Detecting Suicidality From Facial Appearance

Abstract

Suicide is a pervasive problem worldwide. In this investigation, we show that individuals can perceive suicidality from facial appearance with accuracy that is significantly greater than chance guessing. Inferences of expected or obvious cues, such as how depressed a person seems, did not lead to accurate judgments. Rather, perceptions of how impulsive an individual appears differentiated suicide victims from living controls. Teasing apart various forms of impulsivity revealed that perceptions of impulsive aggression, distinct from other forms of impulsive behavior (e.g., impulsive buying), distinguished suicide victims from controls. Finally, experienced mental health clinicians did not perform significantly better than laypersons at judging suicidality. Facial appearance may therefore hold cues to suicidality, expanding what is known about the expression and perception of social cues from the face and providing new insights into the relationship between mental health and nonverbal cues.

Identifying Mental Disorder from the Faces of Women with Borderline Personality Disorder

“…More recent studies have suggested that faces alone express cues to depression (Scott et al 2013), and that people can determine who commits suicide from photos of suicidal and non-suicidal targets' faces (Kleiman and Rule 2013). Scott et al (2013) created composite faces (digital averages of multiple independent faces) of individuals who had scored high and low on a self-report measure of depression, finding that perceivers could perceive differences in depression from these computer-synthesized faces.…”

“…Scott et al (2013) created composite faces (digital averages of multiple independent faces) of individuals who had scored high and low on a self-report measure of depression, finding that perceivers could perceive differences in depression from these computer-synthesized faces. Using more natural stimuli (yearbook and family photos), Kleiman and Rule (2013) found that individuals could distinguish people who had died by suicide from matched controls significantly better than chance and determined that perceptions of the targets' impulsive aggression supported this difference (though they did not have information about the individuals' actual impulsivity or aggression to corroborate these perceptions).…”

“…Previous research suggests that dynamic information about an individual can be extrapolated from static photos (Naumann et al 2009). Whereas research employing videos or brief in vivo interactions may be more accurate, given the increased behavioral information available (Borkenau and Liebler 1992;Waxer 1976), other research suggests that ''thin-slices'' of behavior from photographs may yield similar levels of accuracy (Kleiman and Rule 2013;Penton-Voak et al 2006;Scott et al 2013). In addition, individuals' emotional experiences seem to impact their facial appearance through physiological processes over the lifespan (Ekman 1978;Malatesta et al 1987).…”

Thin-Slice Judgments in the Clinical Context

“…State anxiety is more reliably detected in audio channels, whereas trait anxiety is more reliably detected in video channels (Harrigan et al, 2004). Even suicide attempts and outcomes can be predicted from thin slices Kleiman & Rule, 2013; see also Ekman & Friesen, 1969).…”

“…A recent set of studies examined the ability to predict from just a face, whether a particular person committed suicide (Kleiman & Rule, 2013). This set of studies had the advantage of having an impressively larger sample size of targets.…”