to strict DSM-5 diagnoses, but also of considering mental illnesses in perpetrators as the result of a dynamic process in which individual socio-environmental, neurobiological and cultural factors interact. ^{Reference Sarma, Carthy and Cox1,Reference Ho, Quek, Ho and Choo3} Tools for risk assessment in potential terrorist perpetrators should be based on the widespread 4P model, which distinguishes between predisposing, precipitating, protective and perpetuating factors, which may vary over time. ^{Reference Bolton30}