Feedback Processing as it Relates to Suicidal Ideation and Behavior Using a Time-Frequency Approach

Abstract

AbstractBackgroundSuicidal ideation (SI) and suicidal behavior (SB) may have different etiological pathways, particularly related to feedback processing. Identifying ERP components related to these suicide states may help researchers localize and understand these differences.Objective.Our primary purpose was to utilize time-frequency decomposition techniques to assess neurophysiological correlates of SI and SB during positive and negative feedback processing.Methods271 subjects (55.7% female; mean age=35.75, SD=16.07; 30.6% veterans) completed a questionnaire battery and a multi-task protocol, including a gambling feedback task (Gehring & Willoughby, 2002), while electroencephalography (EEG) was collected using a 96-channel EEG system. Exploratory factor analysis (EFA) was conducted across 20 self-report suicide items selected to index SI and SB from several suicide-relevant questionnaires. EFA results produced 2 factors, corresponding to SI and SB. Time-frequency principal component analyses produced measures across all frequency bands (i.e., delta-FN/P3, delta-SW, theta-FN, alpha, beta-1, beta-2, and gamma).ResultsExcepting alpha amplitude, all measured frequencies were related to SB, and not SI (across loss and gain, with some loss-gain differences). Alpha, uniquely, showed a relationship to SI and not to SB (across loss and gain, no loss-gain differences). Robust regressions confirmed that the delta, theta, and high-frequency (HF; beta-1, beta-2, and gamma as combined HF loss, gain, and loss-gain difference factors) measures were each independently related to SB.DiscussionBroadly, the results indicate that individuals with SB showed heightened neurophysiological response across multiple ERP components (except alpha) to gambling feedback, with significantly greater increases to loss stimuli relative to gain stimuli.General Scientific SummaryThis study, using time-frequency EEG/ERP measures from a gambling feedback task, supports the theoretical framework that suicidal behavior and suicidal ideation can be indexed separately. Generally, stronger reactivity to all feedback (i.e., loss and gain feedback during a gambling task) is seen in those with suicidal behavior and not in those with suicidal ideation, but with a small separate effect related to suicidal ideation.