Differences in suicide-specific attentional bias based on stimuli across the suicide Stroop and disengagement tasks

Background: Evidence of suicide-specific attentional bias has been mixed, including the effect of specific suicide words (e.g., "suicide" versus "overdose"). The current study examined the psychometric properties and convergent validity of specific suicide words on the Stroop and disengagement tasks.

Methods: Adults (79.8 % male, mean age = 42.19 ± 12.66) on a psychiatric inpatient unit completed both tasks, as well as clinical and demographic measures, following a recent suicide attempt (SA) (n = 55) or recent suicidal ideation (SI) with (n = 29) or without (n = 26) a SA history.

Results: Mean reaction times (RTs) on both tasks demonstrated good internal consistency and were moderately correlated across tasks, while interference scores demonstrated poor reliability and were not correlated. Attentional bias to specific suicide words did not differentiate SA groups, except for a small, unexpected effect on the Stroop task showing greater interference to the word "funeral" among the recent SA group. Correlations with suicide risk factors differed for males and females.

Limitations: Participants all endorsed elevated SI, limiting generalizability beyond a recent suicidal crisis. Different suicide words were used on each task limiting direct comparisons. Groups based on SA history and gender were uneven, and correlations need to be replicated. Effects involving interference scores should be interpreted with caution given poor internal consistency.

Conclusion: The suicide Stroop and disengagement tasks partially capture distinct aspects of suicide-relevant attentional bias. Greater attention is needed regarding the particular stimuli and scoring methods as well as the role gender plays in suicide-relevant attentional bias.