The Promise and the Challenge of Technology-Facilitated Methods for Assessing Behavioral and Cognitive Markers of Risk for Suicide among U.S. Army National Guard Personnel

Brian Baucom, Panayiotis Georgiou, Craig J. Bryan, Eric L. Garland, Feea R. Leifker, Alexis M. May, Alexander E. Wong, Shrikanth NarayananInternational Journal of Environmental Research and Public Health 2017 · DOI

TL;DR

This paper explores novel technology-facilitated methods, specifically Behavioral Signal Processing (BSP) and computer-administered cognitive tasks, for assessing suicide risk in U.S. Army National Guard personnel. These methods aim to overcome limitations of traditional screening by enabling remote, non-disclosure-based assessment of behavioral and cognitive markers, offering a promising avenue for more effective suicide prevention.

What Problem Does This Paper Address?

Suicide rates among U.S. military personnel, particularly the National Guard, are alarmingly high and continue to rise. Existing suicide prevention efforts face significant hurdles: the geographic dispersion of personnel limits access to healthcare, and professional and social stigma discourages individuals from disclosing psychological distress or suicidal ideation. Traditional screening methods are often infrequent, rely on self-report, and may miss individuals at acute risk due to the timing of assessment or concealment of intent. This leaves a critical gap in identifying individuals experiencing heightened suicide risk, especially in a population that may be hesitant to seek help through conventional channels.

Methods

The paper proposes two primary technological approaches:

  • Behavioral Signal Processing (BSP): This involves using computational tools to automatically measure and analyze behaviors during recorded conversations. BSP extracts mathematical features from audio recordings to quantify aspects like negative affect, conflict, and emotional responsiveness. Key techniques include:
  • Voice Activity Detection (VAD): Distinguishing speech from background noise.
  • Diarization: Identifying who is speaking when.
  • Automatic Speech Recognition (ASR): Transcribing spoken words.
  • Feature Extraction: Deriving acoustic and linguistic features (e.g., fundamental frequency, energy, prosody).
  • Automated Observational Coding: Replicating human judgments of behaviors like anger, blame, or support.
  • Feature-Derived Behavioral Markers: Mathematical summaries of vocal behavior (e.g., entrainment, affective markers, saliency) that capture subtle emotional and interactional dynamics.
  • Computer-Administered Cognitive Performance Tasks: These tasks assess implicit cognitive biases related to suicide. They are designed to be taken remotely and do not require direct disclosure of suicidal thoughts. Examples include:
  • Suicide Implicit Association Test (S-IAT): Measures the strength of automatic associations between concepts like "death" and "me" versus "life" and "not me" through reaction times.
  • Dot Probe Task: Assesses attentional bias by measuring reaction times to identify a probe following the presentation of emotional and neutral stimuli, indicating a tendency to focus on suicide-related words.
  • Partner GNAT (Go/No-Go Association Task): Measures implicit beliefs about relationship satisfaction, which is a known factor in suicide risk.

Datasets

Not discussed in the paper. The paper focuses on the methodologies and their potential application, rather than presenting results from specific datasets. It outlines the types of data (audio recordings of conversations, performance on cognitive tasks) that would be used for evaluation.

Key Insights and Ideas

  • Bridging the Disclosure Gap: BSP and implicit cognitive tasks offer a way to assess risk without requiring individuals to explicitly admit to suicidal thoughts or distress, which is crucial for populations with high stigma.
  • Remote and Scalable Assessment: These technologies can be deployed remotely (e.g., via the internet or personal devices), overcoming the geographic dispersion challenges faced by the National Guard.
  • Capturing Subtle Cues: BSP can detect subtle behavioral and emotional cues in speech and interaction that might be missed by human observers or traditional self-report measures.
  • Chronic vs. Acute Risk: The paper highlights the importance of distinguishing between chronic (long-term) risk factors (like relationship problems) and acute (short-term) fluctuations in risk, suggesting that BSP and cognitive tasks can help identify both.
  • Spouses as Key Observers: The paper emphasizes the unique position of spouses in noticing acute warning signs, and how BSP can analyze their interactions to identify risk.

Key Results

This paper is primarily a conceptual and methodological review, outlining the potential of these technologies. It does not present empirical results from their application to suicide risk assessment. Instead, it discusses existing research that supports the underlying principles:

  • Research on BSP has shown its ability to measure behavioral markers of relationship distress and emotional arousal during couple and soldier-therapist interactions, with some studies showing correlations with human ratings of up to 0.80.
  • Studies on implicit measures (like IATs and dot probe tasks) have demonstrated their utility in predicting various psychological outcomes, including suicidal behavior, and their ability to explain variance beyond explicit measures.

Limitations and Future Work

The authors acknowledge several limitations and suggest future research directions:

  • Empirical Validation: The most significant limitation is the need for rigorous empirical testing of these methods to demonstrate their ability to:
  • Accurately differentiate between different risk groups (e.g., history of attempts, ideation only, controls).
  • Detect time-varying risk and predict future suicide attempts.
  • Maintain robustness with data collected under varying conditions and quality (e.g., non-laboratory settings, consumer-grade devices).
  • Iatrogenic Effects: While the likelihood is considered low, the potential for unintended negative side effects from discussing relationship issues needs thorough assessment.
  • Practical Implementation: Further work is needed to refine algorithms, ensure scalability, and develop user-friendly interfaces for widespread adoption.
  • Specific Application: The paper calls for testing these methods specifically within a National Guard sample to establish their validity for this population.
  • Combination of Methods: The paper suggests that combining BSP-based behavioral markers with computer-based cognitive bias measures will likely yield the most comprehensive assessment.

Summary updated

Abstract

Suicide was the 10th leading cause of death for Americans in 2015 and rates have been steadily climbing over the last 25 years. Rates are particularly high amongst U.S. military personnel. Suicide prevention efforts in the military are significantly hampered by the lack of: (1) assessment tools for measuring baseline risk and (2) methods to detect periods of particularly heightened risk. Two specific barriers to assessing suicide risk in military personnel that call for innovation are: (1) the geographic dispersion of military personnel from healthcare settings, particularly amongst components like the Reserves; and (2) professional and social disincentives to acknowledging psychological distress. The primary aim of this paper is to describe recent technological developments that could contribute to risk assessment tools that are not subject to the limitations mentioned above. More specifically, Behavioral Signal Processing can be used to assess behaviors during interaction and conversation that likely indicate increased risk for suicide, and computer-administered, cognitive performance tasks can be used to assess activation of the suicidal mode. These novel methods can be used remotely and do not require direct disclosure or endorsement of psychological distress, solving two challenges to suicide risk assessment in military and other sensitive settings. We present an introduction to these technologies, describe how they can specifically be applied to assessing behavioral and cognitive risk for suicide, and close with recommendations for future research.
BibTeX
@article{baucom2017promise,
  author = {Brian Baucom and Panayiotis Georgiou and Craig J. Bryan and Eric L. Garland and Feea R. Leifker and Alexis M. May and Alexander E. Wong and Shrikanth Narayanan},
  title = {{The Promise and the Challenge of Technology-Facilitated Methods for Assessing Behavioral and Cognitive Markers of Risk for Suicide among U.S. Army National Guard Personnel}},
  journal = {International Journal of Environmental Research and Public Health},
  year = {2017},
  doi = {10.3390/ijerph14040361},
}
Figure